Professional Documents
Culture Documents
VOLUME ONE
VOLUME LV
IN SYNCHRONY WITH THE HEAVENS
Studies in Astronomical Timekeeping and Instrumentation
in Medieval Islamic Civilization
(STUDIES I-IX)
VOLUME ONE
THE CALL OF THE MUEZZIN
BY
DAVID A. KING
BRILL
LEIDEN BOSTON
2004
Illustration on the cover: see fig. 2.1, page 861.
ISSN 0169-8729
ISBN 90 04 12233 8
All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without prior written
permission from the publisher.
Individual parts of this work are dedicated to the teachers and colleagues
who helped me appreciate the achievements of these scholars
in their cultural contexts.
vi part ii, chapter one
TABLE OF CONTENTS
Preface 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Preface 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
Statement on previous publication of parts of this work . . . . . . . . . . . . . . . . . . . . . . . . . xviii
Bibliography and bibliographical abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xx
Part I A survey of tables for timekeeping by the sun and stars . . . . . . . . . . . . . . . 1
Part II A survey of tables for regulating the times of prayer. . . . . . . . . . . . . . . . . . 191
Part III A survey of arithmetical shadow-schemes for time-reckoning . . . . . . . . . . . 457
Part IV On the times of Muslim prayer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
Part V On the role of the muezzin and the muwaqqit in medieval Islamic
societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
Part VIa Universal solutions in Islamic astronomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
Part VIb Universal solutions to problems of spherical astronomy from Mamluk
Syria and Egypt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711
Part VIIa The orientation of medieval Islamic religious architecture and cities . . . . . 741
Part VIIb Architecture and astronomy: the ventilators of medieval Cairo and their
secrets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 773
Part VIIc Safavid world-maps centred on Mecca . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 825
Part VIII Aspects of practical astronomy in mosques and monasteries . . . . . . . . . . . . 847
Part IX When the night sky over Qandahar was lit only by stars . . . . . . . . . . . . . . 881
Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 897
Note: Vol. 2, entitled Instruments of Mass Calculation and containing Studies X-XV, is to be published
by Brill Academic Publishers in 2004. This contains:
Preface
Statement on previous publication of parts of this work
Bibliography and bibliographical abbreviations
Part X Astronomical instrumentation in the Islamic world
Part XI An approximate formula for timekeeping (750-1900)
Part XIIa On the universal horary quadrant for timekeeping by the sun
Part XIIb On universal horary dials for timekeeping by the sun and stars
Part XIIIa The oldest astrolabe in the world, from 8th-century Baghdad
Part XIIIb Instruments of mass calculation from 10th-century Baghdad
Part XIIIc Astronomical instruments from medieval Syria
Part XIIId A medieval Italian testimonial to a forgotten Islamic tradition of non-standard astrolabes
Part XIIIe An astrolabe from medieval Spain with inscriptions in Hebrew, Arabic and Latin
Part XIV The geographical data on early medieval Islamic instruments
Part XV An ordered list of early Islamic and early European instruments (to ca. 1550)
Indexes of instruments and personal names
viii part ii, chapter one
PREFACE
The obligation of Muslims to pray at specific times in a specific direction gave rise to a
substantial literature during the period between ca. 750 and 1900. The prescriptions were
interpreted at two different levels, namely, by the astronomers, who proposed mathematical
solutions, and by the scholars of the religious law, who proposed non-mathematical solutions.
This dichotomy between mathematical science and folk science, now to some extent docu-
mented, is not known to have led to any strife.
This particular aspect of the activity of the Muslim astronomers belongs to what I have called
Astronomy in the Service of Islam.1 But it might also be considered as Islams service to
the astronomers:2 it brought them work and perhaps occasionally recognition. But today, even
in the Islamic world, their activities and achievements have been completely forgotten. Muslim
readers in particular should appreciate that virtually all of the materials presented in this book
have never been researched before in modern times: therefore, they should not be surprised
to find that some of them have called for a reinterpretation of hypotheses proposed by previous
generations who did not have access to these materials. In fact, only those Muslim astronomers
whose works were used in the medieval Europe have received much recognition in modern
times. It is my firm opinion that the history of Islamic astronomy merits study for its own sake,
as part of the history of Islamic civilisation generally.3 Furthermore, that history cannot be
written from texts alone. In this study I have recourse to not a few texts, but my main sources
are astronomical tables, most without any accompanying text, and also astronomical instru-
ments.
Virtually none of the materials I present in this book were known in medieval Europe.
Nevertheless, as the reader will learn, European astronomers from the 16th to the 20th century
produced tables of the same kind as the earlier Islamic ones, without any inkling that they were
not the first to do so.
The literature of the Muslim astronomers relating to timekeeping in general and the
regulation of the prayer-times in particular is treated here first (Parts I-II), because at the time
I worked on it over 20 years ago I had no conception that the legal scholars addressed the
same topics at a different level. However, I have since included at least the arithmetical schemes
for timekeeping proposed by the legal scholars and some of their comments on the definitions
of the times of prayer (III-IV). I also investigate the social background of the Muslim
astronomers charged with the regulation of the times of prayer (V) and a significant technical
aspect of astronomical timekeeping and instrumentation, namely, universalism (VI). Some of
the material in the scientific sources in this case, a table for orienting massive ventilators
1
This is the title of one of the volumes of reprints of some of my early publications (Studies, C).
2
This point is made in Ragep, Freeing Astronomy from Philosophy, p. 51.
3
See the surveys in King, Islamic Astronomy, and Saliba, Islamic Astronomy and Astrology, as well
as idem, The Astrologer in Islamic Society. None of our generation has produced anything to really replace
Nallino, Islamic Astronomy (1921).
x preface
of some 125 which Kennedy was able to identify. In the 1970s, mainly during our Cairo days,
Kennedy and I upped the number of known zjes to over to 200. Also I had started to unearth
vast numbers of tables that were not at all of the kind standard in zjes.5 Yet today, half a century
after Kennedy started work on zjes, not one more zj has been published in its entirety. Now
each zj contains a couple of hundred pages of tables and text, so there are good reasons why
only the hardy would get involved with a zj. We were all delighted when Benno van Dalen
undertook to prepare a much more detailed survey of the some 225 Islamic zjes now known,
and his Frankfurt-based project has made very satisfactory progress, so that a publication is
in sight. In the meantime, however, the Encyclopaedia of Islam was progressing to the letter
Z, and I undertook to write the article on zjes, with Julio Sams contributing the early Eastern
Islamic and Andalus sections, and Bernard Goldstein the section on Hebrew zjes. Such works
contain:
Tables for calendar conversion;
Tables of trigonometric functions;
Tables of stellar coordinates;
Tables of geographical longitudes and latitudes;
Tables of functions relating to spherical astronomy, such as solar declination and right and
oblique ascensions; tables for timekeeping (not common in zjes);
Tables of solar, lunar and planetary mean-motions and equations; lunar and planetary
latitudes; planetary stations; lunar and planetary visibility; equation of time;
Tables for computing conjuctions and eclipses; and, last but not least,
Tables for mathematical astrology.
Someone could (and should) write a book on each of these categories. Inevitably, what we
wrote for the Encyclopaedia was far longer than what the Editors wanted or could accept. A
short version was prepared for the Encyclopaedia, and was published with several illustrations.
The longer version was published in Suhayl (Barcelona) in 2001, without any illustrations.6
It was subtitled An Interim Report because Benno van Dalen has a much more detailed survey
in store for us, and for the time being I refer the interested reader to his website.7 My point
is that the subject of this book, Islamic astronomical timekeeping, is just one part of a much
greater entity, astronomy in medieval Islamic civilization.8
In two parallel studies, I have investigated the ways in which Muslims handled the
determination of the sacred direction towards Mecca. In the first, I show how the Muslim
astronomers treated the problem as one of mathematical geography, using formulae, tables and
5
King, Islamic Astronomical Tables.
6
King & Sams (with Goldstein), Islamic Astronomical Handbooks and Tables.
7
www.rz.uni-frankfurt.de/~dalen/index.htm (2003).
8
See the overviews mentioned in n. 3 above. The interested reader will find over a hundred extracts from
Islamic scientific manuscripts in Cairo ENL Survey, pp. 195-331, but should bear in mind that these are mainly
from manuscripts post-dating ca. 1300. More representative pictorial overviews could be prepared from the
collections in Turkey, Iran or Europe. UNESCO is about to publish a volume on illustrated scientifio manuscripts
in Iranian collections, edited by Ziva Vesel. This is a step in the right direction, but it should be remembered
that some of the most important scientific manuscripts are not illustrated at all. Nobody has taken seriously my
list of Islamic scientific works which should be published in facsimile editions using the best manuscripts (King,
Islamic Scientific Manuscripts, pp. 128-134).
xii preface
maps.9 In the second, not yet published, I describe the ways in which the Muslim legal scholars
developed a sacred folk geography based on astronomical horizon phenomena for the same
ultimate purpose.10 (I have not included the materials presented in the valuable 2000 study by
Mnica Rius dealing in detail with materials relating to the qibla in al-Andalus and the Maghrib:
the author is currently engaged in preparing an English version of her original Spanish text.11)
We have also gained a good measure of control over medieval Islamic instruments: my
catalogue of them, still in preparation, provides a useful research tool but is a long way from
publication.12 A task for the future is to confront the vast corpus of Islamic manuscripts dealing
with different aspects of instrumentation.13
Around 1990 my assistant Kurt Maier in Frankfurt valiantly typed the text of I and II, but
for all the symbols and formulae, into a computer. The formulae I typed in myself, with im-
posing curly brackets around multi-tiered numerators and denominators. I did not look at the
new text of I and II for close to 10 years. In 1990 I published III as a separate article, and
in 1996 I published V. Taking advantage of a sabbatical in the autumn of 1999 I spent some
weeks on the entire text, inserting all the symbols and formulae (for the second time, for the
first disappeared when I upgraded the software),14 and cross-references and updating the notes
9
King, Mecca-Centred World-Maps, pp. 47-127, is devoted to a survey of solutions to the qibla problem
(see also n. 17).
10
The study listed as King, The Sacred Geography of Islam, is not yet published, but a summary has appeared
in the article Makka. iv. As centre of the world in EI2, repr. in King, Studies, C-X.
11
Rius, La Alquibla en al-Andalus y al-Magrib (in Spanish, with original texts edited in Arabic). An English
version is planned.
12
See King, Medieval Instrument Catalogue, A-C, and also VIII. A table of contents from 1991 is to be
found at www.uni-frankfurt.de/fb13/ign/instrument-catalogue.html.
13
See already Sdillot-pre and fils, al-Marrkush, for a much-neglected medieval encyclopaedic work on
instrumentation, and now Charette, Mamluk Instrumentation, on a remarkable 14th-century treatise describing
over 100 instrument-types, and idem & King, The Universal Astrolabe of Ibn al-Sarrj: Innovation in Medieval
Islamic Science (forthcoming), on the most spectacular Islamic astronomical instrument also from the 14th
century and the associated texts.
14
It was nave of me to think that formulae of all things would survive successive versions of software over
several years. But I could not predict in advance the other computer problems I would have. Since many readers
with a sense of humour have enjoyed my previous horror stories about computers (Mecca-Centred World-Maps,
p. xxi, and Ciphers of the Monks, p. 19), I am confident that the following tales will also amuse them.
Many of my early problems were caused mainly by the fact that MSWord cannot handle large texts. Ever
cautious, I had the text of this book saved in three separate files. I wonder how many modern people know
that with software available in the mid 1990s it was not possible to insert more than a few dozen automatic
cross-references in a text of such size before the computer would freeze: the other few hundred cross-references
had to be inserted manually and fixed up at the stage of printing. Or that the English and German versions of
MSWord 6 are not compatible and switching computers with different versions during a project can lead to
unexpected consequences. But my favorite story this time round is what happened to my iMac bought in Germany
in early 2000 with an English system and German software: over the months, the thing invariably froze at some
stage of my work. I learned to live with this, not least because the same thing happens with my iBook purchased
in the US. But when the time before the freezing of the iMac became intolerably short, I sought assistance from
Apple, only to learn to my horror that I had purchased a computer without a serial number, one of that kind
which fell off some truck. So it was not guaranteed and Apple was not interested to help me with an unregistered
computer. But my dealer had forgotten to send me a bill for the computer anyway, and his bill arrived only
in December, 2000, in the very same week that I found out that the computer was not for real. I ignored the
bill, and he ignored my telephone call to tell him that there was a severe problem with the computer. If and
when his lawyer contacts me I shall send him this footnote.
The Frankfurt-based company Mac Org, the new Apple dealers in Frankfurt and the best I have ever worked
with, kindly reinserted a new system and a new English version of Word 98. But the computer crashed upon
preface xiii
opening (fortunately before I took it home), and it was ascertained that this English version of Word was
defective, so I had to be content with a German version. At least this did not cause the computer to crash upon
opening, but it had the annoying habit of always beginning spelling checks whenever files were opened, and
then freezing. When one is dealing with hundreds of pages of coded text (such as a$f@$@ to produce oblique
ascensions as ) one does not need to be told that there are too many misspelled words for Word to control.
It took me two days to turn off the spelling check (there are two places where this can be done, and the more
obvious of the two was jammed because defective). Also the version I acquired had no subprogram for inserting
formulae at all. Since my formulae were by now all one-liners (see the text to n. I-1:38) I conceded myself
a victory against the system. It was about the same time that my Epson 760 Color Printer developed the singularly
annoying habit of printing only the top half of the page for about one-third of every assignment.
In January 2001, some two weeks after converting the text to Word 95, I discovered that the three files on
which the book was preserved had self-destructed: the texts of all three files were punctuated by large chunks
of gibberish. I had to go back to the last backups on MSWord 6 and try to remember what changes I had made
... ... .
In August 2001, when both my iBook and iMac were defunct (constantly crashing) and I was spending some
weeks in France, I ordered an English iMac with English keyboard from Apple in Ireland to be delivered in
France (delivery within three days). Apple kindly sent me a French iMac with French keyboard, which I could
have purchased locally had I wanted. It took a total of four weeks and several phone-calls a day, each time to
different people of divers nationalities, each talking a multiplicity of languages, and most of them in Cork,
Ireland, to get rid of the French computer and, only thereafter, to receive what I had ordered: the real English
Apple arrived (from the Czech Republic) a week before I left France to return to Frankfurt. But just before I
left, a French friend who is an Apple wizard came over to organize for me all new installations: he promptly
changed the system to a more recent French one, also installing a French version of MSWord 2000. This
combination actually seems to work, although frankly I prefer MSWord 5 because it was adequate for all sensible
purposes and it did what I wanted, not vice versa. My Epson printer by this time had developed the habit of
registering almost-new ink-cartridges as empty, and thirsting for new cartridges before it would print anything.
Nevertheless I was able to squeeze the entire text out of it before returning to Frankfurt for the Winter Semester.
The final version of the book submitted to E. J. Brill in the Summer of 2002 was somehow achieved under
the above-mentioned circumstances. But their printers could not receive the text by email, nor could they open
the superdisk I sent them containing the text in four chunks. Then we found to our joint surprise that the powers
that be do not make superdisk readers anymore, no doubt because like Eurochecks, they were so useful. Only
in April, 2003, could the printers access the text.
My Epson Stylus Color 760 printer died of exhaustion in October, 2003, just when I had finished proof-
reading the text and when I wanted to print a collection of chunks of corrected text and the indexes for Brill.
I spent half a day travelling from my home in rural France to the city of Aubenas, where I purchased a C42
Plus. This turned out to be missing the CD to install it on my computer. The installation package from the Epson
website is unusable. I spent another half day going to Nmes to buy another printer.
A few days later, my iBook expired, taking with it half of the files saved on it, for which the back-ups were
in Frankfurt. The materials for Volume 2 are all somehow retrievable somewhere, and interested readers will
find new stories about this in my next publication.
The various chunks of Arabic text were all prepared in Cairo on a typewriter in the 1970s, then retyped on
a computer in New York in the 1980s using al-Kaatib software. They all had to be retyped in the 1990s and
early 2000s in various parts of the world using hardware unknown to me and software whose use is beyond
me. This was achieved by Michael Kreutz in Bochum (IV), Sally Ragep in Norman, Oklahoma (V), Mnica
Herrera in Frankfurt (VIIb), and Franois Charette in Frankfurt (XIIa). To dwell on the problems we encountered
formatting these texts and making final changes would fill another page: my favourite was when the Arabic
in the footnotes came out backwards. The formidable task of preparing the final printouts from the extremely
volatile computer files was achieved by Mohammed Abu Zayed and Mnica Herrera in Frankfurt.
There were days at the turn of the Millennium when I yearned for the 1970s: at that time I was privileged
to have the impeccable combination of a highly competent Egyptian secretary with two splendid Italian
typewriters for English and Arabic.
15
In 1973 I gave a lecture entitled Medieval Islamic Prayer-Tables at the American University of Beirut.
xiv preface
I had a much more broader respect for manuscripts than I had in my youth, so I have tried
to add as much as is feasible about the provenance and date of the manuscripts. If such
information is not given, the reader may assume that I am no longer prepared to hazard a guess
from memory. I note that most colleagues do not care much about the origin of the manuscripts
(or the provenance of the instruments) they are studying, being happy enough to have a copy
or two from which to work. Nevertheless, the original environments of the manuscripts (and
instruments) are extremely important. This study was much facilitated by the availability of
some first-rate catalogues of Arabic scientific manuscripts, of which those by Wilhelm
Ahlwardt (1893) for Berlin, McGulkin de Slane (1883-95) for Paris, and Henri-Paul-Joseph
Renaud (1941) for El Escorial, are exemplary.
Apart from various studies by Glen Van Brummelen and Franois Charette,16 the only other
contributions to the subject of astronomical timekeeping over the 20-or-so years during which
most of the text of this book (I mean, I and II) lay dormant are a series of derivative publications
of my own mostly surveys and encyclopaedia articles which all contain references to, and
promises concerning a mysterious SATMI. The fact that I and II were not published years
ago, as they should have been, and that III, V, VIa, and VIb, were published without I and
II being accessible, has resulted in the rather cumbersome form in which the present work is
presented. And the reader should also be aware that Franois Charettes Mathematical
Instrumentation in Fourteenth-Century Egypt and Syria, published by E.J. Brill in 2003, could
be considered as an instant supplement to the present work, since it deals with tables as well
as instruments. Indeed, it features numerous tables and instruments that are not represented
here at all. In the last few years, numerous Internet-sites have appeared dealing with Islamic
prayer-times on the one hand, and historical instruments on the other: as far as I am aware,
none add anything to the historical materials contained in this book. They do, however, offer
the means to generate prayer-times for any date and any locality, and to purchase copies of
medieval instruments. To Franois Charette too go my heart-felt thanks for numerous
corrections to the penultimate version of this text, and no less for preparing all of the computer
graphics.
None of my research over the last 20 years would have been possible without the infra-
structure of the Institute for the History of Science in Frankfurt. I have to thank our Institutes
librarian, Ryszard Dyga, and my CEO, Wolf-Dieter Wagner, for countless favours over many
years, not least with regard to the acquisition of research materials letters requesting this
or that, library searches for this or that, ordering photos of this, preparing photocopies of that,
... ... . Inevitably, this often had to be done in a hurry so that I could disappear for weeks on
end and work in isolation, ... ... .
Most of the photographs used to illustrate the two volumes of this book were kindly provided
by the Deutsche Staatsbibliothek in Berlin; the Egyptian National Library in Cairo; the Chester
Beatty Library in Dublin; the Museo di Storia della Scienza in Florence; the National Maritime
A young lady walked out a couple of minutes after I started: I later found out that she was interested in medieval
Islamic furniture.
16
See Van Brummelen, al-Khalls Auxiliary Tables; and Charette, Najm al-Dns Monumental Table,
and idem, Mamluk Instrumentation.
preface xv
Museum in Greenwich; the Sleymaniye and Topkap Libraries in Istanbul; the British Museum
and the British Library in London; the Bodleian Library and the Museum of History of Science
in Oxford; the Bibliothque Nationale de France in Paris; and Princeton University Library.
It is a pleasure to record my gratitude to these splendid institutions. I also wish to thank Trudy
Kamperveen and Tanja Cowall of Brill Academic Publications for their patience and under-
standing, and especially Ciska Palm of Palm Produkties, who somehow produced this book
from a few computer files and a pile of photographs.
Whilst the materials presented here will be new to most Western readers, they will be
completely new to Muslim readers unfamiliar with Western writings on the history of Islamic
science.17 (Alas virtually nothing of our research over the past 50 years is known in the Islamic
world outside Turkey, Iran and India; neither is it ever mentioned by those Muslim savants
who write in the West on the new Islamic science.) There is nothing on the subject of
medieval Islamic astronomical timekeeping available for the modern Arabic reader other than
a short article of mine in a general work on Islamic science (Maws{at ta}rkh al-{ulm al-
{arabiyya) published in Beirut in 1997 and not widely available in the Muslim world anyway.
In any case, no Muslim readers should be surprised that a study based on over 500 medieval
manuscripts that no-one in modern times had ever looked at previously, should contain
materials that are new and exciting, and which radically alter I venture to say, improve
our picture of science in the medieval Islamic world.
17
My book on the qibla (n. 9) was generously sponsored by the al-Furqan Islamic Heritage Foundation in
London. The present work could not be sponsored by the Foundation because a colleague from C.N.R.S., Paris,
reported to them that it contained not enough new material to merit their support. No other foundation or
interested individual could be found to sponsor the publication, hence this book is somewhat more expensive
than my book on the qibla, which was likewise mainly based on materials that had never been identified or
studied previously.
xvi preface
Some out-dated notions wide-spread amongst the informed public and even amongst
historians of science are that:
(1) The Muslims were fortunate enough to be the heirs to the sciences of Antiquity.
(2) They cultivated these sciences for a few centuries but never really achieved much that
was original.
(3) They provided, mainly in Islamic Spain, a milieu in which eager Europeans emerging
out of the Dark Ages could benefit from these Ancient Greek sciences once they had learned
how to translate them from Arabic into Latin.
Islamic science, therefore, one might argue, is of no consequence per se for the development
of global science and is important only insofar as it marks a rather obscure interlude between
a more sophisticated Antiquity and a Europe that later became more civilized.
What happened in fact was something rather different. The Muslims did indeed inherit the
sciences of Greek, Indian and Persian Antiquity. But within a few decades they had created
out of this potpourri a new science, now written in Arabic and replete with new Muslim
contributions, which flourished with innovations until the 15th century and continued thereafter
without any further innovations of consequence until the 19th. Most of this activity took place
in the Islamic East. Not much of this Islamic science was known in al-Andalus, and even less
of it was available for any Europeans, however eager. In astronomy, the Europeans rescued
from al-Andalus outdated and mutually inconsistent works such as the astronomical handbooks
of al-Khwrzm and al-Battn (respectively representing the first Arabic redactions of the
Indo-Persian and Greek traditions) and produced therefrom the unhappy mixture known as the
Toledan Tables, influential all over Europe for a few centuries, including England. Indeed when
I was a graduate student some 30 years ago I read a 1952 paper by Otto Neugebauer and Olaf
Schmidt entitled Indian Astronomy at Newminster in 1428. The main tables discussed were
some tables of the tangent of the solar declination multiplied by a very curious coefficient.
They were very much out of place in medieval England (see now I-7.1.1 and 10.1).
In 1997 my colleague, the eminent British historian of medieval science, Charles Burnett,
published a book entitled The Introduction of Arabic Learning into England. It is an excellent
book, not least because it deals with a subject on which Burnett is more qualified to write than
anyone else. I take issue only with the title, my problem being the expression Arabic learning
in the context of the ultimate destination, medieval England. Since, as it happens, only a
modest fraction of the grand total of Muslim scholarship was available in al-Andalus, and only
a very small fraction of that ever made it to England in the Middle Ages, Burnetts title might
be construed as a little misleading: in fact, apparently no Englishman in the Middle Ages ever
came into contact with the full force of Muslim scholarship or Arabic learning, for which
he would have had to go to Baghdad or Cairo or Rayy, rather than, say, Toledo. When English
scholars looked at Arabic scientific manuscripts in Oxford in the 17th century they were in for
preface 2 xvii
a few surprises: see Gerald Toomers Eastern Wisedome and Learning: The Study of Arabic
in Seventeenth-Century England (1996) and Gl Russells Arabick Interest in 17th-Century
England (1994). But by that time, it was already a bit late for actually exploiting the scientific
bounty; the materials discovered tended to be more of historical interest. The real surprises
came during the 19th and 20th centuries when historians from a multiplicity of national
backgrounds investigated Islamic scientific manuscripts in libraries all over Europe and then
in the Near East. Their investigations revealed a intellectual tradition of proportions that no
medieval or Renaissance European could ever have imagined: anyone who might doubt this
should look at the monumental bio-bibliographical writings of Heinrich Suter, Carl Brockel-
mann and Fuat Sezgin,1 and, in the case of astronomy, also the early studies of J.-J. and L.-
A. Sdillot, C. A. Nallino, H. Suter or E. S. Kennedy. These investigations are by no means
over, and there is much more material to be unearthed in manuscript libraries and museums
around the world. When investigating manuscripts that nobody has looked at for centuries it
highly likely that one will unearth materials that call for a complete re-evaluation of some
aspects of Muslim activity in a particular branch of knowledge. The task of the historian is
to organize and present these materials. The task of the reader is to be prepared for a few
surprises.
The present study introduces some of the materials from a very special trend in the Muslim
scientific heritage. These materials have become known only in the past 30 years. Indeed, it
has been my priviledge to look at most of them for the very first time. Inevitably, they modify
the overall picture we have of Islamic science. And it so happens that the particular intellectual
activity that inspired these materials is related to the religious obligation to pray at specific
times. The material presented here makes nonsense of the popular modern notion that religion
inevitably impedes scientific progress, for in this case, the requirements of the former actually
inspired the progress of the latter for centuries. But the same material also enables us to witness
the eventual drying up of initiative in Islamic science.2 The reason for this phenomenon is
seldom ever mentioned by historians of science who write on the decline of the scientific
tradition in the Islamic world: within the medieval context, all of the problems had been solved,
some many times over. Nowhere is this clearer than in the kind of sources that I describe in
this book.
1
As I was reading the proofs of this book, a monumental new publication Boris A. Rosenfeld and
Ekmeleddin ~hsanolu, Mathematicians, Astronomers and Other Scholars of Islamic Civilisation and their Works
(7th-19th C.), Istanbul, 2003 appeared. This is an English version of Galina P. Matvievskaya and Boris A.
Rosenfeld, Mathematicians and Astronomers of the Islamic Middle Ages and Their Works, Moscow, 1983 (three
volumes in Russian), supplemented with information from Ekmeleddin ~hsanolu et al., Ottoman Astronomical
/ Mathematical / Geographical Literature, Istanbul, 1997-2000 (altogether six volumes in Turkish). This new
work is a mine of bio-bibliographical information on over 1,700 Muslim scientists and is of course an important
contribution to our field. It is a useful addition to the standard bio-bibliographical works of Suter, Brockelmann
and Sezgin, which should still be used alongside it. In the present volumes I have not included references to
this new work.
2
A useful essay by a scholar who is familiar with the scope and depth of Islamic science is Al-Hassan, Decline
of Islamic Science (1994). It is significant that this essay deals with the decline that took place in and after
the 16th century, not least because the available sources available now to us, but not known to our predecessors
or not previously studied show that scientific initiative had not dried up before that time.
xviii preface 2
Parts of this work have been published previously. They are included here because they are
all essential to the overall picture provided by the ensemble of studies. They have been revised
and brought into line with the other studies; in particular the references and bibliography have
been updated.
Volume 1:
General remarks:
Parts I, II, IV, VIIc and IX are published here for the first time.
Part VIII is here published in English for the first time.
The reader should please keep in mind that Parts I-II were written almost 30 years ago,
Parts IV and VIIb some 20 years ago, Parts III and VI some 10 years ago. Parts VIIc
and IX were written just a few months before this volume went to press.
Volume 2:
General remarks:
Parts XI, XIIa-b, XIIIa and XV have not been published previously.
Parts X and XIIIc are here published in English for the first time.
Notes: The reader will find here the main modern sources for the history of Islamic astronomy
and astronomical instrumentation, including conference proceedings and exhibition catalogues.
Some additional sources on instruments will be included in Vol. 2. Many studies have been
reprinted in various collected volumes, notably those published by Variorum; those individual
studies related to astronomical timekeeping and related aspects of mathematical astronomy
which are particularly relevant to this study are also listed here. The abbreviations introduced
below are used throughout this book. Further bibliographical lists are given in VIII (comparison
of aspects of Islamic and medieval European science) and X (in-strumentation), and the works
cited there are listed below without abbreviated titles. Earlier publications reprinted in this
volume are indicated by diamonds. Detailed references to other works cited which are not
directly related to our main subject are presented ad loc.
Aaboe Festschrift: From Ancient Omens to Statistical Mechanics: Essays on the Exact Sciences Presented to
Asger Aaboe, J. Lennart Berggren and Bernard R. Goldstein, eds., Acta Historica Scientiarum Naturalium
et Medicinalium (Copenhagen), 39 (1987).
Ackermann, Instruments of Humphrey Cole: Silke Ackermann, Scientific Instruments, in eadem, ed.,
Humphrey Cole, pp. 29-95.
, London BM Quadrans vetus: see EPACT.
, Sloane Astrolabe: see EPACT.
, Umarbeitung eines Astrolabs: eadem, Mutabor: Die Umarbeitung eines mittelalterlichen Astrolabs im
17. Jahrhundert, in Frankfurt IGN Festband, pp. 193-209.
, Zinner Archive: eadem, Dormant Treasures The Zinner-Archive at Frankfurt University, Nuncius
Annali di Storia della Scienza (Florence) 16:2 (2001), pp. 711-722.
, ed., Humphrey Cole: eadem, ed., Humphrey Cole: Mint, Measurement and Maps in Elizabethan England,
(British Museum, Occasional Paper, no. 126), London: The British Museum, 1998.
& Cherry, Three Medieval English Quadrants: eadem and John Cherry, Richard II, John Holland and
Three Medieval Quadrants, AS 56 (1999), pp. 3-23.
Advar, Adnan, La science chez les Turcs Ottomans, Paris: G. P. Maisonneuve, 1939.
AHES: Archive for History of Exact Sciences.
@gh Buzurg, al-Dhar{a: @gh Buzurg Muammad Musin al-Tihrn, al-Dhar{a il tanf al-Sh{a, 25 vols.,
Najaf, later Tehran, 1965-1978.
Ahlwardt: see Berlin Catalogue.
AIHS: Archives Internationales dHistoire des Sciences.
AIOS: Arabische Instrumente in orientalistischen Studien, Fuat Sezgin et al, eds., 6 vols., Frankfurt am Main:
Institut fr Geschichte der Arabisch-Islamischen Wissenschaften, 1990-91, repr. as Islamic Astronomy and
Mathematics, vols. 85-90 (1998), with 6 further volumes ibid., vols. 91-96 (1998). [Reprints of studies from
the 19th and early 20th century: a monumental total of over 5,000 pages.]
Aleppo 1976 Symposium Proceedings: Proceedings of the First International Symposium for the History of
Arabic Science, Ahmad Y. al-Hassan et al., eds., 2 vols., Aleppo: Institute for the History of Arabic Science,
1978. Contains:
D. A. King Astronomical Timekeeping in Fourteenth-Century Syria, in II, pp. 75-84, repr. in idem, Studies, A-X.
Fritz W. Zimmermann, The Dustr al-Munajjimn of MS Paris, BN ar. no. 5968, in II, pp. 184-192.
Algiers BN Catalogue: E. Fagnan, Catalogue gnral des manuscrits de la Bibliothque Nationale dAlgrie
... , first published Paris, 1893, 2nd edn., Algiers: Bibliothque Nationale dAlgrie, 1995.
Ali: see al-Brn, Geography.
al-{@mil, A{yn al-Sh{a: Musin al-Amn al-usayn al-{@mil, A{yn al-Sh{a, 56 vols., Damascus: Maba{at
Ibn Zaydn, 1935-1962.
bibliography and bibliographical abbreviations xxi
Amsterdam NK 1990 Exhibition Catalogue: Anthony J. Turner, ed., time, Amsterdam: Tijd voor Tijd Foundation,
1990. [Catalogue of an exhibition at the Nieuwe Kirk, Amsterdam, in 1990.]
Anagnostakis, Astrolabe Ecliptic Divisions: Christopher Anagnostakis, How to Divide the Ecliptic on an
Astrolabe, in Aaboe Festschrift, pp. 133-144.
Analema Boletn de la Asociacin de Amigos de los Relojes de Sol (Madrid).
Ankara (ADTCF) Dergisi and Yaynlari: Ankara niversitesi Dil ve Tarih-Corafya Fakltesi Dergisi and
Yaynlar.
Ansari & Ghori, Two Treatises on Astronomical Instruments: S. M. Raza Ansari and S. A. Khan Ghori, Two
Treatises on Astronomical Instruments by {Abd al-Mun{im al-{@mil and Qsim {Al al-Qyin, Delhi 1985
Colloquium Proceedings, pp. 215-225.
Anthiaume & Sottas, Quadrans novus: A. Anthiaume and J. Sottas, Lastrolabe-quadrant du Muse des
antiquits de Rouen recherches sur les connaissances mathmatiques, astronomiques et nautiques au moyen
ge, Paris: Librairie G. Thomas, 1910, repr. Paris: Librairie Alain Brieux, 1988.
Apian: see Rttel.
Aratrma: Aratrma, published by Ankara niversitesi Dil ve Tarih-Corafya Fakltesi, Felsefe Aratrmalar
Enstits.
Arcari & Costamagna, Acquafredda Sundial, A-B: Barbara Arcari and Andrea Costamagna, Lorologio a ore
canoniche dellAcquafredda, LAstronomia (Rome) 211 (iuglio 2000), pp. 44-50 (A); and eidem, Il
quadrante solare dellabbazia dellAcquafredda, in eidem, Labbazia dellAcquafredda , Lenno (Como),
(Comune di Lenno, Biblioteca Comunale Vittorio Antonini, Quaderno n. 3), 2002, pp. 115-121, and fig. 18
(B).
Archinard, Cadrans solaires rectilignes: Margarida Archinard, Les cadrans solaires rectilignes, Nuncius
(Florence) 3:2 (1988), pp. 149-181.
, Geneva Navicula: eadem, Navicula de Venetiis : une acquisition prestigieuse du Muse dhistoire des
sciences [Genve], Genava, N.S., 43 (1995), pp. 87-94, published separately with the same pagination,
Geneva: Muse dhistoire des sciences, n.d. [1995].
, Unequal Hour Diagram: eadem, The Diagram of Unequal Hours, AS 47 (1990), pp. 173-190.
Arnaldi & Schaldach, Roman Cylinder Dial: Mario Arnaldi and Karl-Heinz Schaldach, A Roman Cylinder
Dial: Witness to a Forgotten Tradition, JHA 28 (1997), pp. 107-117.
Arvanitaki, Lheure arabe: G. Arvanitaki, Thorie de lheure arabe, Memoires de lInstitut gyptien 8 (1915),
pp. 21-55.
AS: Annals of Science.
al-Aba, al-Yawqt: References in IV are to MS Cairo Dr al-Kutub mqt 948,1, fols. 1v-53r, of al-Abas
Kitb al-Yawqt f {ilm al-mawqt.
ASP: Arabic Science and Philosophy (Cambridge University Press).
Assemani, Globus arabicus: Simon Assemani, Globus caelestis cufico-arabicus Veliterni Musei Borgiani, Padua,
1790, repr. in IMA, vol. 91 (1998), pp. 63-305, plus fold-outs.
Astrolabica: Astrolabica (Paris: Socit Internationale de lAstrolabe). [See especially no. 5: Astrolabica No 5:
Etudes 1987-1989, Anthony J. Turner, ed., Paris: Institut du Monde Arabe & Socit Internationale de
lAstrolabe, 1989.]
Atkinson, Points of the Compass: T. D. Atkinson, Points of the Compass, in ERE, X, pp. 73-88.
A{yn al-Sh{a: see al-{@mil, A{yn al-Sh{a.
Azzawi, History of Astronomy in Iraq: {Abbs al-{Azzw, Ta}rkh {ilm al-falak fi l-{Irq wa-{alqtihi bi-l-
aqr al-islmiyya wa-l-{arabiyya fi l-{uhd al-tliya li-ayym al-{Abbsiyyn min sanat 656 H = 1258 M
il sanat 1335 H = 1917 M [= A History of Astronomy in Iraq and its relations with Islamic and Arab regions
in the period following the Abbasids, from 1258 A.D. to 1917 A.D.], Baghdad: Iraqi Scientific Academy (al-
Majma{ al-{ilm al-{Irq, 1958. [This Arabic work is independent of any earlier European bio-bibliographical
works on the subject.]
Bagheri, al-Ksh on the Visible Horizon: Mohammad Bagheri, A New Treatise by al-Ksh on the Depression
of the Visible Horizon, in Bellagio 2000 Conference Proceedings, pp. 357-368.
Bagvi, M. B. A. Determination of the Direction of Qibla and the Islamic Timings, Lahore: Sh. Muhammed Ashraf,
1972.
Baloch, Measurement of Space and Time in Lower Sind: N. A. Baloch, Measurement of Space and Time
in the Lower Indus Valley of Sind, in Islamabad 1979 Conference Proceedings, pp. 168-196.
Barcel: see Labarta & Barcel.
Barmore, Turkish Mosque Orientation: Frank E. Barmore, Turkish Mosque Orientation and the Secular
Variation of the Magnetic Declination, Journal of Near Eastern Studies 44 (1985), pp. 81-98.
Barthold, Orientierung der ersten Moscheen: Wilhelm Barthold, Die Orientierung der ersten muhammedani-
schen Moscheen, Der Islam 18 (1929), pp. 245-250.
xxii bibliography and bibliographical abbreviations
Bashear, Qibla Musharriqa: Suliman Bashear, Qibla Musharriqa and Early Muslim Prayer in Churches, The
Muslim World 81 (1991), pp. 267-282.
BEA: Biographical Encyclopedia of Astronomers, Thomas Hockey, ed., to be published by Kluwer Academic
Publishers, Dordrecht.
Bennett, Oxford Navicula: see EPACT.
Beer, Astronomical Dating of Works of Art: Arthur Beer, Astronomical Dating of Works of Art, Vistas
in Astronomy 9 (1967), pp. 177-223.
Bellagio 2000 Conference Proceedings: From China to Paris: 2000 Years Transmission of Mathematical Ideas,
(Boethius Texte und Abhandlungen zur Geschichte der Mathematik und der Naturwissenschaften, Band
46), Yvonne Dold-Samplonius, Joseph W. Dauben, Menso Folkerts and Benno van Dalen, eds., Stuttgart:
Franz Steiner, 2002. [Proceedings of a conference held at the Rockefeller Foundationss Research and
Conference Center, Bellagio, during May, 2000.]
Benoit, Cadran solaire monastique: J. L. Benoit, Un cadran solaire monastique du XIIe sicle, Le Cosmos
59/1238 (1908), pp. 435-437.
Berggren, Archimedes among the Ottomans: J. Lennart Berggren, Archimedes among the Ottomans, in
Aaboe Festschrift, pp. 101-109.
, Al-Brn on Plane Maps of the Sphere: idem, Al-Brn on Plane Maps of the Sphere, JHAS 6 (1982),
pp. 47-96.
, Four Analemmas for the Qibla: idem, A Comparison of Four Analemmas for Determining the Azimuth
of the Qibla, JHAS 4 (1980), pp. 69-80.
, abashs Analemma for Azimuth Curves: idem, abashs Analemma for Representing Azimuth Circles
on the Astrolabe, ZGAIW 7 (1991/92), pp. 23-30.
, Islamic Mathematics: idem, Episodes in the Mathematics of Medieval Islam, New York, etc.: Springer, 1986.
, al-Kh on Astrolabe Construction: idem, Ab Sahl al-Khs Treatise on the Construction of the
Astrolabe with Proof: Text, Translation and Commentary, Physis 31 (1994), pp. 141-252.
, Origins of the Method of the Zjes: idem, The Origins of al-Brns Method of the Zjes in the Theory
of Sundials, Centaurus 28 (1985), pp. 1-16.
, Spherical Trigonometry in the Jmi{ Zj: idem, Spherical Trigonometry in Kshyr ibn Labbns Jmi{
Zj, in Kennedy Festschrift, pp. 15-33.
Berggren: see also Aaboe Festschrift.
Bergmann, Innovationen im Quadrivium: Werner Bergmann, Innovationen im Quadrivium des 10. und 11.
Jahrhunderts: Studien zur Einfhrung von Astrolab und Abakus im lateinischen Mittelalter, Stuttgart: Franz
Steiner, 1985.
Berlin Catalogue: Wilhelm Ahlwardt, Die Handschriften-Verzeichnisse der Kniglichen Bibliothek zu Berlin,
5. Band: Verzeichnis der arabischen Handschriften, Berlin: A. Asher, 1893.
Biarne, Le temps du moine: J. Biarne, Le temps du moine daprs les premires rgles monastiques dOccident
(IVe-VIe sicles), in Le temps chrtien de la fin de lAntiquit au Moyen-Age, IIIe-XIIIe sicles, (Colloques
Internationaux du Centre National de la Recherche Scientifique, no. 604), Paris, 1984, pp. 99-128. [Not
consulted.]
Bielefeld Catalogue: Wissenschaftliche Instrumente und Sonnenuhren Kunstgewerbesammlung der Stadt
Bielefeld / Stiftung Huelsmann, Dirk Syndram, ed., Munich: Callwey, 1989.
Bilfinger, Die mittelalterlichen Horen: Gustav Bilfinger, Die mittelalterlichen Horen und die modernen Stunden
Ein Beitrag zur Kulturgeschichte, Stuttgart: W. Kohlhammer, 1892.
al-Brn, Astrolabe Construction: al-Brn, Ist{b al-wujh al-mumkina f an{at al-aurlb, al-Sayyid
Muammad Akbar Jawd al-usayn, ed., with introductory notes by Ayatallh asan asan Zde al-@mul,
Mashhad: Islamic Research Foundation, Astan Quds Razavi, 1380 HS [= 2001/02].
, Astrology: R. Ramsay Wright, transl., The Book of Instruction in the Elements of the Art of Astrology by
... al-Brn, London: Luzac & Co., 1934, repr. Baghdad: al-Muthann, n.d.
, Chronology: C. Edward Sachau, Chronologie orientalischer Vlker von Alberuni, Leipzig: F. A. Brockhaus
& Otto Harrassowitz, 1923, repr. Baghdad: Maktabat al-Muthann, n.d., translated as C. E. Sachau, The
Chronology of Ancient Nations An English Version of the Arabic Text of the Athr-ul-bkiya of Albrn,
or Vestiges of the Past, collected and reduced to writing by the author in A.H. 390-1, A.D. 1000, London,
1879, repr. Frankfurt am Main: Minerva, 1984.
, India: E. C. Sachau, Alberunis India, Delhi, etc.: S. Chand & Co., 1964 reprint.
, Maqld: Marie-Thrse Debarnot, ed. and trans., al-Brn: Kitb Maqld {ilm al-hay}a. La trigonomtrie
sphrique chez les Arabes de lest la fin du Xe sicle, Damascus: Institut Franais de Damas (Publication
no. 114), 1985.
, Qnn: al-Brn, al-Qnn al-Mas{d, [Max Krause, ed.], 3 vols., Hyderabad, Deccan: Osmania Oriental
Publications, 1954-1956.
bibliography and bibliographical abbreviations xxiii
, Shadows: al-Brn, Ifrd al-maql f amr al-ill, no. 2 in Ras}il al-Brn, Hyderabad: Osmania Oriental
Publications, 1948; with translation and commentary in Edward Stewart Kennedy, The Exhaustive Treatise
on Shadows by Ab al-Rayn Muammad b. Amad al-Brn, 2 vols., Aleppo: Institute for the History
of Arabic Science, 1976.
, Tadd: al-Brn, Kitb Tadd nihyt al-amkin, ed. P. Bulgakov, Majallat Ma{had al-makht al-
{arabiyya (Cairo) 8 (1962). [I have not seen the Ankara 1962 edn. by Muammad ibn Twt al-anj.]
Translation in Jamil Ali, The Determination of the Coordinates of Cities ... by al-Brn, Beirut: American
University of Beirut Press, 1967. Commentary in Edward Stewart Kennedy, A Commentary upon Brns
Kitb Tadd [nihyt] al-amkin, An 11th Century Treatise on Mathematical Geography, Beirut: American
University of Beirut Press, 1973.
Blachre et al., Dictionnaire Arabe-Franais-Anglais: Rgis Blachre et al., Dictionnaire Arabe-Franais-
Anglais, 3 vols. to date, Paris: G.-P. Maisonneuve et Larose, 1967 to present.
Bonine, The Morphogenesis of Iranian Cities: Michael E. Bonine, The Morphogenesis of Iranian Cities,
Annals of the Association of American Geographers 69 (1979), pp. 208-224.
, Sacred Direction and City Structure in Morocco: idem, The Sacred Direction and City Structure: A
Preliminary Analysis of the Islamic Cities of Morocco, Muqarnas: An Annual on Islamic Art and
Architecture 7 (1990), pp. 50-72.
Borst, Astrolab und Klosterreform: Arno Borst, Astrolab und Klosterreform an der Jahrtausendwende,
Heidelberg: Carl Winter, 1989. [Problematic.]
Bosworth, Islamic Dynasties: C. Edmund Bosworth, The New Islamic Dynasties A chronological and
genealogical manual, New York: Columbia University Press, 1996.
Bremner, Shadow Table in Mul.Apin: Robert W. Bremner, The Shadow Length Table in Mul.Apin, in Graz
1991 Conference Proceedings, pp. 367-382.
Brentjes, The Republic of Letters in the Middle East: Sonja Brentjes, The Interests of the Republic of Letters
in the Middle East, 1550-1700, Science in Context 12:3 (1999), pp. 435-468.
Brentjes: see also Frankfurt 2001 Conference Proceedings.
Brice, Atlas: Brice, William C., ed., An Historical Atlas of Islam, Leiden: E. J. Brill, 1981.
Brice & Imber & Lorch, Equatorial Semicircle: William Brice, Colin Imber and Richard P. Lorch, The D}ire-
yi Mu{addel of Seyd {Al Re}s, Seminar on Early Islamic Science (University of Manchester) 1 (1976).
[Reviewed in King, Studies, B-XIII.]
Brieux, Authenticit des astrolabes: Alain Brieux, Les astrolabes: tests dauthenticit, Art et curiosit, sept.
1974 (unpaginated).
Brieux & Maddison, Rpertoire: Alain Brieux and Francis R. Maddison, Rpertoire des facteurs dastrolabes
et de leurs uvres, in press (Paris: C.N.R.S.).
Brockelmann, GAL: Carl Brockelmann, Geschichte der arabischen Litteratur, 2 vols., 2nd edn., Leiden: E. J.
Brill, 1943-49, and Supplementbnde, 3 vols., Leiden: E. J. Brill, 1937-42.
Bruin, al-Brn Newsletter: A series of 38 mimeographed articles prepared between 1966 and 1971 by the late
Professor Frans Bruin, then of the American University of Beirut. [For a complete list see the introduction
to XIIa. Copies are available at the Department of History of Science and Medicine at Yale University; the
University Library at the American University of Beirut; and the Institutes for History of Science in Frankfurt
and Hamburg.] In this book the following are cited:
The Astronomical Observatory of Ulugh Beg in Samarqand, al-Brn Newsletter 9 (1967).
The Observational Instruments of al-{Ur ... , ibid. 10 (1968).
The Astronomical Observatory of Nar al-Dn al-s, ibid. 13 (1968).
The Construction of Instruments ... by ... al-{@mil ... , ibid. 15 (1968).
The Fakhr Sextant in Rayy, ibid. 19 (1969).
The Instruments of Taq al-Dn ... , ibid. 29 (1970).
A Treatise on Small Instruments by ... al-Khzin ... , ibid. 30 (1970).
The Planispheric Astrolabe: Its History, Theory, and Use, ibid. 33, 34 and 35 (1970).
Brusa, Navicelle: Giuseppe Brusa, Le navicelle orarie di Venezia, Annali dellIstituto e Museo di Storia
della Scienza di Firenze 5 (1980), pp. 51-59.
Brussels 1984 Exhibition Catalogue: La mesure du temps dans les collections Belges, Brussels: Socit Gnrale
de Banque, 1984. [Catalogue of an exhibition at the Socit Gnrale de Banque during 26.1.-7.4.1984.]
Brussels BR 2001 Exhibition Catalogue: Des agrimensores romains aux arpenteurs du XVIe sicle, Jan De
Graeve and Jean Mosselmans, eds., Brussels: UGEB, 2001. [Catalogue of an exhibition organized by the
Union des Gomtres-Experts immobiliers de Bruxelles held at the Bibliothque Royale de Belgique during
22.11-21.12.2001.]
BSIS: Bulletin of the Scientific Instrument Society.
Buchner, Antike Reiseuhren: Edmund Buchner, Antike Reiseuhren, Chiron 1 (1971), pp. 457-482.
, Die Sonnenuhr des Augustus: idem, Die Sonnenuhr des Augustus, Mainz: Philipp von Zabern, 1982.
xxiv bibliography and bibliographical abbreviations
Burnett, {Abd al-Mas of Winchester: Charles Burnett, {Abd al-Mas of Winchester, in North Festschrift,
pp. 159-169.
, Antioch as a Link between Arabic and Latin Culture: idem, Antioch as a Link between Arabic and Latin
Culture in the Twelfth and Thirteenth Centuries, in Occident et Proche-Orient : Contacts scientifiques au
temps des Croisades, Actes du colloqie de Louvain-la-Neuve, 24 et 25 mars 1997, Isabelle Draelants, Anne
Tihon, Baudouin van den Abeele, eds., Turnhout (B): Brepols, 2000, pp. 1-78.
, The Introduction of Arabic Learning into England: idem, The Introduction of Arabic Learning into England,
(The Panizzi Lectures, 1996), London: The British Library, 1997.
, The Translation of Arabic Science into Latin: idem, The Translation of Arabic Science into Latin: A
Case of Alienation of Intellectual Property?, Bulletin of the Royal Institute for Inter-Faith Studies (Amman,
Jordan) 4:1 (2002), pp. 145-157.
Burnett, ed, Adelard of Bath: Adelard of Bath An English Scientist of the Early Twelfth Century, Charles
Burnett, ed., (Warburg Institute Surveys and Texts, XIV), London: The Warburg Institute, University of
London, 1987.
Cairo ENL Catalogue: David A. King, A Catalogue of the Scientific Manuscripts in the Egyptian National
Library (in Arabic), 2 vols., Cairo: General Egyptian Book Organization, 1981 and 1986.
Cairo ENL Survey: idem, A Survey of the Scientific Manuscripts in the Egyptian National Library (in English),
(American Research Center in Egypt, Catalogs, vol. 5), Winona Lake, Ind.: Eisenbrauns, 1987.
Caldo, Astrolabi di Palermo: L. Caldo, Astrolabi del Museo Nazionale di Palermo, Pubblicazione dell
Osservatorio astronomico di Palermo, Memorie, no. 65, 1936, repr. in Coelum 8 (June 1938), pp. 101-108,
and 8 figs. [Not reprinted in AIOS or IMA.]
Calvo, Ibn Bo and his Universal Plate: Emilia Calvo Labarta, Ab {Al al-usayn ibn Bo (m. 716/1316):
Rislat al-afa al-y mi{a li-y am{ al-{ur (Tratado sobre la lmina general para todas las latitudas),
Madrid: Consejo Superior de Investigaciones Cientficas and Instituto de Cooperacin con el Mundo rabe,
1993.
, Ibn Bos Universal Plate: eadem, Ibn Bos Astrolabe [Plate] in the Maghrib and East, in Vernet
Festschrift, II, pp. 755-767.
, Ibn Bos afa Mujayyaba: eadem, Transformation of Coordinates in Ibn Bos al-Risla fi l-afa
al-mujayyaba dht al-awtr, JHAS 12 (2001), pp. 3-21.
, La resolution graphique des questions astrologiques: eadem, La resolution graphique des questions
astrologiques al-Andalus, Actes du 3me Colloque Maghrbin sur lHistoire des Mathmatiques Arabes
Tipaza (Alger, Algrie) 1-3 Dcembre 1990, Algiers: Association Algrienne dHistoire des Math-
matiques, 1998, pp. 31-44.
, The Use of Ibn Bos Universal Plate: eadem, A Study of the Use of Ibn Bos Universal Astrolabe
Plate, AIHS 50 (2000), pp. 264-295.
Cambridge DI 1998 Conference Proceedings: selected papers given at a conference New perspectives on science
in medieval Islam held at the Dibner Institute, Cambridge, Mass., during Nov. 6-8, 1998, to appear.
Cambridge History of Iran, IV-VI: The Cambridge History of Iran, vol. IV: The Period from the Arab Invasion
to the Saljuqs, (Richard N. Frye, ed.), vol. V: The Saljuq and Mongol Periods (John A. Boyle, ed.), and
vol. VI: The Timurid and Safavid Periods (Peter Jackson and Laurence Lockhart, eds.), Cambridge:
Cambridge University Press, 1975, 1968 and 1986. [See Kennedy, Exact Sciences in Iran.]
Cambridge WMHS Catalogue: David J. Bryden, The Whipple Museum of the History of Science, Catalogue 6:
Sundials and Related Instruments, Cambridge: The Whipple Museum of the History of Science, 1988.
Canobbio, Fragment of a Spherical Astrolabe: Ernesto Canobbio, An Important Fragment of a West Islamic
Spherical Astrolabe, Annali dellIstituto e Museo di Storia della Scienza di Firenze 1 (1976), pp. 37-41.
Cantor, Geschichte der Mathematik: Moritz Cantor, Vorlesungen ber die Geschichte der Mathematik 2nd edn.,
4 vols., repr. Leipzig: B. G. Teubner, 1913.
Carandell, Analemma for the Qibla: Joan Carandell, An Analemma for the Determination of the Azimuth
of the Qibla in the Risla f {ilm al-ill of Ibn al-Raqqm, ZGAIW 1 (1984), pp. 61-72.
, Ibn al-Raqqm sobre los cuadrantes solares: idem, Risla f {ilm al-ill de Muammad Ibn al-Raqqm
al-Andalus: edicion, traduccin y comentario, Barcelona: Universidad de Barcelona, Instituto Mills
Vallicrosa de Historia de la Ciencia rabe, 1988.
Carboni, Kitb al-bulhn: Stefano Carboni, Il Kitb al-bulhn di Oxford, (Eurasiatica - Quaderni del Diparti-
mento di Studi Eurasiatici, Universit degli Studi di Venezia, 6), Turin: Tirrenia, 1988.
Carra de Vaux, Lastrolabe linaire: Bernard Carra de Vaux, Lastrolabe linaire ou bton dal-Tousi,
Journal Asiatique, 11. sr., 5 (1895), pp. 464-516, repr. in AIOS, III, pp. 181-233.
Casanova, Cadran solaire syrien: Paul Casanova, La montre du Sultan Nour ad Din (554 de lHgire = 1159-
1160), Syria 4 (1923), pp. 282-299, repr. in AIOS, IV, pp. 242-262.
Casulleras, Andalus Sundial: Josep Casulleras, El ltimo captulo del Kitb al-Asrr f nat}iy al-afkr,
in Vernet Festschrift, II, pp. 613-653.
bibliography and bibliographical abbreviations xxv
Ctedra & Sams, Astrologa de Enrique de Villena: Pedro M. Ctedra (ed. and transl.) with Julio Sams (comm.),
Tratado de astrologa atribuido a Enrique de Villena, Madrid: Ro Tinto Minera, 1980.
Caussin, Table Hakmite: A. P. Caussin de Perceval, Le livre de la grande table Hakmite, in Notices et
extraits des manuscrits de la Bibliothque nationale, 7 (An XII [= 1804]), pp. 16-240, with separate pagination
in the separatum.
Chabs, Astronomy in 15th-Century Salamanca: Jos Chabs, Astronomy in Salamanca in the Mid-Fifteenth
Century: The Tabulae resolutae, JHA 29 (1998), pp. 167-175.
, Cahier dun croisier: idem, Le cahier dastronomie dun croisier du XVe sicle, Nuncius (Florence) 12:1
(1997), pp. 3-16.
& Goldstein, Zacut: idem and Bernard R. Goldstein, Astronomy in the Iberian Peninsula: Abrahan Zacut
and the Transition from Manuscript to Print, (Transactions of the American Philosophical Society 90:2),
Philadelphia, Pa.: 2000.
Chapoutot-Remadi: see Souissi & Chapoutot-Remadi.
Charette, Der geflgelte Quadrant: Franois Charette, Der geflgelte Quadrant: Ein ungewhnliches
Sinusinstrument aus dem 14. Jh., in Saltzer Festschrift, pp. 25-36.
, al-Khwrizm on the Astrolabe: idem with Petra Schmidl, Scientific Initiative in 9th-Century Baghdad:
al-Khwrizm on the Astrolabe and Related Treatises, to appear.
, The Library of Muaf idq: idem, A View of Scientific Life in Mid-Eighteenth Century Istanbul: The
Library of Muaf idq [revised version of a paper presented at the DAVO Annual Congress (German
Middle East Studies Association for Contemporary Research and Documentation), Hamburg, December
1999], forthcoming.
, Mamluk Instrumentation: idem, Mathematical Instrumentation in Fourteenth-Century Egypt and Syria
The Illustrated Treatise of Najm al-Dn al-Mir, Leiden: E. J. Brill, 2003.
, Najm al-Dns Monumental Table: idem, A Monumental Medieval Table for Solving the Problems of
Spherical Astronomy for all Latitudes, AIHS 48 (1998), pp. 11-64.
, Orientalisme et histoire des sciences: lhistoriographie europenne des sciences islamiques et hindoues,
1784-1900, Masters thesis, Dpartment dhistoire, Universit de Montral, 1995.
& King, Survey of Islamic Tables for Astrolabe Construction: idem and David A. King, A Survey of
Medieval Islamic Tables for Astrolabe Construction, to appear.
& King, The Universal Astrolabe of Ibn al-Sarrj: idem and David A. King, The Universal Astrolabe of Ibn
al-Sarrj, to appear.
& Schmidl, Habashs Universal Plate: idem and Petra G. Schmidl, A Universal Plate for Timekeeping
with the Stars by abash al-sib: Text, Translation and Preliminary Commentary, Suhayl 2 (2001), pp.
107-159.
Charette: see also Frankfurt 2001 Conference Proceedings and King & Charette.
Cheikho, al-Mukula: Louis Cheikho (= Shaykh), al-Mukula (in Arabic), al-Machriq [= al-Mashriq]
(Beirut) 10 (1907) [repr. Munich: Kraus-Thomson Organization G.m.b.H., 1980], pp. 75-80, repr. in IMA,
vol. 91 (1998), pp. 370-375.
Chidyaq, Indicateur de qibla: Hana Chidyaq, Un indicateur de qibla sign Muammad {Izz al-abbgh,
Astrolabica 5 (1989), pp. 27-36.
Christies London 08.04.1998 Catalogue: Christies South Kensington Exceptional Scientific and Engineering
Works of Art, Instruments and Models 8 April 1998, London: Christies, 1998.
Colin & Renaud, Ab Miqra{: Georges Colin and Henri-Paul-Joseph Renaud, Note sur le muwaqqit
marocain Ab Muqri{ ou mieux Ab Miqra{ al-Baw (XIIIe s. J.C.), Hespris 25 (1933), pp. 94-96.
Comes, Ecuatorios andaluses and Ecuatorios andaluses: Merc Comes, Ecuatorios andaluses Ibn al-Sam,
al-Zarqlluh y Ab-l-alt, Barcelona: Universidad de Barcelona, Facultad de Filologa, etc., 1991; and a
summary in the chapter Los ecuatorios andaluses in Madrid MAN 1992 Exhibition Catalogue, pp. 75-
87.
The Compendium: The Compendium Journal of the North American Sundial Society.
Cook, Muslim Material on Comets and Meteors: David Cook, A Survey of Muslim Material on Comets and
Meteors, JHA 30 (1999), pp. 131-160.
Cotter, Nautical Tables: Charles H. Cotter, The Development of Nautical Astronomical Inspection Tables
in the Period from 1770 to 1919, Vistas in Astronomy 20 (1976), pp. 245-247.
Crombie, Griechisch-arabische Naturwissenschaften und Abendlndisches Denken: Alistair C., Crombie,
Griechisch-arabische Naturwissenschaften und Abendlndisches Denken, in Gereon Sievernich and
Hendrik Budde, eds., Europa und der Orient 800-1900, Gtersloh and Munich: Bertelsmann Lexikon Verlag
GmbH, 1989, pp. 102-131.
Curtze, Urkunden zur Geschichte der Trigonometrie: Maximilian Curtze, Urkunden zur Geschichte der
Trigonometrie im christlichen Mittelalter, Bibliotheca Mathematica, 3. Folge, 1 (1900), pp. 321-416.
xxvi bibliography and bibliographical abbreviations
Al-Daffa & Stroyls, Studies: Ali A. Al-Daffa and John S. Stroyls, Studies in the Exact Sciences in Medieval
Islam, New York: John Wiley & Sons, 1984. [To be used with caution.]
Dallal, Al-Brn on Climates: Ahmad Dallal, Al-Brn on Climates, AIHS 34(1984), pp. 3-18.
, Ibn al-Haythams Universal Solution for the Qibla: idem, Ibn al-Haythams Universal Solution for Finding
the Direction of the Qibla by Calculation, ASP 5 (1995), pp. 145-193.
, Islamic Science, Medicine, and Technology: idem, Science, Medicine, and Technology The Making
of a Scientific Culture, in Esposito, ed., The Oxford History of Islam, pp. 154-213.
Damascus Catalogue: Ibrahim Khoury, Fihris makht Dr al-Kutub al-hiriyya, vol. VI: Ilm al-haya wa-
mulaqtuhu, Damascus: Majma al-lugha al-Arabiyya, 1389 H [= 1969].
Daniel, Greenwich Navicula: Christopher Daniel, The Fifth Navicula, Clocks (Hemel Hempstead, U.K.) 14
(1992), p. 37. [On the navicula now in Greenwich.]
Davidian, al-Brn on the Time of Day: Davidian, Marie-Louise, al-Brn on the Time of Day from Shadow
Lengths, JAOS 80 (1960), pp. 330-335, repr. in Kennedy et al., Studies, pp. 274-279.
Davidian: see also Kennedy & Davidian.
Debarnot, al-Brns Maqld: see al-Brn, Maqld.
, Islamic Trigonometry: Marie-Thrse Debarnot, Trigonometry, in EHAS, II, pp. 495-538.
, Zj of abash: eadem, The Zj of abash al-sib: A Survey of MS Istanbul Yeni Cami 784/2, in
Kennedy Festschrift, pp. 35-69.
Dekker, Astrolabes and Dates and Dead Ends: Elly Dekker, Of Astrolabes and Dates and Dead Ends, AS
49 (1992), pp. 175-184. [A critical study of Torode, Dating Astrolabes.]
, Cartographic Grids from Iran: eadem, Cartographic Grids from Iran: An Early Version of the Retro-
Azimuthal Orthographic Projection, The Cartographical Journal (The British Cartographic Society,
London) 37:2 (2000), pp. 109-116. [A discussion of the grids on the maps described in King, Mecca-Centred
World-Maps see now VIIc.]
, Quadrans novus: eadem, An Unrecorded Medieval Astrolabe Quadrant from c. 1300, Annals of Science
52 (1995), pp. 1-47.
Delambre, HAMA: Jean-Baptiste Delambre, Histoire de lastronomie du moyen ge, Paris, 1819, repr. New York:
Johnson Reprints, 1965.
Delehar, Illustrations: Peter Delehar, Illustrations of Scientific Instruments in the Gentlemans Magazine
1746-1796, in Turner Festschrift, pp. 383-394.
Delphin, Lastronomie au Maroc: Getan Delphin, Lastronomie au Maroc, Journal Asiatique, 8e ser., 17
(1891), pp. 177-201.
De Smet: see Poulle & De Smet.
Destombes, Astrolabe carolingien: Marcel Destombes, Un astrolabe carolingien et lorigine de nos chiffres
arabes, AIHS 58-59 (1962), pp. 1-45, repr. in idem, Contributions, pp. 153-200.
, Chiffres coufiques: idem, Les chiffres coufiques des instruments astronomiques arabes, Physis 2:3
(1960), pp. 197-210, repr. in idem, Contributions, pp. 127-140.
, Contributions: G. Schilder et al., M. Destombes (1905-1983): Contributions slectionnes lHistoire de
la Cartographie et des Instruments scientifiques, Utrecht: Hes & Paris: A. G. Nizet, 1987.
& Kennedy, al-f on the Astrolabe: idem and Edward Stewart Kennedy, Introduction to Kitb al-{Amal
bil Asurlb [of al-f], repr. in Kennedy et al., Studies, pp. 405-447.
de Vries et al., Hafr and Halazn: Fer J. de Vries, William S. Maddux and Mac Oglesby, Hafr and Halazn,
The Compendium Journal of the North American Sundial Society 6:2 (1999), pp. 23-26.
de Vries et al., Regiomontanus Sundials: Fer J. de Vries, Mac Oglesby, William S. Maddux and Warren Thom,
Regiomontanus, Apian and Capuchin Sundials, The Compendium 5:4 (1998), pp. 13-19.
al-Dhar{a: see @gh Buzurg, al-Dhar{a.
DHaenens, La clepsydre de Villiers: Albert DHaenens, La clepsydre de Villiers (1267) Comment on
mesurait et vivait le temps dans une abbaye cistercienne au XIIIe sicle, in Harry Khnel, ed., Klsterliche
Sachkultur des Sptmittelalters, Vienna: sterreichische Akademie der Wissenschaften, 1980.
dHollander, LAstrolabe: Raymond dHollander, LAstrolabe Histoire, thorie et pratique, Paris: Institut
ocanographique, 1999.
Dicks, Ancient Astronomical Instruments: D. R. Dicks, Ancient Astronomical Instruments, Journal of the
British Astronomical Association 64 (1954), pp. 77-85.
Dizer, Equatorial Semicircle: Muammer Dizer, The d}irat al-mu{addal in Kandilli Observatory ... , JHAS
1 (1977), pp. 257-260.
Dizer: see also Kandilli Instrument Handbook.
Djebbar, Une histoire de la science arabe: Ahmed Djebbar, Une histoire de la science arabe Introduction
la connaissance du patrioine scientifique des pays dIslam, Paris: Seuil, 2001.
DMA: Dictionary of the Middle Ages, 13 vols., New York: Charles Scribners Sons, 1982-89.
Dobrzycki: see Kremer & Dobrzycki.
bibliography and bibliographical abbreviations xxvii
Esposito, ed., The Oxford History of Islam: The Oxford History of Islam, John L. Esposito, ed., Oxford & New
York: Oxford University Press, 1999.
Evans, Ancient Astronomy: James Evans, The History and Practice of Ancient Astronomy, New York & Oxford:
Oxford University Press, 1998.
Fantoni, Orologi solari: Girolamo Fantoni, Orologi solari Trattato completo di gnomonica, Rome: Tech-
nimedia, 1988.
Fehrvri, Quadrant of al-Mizz: Gza Fehrvri, An Eighth/Fourteenth-Century Quadrant of the Astrolabist
al-Mizz, Bulletin of the School of Oriental and African Studies (London) 36 (1973), pp. 115-117 and 2
pls.
Ferrari & Severino, Meridiane islamiche: Gianni Ferrari and Nicola Severino, Appunti per uno studio delle
meridiane islamiche, privately distributed, 1997.
Field & Wright, Gears from the Byzantines: Judith V. Field and Michael T. Wright, Gears from the
Byzantines: A Portable Sundial with Calendrical Gearing, Annals of Science 42 (1985), pp. 87-138, repr.
in eidem and Hill, Byzantine and Arabic Gearing.
& Hill & Wright, Byzantine and Arabic Gearing: eidem and Donald R. Hill, Byzantine and Arabic
Mathematical Gearing, London: The Science Museum, 1985
Fierro & Sams, eds., The Formation of al-Andalus, II: Maribel Fierro and Julio Sams, eds., The Formation
of al-Andalus, Part 2: Language, Religion, Culture and the Sciences, Aldershot, Brookfield, Vt.: Ashgate
Variorum, 1998.
Fletcher et al., Index of Mathematical Tables: A. Fletcher, J. C. P. Miller, L. Rosenhead, and L. J. Comrie, An
Index of Mathematical Tables, 2 vols., Oxford: Blackwell, 1962.
Forcada, Anw} Books in al-Andalus: Miquel Forcada, Books of Anw} in al-Andalus, in Fierro & Sams,
eds., The Formation of al-Andalus, II, pp. 305-327.
, Folk Astronomy of Amad b. Fris: idem, Astrology and Folk Astronomy: the Mukhtaar min al-Anw}
of Amad b. Fris, Suhayl 1 (2000), pp. 107-205.
, Mqt en los calendarios andaluses, idem, Mqt en los calendarios andaluses, al-Qanara 11 (1990),
pp. 59-69.
, The Mukhtaar min al-Anw{ of Amad ibn Fris: A New Andalusian Astronomical Source from the
Fourth/Tenth Century: The Mukhtaar min al-Anw{ of Amad ibn Fris, in Vernet Festschrift, II, 769-
780.
, Les sources andalouses du Calendrier dIbn al-Bann, idem, Les sources andalouses du Calendrier dIbn
al-Bann, in Actas del II Coloquio hispano-maroqu de ciencias histricas Historia, ciencia y sociedad,
Granada, 6-10 nov. 1989, pp. 183-196.
Frank, Zur Geschichte des Astrolabs: Joseph Frank, Zur Geschichte des Astrolabs, Erlangen: Universitts-
Buchdruckerei, 1920, repr. in AIOS, IV, pp. 31-62.
, al-Khwrizm ber das Astrolab: idem, Die Verwendung des Astrolabs nach al-Chwrizm, Abhand-
lungen zur Geschichte der Naturwissenschaften und der Medizin 3 (1922), repr. in AIOS, IV, pp. 125-159,
with a review by Carl Schoy repr. ibid, pp. 160-161.
, Zwei arabische Instrumente: idem, ber zwei astronomische arabische Instrumente, Zeitschrift fr
Instrumentenkunde 41 (1921), pp. 193-200, repr. in AIOS, IV, pp. 63-70.
& Meyerhof, Mogulisches Astrolab: idem and Max Meyerhof, Ein Astrolab aus dem indischen
Mogulreiche, Heidelberger Akten der von-Portheim-Stiftung 13 (Heidelberg: Carl Winter), 1925, reprinted
in AIOS, IV, pp. 307-356.
Frank: see also Wiedemann & Frank.
Frankfurt IGN Festband: Ad radices Festband zum fnfzigjhrigen Bestehen des Instituts fr Geschichte der
Naturwissenschaften Frankfurt am Main, Anton von Gotstedter, ed., Stuttgart: Franz Steiner, 1994.
Frankfurt 2001 Conference Proceedings: Sonja Brentjes, Benno van Dalen and Franois Charette, eds., Certainty,
Doubt, Error: Aspects of the Practice of Pre- and Early Modern Science. ... , in Early Science and Medicine
(E. J. Brill, Leiden) 7:3 (2002), pp. 173-310. [Proceedings of a conference held in Frankfurt am Main during
17-18.11.2001.]
Fck, Arabische Studien: Johann Fck, Die arabischen Studien in Europa bis in den Anfang des 20. Jahrhunderts,
Leipzig: Otto Harrassowitz, 1955.
Fuller, Universal Rectilinear Dials: A. W. Fuller, Universal Rectilinear Dials, The Mathematical Gazette
(London) 41 (1957), pp. 9-24.
Garbers, Thbit ibn Qurra ber Sonnenuhren: Karl Garbers, Ein Werk Tbit b. Qurras ber ebene Sonnen-
uhren, Quellen und Studien zur Geschichte der Mathematik, Astronomie und Physik, Abt. A, Bd. 4 (1936).
Garca Avils, El tiempo y los astros: Alejandro Garca Avils, El tiempo y los astros Arte, ciencia y religion
en la Alta Edad Media, Murcia: Unversidad de Murcia, 2001.
Garca Franco, Astrolabios en Espaa: Salvador Garca Franco, Catlogo crtico de los astrolabios existentes
en Espaa, Madrid, Instituto Histrico de Marina, 1945.
bibliography and bibliographical abbreviations xxix
Gunella et al., Antecedentes del reloj de Rodrigo Zamorano: Alessandro Gunella, Antecedentes del reloj de
Rodrigo Zamorano, Giovanni Paltrinieri, Sobre un instrumento de Oroncio Fin, and Manuel Mara
Valds, Antecedentes del reloj universal de Rodrigo Zamorano (1581), Analema no. 24 (Septiembre-
Diciembre 1998), pp. 3-4, 5-7, and 8-9.
Gunther, Astrolabes: Robert T. Gunther, The Astrolabes of the World, 2 vols., Oxford: Oxford University Press,
1932, repr. in 1 vol., London: The Holland Press, 1976.
, Early Science in Oxford, II and V: idem, Early Science in Oxford, vol. II: Astronomy, Oxford: Clarendon
Press, 1923, and vol. V: Chaucer and Messahalla on the Astrolabe, Oxford: University Press, 1929.
Haddad: see Kennedy & Pingree & Haddad.
Hrtel, Ludolfus Borchdorp de Brunswik: Helmar Hrtel, Ludolfus Borchdorp de Brunswik, Braunschweigi-
sches Jahrbuch 68 (1987), pp. 113-120. [Establishes the identity of the maker of an astrolabe from Einbeck,
ca. 1330.]
Hahn, Medieval Mensuration: Nan L. Britt Hahn, Medieval Mensuration: Quadrans vetus and Geometrie due
sunt partes principales, Transactions of the American Philosophical Society 72:8 (1984).
Hakim Mohammed Said: see Islamabad 1979 Conference Proceedings.
Hamadanizadeh, Medieval Islamic Interpolation Schemes: Javad Hamadanizadeh, A Survey of Medieval
Islamic Interpolation Schemes, in Kennedy Festschrift, pp. 143-152.
Hartner, Astrolabe, A-B: Willy Hartner, The Principle and Use of the Astrolabe, a chapter in A Survey of
Persian Art from prehistoric times to the present, Arthur Upham Pope, ed., 9 vols., London & New York:
Oxford University Press, 1938-39, III, pp. 2530-2554, reprinted with the original plates in Astrolabica, no.
1, Feb. 1978, Paris: Socit Internationale de lAstrolabe, and article Asurlb in EI2, both reprinted in
idem, Oriens-Occidens, Hildesheim: Georg Olms, 1968, pp. 287-311 and pp. 312-318.
, Oriens-Occidens, I-II: idem, Oriens-Occidens Ausgewhlte Schriften zur Wissenschafts- und Kulturge-
schichte Festschrift zum 60. Geburtstag, Hildesheim: Georg Olms, 1968 (I), and Oriens-Occidens
Ausgewhlte Schriften zur Wissenschafts- und Kulturgeschichte, Band II, Yasukatsu Maeyama, ed., Hildes-
heim, etc.: Georg Olms, 1984 (II).
& Ruska, Katalog: Willy Hartner and Julius Ruska, Katalog der orientalischen und lateinischen Ori-
ginalhandschriften, Abschriften und Photokopien des Instituts fr Geschichte der Medizin und der Natur-
wissenschaften in Berlin, Quellen und Studien zur Geschichte der Naturwissenschaften 7:2/3 (1939),
pp. 1-149.
Hartner Festschrift: Naturwissenschaftsgeschichtliche Studien Festschrift fr Willy Hartner,
Yasukatsu Maeyama and Walter G. Saltzer, eds., Wiesbaden: Franz Steiner, 1977.
Al-Hassan, Decline of Islamic Science: Ahmad Y. al-Hassan, Factors behind the Decline of Islamic Science
after the Sixteenth Century, in Sharifah Shifa al-Attas, ed., Islm and the Challenge of Modernity: Historical
and Contemporary Contexts, Proceedings of the Inaugural Symposium on Islm and the Challenge of
Modernity ... at ... the International Institute of Islamic Thought and Civilization in Kuala Lumpur, August
1-5, 1994, Kuala Lumpur: International Institute of Islamic Thought and Civilization, pp. 351-389.
Hastings: see ERE.
Hawkins & King, Orientation of the Ka{ba: see King, Studies, C-XII.
Heidarzadeh, Astronomy in Iran: Tofigh Heidarzadeh, From the Maragha School to the Darolfonun - A
Historical Review of Astronomy in Iran from the 13th to the 19th Century, paper presented at the Symposium
on Science and Technology in the Turkish and Islamic World, Istanbul, 3-5 June, 1994. [Not printed in the
Proceedings.]
Heinen, Islamic Cosmology: Anton Heinen, Islamic Cosmology: A Study of as-Suys al-Hay}a al-sanya fi-
l-hay}a al-sunnya, Beiruter Texte und Studien (Orient-Institut der Deutschen Morgenlndischen Gesellschaft),
Band 27, Beirut (for Franz Steiner Verlag, Wiesbaden), 1982.
Helmecke, Berliner Astrolab des Muammad Zamn: Gisela Helmecke, Das Berliner Astrolab des Muam-
mad Zamn al-Mahad, Forschungen und Berichte der Staatlichen Museen zu Berlin, Hauptstadt der DDR
(Berlin: Akademie-Verlag) 25 (1985), pp. 129-142 and pls. 31-32.
Higgins, Classification of Sundials: Kathleen Higgins, The Classification of Sundials, Annals of Science
9 (1953), pp. 342-358.
Higton, Portable Sundials: Hester K. Higton, Sundials An Illustrated History of Portable Dials, London: Philip
Wilson Publishers, Ltd., 2001.
Higton: see also Greenwich Sundial Catalogue.
Hill, Al-Brns Mechanical Calendar: Donald R. Hill, Al-Brns Mechanical Calendar, AS 42 (1985),
pp. 139-163, repr. in idem, Studies, XIV, and also in Field & Hill & Wright, Byzantine and Arabic Gearing.
, Islamic Science and Engineering: idem, Islamic Science and Engineering, Edinburgh: Edinburgh University
Press, 1994.
bibliography and bibliographical abbreviations xxxi
, Studies: idem, Studies in Medieval Islamic Technology: From Philo to al-Jazar from Alexandria to Diyr
Bakr, David A. King, ed., Aldershot, Brookfield USA, Singapore and Sydney: Ashgate-Variorum, 1998.
Hill: see also Field & Wright & Hill.
G. F. Hill, Arabic Numerals in Europe: G. F. Hill, The Development of Arabic Numerals in Europe Exhibited
in Sixty-Four Tables, Oxford: Clarendon Press, 1915.
Hinks, Retro-Azimuthal Equidistant Projection: A. Hinks, A Retro-Azimuthal Equidistant Projection of the
Whole Sphere, Geographical Journal 73:2 (1929), pp. 245-247.
Hinz, Islamische Masse und Gewichte: Walter Hinz, Islamische Masse und Gewichte, Leiden: E. J. Brill, 1955.
HM: Historia Mathematica.
History of Cartography: The History of Cartography, J. B. Harley and David Woodward, eds., vol. I: Cartography
in Prehistoric, Ancient, and Medieval Europe and the Mediterranean; vol. II, book 1: Cartography in the
Traditional Islamic and South Asian Societies, Chicago, Ill., & London: The University of Chicago Press,
1987 and 1992. [See also the review of II:1 by Karen C. Pinto in Harvard Middle Eastern and Islamic Review
1:1 (1994), pp. 109-115.]
History of Oriental Astronomy: see New Delhi 1985 Colloquium Proceedings.
Hogendijk, Al-Khwrizms Sine of the Hours: Jan P. Hogendijk, Al-Khwrizms Tables of the Sine of
the Hours and the Underlying Sine Table, Historia scientiarum 42 (1991), pp. 1-12.
, Mathematical Astrology: idem, Mathematical Astrology in the Islamic Tradition (dealing with houses,
rays and progressions), in Cambridge DI 1998 Conference Proceedings, to appear.
, al-Mu{taman ibn Hd: idem, al-Mu}taman ibn Hd. 11th Century King of Saragossa and Brilliant
Mathematician, HM 22 (1995), pp. 1-18.
, al-Nayrz on the Qibla at Baghdad: idem, Al-Nayrzs Mysterious Determination of the Azimuth of
the Qibla at Baghdad, Sciamvs 1 (2000), pp. 49-70.
, Seasonal Hour Lines on Astrolabes and Sundials: idem, The Contributions by Ab Nar ibn {Irq and
al-ghn to the Theory of Seasonal Hour Lines on Astrolabes and Sundials, ZGAIW 14 (2001), pp. 1-
30.
, Tables for Casting the Rays: idem, The Mathematical Structure of Two Islamic Astrological Tables for
Casting the Rays, Centaurus 32 (1989), pp. 171-202.
, Three Islamic Lunar Crescent Visibility Tables: idem, Three Islamic Lunar Crescent Visibility Tables,
JHA 19 (1988), pp. 29-44.
Hogendijk: see also Luckey, Thesis.
Hoyland, Seeing Islam as Others Saw it: Robert G. Hoyland, Seeing Islam as Others Saw it: A Survey and
Evaluation of Christian, Jewish and Zoroastrian Writings on Early Islam, Princeton, NJ: The Darwin Press,
1997.
HM: Historia Mathematica.
HS: History of Science.
IAM: see Islamic Astronomy and Mathematics.
Ibish, ed., Editing Islamic Manuscripts on Science: Editing Islamic Manuscripts on Science, Yusuf Ibish, ed.,
London: Al-Furqan Islamic Heritage Foundation, 1999.
Ibn al-Nadm, Fihrist: Ibn al-Nadm, Kitb al-Fihrist, ed. Gustav Flgel (1871), repr. Beirut: Khayats, 1964,
transl. by Bayard Dodge as The Fihrist of al-Nadm, 2 vols., New York and London: Columbia University
Press, 1970.
Ibn Nujaym, Raf{ al-ghish{: References in IV are to MS Cairo al}at majm{ 200,5 (fols. 27r-30r, copied 1733/
34) of his treatise Raf { al-ghish} {an waqtayi l-{ar wa-l-{ish}.
Ibn Raq, al-Mawqt: References in IV are to MS Berlin Ahlwardt 5664 of his treatise on folk astronomy.
Ibn al-Raqqm: see Carandell.
Id, Analemma for Ascensions: Yusuf Id, An Analemma Construction for Right and Oblique Ascensions,
The Mathematics Teacher 62 (1969), pp. 669-672, repr. in Kennedy et al., Studies, pp. 495-498.
~hsanolu, Western Science in the Ottoman World: Ekmeleddin ~hsanolu, Introduction of Western Science
to the Ottoman World: a Case Study of Modern Astronomy (1660-1860), in Istanbul 1987 Symposium
Proceedings, pp. 67-120.
et al., Ottoman Astronomical Literature: idem et al., Osmanl astronomi literatr tarihi History of
Astronomy Literature during the Ottoman Period (in Turkish), 2 vols., (Studies and Sources on the History
of Science, Series No. 7), Istanbul: Research Centre for Islamic History, Art and Culture (IRCICA), 1997.
, Ottoman Geographical Literature: eidem, Osmanl corafya literatr tarihi History of Geographical
Literature during the Ottoman Period (in Turkish), 2 vols., (Studies and Sources on the History of Science,
Series No. 9), Istanbul: Research Centre for Islamic History, Art and Culture (IRCICA), 2000.
, Ottoman Mathematical Literature: eidem, Osmanl matematik literatr tarihi History of Astronomy
Literature during the Ottoman Period (in Turkish), 2 vols., (Studies and Sources on the History of Science,
Series No. 8), Istanbul: Research Centre for Islamic History, Art and Culture (IRCICA), 1999.
xxxii bibliography and bibliographical abbreviations
~hsanolu: see also Istanbul 1987 Symposium Proceedings, and Istanbul 1991 and 1994 Symposia Proceedings.
Ilyas, Astronomy of Islamic Times: Mohammad Ilyas, Astronomy of Islamic Times for the Twenty-first Century,
London & New York: Mansell, 1988.
, Islamic Calendar, Times & Qibla: idem, A Modern Guide to the Astronomical Calculations of Islamic
Calendar, Times & Qibla, Kuala Lumpur: Berita, 1984.
Imad-ad-Dean Ahmad, Signs in the Heavens: A Muslim Astronomers Perspective on Religion and Science,
Beltsville, Md.: Writers Inc.-International, 1992.
Index Islamicus: Index Islamicus, numerous vols. to date, Cambridge: W. Heffer & Sons, starting in 1958,
continued up to the present as The Quarterly Index Islamicus, London: Bowker-Saur, 4th edn. (including 1906-
1999) available on CD-ROM from E. J. Brill, Leiden.
Ingolstadt Apian Exhibition: see Rttel.
Irani, Arabic Numeral Forms: Rida A. K. Irani, Arabic Numeral Forms, Centaurus 4 (1955), pp. 1-12, repr.
in Kennedy et al., Studies, pp. 710-721.
, The Jadwal al-Taqwm of abash al-sib, unpublished Masters dissertation, American University of
Beirut, 1956.
Islamabad 1979 Conference Proceedings: History and Philosophy of Science Proceedings of the International
Congress of the History and Phlosophy of Science, Islamabad, 8-13 December, 1979, Hakim Mohammed
Said, ed., Nazimabad, Pakistan: Hamdard Foundation Press, n.d. (ca. 1980).
Islamic Geography: Islamic Geography, Fuat Sezgin, with Mazen Amawi, Carl Ehrig-Eggert and Eckhard
Neubauer et al., eds., 278 vols. to date, Frankfurt am Main: Institut fr Geschichte der Arabisch-Islamischen
Wissenschaften, 1992 to present. [Facsimile reprints of early writings, mainly 19th and early-20th century.]
Islamic Mathematics and Astronomy: Islamic Mathematics and Astronomy, eidem, eds., 112 vols. to date,
Frankfurt am Main: Institut fr Geschichte der Arabisch-Islamischen Wissenschaften, 1997 to present.
[Reprints of texts and studies mainly from the 19th and early 20th centuries. Materials relating to al-
Khwrizm feature in vols. 1-7; al-Farghn 9-10, 68; al-Battn 11-13; Thbit ibn Qurra 21-22; Ibn Ynus
25; al-f 26; Ibrhm ibn Sinn 27, 101; Ab Nar 28, 37; al-Brn 29-36, 74; Ibn al-Zarqlluh 39-40;
al-Marrkush 41-42; Ibn al-Bann{ 43-44; {Umar al-Khayym 45-46; al-s et al. 47-51; Ulugh Beg 52-
55; al-Ksh 56, 84; Ibn al-Haytham 57-58, 75; Abu {l-Waf} 60-61; al-Sijz 66; Maimonides 67; Alfonso
X 98-99; calendrics 64-65; astronomical instruments 85-96; Sayls Observatory in Islam 97; Nallinos
Ta}rkh {ilm al-falak 100.]
Istanbul 1977 Symposium Proceedings: Proceedings of the International Symposium on the Observatories in
Islam (Istanbul, 1977), Muammer Dizer, ed., Istanbul: Milli Eitim Basmevi, 1980.
Istanbul 1987 Symposium Proceedings: Transfer of Modern Science & Technology to the Muslim World, Ekme-
leddin ~hsanolu, ed., Istanbul: Reseach Centre for Islamic History, Art and Culture (IRCICA), 1982.
Istanbul 1991 and 1994 Symposia Proceedings: Science in Islamic Civilisation, Ekmeleddin ~hsanolu and Feza
Gnergun, eds., Istanbul (IRCICA) 2000.
Istanbul: see also Kandilli Instrument Handbook.
Jachimowicz, Islamic Cosmology: E. Jachimowicz, Islamic Cosmology, in C. Blacker and M. Loewe, eds.,
Ancient Cosmologies, London: George Allen & Unwin Ltd., 1985, pp. 143-171.
Jackson, Retro-Azimuthal Projections: J. E. Jackson, Retro-Azimuthal Projections, Survey Review 19 (1968),
pp. 319-328.
Janin, Cadran solaire de Damas: Louis A. Janin, Le cadran solaire de la Mosque Umayyade Damas,
Centaurus 16 (1972), pp. 285-298, repr. in Kennedy & Ghanem, eds., Ibn al-Shir, pp. 107-121.
, Un cadran solaire grec dAfghanistan: idem, Un cadran solaire grec A Khanoum, Afghanistan,
LAstronomie (Paris) 92 (1978), pp. 357-362.
, Gnomonique tunisienne: idem, Quelques aspects rcents de la gnomonique tunisienne, Revue de
lOccident musulman et de la Mditerrane 23 (1977), pp. 202-221.
, Projection horizontale: idem, Astrolabe et cadran solaire en projection horizontale, Centaurus 22 (1979),
pp. 298-314.
, Ar-Rdn sur lastrolabe sphrique: idem, Un texte dar-Rdn sur lastrolabe sphrique, Annali
dellIstituto e Museo di Storia della Scienza di Firenze 3 (1978), pp. 71-75. [See also Pellat, Lastrolabe
sphrique dal-Rdn.]
Janin & King, Cadran solaire dIbn ln: see King, Studies, B-XVI.
& , Ibn al-Shirs Compendium: see King, Studies, B-XII.
Janin & Rohr, Deux astrolabes-quadrants turcs, idem and R. R. J. Rohr, Deux astrolabes-quadrants turcs,
Centaurus 19 (1975), pp. 108-124.
JAOS: Journal of the American Oriental Society.
Jaubert, Capteurs de vents dgypte: Olivier Jaubert, Capteurs de vents dgypte Essai de typologie,
Annales islamologiques (Institut Franais dArchologie Orientale du Caire) 25 (1995), pp. 169-231.
bibliography and bibliographical abbreviations xxxiii
Jensen, Ab Nars Table of Minutes: Claus Jensen, Ab Nars Approach to Spherical Trigonometry as
Developed in His Treatise The Table of Minutes, Centaurus 16 (1971), pp. 1-19.
, The Lunar Theories of al-Baghdd: idem, The Lunar Theories of al-Baghdd, AHES 8 (1971-72), pp.
321-328.
JHA: Journal for the History of Astronomy.
JHAS: Journal for the History of Arabic Science (Institute for the History of Arabic Science, Aleppo).
Jiminez, La qibla extraviada: Alfonso Jiminez, La qibla extraviada, in Cuadernos de Madnat al-Zahr{,
vol. III, Cordova, 1991, pp. 189-209.
JNES: Journal of Near Eastern Studies.
Jones, Byzantine Astronomical Manual: Alexander Jones, An Eleventh-Century Manual of Arabo-Byzantine
Astronomy, (Corpus des astronomes byzantins III), Amsterdam: J. C. Gieben, 1987.
JRAS: Journal of the Royal Asiatic Society.
Juschkewitsch, Mathematik im MA: Adolf P. Juschkewitsch, Geschichte der Mathematik im Mittelalter, (first
published in Russian in 1961), Leipzig: B. G. Teubner, 1964. [See also Youschkevitsch, Mathmatiques
arabes.]
Kamal al-Din, Tables: Mathematical tables and graphs by Husayn Kamal al-Din (then Professor in the Civil
Engineering Dept., Riyadh University) in 3 volumes bearing the titles: 1 Kebla [sic] Directions for the Whole
World, 2 Praying Time Tables for the Whole World, 3 Curves of Praying Times for the Whole World, Cairo:
Dr al-Fikr al-{Arab, for The Conference of Islamic Jurisprudence at the Islamic University of Imm
Muammad ibn Sa{d, Saudi Arabia, n.d. [ca. 1970 ?]. [Copy inspected at Aligarh Muslim University in
1978, by which time the book was no longer available at the publishers in Cairo.]
Kandilli Instrument Handbook: Muammer Dizer, Astronomi hazineleri, (B.. Yaynlar no. 404), Istanbul:
Boazii niversitesi and Kandilli Rasathanesi, 1986.
Kaye, The Observatories of Jai Singh, A-B: George R. Kaye, The Astronomical Observatories of Jai Singh,
(Archaeological Survey of India, New Imperial Series, vol. XL), and A Guide to the Observatories at Delhi,
Jaipur, etc., Calcutta, 1918 and 1920, both repr. in IAM, vol. 93.
Kennedy, al-Brn on Prayer-Times: Edward Stewart Kennedy, al-Brn on the Muslim Times of Prayer,
in The Scholar and the Saint: Studies in Commemoration of Abl-Rayn al-Brn and Jall al-Dn al-
Rm, Peter J. Chelkowski, ed., New York: New York University Press, 1975, pp. 83-94, repr. in idem et
al., Studies, pp. 299-310.
, al-Brns Shadows: see al-Brn, Shadows.
, Comets in Islamic Astronomy and Astrology: idem, Comets in Islamic Astronomy and Astrology, JNES
16 (1957), pp. 44-51, repr. in idem et al., Studies, pp. 311-318.
, The Digital Computer and the History of the Exact Sciences: idem, The Digital Computer and the History
of the Exact Sciences, Centaurus 12 (1967), pp. 107-113, repr. in idem et al., Studies, pp. 385-393.
, The Equatorium of Ab al-alt: idem, The Equatorium of Ab al-alt, Physis 12 (1970), pp. 73-81,
repr. in idem et al., Studies, pp. 481-491.
,Exact Sciences in Iran, A-C: idem, The Exact Sciences during the Abbasid Period, Cambridge History
of Iran, IV, pp. 378-395 (A); The Exact Sciences in Iran under the Seljuqs and Mongols, ibid., V, pp.
659-679 (B); The Exact Sciences in Iran under the Timurids, ibid., VI, pp. 568-580.
, History of Trigonometry: idem, Overview of the History of Trigonometry (first published 1969), repr.
in idem et al., Studies, pp. 3-29.
, Ibn al-A{lams Tables: The Astronomical Tables of Ibn al-A{lam, JHAS 1 (1977), pp. 13-23.
, Ibn al-Raqqms Tables: idem, The Astronomical Tables of Ibn al-Raqqm, a Scientist of Granada,
ZGAIW 11 (1997), pp. 35-72.
, Islamic Mathematical Geography: idem, Mathematical Geography, in EHAS, I, pp. 185-201.
, al-Ksh on Instruments: idem, al-Kshs Treatise on Astronomical Observational Instruments, JNES
20 (1961), pp. 98-108, repr. in idem et al., Studies, pp. 394-404.
, al-Kshs Equatorium: idem, The Planetary Equatorium of Jamshd Ghiyth al-Dn al-Ksh (d. 1429) ...
, Princeton N.J.,: Princeton University Press, 1960.
, Spherical Astronomy in al-Kshs Khqn Zj: see idem, Studies, VII.
, Studies: idem, Astronomy and Astrology in the Medieval Islamic World, (Variorum Collected Studies Series:
CS600), Aldershot & Brookfield, Vt.: Ashgate-Variorum, 1998. Contains:
I abash al-sib on the Melon Astrolabe (with Richard P. Lorch), first publication.
VII Spherical Astronomy in Kshs Khqn Zj, ZGAIW 2 (1990), pp. 1-46.
, Al-f on the Celestial Globe: idem, Al-f on the Celestial Globe, ZGAIW 5 (1989), pp. 48-93.
, Zj Survey: idem, A Survey of Islamic Astronomical Tables, Transactions of the American Philosophical
Society, N. S., 46:2 (1956), pp. 123-177, repr. with separate pagination, n.d. [ca. 1990].
& Davidian, Al-Qyin on Twilight: idem and Marie-Louise Davidian, Al-Qyin on the Duration of Dawn
and Twilight, JNES 20 (1961), pp. 145-153, repr. in idem et al., Studies, pp. 284-292.
xxxiv bibliography and bibliographical abbreviations
& Debarnot, Two Mappings of al-Brn: idem and Marie-Thrse Debarnot, Two Mappings Proposed
by al-Brn, Zeitschrift fr Geschichte der arabisch-islamischen Wissenschaften (Frankfurt am Main) 1
(1984), pp. 144-147, repr. in idem, Studies, V.
& Faris, Yay b. Ab Manr on Solar Eclipses: idem and Nazim Faris, The Solar Eclipse Technique
of Yay b. Ab Manr, JHA 1 (1970), pp. 20-38, repr. in idem et al., Studies, pp. 185-203.
& Ghanem, eds., Ibn al-Shir: idem and Imad Ghanem, eds., The Life and Work of Ibn al-Shir: an Arab
Astronomer of the Fourteenth Century, Aleppo: Institute for the History of Arabic Science, 1976.
& Id, abashs Analemma for the Qibla: idem & Yusuf Id, A Letter of al-Brn: abash al-sibs
Analemma for the Qibla, Historia Mathematica 1 (1973), pp. 3-11, repr. in Kennedy et al., Studies, pp.
621-629.
& Kennedy, Islamic Geographical Coordinates: idem and Mary Helen Kennedy, Geographical Coordinates
of Localities from Islamic Sources, Frankfurt am Main: Institut fr Geschichte der Arabisch-Islamischen
Wissenschaften, 1987.
& King, Ibn al-Majds Tables for Calculating Ephemerides: idem and David A. King, Ibn al-Majds
Tables for Calculating Ephemerides, JHAS 4 (1980), pp. 48-68, repr. in King, Studies, A-VI.
, Indian Astronomy in Fourteenth-Century Fez: eidem, Indian Astronomy in Fourteenth-Century Fez:
The Versified Zj of al-Qusunn [sic: read al-Qusunn], JHAS 6 (1982), pp. 3-45, repr. in King, Studies,
A-VIII.
& Krikorian-Preisler, Projecting the Rays: idem and H. Krikorian-Preisler, The Astrological Doctrine of
Projecting the Rays, Al-Abhath 25 (1972), pp. 3-15, repr. in Kennedy et al., Studies, pp. 372-384.
& Kunitzsch & Lorch, Melon Astrolabe: idem, Paul Kunitzsch and Richard P. Lorch, The Melon-Shaped
Astrolabe in Arabic Astronomy, (Boethius: Texte und Abhandlungen zur Geschichte der Mathematik und der
Naturwissenschaften, Band 43), Stuttgart: Franz Steiner, 1999.
& Pingree, Astrological History of Msh}allh: idem and David Pingree, The Astrological History of
Msh}allh, Cambridge, Ma.: Harvard University Press, 1971.
& Pingree & Haddad, al-Hshim on Astronomical Tables: idem, David Pingree and Fuad I. Haddad, The
Book of Reasons behind Astronomical Tables (Kitb {ilal al-zjt) by {Al ibn Sulaymn al-Hshim, Delmar,
N.Y.: Scholars Facsimiles & Reprints, 1981.
Kennedy & Ukashah, The Chandelier Clock of Ibn Ynis: idem and Walid Ukashah, The Chandelier Clock
of Ibn Ynis, Isis 60 (1969), pp. 543-545, repr. in Kennedy et al., Studies, pp. 499-501.
Kennedy et al., Studies: E. S. Kennedy, Colleagues and Former Students, Studies in the Islamic Exact Sciences,
David A. King and Mary Helen Kennedy, eds., Beirut: American University of Beirut, 1983. Contains:
Overview of the History of Trigonometry, 31st Yearbook of the National Council of Teachers of Mathematics, Washington,
D.C., 1969, pp. 333-359, repr. here pp. 3-29.
The Solar Eclipse Technique of Yay b. Ab Manr (with Nazim Faris), JHA 1 (1970), pp. 20-38, repr. here pp. 185-203.
Mark Lesley, Mark, Brn on Rising Times and Daylight Lengths, Centaurus 5 (1957), pp. 121-141, repr. here pp. 253-273.
Marie-Louise Davidian, al-Brn on the Time of Day from Shadow Lengths, JAOS 80 (1960), pp. 330-335, repr. here pp. 274-
279.
Nadi Nadir, Ab al-Waf{ on the Solar Altitude, The Mathematics Teacher 53 (1960), pp. 460-463, repr. here pp. 280-283.
Al-Qyin on the Duration of Dawn and Twilight (with Marie-Louise Davidian), JNES 20 (1961), pp. 145-153, repr. here, pp.
284-292.
Bernard R. Goldstein, A Medieval Table for Reckoning Time from Solar Altitude, Scripta Mathematica 27 (1963), pp. 61-66,
repr. here pp. 293-298.
al-Brn on the Muslim Times of Prayer, in The Scholar and the Saint ... , P. J. Chelkowski, ed., New York, 1975, pp. 83-
94, repr. here pp. 299-310.
Comets in Islamic Astronomy and Astrology, JNES 16 (1957), pp. 44-51, repr. here, pp. 311-318.
The Astrological Doctrine of Projecting the Rays (with H. Krikorian-Preisler), Al-Abhath 25 (1972), pp. 3-15, repr. here pp.
372-384.
The Digital Computer and the History of the Exact Sciences, Centaurus 12 (1967), pp. 107-113, repr. here pp. 385-393.
al-Kshs Treatise on Astronomical Observational Instruments, JNES 20 (1961), pp. 98-108, repr. here pp. 394-404.
The Equatorium of Ab al-alt, Physis 12 (1970), pp. 73-81, repr. here pp. 481-491.
The Chandelier Clock of Ibn Ynis (with Walid Ukashah), Isis 60 (1969), pp. 543-545, repr. here, pp. 499-501.
Yusuf Id, An Analemma Construction for Right and Oblique Ascensions, The Mathematics Teacher 62 (1969), pp. 669-672,
repr. here pp. 495-498.
A Letter of al-Brn: abash al-sibs Analemma for the Qibla (with Yusuf Id), Historia Mathematica 1 (1973), pp. 3-11,
repr. here pp. 621-629.
Rida A. K. Irani, Arabic Numeral Forms, Centaurus 4 (1955), pp. 1-12, repr. here pp. 710-721.
Kennedy Festschrift: From Deferent to Equant: Studies in the History of Science in the Ancient and Medieval
Near East in Honor of E. S. Kennedy, David A. King and George Saliba, eds., Annals of the New York Academy
of Sciences 500 (1987). Contains:
Marie-Thrse Debarnot, The Zj of abash al-sib: A Survey of MS Istanbul Yeni Cami 784/2, pp. 35-69.
Javad Hamadanizadeh, A Survey of Medieval Islamic Interpolation Schemes, pp. 143-152.
D. A. King, Some Early Islamic Tables for Lunar Crescent Visibility, pp. 185-225.
bibliography and bibliographical abbreviations xxxv
G. Rosinska, Tables of Decimal Trigonometric Functions from ca. 1450 to ca. 1550, pp. 419-426.
Kennedy: see also al-Brn, Geography, and idem, Shadows; Destombes & Kennedy; and King, Studies, A-
VI and A-VIII (joint papers).
Kessler: Mecca-Oriented Architecture in Cairo, A-C: Christel Kessler, Mecca-Oriented Architecture and
Urban Growth of Cairo, Atti del Terzo Congresso di Studi Arabi e Islamici (Ravello, 1966), Naples, 1967,
p. 425 (A); Funerary Architecture within the City, Colloque International sur lHistoire du Caire, Cairo,
1969, pp. 257-267 (B); and Mecca-Oriented Urban Architecture in Mamluk Cairo: The Madrasa-Mausoleum
of Sultan Sha{bn II, in In Quest of an Islamic Humanism: Arab an Islamic Studies in Memory of Mohamed
Nowaihi, Cairo: American University of Cairo Press, 1984, pp. 97-108 (C).
Khalili Collection: see London Khalili Collection ... ... .
al-Khwrizm: see Suter, al-Khwrizm; Neugebauer, al-Khwrizm; Goldstein, Ibn al-Muthann on al-
Khwrizm; and also King, al-Khwrizm, and Rosenfeld et al., eds., Al-Khorezmi.
{Abdallh al-Khwrizm, Maft al-{ulm: Gerlof van Vloten, ed., Mafth al-olm explicans vocabula technica
scientiarum tam arabum quam peregrinorum auctore Ab Abdallah Mohammed ibn Ahmed ibn Jsuf al-Ktib
al-Khowarezmi, Leiden, 1895, also pirated edition: Cairo: Maba{at al-Sharq, n. d. [ca. 1960?], and Beirut,
1984 edn. [based on the 1895 edn.?].
Kimerling, Adding the Qibla to Geographic Education: A. Kimerling, Adding the Qibla to Geographic
Education, Journal of Geography 101 (2002), pp. 20-26.
King, Architecture and Astronomy: David A. King, Architecture and Astronomy: The Ventilators of
Medieval Cairo and their Secrets, JAOS 104:1 (1984), pp. 97-133. [A new version is included here as VIIb.]
, Andalus Astronomical Instruments: idem, [Andalus Astronomical Instruments], in Madrid & New York
1992 Exhibition Catalogue, pp. 376-383.
, Andalus Sundials: see idem, Studies, B-XV, also idem, Cuadrantes solares andaluses.
, Aspekte: idem, Aspekte angewandter Wissenschaften in Moscheen und Klstern, Berichte zur
Wissenschaftsgeschichte 18 (1995), pp. 85-95 and 137-149. [A translation of Aspects of Science in Mosques
and Monasteries, unpublished paper presented at the conference Science and Religion in Islam, Judaism
and Christendom organised by Prof. David C. Lindberg, Madison, Wisc., April 15-17, 1993. A new version
is included here as IX.]
, Astronomical Alignments in Medieval Islamic Religious Architecture: see idem, Studies, C-XIII.
, Astronomical Instruments between East and West: idem, Astronomical Instruments between East and
West, in Kommunikation zwischen Orient und Okzident Alltag und Sachkultur, Harry Khnel, ed.,
(Sitzungsberichte der Philosophisch-Historischen Klasse der sterreichischen Akademie der Wissenschaften,
Band 619, Verffentlichungen des Instituts fr Realienkunde des Mittelalters und der Frhen Neuzeit, Nr.
16), Vienna: sterreichische Akademie der Wissenschaften, 1994, pp. 143-198.
, Astronomical Timekeeping in Medieval Cairo: see idem, Studies, A-IX.
, Astronomical Timekeeping in Ottoman Turkey: see idem, Studies, A-XII.
, Astronomical Timekeeping in Medieval Syria: see idem, Studies, A-X.
, Astronomy in Fatimid Egypt: idem, Aspects of Fatimid Astronomy: From Hard-Core Mathematical
Astronomy to Architectural Orientations in Cairo, in Marianne Barrucand, ed., Lgypte Fatimide: son art
et son histoire Actes du colloqie organis Paris les 28, 29 et 30 mai 1998, Paris: Presses de lUniversit
de Paris-Sorbonne, 1999, pp. 497-517.
, Astronomy of the Mamluks: see idem, Studies, A-III.
, Astronomy in the Maghrib: idem, On the History of Astronomy in the Medieval Maghrib, in tudes
Philosophiques et Sociologiques Ddies Jamal ed-Dine Alaoui, Universit Sidi Mohamed Ben Abdallah,
Publications de la Facult des Lettres et des Sciences Humaines Dhar El Mahraz Fs, No Spcial 14
(Dpartement de Philosophie, Sociologie et Psychologie), Fez, 1998 [published 1999], pp. 27-61.
, Astronomie en Syrie: idem, Lastronomie en Syrie lpoque islamique, in Paris IMA 1993-94
Exhibition Catalogue, pp. 386-395 and 432-443.
, Astronomy in Yemen: idem, Mathematical Astronomy in Medieval Yemen A Bio-Bibliographical Survey,
(Publications of the American Research Center in Egypt), Malibu, Ca.: Undena Publications, 1983.
,al-Bazdaw on the Qibla in Tranoxania: idem, al-Bazdaw on the Qibla in Early Islamic Tranoxania,
JHAS 7 (1983/1986), pp. 3-38.
, Byzantine Astronomy: idem, an essay review of Pingree, Astronomical Works of Gregory Chioniades,
and Jones, Byzantine Astronomical Manual, in Isis 82 (1991), pp. 116-118.
, Catalogue of Medieval Astronomical Instruments, A-B: idem, Medieval Astronomical Instruments: A
Catalogue in Preparation, Bulletin of the Scientific Instrument Society no. 31 (Dec., 1991), pp. 3-7, and 1992
A Good Year for Medieval Astronomical Instruments, ibid. no. 36 (March, 1993), pp. 17-18. [For a table
of contents of the Frankfurt catalogue see www.uni-frankfurt.de/fb13/ign/instrument-catalogue.html.]
, Cuadrantes solares andaluses: idem, Los cuadrantes solares andaluses, in Madrid MAN 1992 Exhibition
Catalogue, pp. 89-102.
xxxvi bibliography and bibliographical abbreviations
, Ciphers of the Monks: idem, The Ciphers of the Monks A Forgotten Number Notation of the Middle Ages,
(Boethius: Texte und Abhandlungen zur Geschichte der Mathematik und der Naturwissenschaften, Band 44),
Stuttgart: Franz Steiner, 2001.
, Earliest Muslim Geodetic Measurements: idem, Too Many Cooks ... A Newly-Rediscovered Account
of the First Muslim Geodetic Measurements, Suhayl 1 (2000), pp. 207-241.
, Earliest Qibla Methods: see idem, Studies, C-XIV.
, Geography of Astrolabes: idem, Bringing Astronomical Instruments Back to Earth: The Geographical
Data on Medieval Astrolabes (to ca. 1100), in North Festschrift, pp. 3-53.
, Ibn Ynus: idem, The Astronomical Works of Ibn Ynus, unpublished doctoral dissertation, Yale University,
1972. [Available from ProQuest.com (formerly University Microfilms, Ann Arbor, Mich.), no. 7229740]
, Ibn Ynus and the Pendulum: idem, Ibn Ynus and the Pendulum: A History of Errors, AIHS 29 (1979),
pp. 35-52.
, Islamic Astronomical Tables: see idem, Studies, A-I.
, Islamic Astronomy, A-B: idem, Islamic Astronomy, in Astronomy before the Telescope, Christopher
B. F. Walker, ed., London: British Museum Press, 1996, repr. 1999, pp. 143-174 (A); and Mathematical
Astronomy in Islamic Civilisation, in Selin, Astronomy across Cultures, pp. 585-613 (B).
, Islamic Folk Astronomy, A-B: idem, Folk Astronomy in the Service of Religion: The Case of Islam,
in Astronomies and Cultures, Clive L. N. Ruggles & Nicholas J. Saunders, eds., Niwot, Co.: University Press
of Colorado, 1993 [published 1994], pp. 124-138 (A), with a summary in Archaeoastronomy in the 1990s,
Clive L. N. Ruggles, ed., Loughborough (U.K.): Group D Publications, Ltd., 1993, p. 346; and idem,
Applications of Folk Astronomy and Mathematical Astronomy to Aspects of Muslim Ritual, The Arabist
(Budapest Studies in Arabic), 13-14 (1995) (Proceedings of the XIVth UEAI Congress, Budapest, 1988,
Alexander Fodor, ed., Part I), pp. 251-274 (B).
, Islamic Scientific Manuscripts: idem, Some Remarks on Islamic Scientific Manuscripts and Instrument
and Past, Present and Future Research, in John Cooper, ed., The Significance of Islamic Manuscripts,
London: Al-Furqan Islamic Heritage Foundation, 1992, pp. 115-144 (also available in Arabic).
, Islamic Trigonometric Tables: idem, A Brief Survey of Medieval Islamic Trigonometric Tables, paper
delivered at the Colloquium Innovation et transmission dans les mathmatiques arabes, Paris, Institut du
Monde Arabe, 25-27 Oct., 1990. [As yet unpublished.]
, Italian Astrolabe: idem, A Remarkable Italian Astrolabe from ca. 1300 Witness to an Ingenious Islamic
Tradition of Non-Standard Astrolabes, in Miniati Festschrift, pp. 29-52.
, Al-Khalls Auxiliary Tables: see idem, Studies, A-XI.
, Al-Khalls Qibla Table: see idem, Studies, A-XIII.
, al-Khwrizm: idem, al-Khwrizm and New Trends in Mathematical Astronomy in the Ninth Century,
Occasional Papers on the Near East (New York University, Hagop Kevorkian Center for Near Eastern
Studies), 2 (l983), 43 pp.
, Kuwait Astrolabes: idem, Early Islamic Astronomical Instruments in Kuwaiti Collections, in Arlene
Fullerton & Gza Fehrvri, eds., Kuwait: Art and Architecture Collection of Essays, Kuwait (no publisher
stated), 1995, pp. 76-96.
, Lunar Equation Table: idem, A Double-Argument Table for the Lunar Equation Attributed to Ibn Ynus,
Centaurus 18 (1974), pp. 129-146, repr. in idem, Studies, A-V.
, Mecca-Centred World-Maps: idem, World-Maps for Finding the Direction and Distance to Mecca:
Innovation and Tradition in Islamic Science, Leiden: E. J. Brill, and London: Al-Furqan Islamic Heritage
Foundation, 1999.
, A Medieval Arabic Report on Algebra before al-Khwrizm: idem, A Medieval Arabic Report on Algebra
before al-Khwrizm, Al-Masq: Studia Arabo-Islamica Mediterranea 1 (1988), pp. 25-32.
, Medieval Islamic Multiplication Tables, A-B: see idem, Studies, A-XIV and XV.
, Monumental Syrian Astrolabe: idem, The Monumental Syrian Astrolabe in the Maritime Museum,
Istanbul, in Sayl Memorial Volume, II, pp. 729-735 and 10 pls.
, Muwaqqits: idem, On the Role of the Muezzin and the Muwaqqit in Medieval Islamic Society, in
Oklahoma 1992 & 1993 Conference Proceedings, pp. 285-346. [A new version is presented here as V.]
, Navicula: see idem, Universal Horary Dials.
, Neglected Astrolabe: idem, The Neglected Astrolabe, in Mathematische Probleme in Mittelalter Der
lateinische und arabische Sprachbereich, Menso Folkerts, ed., Wiesbaden: Harrassowitz, 1996, pp. 45-55.
, Nuremberg Astrolabes: idem, Die Astrolabiensammlung des Germanischen Nationalmuseums, in
Nuremberg GNM 1992-93 Exhibition Catalogue, I, pp. 101-114, and II, pp. 568-602, 640-643.
, Orientation of Islamic Architecture: idem, The Orientation of Medieval Islamic Religious Architecture
and Cities, JHA 26 (1995), pp. 253-274. [A new version is presented here as VIIa.]
, Origin of the Astrolabe: see idem, Studies, B-III.
bibliography and bibliographical abbreviations xxxvii
, Quadrans vetus: idem, A Vetustissimus Arabic Treatise on the Quadrans Vetus, JHA 33 (2002), pp.
237-255
, Review of Higton: idem, Towards a History from Antiquity to the Renaissance of Sundials and Other
Instruments for reckoning Times by the Sun and Stars, an essay review of Greenwich Sundial Catalogue,
and Higton, Portable Sundials, to appear in AS 61 (2004), pp. 1-13.
, Review of Hill, al-Jazar: idem, Medieval Mechanical Devices, an essay review of Donald R. Hills
1974 translation of al-Jazars Book of Knowledge of Ingenious Mechanical Devices, HS 13 (1975), pp. 284-
289, repr. in idem, Studies, B-XX.
, Review of London Khalili Collection Catalogue: idem, Cataloguing Islamic Astronomical Instruments,
essay review in Bibliotheca Orientalis 57 (2000), cols. 247-258.
, Review of Sezgin, GAS, V: idem, On the Sources for the Study of Early Islamic Mathematics, an essay
review in JAOS 99 (1979), pp. 450-459.
, Review of Sezgin, GAS, VI: idem, Early Islamic Astronomy, an essay review in JHA 12 (1981), pp.
55-59.
, Review of G. Turner, Elizabethan Instruments: review in Journal of the History of Collections (Oxford)
15:1 (2003), pp. 147-150.
, Review of Varisco & Smith, eds., Al-Afals Anthology, in Yemeni Update 44 (2002) see www.aiys.org/
webdate/kngrev.html.
, The Sacred Geography of Islam: idem, The World about the Kaaba The Sacred Folk Geography of Medieval
Islam, to be submitted to E. J. Brill.
, SATMI: This abbreviation has been used by the author since the mid 1970s to refer to the unpublished
manuscript on which part of the present work (I-II) is based.
, Shadow Schemes: idem, A Survey of Medieval Islamic Shadow Schemes for Simple Timereckoning,
Oriens 32 (1990), pp. 191-249. [A new version is presented here as III.]
, The Shakkzya Quadrant of al-Mridn: see idem, Studies, B-X.
, Star-Names on European Astrolabes: idem, The Stars-Names on Three 14th-Century Astrolabes from
Spain, France and Italy, in Kunitzsch Festschrift, pp. 307-333.
, Strumentazione astronomica nel mondo medievale islamico, a chapter in Gli strumenti, Gerard LE.
Turner, ed., Turin: Giulio Einaudi, 1991, pp. 154-189 and pp. 581-585 (bibliography). [A new version of
the English original is presented here as X.]
, Studies, A: idem, Islamic Mathematical Astronomy, London: Variorum, 1986, 2nd rev. edn., Aldershot:
Variorum, 1993. Includes reprints of:
I On the Astronomical Tables of the Islamic Middle Ages, Studia Copernicana 13 (1975), pp. 37-56.
III The Astronomy of the Mamluks, Isis 74 (1983), pp. 531-555.
V A Double-Argument Table for the Lunar Equation Attributed to Ibn Ynus Centaurus 18 (1974), pp. 129-146.
VI Ibn al-Majds Tables for Calculating Ephemerides (with E. S. Kennedy), JHAS 4 (1980), pp. 48-68.
VII Some Astronomical Observations from Thirteenth-Century Egypt (with Owen Gingerich), JHA 13 (1982), pp. 121-128.
VIII Indian Astronomy in Fourteenth-Century Fez: The Versified Zj of al-Qusunn [sic: read al-Qusann] (with E. S. Kennedy),
JHAS 6 (1982), pp. 3-45.
IX Ibn Ynus Very Useful Tables for Reckoning Time by the Sun, AHES 10 (1973), pp. 342-394.
X Astronomical Timekeeping in Fourteenth-Century Syria, Aleppo 1976 Symposium Proceedings, II, pp. 75-84.
XI Al-Khalls Auxiliary Tables for Solving Problems of Spherical Astronomy, JHA 4 (1973), pp. 99-110.
XII Astronomical Timekeeping in Ottoman Turkey, Istanbul 1977 Symposium Proceedings, pp. 245-269.
XIII Al-Khalls Qibla Table, JNES 34 (1975), pp. 81-122.
XIV On Medieval Islamic Multiplication Tables, HM 1 (1974), pp. 317-323.
XV Supplementary Notes on Medieval Islamic Multiplication Tables, HM 6 (1979), pp. 405-417.
XX Medieval Mechanical Devices (an essay review of Donald R. Hills 1974 translation of al-Jazars Book of Knowledge of
Ingenious Mechanical Devices), HS 13 (1975).
, Studies, B: idem, Islamic Astronomical Instruments, London: Variorum, 1987, repr. Aldershot: Variorum,
1995. Includes reprints of:
II The Medieval Yemeni Astrolabe in the Metropolitan Museum of Art in New York, ZGAIW 2 (1985), pp. 99-122, and
4 (1987/88), pp. 268-269 (corrections).
III On the Origin of the Astrolabe According to the Medieval Arabic Sources, JHAS 5 (1981), pp. 43-83.
VI The {Abd al-A}imma Astrolabe Forgeries (with Owen Gingerich and George Saliba), JHA 3 (1972), pp. 188-198.
VII On the Early History of the Universal Astrolabe in Islamic Astronomy and the Origin of the Term Shakkzya in Medieval
Scientific Arabic, JHAS 3 (1979), pp. 244-257.
VIII The Astrolabe of {Al al-Wad{, previously unpublished.
IX The Astronomical Instruments of Ibn al-Sarrj, previously-unpublished paper presented at the Second International
Symposium for the History of Arabic Science, Aleppo, 1979.
X An Analog Computer for Solving Problems of Spherical Astronomy: The Shakkzya Quadrant of Jaml al-Dn al-Mridn,
AIHS 24 (1974), pp. 219-242.
XI New Light on the Zj al-af}i of Ab Ja{far al-Khzin, Centaurus 23 (1980), pp. 105-117.
xxxviii bibliography and bibliographical abbreviations
XII Ibn al-Shirs andq al-Yawqt: An Astronomical Compendium (with Louis Janin), JHAS 1 (1977), pp. 187-256.
XIV Osmanische astronomische Handschriften und Instrumente, first published 1985.
XV Three Sundials from Islamic Andalusia, JHAS 2 (1978), pp. 358-392.
XVI Le cadran solaire de la mosque dIbn ln au Caire (with Louis Janin), JHAS 2 (1978), pp. 331-357.
XVII The Sundial on the West Wall of the Madrasa of Sultan Qaytbay in Jersualem (with Archibald G. Walls), art and architec-
ture research papers (London) 15 (1979), pp. 16-21.
XVIII A Fourteenth-Century Tunisian Sundial for Regulating the Times of Muslim Prayer, in Hartner Festschrift, pp. 187-202.
, Studies, C: idem, Astronomy in the Service of Islam, Aldershot: Variorum, 1993. Includes reprints of:
I Science in the Service of Religion: The Case of Islam, impact of science on society (U.N.E.S.C.O.), no. 159 (1991), pp.
245-262.
II Some Early Islamic Tables for Lunar Crescent Visibility, in Kennedy Festschrift, pp. 185-225.
III Ibn Ynus on Lunar Crescent Visibility, JHA 19 (1988), pp. 155-168.
IV Lunar Crescent Visibility Predictions in Medieval Islamic Ephemerides, in Quest for Understanding Arabic and Islamic
Studies in Memory of Malcolm H. Kerr, S. Seikaly, R. Baalbaki, P. Dodd, eds., Beirut: American University of Beirut, 1991,
pp. 233-251.
VI Universal Solutions in Islamic Astronomy, in Aaboe Festschrift, pp. 121-132. [A new version is included here as VIa.]
VII Universal Solutions to Problems of Spherical Astronomy from Mamluk Egypt and Syria, in The Way Prepared: Essays
on Islamic Culture in Honor of R. Bayly Winder, Farhad Kazemi and Robert B. McChesney, eds., New York: New York
University Press, 1987, pp. 153-184. [A new version is included here as VIb.]
XII On the Orientation of the Ka{ba (with Gerald S. Hawkins), JHA 13 (1982), pp. 102-109.
XIII Astronomical Alignments in Medieval Islamic Religious Architecture, Annals of the New York Academy of Sciences 385
(1982), pp. 303-312.
XIV The Earliest Mathematical Methods and Tables for Finding the Direction of Mecca, ZGAIW 3 (1986), pp. 82-149, with
corrections listed ibid. 4 (1987/88), p. 270.
, Tables for Lunar Crescent Visibility: see idem, Studies, C-II.
, Three Sundials from Islamic Andalusia: see idem, Studies, B-XV.
, The Toledo Astrolabe: idem, An Astrolabe from 14th-Century Christian Spain with Inscriptions in Latin,
Hebrew and Arabic A Unique Testimonial to an Intercultural Encounter, Suhayl 3 (2002/03), pp. 9-156.
, Tunisian Sundial: see idem, Studies, B-XVIII.
, Universal Astrolabe: see idem, Studies, B-VII.
, Universal Horary Dials, idem, On Universal Horary Dials for Determining Time by the Sun and Stars,
in XIIb. [A summary is in King, Navicula: 14th-Century England or 9th-Century Baghdad? New Insights
on the Origins of the Elusive Astronomical Instrument Called the Navicula de Venetiis, in Goldstein
Festschrift.]
, Universal Horary Quadrant: idem, The Universal Horary Quadrant for Timekeeping by the Sun,
in IXa. [A shorter version is A Vetustissimus Arabic Treatise on the Quadrans Vetus, JHA 33 (2002), pp.
237-255.]
, Universal Quadrant of al-Mridn: see idem, Studies, B-X.
, Universal Solutions in Islamic Astronomy: see idem, Studies, C-VI.
, Universal Solutions from Mamluk Egypt and Syria: see idem, Studies, C-VII.
, Urbino Astrolabe: idem, The Astrolabe Depicted in the Intarsia of the Studiolo of Archduke Federico
in Urbino, in La scienza del Ducato di Urbino The Science of the Dukedom of Urbino, Flavio Vetrano,
ed., Urbino: Accademia Raffaello, 2001, pp. 101-139.
, Yemeni Astrolabe: see idem, Studies, B-II.
, The Zj al-af}i : see idem, Studies, B-XI.
& Charette, The Universal Astrolabe of Ibn al-Sarrj: idem and Franois Charette, The Universal Astrolabe
of Ibn al-Sarrj: Innovation in Medieval Islamic Science, to appear.
King & Gingerich, Astronomical Observations from 13th-Century Egypt: idem and Owen Gingerich, Some
Astronomical Observations from Thirteenth-Century Egypt, JHA 13 (1982), pp. 121-128, repr. in King,
Studies, A-VII.
with Dorothea Girke, An Approximate Trigonometric Formula for Astronomical Timekeeping and
Related Tables, Sundials and Quadrants from ca. 500 to ca. 1900, Johann Wolfgang Goethe-Universitt
Frankfurt am Main, Institut fr Geschichte der Naturwissenschaften, Preprint Series, 1 (1988). [A new
version is included here as XI.]
& Kennedy, Ibn al-Majds Tables: idem and E. S. Kennedy, Ibn al-Majds Tables for Calculating
Ephemerides, JHAS 4 (1980), pp. 48-68, repr. in King, Studies, A-VI.
& Maier, London Catalan Astrolabe: idem and Kurt Maier, The Medieval Catalan Astrolabe of the Society
of Antiquaries, London, in Vernet Festschrift, II, pp. 673-718.
& Sams, Islamic Astronomical Handbooks and Tables: idem and Julio Sams, with a contribution by
Bernard R. Goldstein, Astronomical Handbooks and Tables from the Islamic World (750-1900): An Interim
Report, Suhayl 2 (2001), pp. 9-105. [A summary is in the article Zdj in EI2 (illustrated).]
bibliography and bibliographical abbreviations xxxix
& Turner, Regiomontanus Astrolabe: idem and Gerard LE. Turner, The Astrolabe Presented by
Regiomontanus to Cardinal Bessarion in 1462, Nuncius Annali di Storia della Scienza (Florence) 9 (1994),
pp. 165-206, repr. in Turner, Studies.
& Walls, Jerusalem Sundial: see idem, Studies, B-XVII.
King: see also Cairo ENL Catalogue and Survey, and Kennedy Festschrift.
King-pre, Geared to the Stars: Henry C. King with John R. Milburn, Geared to the Stars, Toronto: University
of Toronto Press, 1978.
Klemm, Georg Hartmann: Hans Gunther Klemm, Georg Hartmann aus Eggolsheim (1489-1564) Leben und
Werk eines frnkischen Mathematikers und Ingenieurs, Forschheim (D): Ehrenbrg-Gymnasium (Wissen-
schaftliche und knstlerische Beitrge, Heft 8), 1990, 104 pp.
Knobel, Ulughbegs Catalogue of Stars: E. B. Knobel, Ulughbegs Catalogue of Stars, Washington, D.C.:
Carnegie Institution, 1917.
Knorr, Quadrans Vetus: Wilbur R. Knorr, The Latin Sources of Quadrans Vetus, and What They Imply for
Its Authorship and Date, in Edith Sylla and Michael McVaugh, eds., Texts and Contexts in Ancient and
Medieval Science Studies on the Occasion of John E. Murdochs Seventieth Birthday, Leiden, etc.: E. J.
Brill, 1997, pp. 23-67.
, Sacroboscos Quadrans: idem, Sacroboscos Quadrans: Date and Sources, Journal for the History of
Astronomy 28 (1997), pp. 187-222.
Kragten, Greenwich Navicula: Jan Kragten, The Venetian Ship of Sibton Abbey, Eindhoven (NL), published
privately, 1992.
, Navicula: idem, The Little Ship of Venice [sic] (Navicula de Venetiis), Eindhoven (NL), published privately,
1989.
Krause, Stambuler Handschriften: Max Krause, Stambuler Handschriften islamischer Mathematiker, Quellen
und Studien zur Geschichte der Mathematik, Astronomie und Physik, B 3:4 (1936), pp. 437-532.
Krause: see also al-Brn, Qnn.
Kremer & Dobrzycki, Peurbach and Margha Astronomy?: Richard L. Kremer and Jerzy Dobrzycki, Peurbach
and Margha Astronomy? The Ephemerides of Johannes Angelus and their Implications, JHA 27 (1996),
pp. 187-237.
Kren, Bibliography of Medieval Science and Technology: Claudia Kren, Medieval Science and Technology
A Selected, Annotated Bibliography, New York and London: Garland Publishing, Inc., 1985.
, The Travellers Dial: eadem, The Travellers Dial in the Late Middle Ages: The Chilinder, Technology
and Culture 18 (1977), pp. 419-435;
Kunitzsch, Anw}-Tradition: Paul Kunitzsch, ber eine anw}-Tradition mit bisher unbekannten Sternnamen,
Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Phil.-hist. Klasse, Heft 5, (Beitrge zur
Lexikographie des Classischen Arabisch, Nr. 4), Munich, 1983.
, Arabische Sternnamen in Europa: idem, Arabische Sternnamen in Europa, Wiesbaden: Otto Harrassowitz,
l959.
, Fachausdrcke der Astrolabliteratur: idem, Glossar der arabischen Fachausdrcke in der mittelalterlichen
europischen Astrolabliteratur, Nachrichten der Akademie der Wissenschaften in Gttingen, I. Philologisch-
Historische Klasse, Jahrgang 1982, Nr. 11, pp. 455-571. [The separatum is paginated 1-117.]
, Ibn abbs Book on the Stars: idem, {Abd al-Malik ibn abbs Book on the Stars, ZGAIW 9 (1994),
pp. 161-194.
, Al-Khwrizm as a Source for the Sententie astrolabii: idem, Al-Khwrizm as a Source for the Sententie
astrolabii, in Kennedy Festschrift, pp. 227-236, repr. in idem, Studies, IX.
, Messahalla on the Astrolabe: idem, On the Authenticity of the Treatise on the Composition and Use of
the Astrolabe Ascribed to Messahalla, AIHS 31 (1981), pp. 42-62.
, Sternkatalog des Almagest: idem, Claudius Ptolemus Der Sternkatalog des Almagest Die arabisch-
mittelalterliche Tradition, 3 vols., Wiesbaden: Otto Harrassowitz, 1986-1991.
, Sternnomenklatur der Araber: idem, Untersuchungen zur Sternnomenklatur der Araber, Wiesbaden: Otto
Harrassowitz, 1961.
, Sternverzeichnisse: idem, Typen von Sternverzeichnissen in astronomischen Handschriften des zehnten bis
vierzehnten Jahrhunderts, Wiesbaden: Otto Harrassowitz, 1966.
, Studies: idem, The Arabs and the Stars, Northampton: Variorum, 1989.
, Al-f and the Astrolabe Stars: idem, Al-f and the Astrolabe Stars, ZGAIW 6 (1990), pp. 151-166.
& Dekker, The Stars on the Carolingian Astrolabe: idem and Elly Dekker, The Stars on the Rete of the
So-Called Carolingian Astrolabe, in Vernet Festschrift, II, pp. 655-672.
Kunitzsch & Lorch, Matli{ al-samt: idem and Richard Lorch, Ab Nar and abash on matli{ al-samt,
ZGAIW 9 (1994), pp. 43-82.
xl bibliography and bibliographical abbreviations
Kunitzsch Festschrift: Sic itur ad astra. Studien zur Geschichte der Mathematik und Naturwissenschaften.
Festschrift fr den Arabisten Paul Kunitzsch zum 70. Geburtstag, Menso Folkerts and Richard P. Lorch, eds.,
Wiesbaden: Otto Harrassowitz, 2000.
Kunitzsch: see also Kennedy & Kunitzsch & Lorch.
Kurz, European Clocks: Otto Kurz, European Clocks and Watches in the Near East, London: The Warburg
Institute, and Leiden: E. J. Brill, 1975.
Labarta & Barcel, Un nuevo fragmento de reloj de sol andalus: Ana Labarta and Carmen Barcel, Un nuevo
fragmento de reloj de sol andalus, al-Qanara (Madrid) 16 (1995), pp. 147-150.
, Ocho relojes de sol hispanomusulmanes, eaedem, Ocho relojes de sol hispanomusulmanes, al-Qanara
(Madrid) 9 (1988), pp. 231-247.
Lacroix, Science in the Middle Ages: Paul Lacroix, Science and Literature in the Middle Ages and Renaissance,
translated from the French, first published in 1878, repr. New York: Frederick Ungar, 1964.
de La Lande, Bibliographie Astronomique: Jrme de La Lande, Bibliographie Astronomique ; avec LHistoire
de lAstronomie jusqu 1802, Paris: Imprimerie de la Rpublique, An XI = 1803.
Landes, Revolution in Time: David S. Landes, Revolution in Time Clocks and the Making of the Modern World,
Cambridge, Mass, and London: The Belknap Press of Harvard University Press, 1983
Lane, Lexicon: Edward W. Lane, An Arabic-English Lexicon, 8 pts., London: Williams and Norgate, 1863, repr.
Beirut: Librairie du Liban, 1968.
Langermann, Ibn al-Haythams Cosmology: Y. Tzvi Langermann, Ibn al-Haythams On the Configuration of
the World, (Series: Harvard Dissertations in the History of Science), New York & London: Garland
Publishing, Inc., 1990.
, Studies: idem, The Jews and the Sciences in the Middle Ages, Aldershot: Variorum, 1999.
Lee, Astronomical Tables of Al Farsi: Samuel Lee, Notice of the Astronomical Tables of Mohammed Abibekr
Al Farsi ... ... , Transactions of the Cambridge Philosophical Society 1 (1822), pp. 249-265.
Lehr, De Geschiedenis van het Astronomisch Kunstuurwerk: Andr Lehr, De Geschiedenis van het Astronomisch
Kunstuurwerk Zijn techniek en muziek, The Hague: Martinus Nijhoff, 1981.
Leipzig Catalogue: K. Vollers, Katalog der islamischen ... Handschriften der Universittsbibliothek zu Leipzig
... , Leipzig: Otto Harrassowitz, 1906, reprinted Osnabrck: Biblio Verlag, 1975.
Le Strange, Lands of the Eastern Caliphate: Guy Le Strange, The Lands of the Eastern Caliphate, Mesopotamia,
Persia and Central Asia from the Moslem Conquest to the Time of Timur, Cambridge: Cambridge Universty
Press, 1905, repr. London: Frank Cass & Co. Ltd., 1966, and again in Islamic Geography, vol. 85 (1993).
Lesley, Mark, Brn on Rising Times and Daylight Lengths, Centaurus 5 (1957), pp. 121-141, repr. in
Kennedy et al., Studies, pp. 253-273.
Levin, Direction-Facing in Judaism and Islam: Daniel Z. Levin, Which Way is Jerusalem? Which Way is
Mecca? The Direction-Facing Problem in Religion and Geography, Journal of Geography (National Council
for Geographic Education, Washington, D.C.) 101 (2002), pp. 27-37.
Lexikon des Mittelalters, 9 vols., Munich and Zurich: Artemis, 1980-99, especially articles Astrologie,
Astronomie, Instrumente, Komputistik, Mathematik.
Lindberg, David C., The Beginnings of Western Science, Chicago: University of Chicago Press, 1992.
Linton Collection Catalogue: Scientific Instruments and Rare Books from the Collection of Leonard Linton, Point
Lookout, N.Y., Paris: Alain Brieux, 1980.
Lippincott, Navicula: Kristen Lippincott, The Navicula Sundial, Bulletin of the Scientific Instrument Society,
no. 35 (Dec., 1992), p. 22. [On the navicula now in Greenwich.]
Livingston, Islamic Conical Sundial: John W. Livingston, The Mukula, an Islamic Conical Sundial,
Centaurus 16 (1972), pp. 299-308.
London BL Catalogue: (W. Cureton and C. Rieu), Catalogus codicum manuscriptorum orientalium qui in Museo
Britannico asservantur, Part II: Codices arabicos amplectens, London: British Museum, 1871.
London BM Catalogue: F. A. B. Ward, A Catalogue of European Scientific Instruments in the Department of
Medieval and Later Antiquities of the British Museum, London: British Museum Publications, 1981. [See
also the essay review in Price, Instruments in the British Museum.]
London Khalili Collection Catalogue: The Nasser D. Khalili Collection of Islamic Art, vol. XII: Science, Tools
and Magic, by Francis Maddison and Emilie Savage-Smith, 2 pts., London: The Nour Foundation, in
association with Azimuth Editions and Oxford University Press, 1997. [See King, Cataloguing Islamic
Instruments, for a review.]
London SM 1976 Exhibition Catalogue: Francis R. Maddison and Anthony J. Turner, Catalogue of an Exhibition
Science and Technology in Islam ... , 1976, [a catalogue of the exhibition The World of Islam, held at The
Science Museum, London, in 1976], privately distributed [but unfortunately never published].
Lorch, Astronomical Instruments of Jbir ibn Afla: see Richard P. Lorch, Studies, XVI.
, Al-Khzins Sphere that Rotates by Itself : see idem, Studies, XI.
bibliography and bibliographical abbreviations xli
Manchester Catalogue: Alois Mingana, Catalogue of the Arabic Manuscripts in the John Rylands Library,
Manchester: Manchester University Press, 1934.
Maranyi, Zwischen {ar and marib in Mekka: Miklos Maranyi, Zwischen {ar and marib in Mekka Ein
Augenzeugenbericht von al-Lait b. Sa{d aus dem Jahr 113/732, Die Welt des Orients: wissenschaftliche
Beitrge zur Kunde des Morgenlandes (Gttingen: Vandenhoek & Ruprecht) 23 (1992), pp. 101-128.
al-Marrkush, A-Z of Astronomical Timekeeping: Fuat Sezgin et al., Comprehensive Collection of Principles
and Objectives in the Science of Timekeeping: Jmi{ al-mabdi} wa-l-ghyt by al-Marrkush, (facsimile
edition of MS Topkap1 Ahmet III 3343), 2 vols., Frankfurt am Main: Institut fr Geschichte der Arabisch-
Islamischen Wissenschaften, 1984. [See also Sdillot-pre, Trait, and Sdillot-fils, Mmoire.]
Mart & Viladrich, Tratados de astrolabio: Ramon Mart and Merc Viladrich, En torno a los tratados de
uso del astrolabio hasta el siglo XIII en al-Andalus, la Marca Hispnica y Castilla, in Vernet, ed., Nuevos
estudios, pp. 9-74.
Matvievskaya & Rosenfeld, MAMS: Galina P. Matvievskaya and Boris A. Rosenfeld, Matematiki i astronomi
musulmanskogo srednevekovya i ikh trudi, 3 vols., Moscow: Nauk, 1983.
S. & K. Maurice, Counting the Hours: Sigrid and Klaus Maurice, Counting the Hours in Community Life
of the 16th and 17th Centuries, in Maurice & Mayr, eds., The Clockwork Universe, pp. 144-156.
Maurice & Mayr, eds., The Clockwork Universe: The Clockwork Universe German Clocks and Automata 1550-
1650, Klaus Maurice and Otto Mayr, eds., New York: Neale Watson, 1980.
Mayer, Islamic Astrolabists: Leo A. Mayer, Islamic Astrolabists and Their Works, Geneva: Albert Kundig, 1956,
with a supplement in Aus der Welt der islamischen Kunst, Richard Ettinghausen, ed., Berlin: Gebrder Mann,
1959, pp. 293-296.
McCluskey, Medieval Astronomy: Stephen C. McCluskey, Astronomies and Cultures in Early Medieval Europe,
Cambridge, etc.: Cambridge University Press, 1998.
McKay, How Far to Mecca?: J. McKay, How Far to Mecca?, Mercators World 5:5 (2002), p. 58.
Meliconi, Brechte Plate and Oxford Quadrans vetus: see EPACT.
Mercier, Astronomical Tables in the 12th Century: Raymond Mercier, Astronomical Tables in the Twelfth
Century, in Burnett, ed, Adelard of Bath, pp. 87-118.
, English Orientalists and Islamic Astronomy: idem, English Orientalists and Mathematical Astronomy,
in Russell, ed., Arabick Interest in 17th-Century England, pp. 158-214.
, Ibn al-A{lams Parameters: idem, The Parameters of the Zj of Ibn al-A{lam, AIHS 39 (1989), pp. 22-
50.
Mestres, Hyderabad MS of the Zj of Ibn Isq: ngel Mestres, Maghrib Astronomy in the 13th Century:
a Description of Manuscript Hyderabad Andra Pradesh State Library 298, in Vernet Festschrift, I, pp. 383-
443.
, Zj of Ibn Isq: idem, Materials andalusins en el Zj dIbn Isq al-Tnis (edited text and tables, with
introduction and commentary in English), doctoral thesis, University of Barcelona, 2000.
Meucci, Globo celeste arabico: Ferdinando Meucci, Il globo celeste arabico del secolo XI nel Gabinetto degli
strumenti antichi di astronomia, di fisica e di matematica del R. Istituto di studi superiori, Florence:
Successori Le Monnier, 1878, repr. in AIOS, II, pp. 126-144 plus pullout.
Meyer, Istanbul Sundials: Wolfgang Meyer, ~stanbuldaki gne saatleri, Sandoz kltr yaynlar (Istanbul) 7
(1985). [See also idem, Sundials of the Osmanic Era in Istanbul, in Istanbul 1977 Symposium Proceedings,
pp. 193-202.]
Meyerhof: see Frank & Meyerhof.
Michel, Lastrolabe linaire: Henri Michel, Lastrolabe linaire dal-Tusi, Ciel et terre 59 (1943), pp. 101-
107, repr. in IAM, vol. 94.
, Images des sciences: idem, Images des sciences: les anciens instruments scientifiques vus par les artisites
de leur temps, Rhode-St-Gense (B): Albert de Visscher, 1977.
, Construction des astrolabes persans: idem, Mthodes de trace et dexcution des astrolabes persans,
Ciel et terre 57 (1941), pp. 481-496.
, Scientific Instruments in Art and History: idem, Scientific Instruments in Art and History, London: Barrie
& Rockliff, 1967.
, Trait de lastrolabe: idem, Trait de lastrolabe, Paris: Gauthier-Villars, 1947, reprinted with a new preface
by Francis R. Maddison, Paris: Librarie Alain Brieux, 1989.
& Ben-Eli, Cadran solaire polaire: idem and A. Ben-Eli, Un cadran solaire remarquable, Ciel et terre
81 (1965), pp. 214-216.
Mielgo, Analyzing Mean Motion Tables: Honorino Mielgo, A Method of Analysis for Mean Motion
Astronomical Tables, in Vernet Festschrift, I, pp. 159-180.
Mielgo: see also Sams & Mielgo.
Milenio, Cuadrante Vetustissimus: M. Milenio, Cuadrante Vetustissimus, Analema no. 30 (Septiembre-
Diciembre 2000), pp. 20-24.
bibliography and bibliographical abbreviations xliii
, Relojes de altura Regiomontano: idem, Sobre el diseo de la forma de los relojes de altura Regiomontano
y Capuchino, Analema no. 27 (Septiembre-Diciembre 1999), pp. 11-14.
Mills Vallicrosa, Assaig: Jos M. Mills Vallicrosa, Assaig dhistria de les idees fsiques i matemtiques a
la Catalunya medieval, vol. I, (Estudis Universitaris Catalans, srie monogrfica I), Barcelona: 1931.
, Azarquiel: idem, Estudios sobre Azarquiel, Madrid & Granada: Consejo Superior de Investigaciones
Cientficas and Instituto Miguel Asn, 1943-50.
, Cuadrante con cursor: idem, La introduccin del cuadrante con cursor en Europa, Isis 17 (1932), pp.
218-258, repr. in idem, Estudios, I, pp. 65-110.
, Don Profeit Tibbn sobre lassafea: idem, Don Profeit Tibbn: Tractat de lassafea dAzarquiel, Barcelona:
Universitat de Barcelona, 1933.
, Un ejemplar de azafea rabe de Azarquiel, Al-Andalus 9 (1944), pp. 11-119.
, Estudios and Nuevos estudios: idem, Estudios sobre historia de la ciencia espaola, Barcelona: Consejo
Superior de Investigaciones Cientficas, Instituto Luis Vives de Filosofa, 1949, and idem, Nuevos estudios
sobre historia de la ciencia espaola, Barcelona: Consejo Superior de Investigaciones Cientficas, Instituto
Luis Vives de Filosofa, and Asociacon para la Historia de la Ciencia Espaola, 1960, both repr., 2 vols.,
Madrid: Consejo Superior de Investigaciones Cientficas, 1987.
, Estudios sobre Azarquiel: El tratado de la azafea, Archeion 14 (1932), pp. 393-419.
Mills-Vallicrosa Festschrift: Homenaje a Mills-Vallicrosa, 2 vols., Barcelona: Consejo Superior de Investiga-
ciones Cientificas, 1959.
Mills Vendrell, Apogeos planetarios en el ecuatorio de Azarquiel: Eduardo Mills Vendrell, Las posiciones
de los apogeos planetarios en el ecuatorio de Azarquiel, in Vernet, ed., Nuevos estudios, pp. 119-123.
Mills: see also Sams & Mills.
Mills, Canterbury Dial: Allan Mills, The Canterbury Pendant: A Saxon Seasonal-Hour Altitude Dial, Bulletin
of the British Sundial Society 95.2 (June 1995), pp. 39-44.
Miniati Festschrift: MUSA MUSAEI: Studies on Scientific Instruments and Collections in Honour of Mara
Miniati, Marco Beretta, Paolo Galluzzi and Carlo Triarico, eds., Florence: Leo S. Olschki, 2003.
Monzawi, Persian Manuscripts: Amad Monzawi, A Catalogue of Persian Manuscripts - Fihrist-i nuskhah-
yi kha-i fris (in Persian), 6 vols., Tehran: Regional Cultural Institute, 1348-1353 HS = 1969-1974.
Morelon, Early Eastern Arabic Astronomy: Rgis Morelon, Eastern Arabic Astronomy between the Eighth
and Eleventh Centuries, in EHAS, I, pp. 20-57.
, Thbit ibn Qurra: idem, Thbit ibn Qurra oeuvres dastronomie, Paris: Les Belles Lettres, 1987.
Morelon: see also EHAS.
Morley, Arabic Quadrant: William H. Morley, Description of an Arabic Quadrant, Journal of the Royal
Asiatic Society 17 (1860), pp. 322-330 and two figs., repr. in AIOS, I, pp. 326-336.
, Astrolabe of Shh usayn: idem, Description of a Planispheric Astrolabe Constructed for Shh Sultn
Husain Safaw ... , London: Williams and Norgate, 1856, repr. as the introduction to Gunther, Astrolabes,
I, and again in AIOS, I, pp. 249-325.
Mortillaro, Astrolabio arabo: Vincenzo Mortillaro, Illustrazione di un astrolabio arabo-sicula [sic], Palermo:
Stamperia Oretea, 1848, also appeared as Illustrazione di un astrolabio arabo del nono secolo, in idem,
Opere, IV, Palermo, 1848, pp. 110-135, repr. from the latter in AIOS, I, pp. 162-191.
Mller, Tables des hauteurs du soleil: Frderic Christophe Mller, Tables des hauteurs du soleil pour toute
lAllemagne et les pays qui la bornent au levant et au couchant, Leipzig: Siegfried Lebereclt Crusius, 1791.
Munich Catalogue: Burkhard Stautz, Die Astrolabiensammlungen des Deutschen Museums und des Bayerischen
Nationalmuseums, Munich: Oldenbourg, 1999.
Murdoch, John E., Album of Science Antiquity and the Middle Ages, New York: Charles Scribners Sons, 1984.
Murray, Alexander, Reason and Society in the Middle Ages, (first published 1978), Oxford: Clarendon Press,
1991 edn.
Nadir, Nadi, Ab al-Waf} on the Solar Altitude, The Mathematics Teacher 53 (1960), pp. 460-463, repr. in
Kennedy et al., Studies, pp., 280-283.
Naffah, Cadran cylindrique: Christiane Naffah, Un cadran cylindrique ottoman du XVIIIme sicle, Astro-
labica 5 (1989), pp. 37-52.
Nallino, al-Battn: Carlo Alfonso Nallino, ed., transl. and comm., al-Battani sive Albatenii opus astronomicum,
3 pts., Pubblicazioni del Real Osservatorio di Brera in Milano, no. XL, pts. I-III, Milan, 1899-1907, repr.
in 1 vol., Hildesheim & New York: Georg Olms, 1977. [Pts. I-II were also reprinted Frankfurt am Main:
Minerva, 1969; pt. III (the Arabic text) was reprinted in an undated, unlabelled pirate edition by [Muthann,
Baghdad], in the 1960s.]
, Islamic Astronomy: idem, Sun, Moon, and Stars (Muhammadan), in ERE, vol. XII (1921), pp. 88-101.
, Scritti, V: idem, Raccolta di scritti editi e inediti (ed. Maria Nallino), vol. V: Astrologia Astronomia
Geografia, Rome: Istituto per lOriente, 1944.
xliv bibliography and bibliographical abbreviations
Nasr, Seyyed Hossein, Islamic Science: An Illustrated Study, London: World of Islam Festival Publishing
Company Ltd., 1976, repr. Istanbul: nsan Yaynlar, 1989. [A Persian translation {Ilm dar Islm was
published in Tehran, 1366 HS [= 1988]. [This work is to be used with caution as far as any technical aspects
of Islamic science are concerned.]
Necipoglu, The Topkap Scroll: Glr Necipoglu, The Topkap Scroll Geometry and Ornament in Islamic
Architecture ... , Santa Monica, Ca.: The Getty Center for the History of Art and the Humanities, 1995.
Neugebauer, Abu Shakers Chronography: Otto Neugebauer, Abu Shakers Chronography Treatise of the
13th Century on Chronological, Calendrical, and Astronomical Matters, written by a Christian Arab,
preserved in Ethiopic, Vienna: sterreichische Akademie der Wissenschaften, 1988.
. Astronomical and Calendrical Data in the Trs Riches Heures: idem, Astronomical and Calendrical Data
in the Trs Riches Heures, (originally published 1974), repr. in idem, Astronomy and History, pp. 507-520.
, Astronomische Papyri: idem, Astronomische Papyri aus Wiener Sammlungen: II. ber griechische
Wetterzeichen und Schattentafeln, Sitzungsberichte der philosophisch-historischen Klasse der ster-
reichischen Akademie der Wissenschaften 240:2 (1962), pp. 29-44.
, Early Greek Astronomy: idem, On Some Aspects of Early Greek Astronomy, Journal of the American
Philosophical Society 116 (1972), pp. 243-251, repr. in idem, Essays, pp. 361-369.
, The Early History of the Astrolabe: idem, The Early History of the Astrolabe, Isis 40 (1949), pp. 240-
256, repr. in idem, Essays, pp. 278-294.
, ESA: idem, The Exact Sciences in Antiquity, reprint of the 1957 2nd edn., New York: Dover, 1969.
, Essays: idem, Astronomy and History Selected Essays, New York, etc.: Springer, 1983. Contains:
Neugebauer & Olaf Schmidt, Hindu Astronomy at Newminster in 1428, AS 8 (1952), pp. 221-228, repr. here pp. 425-432.
, Ethiopic Astronomy and Computus: idem, Ethiopic Astronomy and Computus, (sterreichische Akademie
der Wissenschaften, Philosophisch-Historische Klasse, Sitzungsberichte, 347. Band, Verffentlichungen der
Kommission fr Geschichte der Mathematik, Naturwissenschaften und Medizin, Heft 22), Vienna, Verlag
der sterreichischen Akademie der Wissenschaften, 1979.
, HAMA: idem, A History of Ancient Mathematical Astronomy, Berlin, Heidelberg & New York: Springer,
1975.
, al-Khwrizm: idem, The Astronomical Tables of al-Khwrizm, in Kgl. Danske Vidensk. hist.-fil. Skrifter,
4:2 (1962).
, Studies in Byzantine Astronomy: idem, Commentary on the Astronomical Treatise Par. gr. 2425,
Mmoires de lAcadmie Royale de Belgique 49:4 (1969), pp. 3-45.
& Schmidt, Hindu Astronomy at Newminster: idem and Olaf Schmidt, Hindu Astronomy at Newminster
in 1428, AS 8 (1952), pp. 221-228, repr. in Neugebauer, Essays, pp. 425-432.
New Delhi 1985 Colloquium Proceedings: History of Oriental Astronomy Proceedings of an International
Astronomical Colloquium No. 91, New Delhi, India, 13-16 November, 1985, G. Swaru, A. K. Bag and K.
Shukla, eds., Cambridge, etc.: Cambridge University Press, 1987.
Nicholas of Lynn, Kalendarium: see Eisner.
Nolte, Die Armillarsphre: Friedrich Nolte, Die Armillarsphre, Abhandlungen zur Geschichte der
Naturwissenschaften und der Medizin (Erlangen: M. Mencke), Heft 2, 1922.
North, Astrolabe: John D. North, The Astrolabe, Scientific American 230 (1974), pp. 96-106, repr. in idem,
Essays, no. 14.
, Chaucers Universe: idem, Chaucers Universe, Oxford: Clarendon Press, 1988.
, Essays, A-B:, idem, Stars, Minds and Fate Essays in Ancient and Medieval Cosmology (A) and The
Universal Frame Historical Essays in Astronomy, Natural Philosophy and Scientific Method (B), London
& Ronceverte, W.V.: Hambledon, 1989.
, Horoscopes and History: idem, Horoscopes and History, (Warburg Institute Surveys and Texts, vol. 12),
London: The Warburg Institute, 1986.
, Hour-Line Ritual: idem, Astrolabes and the Hour-Line Ritual, JHAS 5 (1981), pp. 113-114, repr. in
idem, Essays, A, no. 15.
, Meteoroscope: idem, Werner, Apian, Blagrave and the Meteoroscope, British Journal for the History
of Science 3 (1966-67), pp. 57-65 and 1 pl.
, Opus quarundam rotarum mirabilium, Physis 8 (1966), pp. 337-372, repr. in idem, Essays, I, pp. 135-
169 (no. 11).]
, Review of King, Mecca-Centred World-Maps: idem, in Bibliotheca Orientalis (Leiden), 57 (2000), cols.
747-750.
, Richard of Wallingford: idem, Richard of Wallingford: An edition of his writings with introductions, English
translation and commentary, 3 vols., Oxford: The Clarendon Press, 1976.
North Festschrift: Between Demonstration and Imagination: Essays in the History of Science and Philosophy
Presented to John D. North, Arjo Vanderjagt and Lodi Nauta, eds., Leiden: E. J. Brill, 1999.
on islamic tables for timekeeping 1
Part I
A survey of tables
for timekeeping by
the sun and stars
2 part i, chapter one
To Ted Kennedy
4 part i, chapter one
The present study (I-II) was made possible by the cooperation of the directors of numerous
manuscript libraries in Europe, the Near East and the United States, who allowed me access
to their collections of Arabic scientific manuscripts or provided me with microfilms of specific
manuscripts. I should like to express my particular gratitude to the Sleymaniye Library in
Istanbul and the Egyptian National Library in Cairo.
In 1972 I received a two-year research grant to catalogue the scientific manuscripts in the
Egyptian National Library. When I arrived at the Library I did not know how many manuscripts
there were or that, for security reasons, it would take me two years to obtain unlimited access
to them. There was a sense in which I had to justify my existence. Whilst preparing my doctoral
thesis on the astronomical works of Ibn Ynus during 1969-72 I had written a description of
the tables for timekeeping attributed to him in a Dublin manuscript, which was published in
1973. Even before that paper was published I realized that the tables were not all computed
by Ibn Ynus. It seemed like a good idea at the time to try to figure out just who had computed
them. In the course of this pursuit, and especially when I found the tables of al-Khall and
Najm al-Dn al-Mir and realized their scope, I decided to write an article on the development
of all such tables from the Islamic Middle Ages. Fortunately, my grant was extended for an
additional five years, and the catalogue was indeed prepared and eventually published. Slowly
but surely, my article on tables for timekeeping also grew and grew, and by the time I left
Cairo for New York an extensive monograph was finally finished. But in New York my research
interests changed, and I started to work on the qibla, on the notion of the world centred on
the Ka{ba, and on mosque orientations. In fact, when I left New York for Frankfurt two
monographs were completed, one from Cairo on timekeeping and one more recent from New
York on the sacred geography of Islam. At least summaries of these were published in two
articles (M_t. ii and Makka. iv) in the Encyclopedia of Islam.
The research on medieval Islamic science conducted at the American Research Center in
Egypt during the years 1972-79 was supported by the Smithsonian Institution and National
Science Foundation, Washington, D.C. (1972-79), and by grants from the Penrose and Johnson
Funds of the American Philosophical Society (1972-74) and the Ford Foundation (1977-79).
The support of each of these five institutions is gratefully acknowledged. Professor Owen
Gingerich of the Smithsonian Astrophysical Observatories and Harvard University generously
sponsored this research. All of the various drafts of the text were typed by Nabila Hajj, who
was the mainstay of the project from its inception right through to 1985, when the Arabic
catalogue of the scientific manuscripts was finally published. Kurt Maier of Frankfurt typed
much of the new text for the computer in the early 1990s. Franois Charette controlled the
penultimate version for consistency but should not in any way be held responsible for a text
penned by someone who at the time he wrote it was younger than Charette is now.
Two sets of tables have been investigated by the next generation of researchers and some
6 part i, acknowledgements and notes
of their results are noted in this book. First, Glen Van Brummelen attacked the tables of al-
Khall with statistical techniques that I could not hope to understand. Second, Franois Charette
has produced masterly studies of the monumental triple-argument universal auxiliary table of
Najm al-Dn al-Mir and the same authors remarkable treatise on instruments. He also helped
me insert new materials by Najm al-Dn in the final version of the present study.
This first part of the main study (I) is dedicated to the one man who influenced me most
in my adult life. Ted Kennedy is a man for whom I have an undying feeling of gratitude. He
is a man whom I can both love and respect and try if in vain in some ways to emulate.
His achievements were to no small extent the result of the support of his wife, Mary Helen.
I first met Ted Kennedy at Brown University in the autumn of 1969 when I was a graduate
student at Yale. Asger Aaboe and Bernie Goldstein organized monthly expeditions there to
gather with the Brown crowd, which then consisted of Otto Neugebauer, David Pingree, Abe
Sachs and Gerald Toomer, with Ted Kennedy on sabbatical leave from the American University
of Beirut. I had already encountered Ted Kennedy on paper, and I was fascinated by the fact
that he was active in Beirut. It also seemed to me absurd that an Englishman, manqu Near
East man, most at home in Drfr or Cairo, should be languishing in New Haven, Connecticut,
when he too could be in Beirut. Sure enough, my wife and I spent the academic year 1970-
71 there, getting to know the Kennedys and their world. I also worked on my dissertation, and
enjoyed unlimited access to Teds library, extensive microfilm collection and unpublished
materials. Teds generosity was boundless: one day George Saliba, the other of Teds terrible
twins, discovered that Kodak had a special offer; we taped all of Teds microfilms together
and had a copy made for each of us, on several enormous spools.
In our early years in Cairo my wife and I went to Beirut regularly to see the Kennedys.
During 1976-77 the Kennedys were in Cairo, living in an apartment next to ours. Ted and I
braved the rigours of the Dr al-Kutub together; even with a sense of humour and a few
provisions against the cold, this was not easy. When we moved to Frankfurt in 1985, the
Kennedys were there too, thanks to the generosity of Professor Fuat Sezgin, for a couple of
years. So our paths have crossed many times in Providence, Beirut, Aynab, Aleppo, Cairo,
Frankfurt, Barcelona, Princeton, New York City, and Doylestown, Pa., as well as various
conference-sites such as Bucharest, Hamburg, Lige and Cambridge, Ma.
In November, 1999, I spoke at a dinner in Pittsburgh organised by a splinter group of the
(U.S.) History of Science Society and attended by colleagues in the history of ancient and
medieval science. I was waxing enthusiastically about the exciting materials available to
historians of Islamic science. I happened to mention Ted Kennedy. A certain distinguished
female colleague, versed in medieval European science of the Aristotelian variety and in
American personalities, voiced the question: You mean the Ted Kennedy?. The only answer
was: Of course!.
In this new version I have not added many new materials, since, after I stopped looking
in the 1980s, I did not find any save those that more or less fell into my lap.1 The few new
1 But notice that few new materials of consequence appear to be documented in the monumental, newly-
published volumes listed as ~hsanolu et al., Ottoman Astronomical Literature, and Ottoman Mathematical
Literature. For details see n. 1:14 below.
acknowledgements and notes 7
materials are obvious from the table of contents in Parts I-II for they bear an asterisc appended
to the section number. Several of these are due to Najm al-Dn al-Mir and have now been
studied by Franois Charette.2 Others have been found in two manuscripts which I saw for
the first time in 2001, namely, Bursa Haraiolu 1177,4 and the precious Istanbul U.L. A314.
On the other hand, much more could be written on aspects of timekeeping using astronomical
instruments, a topic which I have pursued since arriving in Frankfurt and which has more
recently been taken up by Charette. Also I did not add much more information about the
manuscripts themselves. I wrote at least Parts I-II of this book when I was young and innocent;
with increasing years I realize that the lack of information such as date and location of copying
is unfortunate, but I am not inclined to add it now since I no longer trust my own notes and
the majority of the manuscripts are not properly catalogued anyway. The reader may assume
that the manuscripts were copied in the locations for which the tables were compiled; if they
were copied elsewhere, unless they were universal, the tables would be useless anyway, apart
from their inherent academic interest. It was in Cairo that a very high proportion of the tables
were compiled, and it is not an accident that it was in Cairo that I wrote most of this book.
Most researchers will find my English survey of the Cairo scientific manuscripts more
accessible than the manuscripts themselves. However, the survey is full of references to a
certain SATMI. Here, finally, it is.
TABLE OF CONTENTS
2.8 Tables of the time of night as a function of the longitude of the horoscopus, for a
specific latitude 73
2.8.1 Anonymous: Istanbul 73
2.8.2 Muammad Zamn Mashhad: Meshed 74
2.8.3 Anonymous: unspecified locality 74
Chapter 3: Tables of the longitude of the horoscopus as a function of solar and stellar altitude 75
3.0 Introductory remarks 75
3.1 Tables of the longitude of the horoscopus as a function of the instantaneous solar
altitude and solar longitude, for a specific latitude 75
3.1.1 Abu l-{Uql: Taiz 75
3.2 Tables of the longitude of the horoscopus as a function of stellar altitude, for a specific
latitude 76
3.2.1 Anonymous: Qandahar 76
3.2.2 Anonymous: Baghdad / Damascus 78
3.2.3 Ibn D{ir: Sanaa 79
3.3 Tables of the longitude of the horoscopus as a function of time and solar longitude,
for a specific latitude 80
3.3.1 al-Suln al-Ashraf: Sanaa 80
3.3.2 Abu l-{Uql: Taiz 80
3.3.3 Anonymous: Anatolia 82
3.4 Tables of the longitude of the horoscopus as a function of the time since rising of the
stars, for a specific latitude 83
3.4.1 Ibn D{ir: Sanaa 83
Chapter 4: Tables of solar and stellar altitude 84
4.0 Introductory remarks 84
4.1 Tables of solar altitude at the solstices as a function of time, for a specific latitude 84
4.1.1 Pseudo-al-Khwrizm / abash: various latitudes 84
4.1.2 al-Battn: Raqqa 88
4.1.3 al-Maqs and al-Marrkush: various latitudes 88
4.1.4 al-Suln al-Ashraf: Yemen and Hejaz 88
4.1.5 Anonymous: Istanbul 88
4.2 Tables of solar altitude as a function of time since sunrise and solar longitude, for a
specific latitude 89
4.2.1 Anonymous: Baghdad 89
4.2.2 Anonymous: Baghdad 89
4.2.3 Sa{d ibn Khaff al-Samarqand or Ibn al-@dm: Baghdad 89
4.2.4 al-Marrkush: Cairo 90
4.2.5 al-Suln al-Ashraf: Sanaa 91
4.2.6 Abu l-{Uql: Taiz 91
4.2.7 Najm al-Dn al-Mir: 4th climate (?) 91
4.3 Tables of solar altitude as a function of time since sunrise and meridian altitude, for
all latitudes 91
4.3.1 Anonymous (Baghdad) 92
4.3.1* Muyi l-Dn al-Maghrib (?) (Maragha) 92
4.3.2 al-Marrkush (Cairo) 92
4.3.3 Anonymous (Yemen) 92
4.3.4 Anonymous (Yemen) 93
4.3.5 al-li (?) (Damascus) 94
4.3.6 usayn Qu{a (?) (Tunis) 94
4.3.7 Zacuto (?) (Salamanca) / Anonymous (Yemen) 94
4.4 Tables of solar altitude as a function of the semi diurnal arc and the time since sunrise 94
4.4.1 Najm al-Dn al-Mir (?): Cairo 94
4.5 Tables of stellar altitude as a function of the longitude of the horoscopus, for a specific
latitude 96
4.5.1 Abu l-{Uql: Taiz 96
4.6 Tables of stellar altitude as a function of time of night, for a specific latitude 96
4.6.1 Ibn D{ir: Sanaa 96
table of contents 11
4.7 Tables of solar altitudes in certain azimuths as a function of solar longitude, for a
specific latitude 98
4.7.1 Ibn Ynus: Cairo 99
4.8 Tables of solar altitude in the prime vertical, for a specific latitude 99
4.8.1 Ibn Ynus: Cairo 100
4.8.2 al-Mizz: Damascus 100
4.8.3 al-Khall: Damascus 100
4.8.4 Anonymous: Cairo 100
4.8.5 al-Asy: Assiut 100
4.8.6 al-Maall: Damietta 100
4.8.7 al-Ds: Lattakia 101
4.8.8 Anonymous: Alexandria 101
4.8.9 Ysuf Kilrj: Crete 101
4.8.10 Anonymous: Nablus 101
Chapter 5: Tables of solar azimuth 102
5.0 Introductory remarks 102
5.1 Tables of solar azimuth as a function of solar altitude and solar longitude, for a specific
latitude 102
5.1.1 Ibn Ynus: Cairo 102
5.1.2 Shihb al-Dn al-alab: Damascus 105
5.1.3 Anonymous: Alexandria 105
5.1.4 Anonymous: Damietta (?) 106
5.2 Tables of solar azimuth as a function of solar longitude and solar altitude, for a specific
latitude 106
5.2.1 Anonymous: Cairo 106
5.3 Tables of solar azimuth as a function of meridian altitude and instantaneous altitude 106
5.3.1 al-Khall (Damascus) 107
5.4 Tables of solar azimuth as a function of time and solar longitude, for a specific latitude 107
5.4.1 Pseudo-al-Khwrizm (abash): various latitudes 107
5.4.2 al-Battn: Raqqa 107
5.4.3 Ibn al-@dm or Sa{d ibn Khaff al-Samarqand: Baghdad 107
5.4.4 al-Maqs: Cairo 107
5.4.5 al-Marrkush: Cairo 107
5.4.6 al-Suln al-Ashraf 107
5.5 Tables of solar azimuth at the equinoxes as a function of time, for a specific latitude 108
5.5.1 Anonymous: Cairo 108
5.5.2 Ibn Abi l-Fat al-f: all latitudes 108
5.6 Tables of solar rising amplitude, for a specific latitude 109
5.6.1 Anonymous: Baghdad 109
5.6.2 Ibn Ynus: Cairo 109
5.6.3 Muyi l-Dn al-Maghrib: Maragha 109
5.6.4 al-Marrkush: Cairo 109
5.6.5 al-Mizz: Damascus 110
5.6.6 al-Khall: Damascus 110
5.6.7 Anonymous: Tunis 110
5.6.8 Anonymous: Isfahan 110
5.6.9 Miscellaneous: Cairo 110
5.6.10 al-Asy: Assiut 111
5.6.11 al-Ds: Lattakia 111
5.6.12 Anonymous: Alexandria 111
5.6.13 Ysuf Kilrj: Crete 111
5.7 Tables of rising amplitude as a function of declination, for a specific latitude 111
5.7.1 Anonymous: Cairo 111
5.8 Tables of maximum solar rising amplitude as a function of latitude 111
5.8.1 al-Battn (Raqqa) 112
5.8.2 Abu l-Waf} (Baghdad) 113
5.8.3 Sanjar al-Kaml (Shiraz) 113
12 part i, table of contents
CHAPTER 1
That cultural collaboration had ... become possible for the most different nations Al-Brn
appreciates with full consciousness ... . For him Islam was culture rather than religion, and
the Arabic language the language of Science rather than that of the Qur}n. A. Z. Validi,
Islam and Geography (1934), p. 519.
Wherever in the medieval world there were tables, real astronomy was practiced; where
tables were lacking there were only dilettantes and dabblers. James Evans, Ancient
Astronomy (1998), p. viii.
In this study I present a survey of all known examples of a category of Islamic astronomical
tables1 preserved in scientific manuscripts located mainly in libraries in Europe and the Near
East. Since many of these precious sources are uncatalogued or not properly catalogued, and
also since many of the tables are not clearly associated with an author or do not boast a clear
title, the vast majority of the tables have not been investigated previously in modern times.2
The tables were intended as aids to the solution of the basic problems of spherical astronomy:
given the altitude of the sun or any star, to find the time of day or night, the azimuth of the
celestial body, and the longitude of the ascendant or horoscopus. Such tables, many of which
contain several thousand entries, are generally not contained in the Islamic astronomical
handbooks known as zjes, the profusion and diversity of which were shown by Ted Kennedy
already in 1950s.3 The zjes generally tend to have no tables for timekeeping beyond a set of
1 In the context of the tables and Islamic science in general the term medieval refers to the period from
ca. 750 to ca. 1900. The renaissance of the mathematical sciences took place in the Islamic world in the late
8th, 9th and 10th centuries. Muslim interest in these sciences was progressive until the 15th century, and thereafter
and until the 19th century the Muslims continued to practice them but without further original input of
consequence. This is nowhere more evident than in the tables discussed in this study. See also n. II-1:1.
2 Prior to 1970 the only study of an Islamic table for astronomical timekeeping was Goldstein, Medieval
Table for Reckoning Time from Solar Altitude, which contains an analysis of the table described in 2.3.4. King,
Astronomical Timekeeping in Medieval Cairo, includes an analysis of the tables from the Cairo corpus
described in 2.1.1, 4.7.1, and 5.1.1; idem, al-Khwrizm, pp. 7-11, a discussion of the earliest prayer-tables
for Baghdad (II-3.1); idem, Astronomical Timekeeping in Medieval Syria, and Astronomical Timekeeping
in Ottoman Turkey, and Astronomy in the Maghrib, pp. 37-40, for three regional surveys. An overview is
in the article M_t, ii. Astronomical aspects in EI2, repr. in King, Studies, C-V. An overview in Arabic
the only one of its kind is in the Arabic edition of EHAS, I, pp. 173-238. Several of the tables listed in 7.1
and the auxiliary tables discussed in 9.1, 9.3, 9.5 and 9.6 have been analyzed before: references are given ad
loc. The existence of the tables described in 2.3.1 and 2.3.2 was noted in Suter, MAA, p. 50, and Kennedy, Zj
Survey, p. 161, respectively.
3 Kennedy, Zj Survey, published in 1956 and reprinted ca. 1990, contains a survey of about 120 Islamic
zjes and a partial analysis of several notable examples. The number of individual zjes compiled by Muslim
16 part i, chapter one
standard spherical astronomical tables for the latitude for which the work was compiled: with
a few exceptions I have not treated these standard tables here. I shall, however, have occasional
recourse to the data-base of coordinates of the extensive geographical tables in Islamic zjes
prepared by Ted and Mary Helen Kennedy, which is useful for identifying the localities
associated with latitudes underlying tables when these are not noted in the sources.4
Timekeeping called {ilm al-mqt in Arabic was an important branch of Islamic
astronomy (see Fig. 1.0a).5 In addition to the obvious need for the professional astronomer
to be able to determine the time of day or night accurately, the times of the five daily prayers
in Islam are astronomically defined.6 In a parallel study (II) I discuss numerous medieval tables
for determining the prayer times in different localities, some of which supplement the tables
for timekeeping considered in this study (here occasionally referred to as I).
Previous investigations by Otto Neugebauer have called attention to the simple arithmetical
schemes known from Hellenistic, Byzantine and Ethiopic sources, which display midday
shadows throughout the year or, less frequently, the shadow lengths for each seasonal hour
of daylight.7 In another study (III) I have analyzed over 50 Islamic shadow-schemes of one
kind or another; these are characteristic of the tradition of non-mathematical folk astronomy
which flourished in the Muslim world throughout medieval times.8 Likewise, David Pingree
has drawn attention to tables compiled by Indian astronomers displaying the length of daylight
as a function of solar longitude.9 The Islamic tables investigated in this study represent a far
astronomers between the 8th and the 19th century is closer to 200. King, Islamic Astronomical Tables, first
published in 1975, contains a preliminary account of various categories of Islamic astronomical tables which
are not generally found in zjes. See now the interim report in King & Sams, Islamic Astronomical Handbooks
and Tables, with a summary in the article Zdj in EI2. New insights are to be anticipated from the new survey
of zjes currently in preparation by Benno van Dalen of Frankfurt. On the calendars used by Muslim astronomers
see van Dalens article Ta}rkh. 2. Era chronology in astronomical handbooks in EI2.
4 Kennedy & Kennedy, Islamic Geographical Coordinates, presents ca. 15,000 sets of coordinates for
ca. 2,500 localities from some 75 medieval works, organised according to locality, source, and increasing
longitudes, and increasing latitudes. This important reference work is currently being updated by Merc Comes
in Barcelona.
For the purposes of the identification of localities that could be served by a given medieval value of a latitude,
modern latitudes are often irrelevant and, more often than not, misleading.
5 The best basic survey of Islamic astronomy remains that of Nallino, Islamic Astronomy, published in
1921, which must be read in the light of the later contributions of E. S. Kennedy and his school. For surveys
of Islamic astronomy and astronomy in the service of Islam written for a general audience, see Saliba,
Islamic Astronomy and Astrology; and King, Islamic Astronomy, A-B, and idem, Science in the Service
of ... Islam (first published in 1991), repr. in idem, Studies, C-I. For the broader picture see Hill, Islamic Science
and Technology, and Dallal, Islamic Science, Medicine, and Technology. A useful survey of Islamic scientific
literature in general is Endre, Wissenschaftliche Literatur. A more popular overview in French by a specialist
is Djebbar, Une histoire de la science arabe. It has taken an interested layman to produce the first overview
of Islamic science that could be used in schools or colleges in the English-speaking world: see H. Turner, Science
in Medieval Islam.
6 See further II-1.1 on these definitions, and IV on the motivation behind those for the daylight prayers.
I use the expression timekeeping rather than, say time-measurement (as in msure du temps or mesura
del tiempo) or time-reckoning (as in Zeitrechnung), because, at least as far as the prayers were concerned,
the times were pre-defined (mawqt al-alt), and Muslims have always endeavoured to keep those times.
7 See Neugebauer, HAMA, II , pp. 736-747, for an overview.
8 See King, Islamic Folk Astronomy, and, most recently, Varisco, Islamic Folk Astronomy.
9 See Schmidt, Daylight in Hindu Astronomy, for a detailed description of one such table, and Pingree,
on islamic tables for timekeeping 17
Fig. 1.0a: A beautiful miniature of the astronomers of the Istanbul Observatory ca. 1577. The men are depicted
with various instruments (see X), none of which are known to have survived the vicissitudes of time, but some
of the manuscripts on the bookshelf behind the men holding the astrolabe are now in the University Library,
Leiden. Several manuscripts in that collection bear the mark of ownership of Taqi l-Dn ibn Ma{rf (see Fig.
II-5.1), who is one of those two men. His compilations on astronomical timekeeping are surveyed in the present
work (see I-2.3.6 and II-14.9). [From MS Istanbul UL Yldz F-1404, fol. 57r, courtesy of Istanbul University
Library.]
Sanskrit Astronomical Tables in the US, pp. 72-73, for references to several others. Most Islamic methods
of determining time and azimuth are developments of methods originally adopted from Indian sources. See, for
example, Davidian, Al-Brn on the Time of Day; E. S. Kennedys commentary to al-Brn, Shadows; and,
most recently, Lorch, Sine Quadrant.
18 part i, chapter one
more sophisticated tradition of mathematical astronomy, and although they have no direct
precursors in Greek astronomy, some of them in a sense represent a development of Ptolemaic
tables for spherical astronomy, which included, for example, tables of ecliptic rising times for
the seven climates of classical mathematical geography.10 The influence of the climates in
medieval astronomy and also astronomical instrumentation was far more widespread than has
been recognized hitherto.11 A significant part of the Islamic tradition which I investigate here
was concerned with what I call universal solutions, by which I mean solutions to problems
of spherical astronomy which serve all latitudes; tables and instruments serving all of the
climates are in a sense universal (see further VIa-b).
It has been possible to locate extensive corpuses of tables for timekeeping computed for
each of Baghdad, Cairo, Taiz, Damascus, Jerusalem, Tunis, Alexandria and Istanbul. The
earliest tables are those for Baghdad: some date from the 9th and 10th centuries, when Muslim
astronomers had assimilated Indian and Ptolemaic astronomical ideas and were already making
significant contributions of their own.12 The principal centres of astronomical timekeeping from
the 13th to the 15th century were Cairo and Damascus,13 and from the 16th century onwards,
Istanbul.14 There was also impressive activity in Taiz and Tunis in the late 13th and 14th
centuries. The growth of interest in timekeeping in Cairo lead to the establishment of the office
of the muwaqqit, that is, the professional astronomer associated with a religious institution,
one of whose functions it was to regulate the times of prayer. I have surveyed the activities
of the muwaqqits in another study (V). But as we shall see, the tables compiled by certain Syrian
and Egyptian muwaqqits between the 13th and 16th centuries relate not only to the times of
prayer but also to astronomical timekeeping in general and they are particularly impressive
from a mathematical point of view. The Yemeni and Tunisian tables form part of sophisticated
traditions of mathematical astronomy in these regions which have only recently been
10In Almagest, II.8 (pp. 100-103) the individual and accumulated rising times of each 10 of the ecliptic
are tabulated, and in the Handy Tables (Stahlman, Handy Tables, pp. 206-242) normed right ascensions and
oblique ascensions are tabulated for each 1 of the ecliptic.
11 See the EI article I_lm by A. Miquel, and on the importance of the climates in instrumentation see
2
King, Astronomical Instruments between East and West, esp. pp. 152, 168-169, and idem, Geography of
Astrolabes, pp. 6-9, as well as VIa-b.
The only Islamic table for timekeeping involving the climates is a 16th-century Egyptian table for twilight
at each half-climate discussed in II-8.2 (illustrated). A much earlier (Andalus?) table for lunar crescent visibility
computed for each of the climates is discussed in King, Islamic Tables for Lunar Crescent Visibility, pp. 197-
207, and also VIa-7 (illustrated).
12 On the foreign influences on early Islamic astronomy see Pingree, Indian Influence, and idem, Greek
Influence, and the same authors article {Ilm al-hay}a in EI2; and Saliba, Arabic Science and the Greek
Legacy. The only survey of early Islamic astronomy is Morelon, Early Eastern Arabic Astronomy.
13 Several Cairo and Damascus astronomers associated with timekeeping and their major works are listed
in Suter, MAA; Brockelmann, GAL; and Cairo ENL Survey. For an overview of their activities see King,
Astronomy of the Mamluks.
14 The Ottoman tradition in mathematical astronomy was until recently documented only in a few specialized
studies by Aydn Sayl and Sevim Tekeli and the art-historian Sheyl nver, and various symposium
publications such as Istanbul 1977 Symposium Proceedings and Istanbul 1991 and 1994 Symposia Proceedings.
An overview of Ottoman activity in timekeeping is in King, Astronomical Timekeeping in Ottoman Turkey.
We now have the six impressive volumes of references to hundreds of Ottoman scientific scholars and their
works listed as ~hsanolu et al., Ottoman Astronomical / Mathematical / Geographical Literature.
on islamic tables for timekeeping 19
investigated for the first time.15/16 We shall be constantly confronted with the fact that
astronomy in Islamic civilization did not progress uniformly, and that eventually, when all the
relevant problems had been solved, some many times over, all initiative dried up. What any
given astronomer might know in the late period was to some extent a matter of chance. It
depended not least on his location, and the authorities locally recognized. The works of
astronomers in the period from the 8th to the 11th period were not generally available, and
some of them are known now thanks only to the labours of a series of orientalists.
The reader will observe that I have been able to locate virtually no material from al-
Andalus17 and rather little from medieval Iran, Central Asia and India. In al-Andalus at least,
mathematical astronomy was apparently directed towards other ends. Yet there must have been
some activity in this field: how else can one explain the appearance in the early 15th century
of a table of solar altitude as a function of the date and the time of day computed for the latitude
of Baez (see 10.1)? Were corpuses of tables for timekeeping available for Rayy and Nishapur,
Bukhara and Samarqand? I have no evidence that they were, but find it hard to believe that
they were not. The rich collections of astronomical manuscripts preserved in Iran remain
virtually untouched by modern scholarship,18 but those manuscripts of Iranian provenance
which I have examined elsewhere (including the former Uzbek S.S.R. and India) contain little
material on timekeeping.19 However, I do not doubt that further investigation of the vast
manuscript sources available for the study of the history of Islamic astronomy will bring to
light other examples of the various categories of tables described in this part of my study.
Indeed, no-one will be happier than myself to see a supplement to the present work.20
Another question that must be posed is: were these tables ever used? I have no evidence
(like a sentence here or there in a historical source) that they were. The fact that some of them
were copied dozens of times over several centuries is all the indirect evidence that we have.
Furthermore, the Muslim astronomers are generally silent about the methods used to compile
their tables, the interpolation schemes they employed,21 and the nature of the tables for
sexagesimal arithmetic22 as well as of the trigonometric tables which they had at their
disposal.23 Some examples are shown in Figs. 1.0b-f. All that we have is the tables for
15 See King, Astronomy in Yemen, based on over 100 Yemeni astronomical manuscripts.
16 For a first attempt to survey the Maghrib tradition of astronomy see the second version of King,
Astronomy in the Maghrib, parts of which are already superceded by various studies of the Barcelona School,
including Sams, Maghrib Zjes, and notably Mestres, Zj of Ibn Isq.
17 I use this term to indicate that part of the Iberian Peninsular under Muslim domination at any given time.
18 This is not to say that they are not catalogued, and much credit is due to Iranian scholars (especially al-
{@mil, A{yn al-Sh{a (56 vols., published 1935-1962), and @gh Buzurg, al-Dhar{a (25 vols., published 1965-
1978), who have surveyed vast quantities of manuscripts mainly in Iranian collections. But tables of the kind
studied here seldom bear titles that a cataloguer would pick up.
19 The standard bibliographical reference work on Iranian and Indian astronomy in Persian, listed as Storey,
PL, II:1, contains no references to any works of consequence dealing with timekeeping. Likewise the A{yn al-
Sh{a and al-Dhar{a il tanf al-Sh{a (see previous note), contain innumerable references to astronomical
works in Persian.
20 See the text to n. 1:51.
21 On Islamic interpolation schemes see Hamadanizadeh, Medieval Islamic Interpolation Schemes, and the
references there cited, as well as Rashed, Mthodes dinterpolation. On interpolation in tables of timekeeping
see II-5.8. Some primitive schemes are discussed in II-8.7 and 11.9. The short notes in MS Cairo Zakiyya 917,4
would be worth investigating: see n. II-10:41.
22 See King, Medieval Islamic Multiplication Tables, A-B.
23 On Islamic trigonometric tables see the ground-breaking studies of Carl Schoy listed as Beitrge zur
20 part i, chapter one
Fig. 1.0b: An extract from a sexagesimal multiplica- Fig. 1.0c: An extract from a table of sexagesimal
tion table displaying products m n for m = 0;1, 0;2, quotients displaying quotients m n for m = 1, 2,
... , 59;59, 60;0 and n = 1 , 2, ... , 60. This extract ... , 60 and n = 1, 2, ... , 120. This extract shows
shows products for m = 14;21, 14;22, ... , 14;40, and quotients for m = 31 and 32. This table is unique
n = 1, 2, ... , 60. This manuscript was copied ca. 1700. of its kind. [From MS Cairo DMF 8,1, fols. 23v-
Smaller multiplication tables with integral m are 24r, courtesy of the Egyptian National Library.]
attested from the 10th century onwards. [From MS
Paris BNF ar. 2552, fols. 44v-45r, courtesy of the
Bibliothque Nationale de France.]
Fig. 1.0d: An extract from the sine tables attributed to the Fig. 1.0e: An extract from some trigonometric tables
10th-century Cairene astronomer Ibn Ynus. The table from Mamluk Egypt. This extract is from a table of
gives values of the sine to base 60 for each minute of each cotangents to base 12 and values are given to three
degree, with first differences for each second. Entries in sexagesimal figures for each minute. This double-
the main table are to five sexagesimal digits, but they are page serves arguments 42-47. The tables were
based on the values for each 0;10, which are taken from copied by the Cairo muwaqqit Sayf al-Dn Silmish
his major work, the kim Zj. Their attribution to Ibn ca. 1450. [From MS Cairo M 80, fols. 23v-24r,
Ynus cannot be considered certain, given the penchant of courtesy of the Egyptian National Library.]
Mamluk astronomers for stretching tables in this way
(see II-7.2). The extract shows the entries for 22 and 23.
[From MS Berlin 5752, fols. 13r-14r, courtesy of the
Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
on islamic tables for timekeeping 21
Fig. 1.0f: An extract from the sine tables compiled by Ulugh Beg in Samarqand ca. 1430. Entries are to five
sexagesimal digits, with values for each minute of arc. This extract shows the first few entries in the table, serving
0-13 minutes over 0, 1 and 2. These tables were often copied separately from the zj. [From an unidentified
Cairo manuscript (one of those listed in Cairo ENL Survey, no. G49, 2.2.1), courtesy of the Egyptian National
Library.]
timekeeping themselves, and detailed analysis of error patterns would doubtless furnish
additional information on medieval computational techniques. The analysis of medieval tables
has made vast strides in recent years.24 See further 1.5 and II-1.5, also II-10.10.
Until a survey has been made of medieval and Renaissance European tables for spherical
astronomy it will not be possible to ascertain whether any of the different categories of Islamic
tables found their way to medieval Europe. There is, alas, no survey of medieval European
astronomical tables in general, arranged either chronologically or regionally, nor has any
arabischen Trigonometrie, Die trigonometrischen Lehren des al-Brn, and Schattentafeln (on the cotangent
and tangent functions), as well as numerous others reprinted in his Beitrge, and the following more recent
studies: Kennedy, Overview of the History of Trigonometry; idem, Zj Survey, pp. 139-140; Berggren,
Islamic Mathematics, pp. 127-156.
24 On the first attempts the reader may consult Gingerich, Applications of Computers to the History of
Astronomy; Kennedy, The Digital Computer and the History of the Exact Sciences; and King, Astronomical
Timekeeping in Medieval Cairo, p. 353.
On more recent activity see van Dalen, Ancient and Mediaeval Astronomical Tables, and idem, Statistical
Method for Recovering Parameters; Van Brummelen, Mathematical Tables in Ptolemys Almagest; idem &
Butler, Interdependence of Astronomical Tables; Mielgo, Analyzing Mean Motion Tables; and King, Mecca-
Centred World-Maps, pp. 163-168.
22 part i, chapter one
attempt been made to prepare one.25 However, the very first medieval European manuscript
that I looked at (in Gotha in 1973) and several others that I have seen in recent years (mainly
during the 1990s) also contain tables for timekeeping, and it is clear that further research in
this field would be worthwhile (see already 10.1-2).
For inadequacies in the presentation of the material in I-II, I beg the indulgence of the reader.
The problems associated with research on sources scattered between Princeton and Tashkent,
and between Dublin and Sanaa, should be obvious. Not all of the libraries where the sources
are housed have facilities for photographing, or even for photocopying manuscripts, and some
of those that do have such facilities are administrated by bureaucrats, in-house or in-capital,
who are too stupid to provide such services to researchers or who are so clever that they charge
outrageous amounts for a single photo. I make no apologies for the organisation of the material
presented: it seemed a good idea at the time to arrange the spherical astronomical tables
according to their structure and the prayer-tables and the corpuses in which they are found
chronologically according to provenance. Also, all this was done in the 1970s, when one worked
with type-writers and revised texts using a technique known as cut and paste. In this first
part of my study the reader should bear in mind that the standard tables of spherical astronomy
also occur in zjes, and for these only unusual tables relating specifically to timekeeping have
been included here.
This study owed its inception to the fact that the manuscripts in Berlin and Paris had long
been catalogued in an excellent fashion, and that the descriptions of tables were sufficiently
detailed to convey a good idea of the nature of the tables themselves. It owed its continuation
and fruition to the fact that for several years I had unlimited access to manuscripts in Cairo
and Istanbul. As I prepare the final version of this text I feel confident that most of the materials
preserved in the libraries of those two cities, as well as the major European libraries, have been
included. Libraries in the Maghrib, Iran and India have been exploited to a much lesser extent,
and it may well be that new tables unknown to me now will one day be found there. Any future
researcher would do well to start with some materials, mainly anonymous, listed by Ekmeleddin
~hsanolu and his colleagues in Istanbul.26 That researcher, if diligent, will find that some of
these are indeed already treated in the present work, others maybe not. In any case, the present
work should serve as a framework for controlling, classifying, and setting a context for any
significant new sources.
25 There are, of course, numerous valuable studies of particular corpuses of tables by scholars familiar with
the manuscript sources. See, for example, Toomer, Toledan Tables, and now F.S. Pedersen, The Toledan
Tables; Poulle, Studies; Goldstein & Chabs, Zacuto. See also North, Horoscopes and History, on the way in
which astronomical tables were used to compile horoscopes, and Chabs, Cahier dun croisier, on an
astronomers notebook.
26 See ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 729, 793-795, 801, 802, 804, 805, 836, 869-
870, for various tables relating to timekeeping. For one table mentioned by ~hsanolu et al. which is highly
important and of a type not attested elsewhere see 8.5.1* and 9.4* below.
on islamic tables for timekeeping 23
In this text notation in the form (1.2.3) stands for (see Section 1.2.3). Occasionally here
in I there are also cross-references to II in the form (II-1.2.3), and vice versa.
In the mathematical analysis of tables I use the notation f(x,y) to indicate that the function
f is tabulated with x as the horizontal argument and y as the vertical argument. The following
notation is used freely in the sequel, On the auxiliary functions for timekeeping (B, C and G),
and for calculating ascensions (e) and the azimuth (k and L) see further Chs. 6-9.
Latin
a azimuth (usually measured from the prime vertical)
d equation of half-daylight (= D 90)
B an auxiliary function for timekeeping, called absolute base, related to cos
cos
C an auxiliary function for timekeeping related to sin sin
D semi diurnal arc (sun) or semi nocturnal arc (stars)
e the auxiliary function tan tan
G an auxiliary function for timekeeping related sec sec
h instantaneous solar or stellar altitude
number of degrees in one seasonal day-hour (= 1/12 2D)
number of degrees in one seasonal night-hour (= 1/12 2N)
ha solar altitude at the beginning of the afternoon prayer
h0 altitude in the prime vertical
-h equinoctial hours (also used in the representation of functions, thus: 2Dh() =
1/
15 D())
H meridian altitude
H the difference between the Sines of the meridian and instantaneous altitudes (H
= Sin h - Sin h)
k an auxiliary function for azimuth calculations, related to sin h tan
L an auxiliary function for azimuth calculations, related to sin = sin / cos
L terrestrial longitude
N semi nocturnal arc
q qibla, measured from the meridian
-q relates to the qibla
R base for trigonometric functions (usually 60, although also 12 and 7 are used
for Cotangents and Tangents)
-sdh seasonal day-hours
-snh seasonal night-hours
t hour-angle (actually an arc in equatorial degrees, rather than an angle)
T for the sun: time since sunrise or time until sunset; for a star: time since rising
or time remaining until setting
z horizonal shadow of a vertical gnomon (base given in parentheses as subscript)
24 part i, chapter one
Greek
right ascensions
normed right ascensions ( = + 90)
oblique ascensions for latitude
r oblique ascensions of the ascendant at daybreak
s oblique ascensions of the ascendant at nightfall
solar declination
* solar declination augmented by 90 (useful for finding H = *-)
stellar declination
obliquity of the ecliptic
independent variable
solar longitude or ecliptic longitude
see below
* see below
H longitude of the ascendant (horoscopus)
longitude of the co-ascendant (point of the ecliptic which rises with a star)
terrestrial latitude
rising amplitude
Miscellaneous
complement of
longitude measured from the nearer equinox
* longitude of the point on the ecliptic opposite the point with longitude (*
= + 180)
[x] integral part of x
In virtually all of the tables inspected the entries are written sexagesimally in standard Arabic
alphabetical notation (abjad).27 In one important table dating from the earliest period of Islamic
astronomy and in most which date from recent centuries the entries are given in Arabic
numerical notation see Figs. II-3.1 on the former and I-2.2.4 for one example of the latter.
In the analysis the term digit means sexagesimal digit,28 unless otherwise stated, so that,
for example, a number written 23;50,21 represents 23 + 50/60 + 21/3600. In the middle of an Arabic
text this number would be written 23 50 21 seconds. In the tables the number would be written
23 50 51, and the headings would indicate that the three figures are degrees (abbreviated
j for daraj), minutes (q for daq}iq) and seconds (y for thawn).
27 On this notation see Irani, Arabic Numeral Forms, and Destombes, Chiffres coufiques.
28 See, for example, Kennedy, Zj Survey, p. 17, and Berggren, Islamic Mathematics, pp. 39-63.
on islamic tables for timekeeping 25
The medieval trigonometric functions are to base R other than unity (see below). The
functions are here denoted by the standard capital notation;29 thus for a general arc :30
Sin = SinR = R sin ,
Cos = Sin = R cos ,
Vers = R - Cos = R (1 - cos ) , etc.
The Cotangent and Tangent functions were generally, but not exclusively, used with the solar
altitude as argument to represent the horizontal shadow of a vertical gnomon length n and the
vertical shadow of a horizontal gnomon of the same length, respectively. Thus:
Cot h = Cotn h = n cot h ,
Tan h = Tann h = n tan h .
I use Z for Cot H and z for Cot h and occasionally z for Tan h, with the base indicated in
parentheses in the subscript, thus, for example, Z(6;40) stands for Cot6;40 H. I also use za, etc.,
for the shadow at the {ar prayer. The z, for Arabic ill, shadow, should serve as a reminder
that the medieval Cotangent function is of a different nature from the medieval Sine function.
The base used for the Sine and related functions is generally 60, although some of the earliest
sources use the Indian value 150.31 The Cotangent and related functions are usually to base
12 or 7, although in some sources we find 60 (see further below). Most of the sources surveyed
in this study use 12 (digits) or 7 (feet) for the gnomon length n. We shall also encounter
occasional use of 6 or 6;40, intended as variants of 7;32 as well as of 60, the last-mentioned
intended to correspond to the tables of the Sine function. The use of base 5 is not to be compared
with these, for the 5 here is simply 6012.33 However, and this is of greater historical interest
(if not significance), we shall also encounter the use of 1 and 10 in early sources, with entries
expressed sexagesimally.34 Very occasionally, we find 20 used as a base.35
In the analysis, it is to be assumed that all intervals of time are expressed in equatorial
degrees unless otherwise indicated. These may be converted to equinoctial hours (denoted
by superscript h) according to the relation:
T h = T / 15 = 0;4 T ,
or to seasonal day-hours (denoted by superscript sdh) according to:
T sdh = 6 T / D .
In certain Ottoman tables the time of day is given in equinoctial hours according to the
convention that sunset is 12 oclock.36 If T represents the equinoctial hours that have elapsed
since sunrise on a day when daylight is 2D equinoctial hours long, then the time of day T
according to this convention is:
T = 12 - 2D + T .
For the convenience of the reader I list in this section the formulae that underlie the major
tables described in the sequel. To simplify the formulae slightly I use modern notation for
trigonometric functions, that is, base R = 1 rather than 60. Each of these formulae is numbered,
and I refer to them in the analysis by these numbers preceded by F. The formulae underlying
the minor tables discussed in the sequel are presented ad loc. Likewise approximate methods
are discussed only in the analysis. As we shall see, the Muslim astronomers used an array of
auxiliary functions to facilitate their computations: these are introduced in the immediate sequel
in special script, which is thereafter abandoned but used again in XIIb. The formulae for which
they provide simplified solutions are presented in rectangular frames.
The formulae derived below represent methods that would be outlined in words in the
astronomical handbooks known as zjes or in the instructions to the tables themselves.37 At
least from the 13th century onwards lists of formulae of the form a : b = c : d, with four quantities
a,b,c,d arranged in columns were available: see Fig. 1.2a and Table 10.3b. Two examples
are:
(1) the Sine of instantaneous altitude : the Cosine of the instantaneous altitude
= the length of the gnomon : the horizontal shadow,
that is, Sin h : Cos h = n : zn ( = n : Cotn h ) .
(8) the absolute base : the difference between the Sine of the meridian altitude and the Sine
of the instantaneous altitude = unity : the Versed Sine of the hour-angle,
that is, B : (Sin H - Sin h) = 1 : Vers t .
It is perhaps appropriate that for technical reasons the formulae in this study are reproduced
on a single line: in the first computer-generated version of this study they were more visually
impressive, with large curly brackets and multi-level division bars, but with updated hard- and
soft-ware these behemoths actually disappeared, their previous existence being indicated on
my computer-screen by the message Fehler!.38 The present mode of rendition of the formulae
should constantly remind the reader that the equivalent procedures in medieval Arabic texts
37 On Islamic methods in spherical astronomy, see, for example, Neugebauer, al-Khwrizm (al-Khwrizm/
al-Majr), and Goldstein, Ibn al-Muthann on al-Khwrizm; Nallino, al-Battn; King, Ibn Ynus; Sams,
Estudios sobre Ab Nar; Berggren, Spherical Trigonometry in the Jmi{ Zj (Kshyr); Schoy, Die
trigonometrischen Lehren des al-Brn; Debarnot, al-Brns Maqld; Villuendas, Trigonometra de Ibn
Mu{d; Sdillot-pre, Trait (al-Marrkush); Vernet, Ibn al-Bann}; Kennedy, Spherical Astronomy in Kshs
Khqn Zj; and Sdillot-fils, Prolgomnes (Ulugh Beg).
On methods for timekeeping in particular see especially Nadir, Ab al-Waf} on the Solar Altitude; and
Davidian, al-Brn on the Time of Day, and E. S. Kennedys commentary on al-Brn, Shadows. See also
the articles Mali{ [= ascensions] and Mayl [= declination] in EI2.
For a modern discussion of spherical astronomy see Smart, Spherical Astronomy, esp. Ch. II.
38 See n. 14 to the main preface.
on islamic tables for timekeeping 27
Fig. 1.2a: A list of 32 sets of four quantities in spherical astronomy such that the ratio of the first to the second
equals the ratio of the third to the fourth. Notice that all expressions are in words, without any recourse to symbols
of any kind. This list is anonymous but was taken from the magnum opus of the late-13th-century Cairo astronomer
Ab {Al al-Marrkush. [From MS Cairo DM 291,3, fols. 4v-5r, courtesy of the Egyptian National Library.]
involved sentences of several lines. The language of the astronomers of the late medieval period
(13th-15th century) considerably facilitated the formulation of complicated mathematical
procedures: the crisp, highly technical Arabic of al-Khall (Damascus, ca. 1360) and al-
Mridn (Cairo and/or Damascus, ca. 1375), let alone the garbled colloquial of Najm al-Dn
al-Mir (Cairo, ca. 1325), is quite different in vocabulary and style from the Arabic of earlier
scholars such as al-Battn (Raqqa, ca. 910) and Ibn Ynus (Cairo, ca. 990).39
All of the formulae are easily derived either from trivial considerations of the celestial sphere
or from not-so-trivial orthogonal projections of the sphere, using the elegant procedure called
the analemma that was known since classical Antiquity.40 Certain Muslim astronomers
preferred solutions to problems in spherical astronomy using the rules of spherical trigonom-
etry,41 but not those who specialised in timekeeping.
39 The text of al-Battns Zj is published in Nallino, al-Battn, III. For an example of this late medieval
scientific Arabic see the texts of al-Mridns treatises on a quadrant of his own invention and on the use of
his auxiliary tables published in King, al-Mridns Universal Quadrant, and al-Khalls instructions on the
use of his various tables translated in II-10.4, 10.7 and 10.8). For the Arabic of Najm al-Dn see Charette, Mamluk
Instrumentation.
40 On the analemma, see, for example, Paul Luckey, Das Analemma von Ptolemus; Id, Analemma for
Ascensions; Kennedy & Id, abashs Analemma for the Qibla; Carandell, Analemma for the Qibla; E.
S. Kennedys commentary to al-Brn, Shadows; King, Ibn Ynus, passim; Berggren, Four Analemmas for
the Qibla; and, most recently, King, Mecca-Centred World-Maps, pp. 15-16.
41 See, for example, Sams, Estudios sobre Ab Nar; and al-Brn, Maqld.
28 part i, chapter one
the axis OP once every 24 hours, where P is the celestial pole. The segment ZP defines the
local meridian, intersecting the horizon at N and S, the north- and south-points, and the arc
PN measures the local latitude (Arabic, {ar al-balad). The perpendicular plane EZW is the
prime vertical, intersecting the horizon at E and W, the east- and west-points. The great circle
EQW with its pole at P is the celestial equator.
The sun and stars participate in the apparent motion of the celestial sphere and are observed
to traverse small circles parallel to the celestial equator, called declination circles or day-circles
(madrt). A given celestial body rises at A, culminates at B on the meridian, sets at C, and
crosses the meridian below the horizon at D; its day-circle is ABCD.
on islamic tables for timekeeping 29
The position of the day-circle relative to the celestial equator is determined by the declination
of the celestial body, measured by arc BQ. I use to denote declination in general or stellar
declination in particular (al-bu{d) and only to denote solar declination (al-mayl). Observe
that the meridian altitude H (ghyat al-irtif{), measured by arc SB, and the meridian altitude
below the horizon H* are related to the local latitude and the declination by:
F1 H = + and H* = - .
Some Muslim astronomers tabulated *() = () + 90, with which H() = *() - .
Two other quantities related to the latitude and declination are the rising amplitude (sa{at
al-mashriq), measured by arc EA on the horizon, and the altitude in the prime vertical h0 (al-
irtif{ alladh l samt lahu), measured by arc EL on the prime vertical or east-west colure,
which is the great-circle ELZW perpendicular to the meridian.
The most important problem of spherical astronomy is, given the instantaneous altitude h
(al-irtif{) of a celestial body at X, to determine the time since rising T (al-d}ir), measured
by the arcs AX on the day-circle or FT on the celestial equator, or the hour-angle t (fal al-
d}ir), measured by the arcs XB or TQ. These times are clearly related by the simple formula:
F2 T+t=D,
where D is half the diurnal arc (nif qaws al-nahr), measured by arcs AB or FQ. The arc
FE is the equation of half-daylight d (ta{dl nif qaws al-nahr) and clearly:
F3 D = 90 + d .
When the sun or star is at G, the intersection of arcs AB and PE, the time since rising is d
and the hour-angle is 90. Another important problem of spherical astronomy is to determine
the azimuth a of the celestial body (samt), measured by the arc EK on the horizon, where ZXK
is the altitude circle of the celestial body.
Solar coordinates
As a result of the motion of the earth about the sun, the sun appears to move against the
background of fixed stars on a great circle called the ecliptic which is inclined to the celestial
equator at an angle 231/2 called the obliquity (al-mayl al-a{am). The sun moves along
the ecliptic in a direction opposite to that of the apparent daily rotation and completes its journey
around the ecliptic in a year.
When the sun is at one or other of the points of intersection of the ecliptic and celestial
equator it rises at the east-point and sets at the west-point, and day is equal to night: these
solar positions are the vernal and autumnal equinoxes. General solar positions on the ecliptic
are determined by the suns longitude (darajat al-shams), which is measured from the vernal
equinox V. In Fig. 1.2c the longitude of the sun at X is measured by the arc VX.
The solar longitude at a given time can be determined either from the tables standard in
zjes, or from an ephemeris showing solar and other planetary positions for each day over a
period of several years.42 Another kind of table (called in Arabic shabaka) was common in
tables for timekeeping: these show the solar longitude for each day of a period of four Syrian
42
On the solar tables in zjes see Kennedy, Zj Survey, p. 141. On Islamic ephemerides see the EI2 article
Ta_wm. i by Michael Hofelich.
30 part i, chapter one
Fig. 1.2c.
Fig. 1.2d: An extract from a Tunisian table showing the solar longitude as a function of the date in the Syrian
(solar) calendar over a four-year cycle. This precedes the set of auxiliary tables for timekeeping described and
illustrated in 9.7. The tables were compiled about 1400 and the manuscript copied about 1600. [From MS Cairo
DM 689, fols. 1v-2r, courtesy of the Egyptian National Library.]
or Coptic years (to cover the leap year) at a particular epoch.43 For an example, see Fig. 1.2d.
As the sun moves around the ecliptic from the vernal equinox its declination to the north
of the celestial equator increases and the length of daylight also increases, reaching a maximum
when = 90 and = + at the summer solstice. Likewise, as the sun moves past the autumnal
equinox its declination to the south of the celestial equator increases and the length of daylight
decreases, reaching a minimum when = 270 and = - at the winter solstice. The declination
is measured by the arc XD drawn perpendicular to the equator. The ecliptic is divided into
twelve zodiacal signs of 30 each, six northerly and six southerly. Tables of () are standard
in zjes.
43 The only published example of a shabaka table is that of al-Marrkush computed for Cairo, 992 Diocletian
[= 1275/76], reproduced in Sdillot-pre, Trait, I, pp. 136-137. See also nn. I-9:26 and II-10:19.
on islamic tables for timekeeping 31
Fig. 1.2e.
Fig. 1.2f: The star table of Ibn Ynus, missing from the surviving fragments of his zj, was discovered during
in the 1970s in an Egyptian manuscript copied ca. 1475 by Amad ibn Timurby. It displays longitudes and
latitudes for 59 stars for epoch 400 Yazdigird, that is 1032 C.E., some 30 years after the death of Ibn Ynus.
(His solar, lunar and planetary tables run up to 1800 Yazdigird, that is, 2432 C.E., so he clearly was in the habit
of planning ahead.) The table has been edited but has not yet been published. [From MS Cairo MM 188,1,
fol. 72v, courtesy of the Egyptian National Library.]
The position of the sun relative to the celestial equator is measured by the declination
perpendicular to the equator and a coordinate called right ascension (mali{ darajat al-
shams) measured as VD along the equator from the vernal equinox V. Tables of () are also
standard in zjes. Another useful notion in spherical astronomy is the second declination 2
(al-mayl al-thn), which is measured perpendicular to the ecliptic and represented by the arc
XC. This quantity, occasionally tabulated in zjes, is of use in the conversion of ecliptic and
equatorial coordinates see below.
Stellar coordinates
The positions of the stars in the heavens are conveniently measured with respect to the ecliptic
or the celestial equator by two systems of orthogonal coordinates. In the first the coordinates
32 part i, chapter one
are longitude and latitude, and , and in the second they are right ascension and declination,
and see Fig. 1.2e. The phenomenon of precession causes stellar longitudes slowly to
increase but does not affect the latitudes; both the right ascension and declination of the stars
vary slowly with time. Islamic star catalogues display coordinates of prominent stars in one
or other system, sometimes both, for a particular epoch: for an example see Fig. 1.2f.44 Zjes
generally contain instructions on methods of converting from the one coordinate system to the
other.45 For calculating the time by night it is necessary to know the coordinates of the stars
in the equatorial system.
Fig. 1.2g.
44 Kennedy, Zj Survey, p. 144a. The earliest Islamic star-catalogues from 9th- and 10th-century Baghdad
are investigated in Girke, Die frhesten islamischen Sternkataloge, alas not yet published. On the most
significant star-catalogue of late Islamic astronomy see Knobel, Ulughbegs Catalogue of Stars.
45 The conversion of stellar coordinates from one system to the other can also be effected using an instrument
such as the afa of the 11th-century Andalus astronomer Ibn al-Zarqlluh (Azarquiel), which was specifically
designed for that purpose. See the article Shakkziyya in EI2 and the references there cited, and also X-5.2.
on islamic tables for timekeeping 33
Now we rotate the ecliptic into the working plane so that V and X move to V2 and X2. Clearly:
arc V2X2 = and OX1 = sin .
Next we rotate the declination circle about its axis EF into the semicircle EHF with midpoint
at H on the working plane so that X moves to X3. Clearly:
arc HX3 = and OY = GX1 = cos sin .
Now in X1OY we have:
O = , OX1 = sin , X1Y = sin and OY = cos sin .
It follows immediately that:
F4 = arc sin (sin sin )
and:
F5 = arc sin { tan / tan } .
Note that the second declination 2() equals the declination corresponding to a longitude
whose right ascension is , that is:
2() = () where () = .
It follows from F5 that:
= arc sin { tan () / tan } = arc sin { tan 2() / tan } ,
whence:
F6 2() = arc tan { tan sin } .
Now the segment A1D below the horizon measures the projection of the day-circle above the
horizon for declination -. For the sun A1D represents the projection of the day-circle for
longitude * = + 180 (nar darajat al-shams), so that:
H*() = H(*) .
Next we rotate the celestial sphere about its axis ZO until X crosses the meridian at Y1 and
draw the perpendiculars X1X2 and Y1Y2 from X1 and Y1 to NS. Then since:
arc Y1S = h ,
we have:
X1X2 = Y1Y2 = sin h and Y2O = cos h .
Similarly:
L1O = sin h0 .
34 part i, chapter one
Fig. 1.2h. The celestial sphere reduced to two dimensions. Compare Fig. VIIa-3.5.
We now rotate the day-circle about its axis G1B into the working plane so that A, G and X
move to A3, G3 and X3. Then:
arc A3B = D, arc A3G3 = d, arc A3X3 = T, and arc X3B = t ,
and since:
G1B = cos ,
we have:
A1B = vers D cos ,
A1G1 = sin d cos ,
on islamic tables for timekeeping 35
so that:
F10 t = arc vers { vers D - sin h / ( cos cos ) }
= arc vers { vers D - sin h vers D / sin H } .
Alternatively, since also:
A1X1 = { sin d + sin (T-d) } cos ,
we have:
F11 T = d + arc sin { sin h / ( cos cos ) - sin d }
= d + arc sin { sin h vers D / sin H - sin d } .
Likewise in X1VG1, where G1V is perpendicular to X1X2:
X1 = and X1G1 = cos t cos ,
so that:
X1V = cos t cos cos .
But:
X1V = X1X2 - G1G2 = sin h - sin sin ,
so that:
F12 t = arc cos { [ sin h - sin sin ] / ( cos cos ) } .
This last is the modern formula for t(h,,). Notice that:
t = arc cos { [ sin h - C ] / B }
Finally, in L1OG1 we have:
L1 = , LO = sin h0, and OG1 = sin ,
whence:
F13 h0 = arc sin { sin / sin } .
Azimuth
To derive the azimuth a as a function of h, and , we project the celestial sphere onto the
horizon plane and rotate the projection about its axis SN into the semicircle SKEDN in the
working plane. Points X and A move to X4 on OK and A4 on the semicircle SEN, respectively,
and clearly:
arc EA4 = and OA1 = sin .
(For convenience we adopt the convention that \ O as \ 0.) But in OA1G1 we have
also:
O = and OG1 = sin ,
whence:
F14 = arc sin { sin / cos } .
Furthermore, arc EK = a (now with the convention that a \ 0 as X is north or south of E),
and since:
OX4 = OY2 = cos h ,
we have:
OX2 = cos h sin a .
Now in X1X2A1:
X1 = and X1X2 = sin h ,
on islamic tables for timekeeping 37
so that:
A1X2 = sin h tan .
This quantity is of importance in medieval azimuth calculations and I denote it by k . With
k = sin h tan ,
we have further:
OX2 = cos h sin a = A1X2 - OA1 = k - sin ,
whence we derive the standard medieval formula:
F15 a = arc sin { [ sin h tan - sin / cos ] / cos h } ,
or, in terms of the auxiliary function:
a = arc sin { [ k - sin ] / cos h }
which is equivalent to the modern formula:
a = arc sin { [ sin h sin - sin ] / ( cos h cos ) } .
Finally, note that in X1ML1, where L1M is perpendicular to X1X2:
X1 = and X1M = X1X2 - L1O = sin h - sin h0 ,
so that:
OX2 = L1M = (sin h - sin h0) tan .
But:
OX2 = cos h sin a ,
so that:
F16 a = arc sin { [ sin h - sin h0 ] tan / cos h } ,
as given in some medieval sources (see 5.3).
Horoscopus
The point of the ecliptic instantaneously rising over the local horizon (H in Fig. 1.2h) is called
the ascendant or horoscopus (al-li{). The point where the ecliptic intersects the meridian is
46 See the EI2 article Mali{.
38 part i, chapter one
called midheaven (wasa al-sam}). The position of the ecliptic relative to the horizon and
meridian is of prime importance in astrology, but also, needless to say, in timekeeping.
If we now denote the longitude of horoscopus H by H when the sun at X has altitude h
and longitude , it follows from a simple consideration of the celestial sphere that:
F20 (H) = () + T(h,) .
Likewise for a star with co-ascendant and declination :
F21 (H) = () + T(h,) .
Note further that if M denotes the longitude of upper midheaven M, then:
F22 (H) = (M) + 90 = (M) ,
where denotes normed right ascensions, measured from the winter solstice rather than the
vernal equinox, and more commonly used in Islamic timekeeping than . See further IX.
In the sequel the reader should bear in mind that the tables of:
(), 2() and () or () ,
as well as tables of:
H(), d(), D() and () for some particular latitude,
were standard in Islamic zjes.50 In these tables values were usually given to two sexagesimal
digits, occasionally three, for each degree of . As noted already, I do not include such common
(not in the sense of vulgar) tables in this survey. I have endeavoured to arrange the analysis
of other Islamic tables for astronomical timekeeping in such a way that new material can be
easily classified and included in any future supplement.51
In Ch. 2 I present examples of five kinds of tables for finding the time of day from the
solar altitude and two for finding the time of night by the stars:
1. T(h,) and t(h,) for specific
2. T(,h) and t(,h) for specific
3. T(H,h) and t(H,h) for specific
4. t(h) for various stars and specific
5. T(H,h) for all
6. T(H,h,D) for all
7. T(,) for specific
In Ch. 3 I present examples of four kinds of related tables for finding the longitude of the
horoscopus from solar or stellar altitudes or from the time of day or night:
1. H(h,) for specific
2. H(h) for various stars and specific
3. H(T,) for specific
4. H(T) for various stars and specific
In Ch. 4 we consider examples of eight different kinds of tables for finding the altitude of
the sun or fixed stars, given the time, or the longitude of the horoscopus, or the azimuth:
1. h(T) at the equinoxes and solstices for specific
2. h(T,) for specific
3. h(T,H) for all
4. h(D,T) for specific
5. h(H) for various stars and specific
6. h(T) for various stars and specific
7. h(a,) for specific
8. h0() for specific
Ch. 5 contains a discussion of examples of seven different kinds of tables for finding the
azimuth of the sun or the stars:
1. a(h,) for specific
2. a(,h) for specific
50 See Kennedy, Zj Survey, pp. 140-141, and King, Ibn Ynus, II.3.
51 See the text to n. 1:20.
40 part i, chapter one
In general the Muslim astronomers took advantage of the symmetry of the functions they were
tabulating, in order to avoid pointless duplication of entries. On the behaviour of the various
functions see my previous study of the Cairo corpus, where various graphs are presented.52
For functions such as (), as well as d() and () for particular latitudes, whose absolute
values are symmetrical about the equinoxes and solstices, 90 entries suffice to display the
function for each degree of . Tables of these functions were usually arranged in three columns
of 30 entries, often with a single vertical argument running from 1 to 30, understood to represent
1-30, or 31-60 or 61-90 as necessary.
For functions such as H() and D() which are symmetrical about the solstices but not the
equinoxes, two sets of 90 entries and an additional (non-zero) equinoctial value display the
function for each degree of . Tables of these functions were usually arranged in six columns
of 30 entries, the first three serving the northern signs and the second three the southern ones.
A vertical argument running from 1 to 30 downwards serves the first and third quarters of the
ecliptic, and a second one running from 0 to 29 upwards serves the second and fourth quadrants.
The equinoctial value is thus omitted (but is generally obvious anyway). Alternatively, since
both H and D increase monotonically between the winter and summer solstice, the tables might
be arranged so that the first three columns serve the southern signs and the second three columns
the northern ones. In this case, if the arguments run from 1 to 30 downwards and 0 to 29
upwards, the value of the function at the winter solstice will be omitted.
The functions t(h,), T(h,), a(h,) and h(a,) display the same symmetry as H() and D(),
and for a given value of the first argument a page or double page of tables serves to display
the function for each degree of . See, for example, Figs. 2.1.1a and II-5.6a. If either of the
first three functions are tabulated with as the horizontal argument, a given page or double
page might serve each degree of from, say, the winter solstice to the summer solstice, but
now the maximum vertical argument is H or [H] and hence varies with . See, for example,
Figs. 2.2.1-3.
52 See King, Astronomical Timekeeping in Medieval Cairo (n. 1:2), esp. pp. 361, 363, 365, 367 and 370.
on islamic tables for timekeeping 41
The functions () and () are not symmetrical about the equinoxes or the solstices. Tables
of these functions were usually arranged in 12 columns of 30 entries, with a single vertical
argument running from 1 to 30 or from 0 to 29, with each column essentially serving a single
zodiacal sign. The tables of H(h,) are of the same kind, and for a given value of h a double
page of tables with 12 columns of 30 entries serves to display the function for each degree
of . See, for example, Fig. 3.1.1.
The various functions f(H,h) discussed in the sequel are computed for a given range of H
and each degree of h up to H. The resulting table, which is usually spread over several pages,
is trapezoidal in shape. Such tables were generally called in medieval Arabic aylasn, which
means shawl.53 See, for example, Figs. 2.3.1-2.
For other large tables the format is devised to best suit the function tabulated, often with
astounding ingenuity. See, for example, Figs. 2.6.1 and 4.4.1.
One mans error is another mans data. Bermans corollary to Roberts axiom.
... (the) non-existence (of al-Khzins qibla-table) does not prevent King from specifying
its characteristics in considerable detail. This level of theoretical construction makes a
reader seriously question imposing on data an order that might not be fully justified by
the medieval texts (some of which are not even extant). E. Savage-Smith, Review of
King, Mecca-Centred World-Maps, p. 34c.
Most of the tables discussed below were recomputed during the period 1970-74 with the
electronic computers and myriads of punched cards at the American University of Beirut, Yale
University or the American University in Cairo.54 In certain cases, the computer was useful
in facilitating the analysis of the tables and the determination of the underlying parameters.
Very occasionally, one may be able to reconstruct a table that does not survive in its original
form, or at least comment on its structure and the accuracy of its entries, using in both cases,
other derivative tables that have survived.55 Such techniques may inevitably be beyond the
53 On this term see Goldstein, Medieval Table for Reckoning Time, p. 61; and Albert Arazi, Noms des
vtements daprs al-Adt al-isn f fal al-aylasn dal-Suy, Arabica 23 (1976), pp. 109-155. In Yedida
Kalfon Stillman, Arab Dress from the Dawn of Islam to Modern Times A Short History, Norman A. Stillman,
ed., Leiden, etc.: E. J. Brill, 2000, pp. 18, 39, 44, 48, 51-52, 71, 73-74, 91, 96 and 179, there no mention of
a trapezoidal-shaped shawl.
The term is uncommon in the medieval astronomical literature, apart from the titles of this kind of table.
In fact, I know of only two other attestations, both by al-Brn (see II-2.2). In his treatise on the construction
of astrolabes entitled al-Ist{b (MS London BL Or. 5593, fol. 56v), al-Brn poses the question whether a
particular variety of non-standard astrolabe leads to the knowledge of the hours and the ascendants in the same
way as the aylasn table and different kinds of sundials do. (This statement is recorded in Wiedemann, Aufstze,
II, pp. 523-524; see also his n. 10 on p. 528.) Again, al-Brn in his Chronology (p. 132 of Sachaus translation)
uses the term aylasn to describe a table for calendar conversion. (This reference is already recorded in Kennedy,
Zj Survey, p. 137, ad no. X201.)
54 Computer time was made available in Beirut by the Mathematics Department, American University of
Beirut; in New Haven by the Department of Near Eastern Languages and Literatures, Yale University; and in
Cairo by the Smithsonian Institution, Washington, D.C.
55 See, for example, Hogendijk, al-Khwrizms Sine Table, and King, Mecca-Centred World-Maps, pp.
71-75.
42 part i, chapter one
56 See various recent studies by Glen Van Brummelen and Benno van Dalen listed in n. 1:24.
tables of time as a function of solar and stellar altitude 43
CHAPTER 2
The functions tabulated in the tables described in this chapter are either the time since rising
of the sun or star T (for altitudes in the east), or the hour-angle t. These are related by the
simple formula (cf. F2):
T+t=D,
where D is the semi diurnal arc. The Arabic terms for the time since rising and hour-angle
are al-d}ir and fal al-d}ir, respectively.
In those tables for the sun in 2.1 and 2.2 the arguments are h and . The tables discussed
in 2.3 and 2.5 have H and h as arguments, and in the case of those for the stars in 2.4 the
argument is h, but a value of H underlies the table for each star. The tables of 2.5 display the
time approximately in seasonal day-hours for all latitudes. The table described in 2.6 has three
arguments H, h and D and serves both the sun and stars and works for any latitude. The
tables for timekeeping by the stars discussed in 2.7 are of a different kind, giving the time
as a function of the ascension of the star that is culminating and the solar longitude . Those
in 2.8 take advantage of a convenient temporary celestial phenomenon. In most of the tables
time is measured in equatorial degrees.
See II for the other, smaller tables which accompany such tables as these for Cairo, Mecca,
Damietta, Jerusalem, Damascus, Aleppo, Anatolia and Istanbul.
2.1 Tables of time as a function of instantaneous solar altitude and solar longitude, for a specific
latitude
In this section I note examples of tables of t(h,) or T(h,) for particular latitudes. Since both
of these functions are dependent on , and since two points on the ecliptic equidistant from
the solstices have the same declination, for a given altitude 181 values suffice to display either
function for each integral degree of solar longitude ( = 0 and = 1, 2, ... , 90, \ 0).
In the tables described below, full advantage is taken of the symmetry of the functions t(h,)
and T(h,). For given h, generally six columns of 30 entries display the tabulated function.
When for certain longitudes the sun does not attain the altitude in question, no entry is given
in the table, and columns that would otherwise be devoid of entries are omitted. This format
results in the omission of either the equinoctial value or a solstitial value. The format of the
tables described in 2.2 obviates the need to prepare additional tables for either the equinoxes
or the solstices (2.1.1 and 5.1.1). Each of the tables of t or T described in this section and in
the next contains over 10,000 entries.
44 part i, chapter two
A few Muslim astronomers used tables of t(h,) or T(h,) to compile other tables for the
duration of twilight,1 based on the assumption that this phenomenon begins or ends when the
sun is at a particular angle of depression below the horizon see Figs. 2.1.1a-b and II, passim.
1
See the article Shafa_ on twilight in EI2.
2
This corpus was analyzed for the first time in King, Ibn Ynus Very Useful Tables for Reckoning Time
by the Sun, AHES 10 (1973), pp. 342-394, repr. in idem, Studies, A-IX (hereafter Astronomical Timekeeping
in Medieval Cairo). However, that early study was based on a set of manuscripts in which the tables are mainly
attributed to Ibn Ynus (see next note). The corpus is investigated anew in II-4-5.
3
On Ibn Ynus see the article in DSB. The introduction and first few chapters of his major work are published
with French translation in Caussin, Table Hakmite, and his treatment of spherical astronomy is analyzed in
King, Ibn Ynus. On his influence in later Egyptian astronomy see idem, Astronomy in Fatimid Egypt, and
idem, Astronomy of the Mamluks.
4
On al-Maqs (Suter, MAA, no. 383) see Cairo ENL Survey, no. C15, and King, Astronomy of the Mamluks,
pp. 540, 541 and 547.
5
On al-Bakhniq see Brockelmann, GAL, SII, pp. 158 and 1019; Cairo ENL Survey, no. C28; and King,
Astronomy in Yemen, no. 13.
6
On Ibn al-Kattn (Suter, MAA, no. 411) see Cairo ENL Survey, no. C32.
tables of time as a function of solar and stellar altitude 45
46 part i, chapter two
Fig. 2.1.1b: An extract from the tables of the time since sunrise for altitudes 81-83. See also Figs. II-5.4a
and 5.6b. [From MS Cairo DM 444, courtesy of the Egyptian National Library.]
Fig. 2.1.1a: An extract from the tables of the hour-angle serving altitude 45. [MS Dublin CB 3673, courtesy
of the Chester Beatty Library.]
tables of time as a function of solar and stellar altitude 47
Fig. 2.1.1c: The table of time since sunrise for solar altitude 16 with the instruction that the table serves to
find the duration of evening twilight if one enters with the opposite point of the ecliptic from the longitude of
the sun. [From MS Cairo DM 444, courtesy of the Egyptian National Library.]
Maqss values of T(h,), Ibn al-Kattns values of t(h,) and Ibn Ynus values of a(h,) were
displayed side by side.
The following list indicates the manuscript sources for the tables of T, t and a for Cairo
and their attribution. (The tables in MS Cairo TR 354 falsely attributed to Ibn Ynus are
computed for Alexandria see 2.1.6.) The notation T/t means that T and t are tabulated on
facing pages, and the notation T/t/a means that T, t and a are tabulated together, with triplets
of values for each pair of arguments. Otherwise the functions listed are tabulated separately.
An asterisk indicates that tables are preceded by an introduction on their use prepared by al-
Maqs, and a double asterisk indicates that this introduction is accompanied by some notes
by al-Bakhniq describing the way in which he had rearranged the tables (see further II-5.6).
There is no indication in any of the sources investigated of the way in which the tables of
T(h,) were compiled. The simplest method would be to compile tables of t(h,) first. A formula
for t(h,) that is not given in Ibn Ynus major work, the kim Zj, but which is outlined
in MS Paris BNF ar. 2513, fol. 62v, of a recension of the 13th-century Egyptian Muala Zj
based mainly on his writings (6.7.1), is the following (cf. F10):
Vers t(h,) = Vers D() - Sin h G() = Vers D() - p(h,),
where p is an auxiliary function not used elsewhere and G is defined by (see 6.0):
G() = G1() G2() = R / Cos R / Cos () .
A table of G2() accompanies the theoretical discussion in MS Paris BNF ar. 2513 (6.7.1).
In the kim Zj itself, Ibn Ynus describes the computation of T(h,) rather than t(h,).7 His
method is the following. Firstly form the product:
p(h,) = Sin h G() ,
where G() is defined by any one of three equivalent expressions (cf. F8):
G() = R2 / ( Cos () Cos ) = Vers D() / Sin H()
= R / { 1/2 [ Sin H() - Sin H(*) ] }
and then (cf. F11):
T(h,) = d() - arc Sin { p(h,) - Sin d() } ,
where d is the equation of half-daylight (= D - 90). In the kim Zj Ibn Ynus tabulated
both d() and Sin d() to three digits for each degree of , giving no reason for tabulating the
latter.
In passing we note that since:
H() = + () and H(*) = - () ,
the identity:
cos cos = 1/2 { cos (-) - cos (+) }
7
Cf. King, Ibn Ynus, III.15.3.
tables of time as a function of solar and stellar altitude 49
underlies the equivalence of the first and the third of the above expressions for G(). This was
noted already by Jean-Baptiste Delambre in his Histoire de lastronomie du moyen ge,
published in Paris in 1820.8 His pronouncement:
Ebn Jounis a chang depuis cos cos en 1/2 {cos(-) - cos(+)}, et cest le premier
exemple quon trouve de cette pratique connue sous le nom de prostaphrse,
led to the notion that Ibn Ynus was:9
the first to propound the prostapherical formula ... which before the invention of loga-
rithms was of great value to astronomers as it transformed the complicated multiplica-
tion of trigonometric functions expressed in sexagesimal fractions into an addition.
This has percolated through both the scholarly and popular literature on Islamic mathematics.10
The fact that Ibn Ynus did not propound any prostaphaeretical formulae does not, however,
detract from his remarkable achievements in mathematical astronomy.
8
Delambre, HAMA, pp. 112 and 164.
9
Quoted from Heinrich Suters article Ibn Ynus in EI1. See also, for example, Sezgin, GAS, V, p. 342.
10
For some comments on this see Al-Daffa & Stroyls, Studies, pp. 27-28. (This work is otherwise to be used
with extreme caution.) See also King, Ibn Ynus, pp. 7 and 149; idem, Astronomical Timekeeping in Egypt,
p. 360; and idem, Astronomy in Fatimid Egypt, p. 508.
11
On Abu l-{Uql see King, Astronomy in Yemen, no. 9. In Varisco, Yemeni Almanac, p. 13, he is identified
as Muammad ibn Amad al-abar, the first teacher appointed by al-Malik al-Mu}ayyad D}d, the brother
and successor of al-Ashraf, to the Mu}ayyadiyya madrasa in Taiz.
12
On al-Afal see King, Astronomy in Yemen, no. 18. The manuscript is now available in facsimile in Varisco
& Smith, eds., al-Afals Anthology.
50 part i, chapter two
Fig. 2.1.2: The tables of Abu l-{Uql displaying the time since sunrise and the hour-angle for altitude 35. [From
MS Berlin Ahlwardt 5720, courtesy of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
these tables of T(h,) and t(h,). However, I know of no auxiliary tables of any kind which
Abu l-{Uql might have used to compile his main tables for timekeeping.
Fig. 2.1.4a-b: al-Khalls hour-angle tables for altitudes 30 (a) and 75-80 (b). Note the different format from
the tables of the Cairo corpus shown in Fig. 2.1.1. [From MS Paris BNF ar. 2558, courtesy of the Bibliothque
Nationale de France.]
the references there cited, to which add Kennedy & Ghanem, eds., Ibn al-Shir, and King, Astronomy of the
Mamluks, pp. 538-539 and 547-548. Several more recent papers reprinted in Saliba, Studies, deal with his
contributions to theoretical astronomy. At the time I wrote the DSB article it was not known that Muslim
astronomers concerned themselves with non-Ptolemaic models for several centuries after the time of Ibn al-
Shir. Several new sources from later centuries, albeit none yet from Syria, have been located and described
in numerous papers by George Saliba: see, for example, his The Astronomical Tradition of Maragha (1991),
Arabic Planetary Theories after the 11th Century (1996), Al-Qshjs Reform of the Ptolemaic Model for
Mercury (1993), A 16th-Century Arabic Critique of Ptolemy (1994), and Arabic Planetary Theories after
the 11th Century (1996).
52 part i, chapter two
several copies including MSS Paris BNF ar. 2558, Cairo MM 71, Damascus hiriyya 3116
and 9233, Oxford Seld. Supp. 100, Oxford Marsh 39 (Uri 1042), fols. 113v-142v, Cairo K 8525
and M 228,5, and Mosul al-Muammadiyya 129. Of these, MS Paris BNF ar. 2558 is the
finest copy, dated 811 H [= 1408] and containing all of the tables of the corpus. See further
4.8.3, 5.6.6, 6.1.2, 6.3.2, 6.5.1, 8.3.2, 9.5 and II-10, especially 10.5.
The format of al-Khalls hour-angle table differs from that used in the earlier Egyptian
tables (2.1.1), the argument domains being:
h = 1, 2, ... , 80 and = 271, 272, ... , 359, 0, 1, ... , 90 .
Note that no entry is given for the winter solstice. Extracts from these tables are shown in Figs.
2.1.4a-b.
al-Khalls computational accuracy is remarkably high. He may have used his first set of
auxiliary tables (9.4) to compile the hour-angle tables for the latitude of Damascus, but he may
also have tabulated the functions B() and C() for this latitude (6.0) and then used the simple
formula (cf. F12):
t(h,) = arc Cos { R [ Sin h - C() ] / B() } .
In MS Cairo MM 71 triplets of entries (t,T,a) for Damascus are given for each pair of arguments
(h,). The values of both t and T are attributed to al-Khall, and the values of a are attributed
to Shihb al-Dn al-alab: see 5.1.2 (illustrated).
In MS Damascus hiriyya 7387, fols. 57v-60r, of an abridgement of the Zj of Ibn al-Shir
by {Abd al-Ram al-Qazwn of Damascus (fl. ca. 1610),16 there is a set of tables displaying
the time remaining till moonset for lunar altitudes h = 7, 8, ... , 16 and each degree of lunar
longitude (symmetrically arranged). The entries are those of al-Khalls tables of T(h,), and
al-Qazwns tables, as stated in the title, can be used to find the arc of visibility (qaws al-
ru}ya) from the arc of tarrying (qaws al-makth) for calculations relating to lunar crescent
visibility.17
16
On al-Qazwn see Cairo ENL Survey, no. D38.
17
On these concepts see my article Ru}yat al-hill [= lunar crescent visibility] in EI2.
tables of time as a function of solar and stellar altitude 53
inclined to think that these fragments are from Ibn al-Rashds hour-angle tables.18 Ibn al-
Rashd also made some corrections to Ibn Ynus azimuth tables where these had been
incorrectly copied (5.1.1).
In MS M 81,1, copied ca. 1900, there is a complete set of tables of the function t(h,)
with values in degrees and minutes, for the same parameters:
= 32;0 (Jerusalem) and = 23;35 .
These tables are copied without the solar longitude arguments, and without the degrees of the
entries, except at the head of each column. I have not been able to check that the entries in
various sub-tables are the same as those in the fragments mentioned above, but it seems unlikely
that there would be two different sets in existence. See also I-2.2.6 and II-6.12 and II-11.12.
see Fig. 2.1.6 as in al-Bakhniqs edition of the Cairo corpus see Figs. II-5.6a-b. The
errors in the sum of corresponding values of t and T are related to those in a table of D()
for these parameters in MSS Oxford Marsh 676 (Uri 944 = 995) and Princeton Yahuda 861,1,
originally compiled by Najm al-Dn al-Mir (2.6.1). I know of no tables of auxiliary functions
based on latitude 31;0 which might have been used to compile the larger set. Several
manuscripts exist of a corpus of prayer-tables based on the same parameters: see further II-
8.5. However, the identity of the compiler of these tables remains obscure.
MS Cairo M 216, copied in 1003 H [= 1594/95], contains a different set of tables of the
functions T(h,) and t(h,) based on the parameters = 31;0 and = 23;35 (?). No compiler
is associated with this set either.
al-Kutub states that he also included a correction for refraction at the horizon. See further II-
7.14.
2.2 Tables of time as a function of solar longitude and instantaneous altitude for a given latitude
These tables differ from those discussed in 2.1 only in that the arguments h and are
interchanged. For each degree of a given table displays entries for each degree of h up to
20
On Amad Efend see Cairo ENL Survey, nos. H28. He does not appear to be listed in ~hsanolu et al.,
Ottoman Astronomical Literature.
21
On al-Kutub see Cairo ENL Survey, no. D76; and ~hsanolu et al., Ottoman Astronomical Literature, I,
pp. 427-429, no. 281.
tables of time as a function of solar and stellar altitude 55
Fig. 2.1.8: al-Kutubs tables for solar altitudes 37-38. These are based on the earlier tables of the Cairo corpus,
although the author gives no indication of this. [From MS Cairo J 811,8, fol. 71r, courtesy of the Egyptian
National Library.]
the maximum for the solar longitude in question. that is, H() or the smaller integral value
[H()]. This arrangement has the advantage that entries can be included both for the equinoxes
and for the solstices.
22
On al-Karak see Cairo ENL Survey, no. C35.
56 part i, chapter two
Fig. 2.2.1a-b: The tables of the time since sunrise for solar longitude Aquarius 12 / Scorpio 18 (a) and the
hour-angle for longitudes 4-6 of a certain sign (b) in the Jerusalem corpus. [From MS Leipzig UB 808, fol.
23v and 119v, courtesy of the Universittsbibliothek.]
Jerusalem. With due respect to al-Karak, it seems to me probable that Ibn al-Rashds tables
of t(h,) were computed for Jerusalem and that al-Karak simply changed the format and added
the values of T(,h). The entries for t(,h) in the Leipzig manuscript are identical with the
corresponding entries in the tables of t(h,) for Jerusalem in MSS Cairo DM 45 and Cairo DM
153, which I suspect were computed by al-Rashd (2.1.5).
In fols. 94v-123v of the Leipzig manuscript there is another set of hour-angle tables in a
different hand see Fig. 2.2.1b. It is stated at the beginning and at the end of the tables that
they are for latitude 32;40 (Ramla), but in fact the tables are simply those of al-Karak for
latitude 32;0.
In MS Princeton Yahuda 861,1, copied ca. 1600, amidst a set of anonymous prayer-tables
for latitude 32, there is an odd table of T(h) and t(h) computed for the equinoxes, which apart
from copyists errors has the same entries as al-Karaks tables of T(,h) and t(,h) for =
0. Likewise the entries in the twilight tables for latitude 32 in MS Princeton Yahuda 861,1
are related to the entries for solar altitudes 20 and 16 in al-Karaks set. See further II-8.1
and 9.5.
tables of time as a function of solar and stellar altitude 57
23
On li Efend see ibid., no. H36; and ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 453-
458, no. 303.
58 part i, chapter two
Fig. 2.2.2: The tables for Capricorn 0-11 / Sagittarius 30-19 in the Edirne corpus. [From MS Oxford arab.
e. 93, fols. 3v-4r, courtesy of the Bodleian Library.]
Fig. 2.2.3: The tables for Taurus 14 / Leo 16 in the corpus of li Efend. [From MS Princeton Yahuda 353,
courtesy of the Special Collections, Princeton University Library.]
Kandilli 440 and 441, and Cairo M 120 contain tables displaying the times before and after
midday when the sun has the given longitude and altitude, with entries in equinoctial hours,
minutes and seconds, expressed according to the Ottoman convention. Likewise the tables in
MSS Istanbul Esat Efendi 1979, Istanbul UL T1963 and T1964 display these same times in
equinoctial hours and minutes. Most of these sets of tables are anonymous and most of them
are labelled in Turkish irtif cedveli < irtif{ jadwali, altitude tables. In MS Istanbul Husrev
Paa 232, also apparently MS Baghdad Awqf 325/12248, the tables are attributed to the
muwaqqit Muammad diq Jihngr. An abridged version of li Efends tables appears
on an Ottoman quadrant.24
24
See Khalili Collection Catalogue, II, p. 268, no. 157, and King, Review, col. 256.
tables of time as a function of solar and stellar altitude 59
Fig. 2.2.4: This extract from usayn usns tables for Mecca shows the time in hours, minutes and seconds
for solar altitudes 35-65 in the east and west and for solar longitudes Pisces 20-24 and Libra 10-6. [From
MS Cairo K 4002, fol. 10v, courtesy of the Egyptian National Library.]
25
On usayn usn see Azzawi, History of Astronomy in Iraq, pp. 268, 284 and 285; Cairo ENL Survey,
no. H47; and ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 581-584, no. 419.
26
On Islamic values for the latitude of Mecca see King, Earliest Muslim Geodetic Measurements, pp. 225-
226.
60 part i, chapter two
Fig. 2.2.6: An extract from some tables for Jerusalem preserved in a late manuscript. These are probably based
on those illustrated in Fig. 2.1.5. [From MS Cairo M 81,1, fols. 1v-2r, courtesy of the Egyptian National
Library.]
2.3 Tables of time as a function of meridian altitude and instantaneous altitude, for a specific
latitude
The tables listed in this section display the function T(H,h) and can in principle be used for
either the sun or the stars. Since h H the tables are trapezoidal in shape, and in medieval
27
On al-anw see Cairo ENL Survey, no. D123; and ~hsanolu et al., Ottoman Astronomical Literature,
II, pp. 646-649, no. 496.
tables of time as a function of solar and stellar altitude 61
Arabic these tables are generally refered to by the term aylasn (1.4). In such tables only the
entries for - H + can be used for the sun. The entries for general arguments H can
be used to find the time elapsed since the rising of a star with declination such that = H
- . Tables of T(H,h) computed for a given latitude for timekeeping by the sun require about
one-third as many entries as tables of T(h,), if entries are given for each degree of argument
in both cases.
No indication is given in the sources of the method of computation of the entries in these
tables, but it seems reasonable to assume that one would first compute the value of D
corresponding to H and then apply the standard formula (cf. F9):28
Vers t(H,h) = [ Sin H - Sin h ] Vers D / Sin H .
Given D and t the computation of T is trivial. Note that the function T(H,h) is independent
of the obliquity .
Other tables of this kind are attested in some copies of the Toledan Tables and in a medieval
Hebrew source. In the former, values are given for each degree of both arguments, and the
meridian altitude runs up to about 80. In the latter, the horizontal arguments are periods of
days of the year (during which the meridian altitude is assumed constant) and the vertical
argument is the solar altitude in 5-intervals. See further 10.1.
Fig. 2.3.1: The sub-tables for solar altitudes 21-27 in the Baghdad corpus of {Al ibn Amjr. Note that the
scribe copied the column for 24 twice, realising his error just before he finished the second set and marking
it mukarrar, repeated. [From MS Paris BNF ar. 2486, 239v-240r, courtesy of the Bibliothque Nationale de
France.]
examined MS Cairo DMF 1, 136 fols., copied at Maragha in 692 H [= 1293], as well as MSS
Florence Medici 269, fols. 150r-152v, Istanbul UL P1418,2, fols. 234r-237v, and Oxford Hunt.
143, fols. 155v-160v, of this table. There the function:
T(H,h)
is tabulated to three digits for the domains:
H = 29;10, 30, 31, ... , 75, 76;10, and h = 1, 2, ... , H ,
based on the parameter:
= 37;20 (Maragha).
II:1, pp. 58-60; and, most recently, Ragep, al-ss Memoir on Astronomy, and the same authors article al-
s. 3. As scientist in EI2. The existence of the lkhn aylasn table was noted in Kennedy, op. cit., p. 161.
tables of time as a function of solar and stellar altitude 63
Fig. 2.3.2a-b: Some of the entries in a splendid late 13th-century copy of al-ss tables for meridian altitudes
41-64 (a) and 65-75 and then 76;10 (b). After this aylasn table there is a table for the equation of time.
From MS Cairo DMF 1, courtesy of the Egyptian National Library.]
Two extracts are displayed in Figs. 2.3.2a-b. The minimum and maximum values of H indicate
that the value assumed for was the lkhn value 23;30, although as noted above the values
of T(H,h) are independent of this parameter. Since the second digits in the entries are often
in error, the third ones are meaningless. Values are also given in equinoctial hours, minutes
and seconds, and these are generally accurate to the nearest minute.
32
On Sanjar al-Kaml and the Ashraf Zj see Kennedy, Zj Survey, no. 4, and Storey, PL, II:1, p. 64.
64 part i, chapter two
Fig. 2.3.5: The sub-tables for solar meridian alti- Fig. 2.3.6: The sub-tables for solar meridian altitudes
tudes 63-64 in the Tunis corpus. [From MS 30-34 from the aylasn table of Taqi l-Dn. [From
Berlin Ahlwardt 5724, fol. 24r, courtesy of the MS Istanbul Kandilli 208, courtesy of Kandilli Obser-
Deutsche Staatsbibliothek (Preuischer Kultur- vatory.]
besitz).]
arguments. The underlying parameters are = 42 and = 23;39, and the table may have
been intended to serve Edirne.
Such tables are clearly analogous to those described in 2.3 since for each star there is an
underlying value of H. See also 6.16.
Fig. 2.4.1: Tables from the Tunis corpus for 14 stars. For each star the ascensions at culmination are also given.
[From MS Berlin Ahlwardt 5724. fols. 55v-56r, courtesy of the Deutsche Staatsbibliothek (Preuischer
Kulturbesitz).]
2.5 Tables of time since sunrise as a function of meridian altitude and instantaneous altitude,
for all latitudes
The tables discussed in this section are also usually called by the name aylasn (2.3) and are
based on an approximate formula:
T(H,h) 1/15 arc Sin { R Sin h / Sin H } .
This is a good approximation to the actual value of T(H,h). The expression is accurate for the
equinoxes (and, of course, when h = 0 or H). I have computed the error incurred by using
the equivalent expression in degrees, namely:
T(H,h) D / 90 arc Sin { R Sin h / Sin H } ,
in the domain:
- H +,
corresponding to solar altitudes for various values of . The approximate values are less than
the accurate values, but the maximum errors for latitudes 15 (Yemen) and 30 (Cairo) are
only about 1 and 11/2, respectively. The tables could also be used for timekeeping by the
stars.
This approximate formula for reckoning time from solar altitude was used in the early 9th
century by astronomers such as al-Khwrizm and abash.36 Certain later Muslim astronomers
36
Ibn al-Muthann (10th century?) attributes this method to al-Khwrizm (Goldstein, Ibn al-Muthann on
al-Khwrizm, pp. 81-82 and 207-208), and states that in addition al-Khwrizm proposed an exact formula (ibid.,
pp. 83). However, the approximate method is not included in the Latin translation of the recension of al-
Khwrizms Zj by al-Majr (ca. 1000) (Kennedy, Zj Survey, no. 21).
The duration of twilight is determined by an equivalent method in MS Berlin Ahlwardt 5750, fol. 153r, of
a recension of the Zj of abash (Kennedy, op. cit., no. 15) but the attribution of the twilight theory in the Berlin
tables of time as a function of solar and stellar altitude 67
used an equivalent formula to compile tables of the duration of twilight: see II-3.3 and 3.7
for tables of this kind for Baghdad and Alamut (?). Others tabulated the solar altitude as a
function of time using this formula: see 4.3. The approximation is further attested in a Byzantine
astronomical treatise,37 and also underlies a table in an English manuscript of the Toledan
Tables (10.1). See also II-1.4 and XI.
It is also worth noting that certain Muslim astronomers in early Abbasid times as well as
later astronomers concerned with timekeeping in Cairo and Damascus used an approximation
ultimately of Indian origin38 involving the increase of a gnomon shadow over its midday
minimum, z = Cotn h - Cotn H, namely:
T(h,H) { 6n / (z + n) } sdh
for various values of the base n. No tables based on this formula are known to me.
manuscript to abash is not certain, since it is not contained in MS Istanbul Yeni Cami 784 of his Zj (ibid.,
no. 16, and Debarnot, Zj of abash). See further XI-5.0.
37
See Neugebauer, Studies in Byzantine Astronomy, pp. 11-12, and Jones, Byzantine Astronomical Manual,
pp. 154-155.
38
On this formula see Pingree, Indian Influence, pp. 121-122; Davidian, Brn on the Time of Day,
pp. 331-332 and 334; and now II-1.4, III-1.3, and XI-1.2.
39
Paris Catalogue, pp. 446-447.
68 part i, chapter two
Fig. 2.5.1: Part of {Al ibn Amjrs aylasn table. Fig. 2.5.2: The part of an anonymous aylasn table
[From MS Paris BNF supp. pers. 1488, fol. 201v, serving solar meridian altitudes 33-34. [From MS Paris
courtesy of the Bibliothque Nationale de France.] BNF ar. 2514, fol. 3r, courtesy of the Bibliothque Na-
tionale de France.]
2.6 Tables of the time since rising of the sun or fixed stars as a function of their altitude, for
all latitudes
To prepare a table which will give the time since rising of a celestial body for all latitudes,
one can either tabulate T(H,h) for each degree of latitude or generate the function T in terms
of h and pairs of arguments such as (,), (,d) or (H,D). The last pair has the advantage that
in modern terms both arguments are always positive.
40
On this zj see King, Astronomy in Yemen, no. 17.
41
On Najm al-Dn al-Mir (Suter, MAA, no. 460, where two different individuals with this name are listed
together; Cairo ENL Survey, no. C16; King, Astronomy of the Mamluks, pp. 540-541) see now Charette, Najm
al-Dns Monumental Table, and idem, Mamluk Instrumentation. I have not updated my text to incorporate
Charettes findings.
70 part i, chapter two
Fig. 2.6.1: An extract from Najm al-Dns monumental table serving meridian altitude 38. Values are shown
for altitudes 27-37 and half arc of visibility 63-87, and then altitudes 1-27 and half arc 108-152. [From
MS Cairo MM 132, fols. 82v-83r, courtesy of the Egyptian National Library.]
since sunrise or since the rising of the star in question. For given H, it is not difficult to show
that D varies between a certain minimum, defined by = - = 1/2 , and 180. The lower limit
D* for D in each table is simply the smallest integer greater than this minimum, i.e.:
D*(H) = [min D(H,)] + 1 .
The compilation of the table is in fact extremely straightforward. The formula for finding T
(= D - t) from H, h and D outlined in Ch. 12 of Najm al-Dns treatise on spherical astronomy
(on which see II-2.5) is the following (cf. F8 and F10):
Vers t(H,h,D) = Vers D - Sin h Vers D / Sin H .
Note that this is the equivalent to:
Vers t = Vers D { 1 - Sin h / Sin H } ,
and that if one had first compiled a table of the function:
H(H,h) = 1 - Sin h / Sin H ,
the table of t(H,h,D) could be generated with considerable facility. Ibn al-Mushrif (6.13.1)
tabulated the related function:
H(H,h) = Sin H - Sin h ,
but no Islamic tables of H have been located.
tables of time as a function of solar and stellar altitude 71
In a sample of 100 random entries in Najm al-Dns main table about 55 showed errors of
3 or less in the second digit, 35 showed errors between 4 and 9, and there were 10 errors
between 10 and 15. It should be borne in mind that 0;15 corresponds to one minute of time,
and I make bold to say that Najm al-Dns table shows the time for any solar or stellar altitude
with the vast majority of entries accurate to the nearest minute.
Only in 1982 with the discovery of an introduction by him in MS Dublin CB Persian 102,1,
did it become clear that Najm al-Dn also intended this table as a universal auxiliary table:
see further 9.3*. For a complete discussion of all aspects of this table, the reader is referred
to the publications of Franois Charette.
2.7 Tables of the time of night as a function of the normed right ascensions of culminating stars
and the solar longitude, for a specific latitude
42
On al-alab (Suter, MAA, no. 434) see Cairo ENL Survey, no. C69.
43
On Muammad ibn Ktib Sinn (Suter, MAA, no. 455) see Cairo ENL Survey, no. H8; and ~hsanolu et
al., Ottoman Astronomical Literature, I, pp. 84-90, no. 46.
72 part i, chapter two
1510]. All of the right ascensions have been increased by 2;0 in the margin of the star
catalogue. The main tables themselves occupy about 500 pages in each of the manuscripts and
contain about 240,000 entries. The horizontal argument is intended to be the normed right
ascension , and the vertical argument is the solar longitude ; the argument increments are
1 for both. Several dozen consecutive pages serve each 30 of solar longitude. Four functions
are tabulated side by side for each pair of arguments:
t1 the time since sunset,
t2 the time remaining until sunrise,
tables of time as a function of solar and stellar altitude 73
2.8 Tables of the time of night as a function of the longitude of the horoscopus, for a specific
latitude
The star al-kaff al-khab (= Cassiopeiae)44 was used by Ottoman and Safavid astronomers
for timekeeping by night because its right ascension was close to zero, that is, it culminated
with the vernal equinox. These tables represent a special case of the theory underlying those
treated in 2.7.
Fig. 2.8.1: An extract from a table for timekeeping by the star al-kaff al-khab, chosen because of its fortuitous
position relative to the vernal equinox in the 16th century. [From MS Cairo ZK 782,5, fols. 25v-26r, courtesy
of the Egyptian National Library.]
45
On Muammad Zamn al-Mashhad (I-2.8.2) see Mayer, Islamic Astrolabists, pp. 78-79 and 87; Cairo
ENL Survey, no. G76; and King, Mecca-Centred World-Maps, pp. 172-175.
46
See n. 3:2.
tables of the longitude of the horoscopus 75
CHAPTER 3
Tables displaying the longitude of the horoscopus are related to the categories described in
Ch. 2 by virtue of the simple formulae (cf. F20):
{H(h,)} = () + T(h,)
for the sun and:
{H(h)} = () + T(h)
for a fixed star. It should be remembered that tables of oblique ascensions () for different
latitudes were standard equipment of the Muslim astronomers.1
The horoscopus is of prime importance in astrology as well as in timekeeping and the Muslim
astronomers also compiled extensive tables displaying the longitudes of the astrological houses
and the projection of the rays.2 These tables are not discussed in the present study.3 See 10.1
on a table displaying the function H(T,) for a particular latitude found in an early-15th-century
manuscript of the Alphonsine Tables written in Germany.
3.1 Tables of the longitude of the horoscopus as a function of the instantaneous solar altitude
and solar longitude, for a specific latitude
Fig. 3.1.1: The part of Abu l-Uqls enormous table displaying the longitude of the horoscopus for solar altitude
43 in the east and west. [From MS Berlin Ahlwardt 5720, courtesy of the Deutsche Staatsbibliothek (Preuischer
Kulturbesitz).]
is tedious but very straightforward. Given also a table of inverse oblique ascensions (), the
compilation of the larger table is made even easier.
3.2 Tables of the longitude of the horoscopus as a function of stellar altitude, for a specific latitude
4
On Kshyr (Suter, MAA, no. 192) see Kennedy, Zj Survey, no. 9, and Sezgin, GAS, V, pp. 343-345,
VI, pp. 246-249 and VII, pp. 182-183. On the theoretical treatment of spherical astronomy in his Zj see Berggren,
Spherical Trigonometry in the Jmi{ Zj.
tables of the longitude of the horoscopus 77
Figs. 3.2.1a-b: The star catalogue with additional information for timekeeping by the stars (a), and the sub-tables
for timekeeping by Sirius and Capella (b). [From MS Berlin Ahlwardt 5751, pp. 188-189 and 190-191, courtesy
of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
78 part i, chapter three
The underlying latitude is found by inspection to be 30;25. A note by the side of these tables
states:
(This is) a table for operations (of timekeeping) with the fixed stars according to the
results of al-Balkh in the year 1310. We recomputed (the entries for) several stars
for the year 1424 and did not find any difference which would affect the determination
of the (co)ascendants, because they (?) do not vary over long periods. We recomputed
values for Sirius and Procyon, Betelgeuze, Spica and Arcturus, and Aldebaran (al-
shi{rayn wa-l-mankib al-ayman wa-l-simkayn wa-{ayn al-thawr),5 and for these
there was no perceptible difference. We used what we derived in the table for finding
the ascendants accurately from the fixed stars. God grants success.
The dates 1310 and 1424, which are written in the abjad notation in the text, must relate to
the calendar of Alexander, and thus correspond to 1000 and 1125 in the Julian calendar. Also
from the table itself (the longitude of Regulus is 136;39) a date of ca. 1050 is suggested.6
At the time of writing this in the 1970s I was not aware of any astronomer named al-Balkh
who was active in Southern al-{Irq or Iran ca. 1000, although {Al ibn Amad al-Nasaw
compiled a zj in central Iran in the early 11th century. The latitude 30;25 was used for
Qandahar (now in Afghanistan) in a recension of the geographical tables of al-Khwrizm
(7.1.1),7 and we can be certain that the table was intended for use there.
The next four pages in the Berlin manuscript (pp. 190-194) contain the only remaining part
of these ascendant tables: see Fig. 3.2.1b. The function tabulated is H(h) for each degree of
eastern and western altitudes for Sirius, Capella, Vega and Arcturus. Entries are given in
zodiacal signs, degrees and minutes, and the underlying latitude appears to be 30;25. Both
star catalogue and ascendant tables seem to be unrelated to the auxiliary tables that immediately
precede them (on pp. 188-189) in the Berlin manuscript (6.16.1), although these auxiliary tables
do indeed serve to compute H(h) for different stars. See 5.8.1 on another table from this
collection in the same manuscript. Both tables are featured again in IX.
5 On these six stars see Kunitzsch, Sternnomenklatur der Araber, pp. 111-112, nos. 289-290; p. 57, no. 158;
p. 105, nos. 269-270; and p. 51, no. 69, respectively.
6 For example Ibn Ynus has Regulus at Leo 16 ca. 1000, and the rate of precession is 1 in ca. 70 years.
7 Kennedy & Kennedy, Islamic Geographical Coordinates, pp. 261-262.
tables of the longitude of the horoscopus 79
Fig. 3.2.2: An extract from the anonymous tables for timekeeping by the stars, serving mankib al-faras and
dhanab al-asad. [From MS Paris BNF ar. 2514, fols. 31v-32r, courtesy of the Bibliothque Nationale de France.]
8 On Ibn D{ir see King, Astronomy in Yemen, no. 31, and now Ihsanolu et al., Ottoman Geographical
Literature, I, pp. 76-78, no. 39. The word D{ir means profligate: see the article Zu{{r by Thierry Bianquis
in EI2.
80 part i, chapter three
3.3 Tables of the longitude of the horoscopus as a function of time and solar longitude, for a
specific latitude
On the astronomical works of al-Ashraf (Suter, MAA, no. 394 and Mayer, Islamic Astrolabists, pp. 83-84)
9
Fig. 3.3.1a: An extract from al-Ashrafs tables giving Fig. 3.3.1b: An extract from al-Ashrafs tables showing
the altitude of the sun at midday and the beginning of the ascendants at the hours as a function of solar
the {ar and the corresponding ascendant. This sub-table longitude, here Gemini. The same entries are for the
serves solar longitudes in Aries. [From MS Oxford night hours for solar longitudes in Sagittarius. [From
Hunt. 233, fol. 63v, courtesy of the Bodleian Library.] MS Oxford Hunt. 233, fol. 82v, courtesy of the Bod-
leian Library.]
An extract is shown in Fig. 3.3.2. The tables are unattributed, but the smaller table described
in 4.2.5 is based on the same distinctive latitude and is attributed to Abu l-{Uql. It seems
that this latitude for Taiz was derived by Abu l-{Uql after he had computed his corpus of
tables for latitude 13;40 (2.1.2). A table of () for the same parameters is preserved in the
anonymous late-14th-century Yemeni zj MS Paris BNF ar. 2523, fol. 95v (2.5.3). Likewise
the tables for marking astrolabes and sundials for Taiz compiled by the Rasulid Sultan al-Ashraf
about 1295 which are extant in MS Cairo TR 105 (4.1.4) are also based on = 13;37, but
the obliquity used is the lkhn value 23;30 (2.3.2).
The function H(T,) is called awli{ al-s{t al-zamniyya, the ascendants of the seasonal
82 part i, chapter three
Fig. 3.3.2: An extract from the anonymous tables of the horoscopus at the hours of day and night for the latitude
of Taiz. [From a manuscript in a private library in Sanaa; taken from the facsimile in Varisco & Smith, eds.,
al-Afals Anthology, p. 114, with permission of Professor Daniel Varisco.]
hours, and the existence of a similar table in a 15th-century German manuscript has been noted
in 3.0 (see also 10.1). Implicit in the headings of the Yemeni table is the fact that the horoscopus
is the same at the nth seasonal hour of the day for solar longitude as it is at the nth seasonal
hour of the night for solar longitude * = + 180. A given sub-table on one page serves one
zodiacal sign, but there are only five sub-tables in the manuscript now. The compilation of
such a table is quite straightforward, since if T = nsdh,
{H(T,)} = () + n/12 {2D(,)} .
are in Persian in the former and in Arabic in the latter (see Fig. IV-5.3). The table displays
the function:
H(Tn,)
for the domains:
= 0, 6, ... , 354 and Tn() = n/8 {2D()} (n = 1, 2, ... , 7)
as well as at daybreak and nightfall. The functions are called simply li{-i awqt, the as-
cendant at different times of day. The underlying latitude is found by inspection to be about
= 38, and it is not possible to determine the value used for ; this latitude perhaps corresponds
to Konya. Entries are given in signs and degrees and it is stated that they were derived using
an astrolabe. The underlying formula is trivial, namely:
{H(T,)} = () + Tn() .
Using an astrolabe fitted with a plate for latitude one sets the ecliptic longitude on the
rete over the local horizon on the plate and can immediately read off () and D(). One then
rotates the rete by the amount Tn() in the direction of the daily rotation and simply reads off
the longitude of the new point where the ecliptic intersects the horizon, which is H(Tn,). See
also II-14.2 and IV-5.3 for further information on this unusual table.
3.4 Tables of the longitude of the horoscopus as a function of the time since rising of the stars,
for a specific latitude
Fig. 3.4.1: The table by Ibn D{ir on the right displays the horoscopus at each seasonal hour of visibility for
18 stars. The one on the left shows the corresponding altitudes (4.6.1). [From MS Berlin Ahlwardt 5720, fols
126v-127r, courtesy of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
84 part i, chapter four
CHAPTER 4
The extensive stellar and solar altitude tables located in various Islamic sources are closely
related to the tables described in Chs. 2 and 3. The numerous smaller solar altitude and azimuth
tables for marking solstitial shadow traces on sundials belong to another category of Islamic
astronomical astronomical tables which merit a separate study,1 and I have intentionally limited
examples of such altitude tables cited below to the tables of abash (Pseudo-al-Khwrizm),
al-Battn, al-Marrkush, al-Maqs and the Yemeni Sultan al-Ashraf. Tables of solar meridian
latitudes for specific latitudes were standard in zjes, and already in the 9th century we find
a small table of H(,) for each of the climates (see Fig. VIa-1.1).
4.1 Tables of solar altitude at the solstices as a function of time, for a specific latitude
Tables of the solar altitude for each seasonal hour at the solstices, and more especially the
corresponding gnomon shadow lengths and azimuths, were used by Muslim astronomers for
constructing the points of intersection of the hour lines with the solstitial shadow traces on
horizontal sundials.2
85
two. [From MS Istanbul Ayasofya 4830, fol. 188v, courtesy of the
Sleymaniye Library.]
86 part i, chapter four
Fig. 4.1.1b: Some sub-tables in the Topkap manuscript, serving latitudes 21, 28;30, (33, 40, 35 and 30).
[From MS Istanbul Topkap A3342, courtesy of the Topkap Library.]
Fig. 4.1.3: The tables for constructing a horizontal Fig. 4.1.4: This extract from al-Ashrafs tables for
sundial for latitude 30 in the treatise of Marrkush. sundial construction serves latitude 21. [From
[From MS Paris BNF 2507, fol. 137r, courtesy of the MS Cairo TR 105, fols. 132v-133r, courtesy of
Bibliothque Nationale de Paris.] the Egyptian National Library.]
tables of solar and stellar altitude 87
6 See King, Mecca-Centred World-Maps, p. 350, for a list of his works in which this latitude is featured,
usually to the exclusion of all others.
7 On Yay ibn Ab Manr and the Mumtaan Zj see Kennedy, Zj Survey, no. 51; Vernet, Las Tabulae
Probatae, the same authors article in DSB; and Sezgin, GAS, V, pp. 227-228, and VI, 136-137. A distorted
facsimile of the Escorial manuscript was published in Frankfurt (IGAIW) in 1986.
8 See, in particular, n. 9:5.
9 On the sine tables see n. 7:10 to 7.1.1.
10 See Sezgin, GAS, VI, p. 137, no 2. The only other early Islamic work mentioned by Ibn al-Nadm which
relates specifically to timekeeping is a treatise by Ibn Karnib (see Sezgin, GAS, V, p. 275), who lived in Baghdad
probably in the latter half of the 9th century, on the determination of the time of day from the solar altitude.
88 part i, chapter four
astronomical tables for Baghdad are contained in the recension of Yays Mumtaan Zj
preserved in MS Escorial r. 927: these are based on parameters = 33;21 and = 23;33.
11 On al-Battn see Willy Hartners fine article in DSB. His Zj is published in Nallino, al-Battn, where
the sundial tables occur in II, p. 188 (see also p. 296).
tables of solar and stellar altitude 89
for each 5 of t at the equinoxes and both solstices, computed for = 41 (Istanbul). There
is also a table displaying values of h(T) for each seasonal hour at the solstices.
4.2 Tables of solar altitude as a function of time since sunrise and solar longitude, for a specific
latitude
The following tables display the solar altitude at the seasonal hours corresponding to a wider
range of values of than in the tables discussed in 4.1. On the computation of these tables
see 4.2.2 below.
12 On the treatises see now Charette & Schmidl, al-Khwrizm on the Astrolabe. On this table see already
King, al-Khwrizm, pp. 25 and 29, and Hogendijk, al-Khwrizms Sine of the Hours.
13 On this treatise and table see ibid., pp. 30-31.
14 On Ibn al-@dam see Sezgin, GAS, VI, pp. 179-180, and on al-Samarqand see ibid., VI, pp. 216-217, and
Cairo ENL Survey, no. B58. The title page and the first few folios of this manuscript are in the hand of the
late-15th-century Egyptian astronomer Ibn Abi l-Fat al-f (9.10), and the main part is in an older hand. On
the title folio al-f wrote that the work was by al-usayn ibn Muammad al-@dam but that in another
manuscript that he had seen, the work was attributed to Sa{d ibn Khaff al-Samarqand.
90 part i, chapter four
15 On al-Marrkush see Suter, MAA, no. 363; Cairo ENL Survey, no. C17; and the article in EI . Sdillot-
2
pre, Trait, contains a translation of the first half of al-Marrkushs treatise on spherical astronomy and
sundials. Sdillot-fils, Mmoire, contains a summary of the second half dealing with instruments. A Topkap
manuscript of the Jmi{ al-mabdi} wa-l-ghyt was published by photo-offset at the Institut fr Geschichte
der Arabisch-Islamischen Wissenschaften, Frankfurt, in 1984.
tables of solar and stellar altitude 91
z(T,) to base 12 for each seasonal day-hour and each 10 interval of , as well as (II.76)
the function h(T,), giving values to two sexagesimal digits for the domains:
T = 1, 2, ... , 6h (1h = 15) and = 0, 30, 60, 90 ( \ 0)
and for the same parameters. The corresponding solar altitudes in the prime vertical and in
the azimuth of the qibla are also given.
4.3 Tables of solar altitude as a function of time since sunrise and meridian altitude, for all
latitudes
A number of sources contain tables of the function h(T,H), where T is measured in seasonal
hours, based on the approximate formula:
h(T,H) arc Sin { Sin H Sin (15 T) / R } .
16 Charette, Mamluk Instrumentation.
92 part i, chapter four
These tables are intended to be valid for all latitudes and the underlying formula is a
transformation of the one discussed in 2.5 above. On a similar table in a 14th-century English
astronomical manuscript of both the Toledan and Alphonsine Tables see 10.1.
Fig. 4.3.3: A table of the tangent of the solar declination (7.1.10) and the anonymous Yemeni table serving solar
meridian altitudes 11-90. [From MS Paris BNF ar. 2523, fol. 86r, courtesy of the Bibliothque Nationale de
France.]
is adequate for use in the Yemen, where the minimum value of the solar altitude is about 50.
See also II-12.3 on this table.
4.4 Tables of solar altitude as a function of the semi diurnal arc and the time since sunrise
Fig. 4.4.1: An extract from the anonymous Mamluk tables of solar altitude serving half arc of daylight 86 (hour-
angle at the {ar 49;12). [From MS Cairo MM 72, fols. 20v-21r, courtesy of the Egyptian National Library.]
pages in the manuscript serves a particular value of D, and the two unlabelled numbers at the
head of each table are in fact the argument D and the associated time from midday to the
beginning of the interval for the afternoon prayer ta. By the side of each sub-table are the solar
altitudes at each seasonal hour for the given value of D. Fig. 4.4.1 shows a pair of facing pages
of these tables.
The title folio and all but the last page of the introduction to these tables are missing from
the manuscript. The copyist was Ibn al-Kattn, who compiled the hour-angle tables for Cairo
in MS Istanbul Kl Ali Paa 684, also penned in his elegant hand (2.1.1 and II-4.1.4).22 I
have previously suggested that these tables might have been computed by Ibn Ynus,23 but
the possibility that the tables in MS Cairo MM 72 were computed by Ibn al-Kattn himself
cannot be entirely excluded. However, it now seems more likely that they were computed by
22 On Ibn al-Kattn (Suter, MAA, no. 411) see Cairo ENL Survey, no. C32.
23 King, Astronomical Timekeeping in Medieval Cairo, pp. 365 and 391.
96 part i, chapter four
the early-14th-century Egyptian astronomer Najm al-Dn al-Mir (2.6.1). In his treatise on
spherical astronomy preserved in MS Milan Ambrosiana 227a (C49), fols. 85v-97r, Najm al-
Dn discusses the determination of h from D and T. The method which he prescribes is
equivalent to the following standard formula used, for example, by Ibn Ynus24 (cf. F10):
h(D,T) = arc Sin { [ Vers D - Vers (D-T) ] Sin H / Vers D } .
For each value of D one must first find the corresponding value of H. Najm al-Dn then adds
a note about a method for finding the solar altitude at the seasonal day-hours: for the nth hour
T = n where is the number of degrees corresponding to one seasonal day-hour. This is
strong evidence that the tables were compiled by Najm al-Dn himself. I see no particular
advantage to tables of solar altitude as a function of the semi diurnal arc, but Najm al-Dn
may already have compiled his enormous table of T(H,h,D) for all latitudes and then thought
it reasonable, and had energy and time, to compile a table of h(D,T) for Cairo.
4.5 Tables of stellar altitude as a function of the longitude of the horoscopus, for a specific latitude
4.6 Tables of stellar altitude as a function of time of night, for a specific latitude
Fig. 4.5.1: An extract from Abu l-{Uqls tables of the altitudes of various stars as a function of the horoscopus,
here the first few degrees of the sign of Leo. [From Varisco & Smith, eds., al-Afals Anthology, p. 173, with
kind permission of Professor Daniel Varisco.]
98 part i, chapter five
Fig. 4.5.2: An extract from Abu l-{Uqls tables of stellar altitudes in the east and west serving each degree
of the horoscopus in Aquarius. [From MS Berlin Ahlwardt 5720, courtesy of the Deutsche Staatsbibliothek
(Preuischer Kulturbesitz).]
4.7 Tables of solar altitudes in certain azimuths as a function of solar longitude, for a specific
latitude
The azimuths for which a knowledge of the corresponding solar altitude is of some importance
in Islamic astronomy are defined by: (i) the meridian; (ii) the prime vertical; and (iii) the qibla
or azimuth of Mecca. In Cairo a fourth direction was of interest to the astronomers because
of the layout of the architecture of the city, namely, (iv) the direction for orienting the ventilators
on the roofs of religious and secular buildings, these being a common feature of the medieval
city.
Tables of the solar meridian altitude H() for a particular latitude are standard in zjes, and
their compilation is trivial since (cf. F1):
H(,) = + () .
Numerous Islamic sources contain tables of the function:
() = 90 + () ,
to be used for finding the meridian altitude for any latitude. Clearly:
H(,) = () - .
Tables of h0(), the solar altitude in the prime vertical, are less common. Those located thus
far are discussed in 4.8 below.
Finally, tables of hq(), the solar altitude in the azimuth of Mecca, are attested for a wider
variety of localities. These are discussed in II-3.6, 3.9, 4.7, 10.6, etc.25
To find h(a,) for a particular azimuth, one can either use interpolation in tables of a(h,)
if these were available (5.1) or calculate the altitude directly. The method of Ibn Ynus is as
25 On Ibn Ynus tables of h () for Cairo see already Schoy, Schattentafeln, p. 42; King, Ibn Ynus, III.28.3,
q
and idem, Astronomical Timekeeping in Medieval Cairo, p. 368. See also II-4.7.
tables of solar azimuth 99
follows.26 First compute the altitude he of the sun at an equinox when it has azimuth a, thus:
he = arc Sin { R Sin a / [ (R Sin / Cos )2 + Sin2 a ] } .
Secondly compute a correction arc C for the solstitial points, thus:
C = arc Sin { Cos he Sin / Sin } ,
and then a correction arc c() for general solar longitudes, thus:
c() = arc Sin { Sin C Sin / R } .
Finally the solar altitude is given by:
h(a,) = he(a) + c() (c \ 0 as \ 0) .
Only one set of tables of h(a,) for general azimuths has come to light.
4.8 Tables of solar altitude in the prime vertical, for a specific latitude
The solar altitude in the prime vertical, h0(), is called in Arabic al-irtif{ alladh l samt lahu
or al-irtif{ al-{adm al-samt, that is, the altitude with no azimuth. It is determined by the
simple formula (cf. F13):
26 See King, Ibn Ynus, III.23.3, on the similar methods of Ibn Ynus and al-Brn.
27 See also ibid., III.22.2, and idem, Astronomical Timekeeping in Medieval Cairo, pp. 364-365.
28
See also ibid., pp. 371-373.
100 part i, chapter four
CHAPTER 5
The calculation of the azimuth of the sun and stars is a standard topic of Islamic spherical
astronomy, and certain Muslim astronomers compiled tables of the solar azimuth as a function
of the altitude. In 5.1 I describe tables having h and as arguments, and in 5.2 a table having
arguments H and h. The tables of azimuth as a function of time discussed in 5.3 are samples
of a category important in Islamic sundial theory.
The function a(h,) displays the same symmetry as t(h,) and T(h,) (2.1). It is generally
defined in zjes1 in terms of two auxiliary functions k(h,) and Sin (,), with which (cf. F15):
a(h,,) = arc Sin { R [ k(h,) - sin (,) ] / cos h } .
Certain Muslim astronomers tabulated k(h) and Sin () for specific latitudes, and at least one
table of the secant function:
G1(h) = R / Cos h ,
especially intended for azimuth calculations is attested in the Islamic sources. al-Khall
tabulated k(h) and Sin () for each degree of latitude, and another auxiliary function for
deriving the azimuth from their difference. The determination of the azimuth with these three
functions tabulated is very simple indeed: see further Ch. 8 and especially Sections 8.5 and
9.4a.
In most Islamic sources the azimuth (Arabic, samt) is reckoned from the prime vertical. The
azimuth of Mecca (Arabic, inirf al-qibla), on the other hand, is generally reckoned from the
meridian.
5.1 Tables of solar azimuth as a function of solar altitude and solar longitude, for a specific
latitude
1 See, for example, Nallino, al-Battn, I, pp. 183-184; Luckey, Beitrge, I, pp. 500-501 (al-Mhn); and
King, Ibn Ynus, III.20.
tables of solar azimuth 103
Fig. 5.1.1a: The additional values of the solar azimuth for solar altitudes from 84 to 88 in some early Mamluk
copies of Ibn Ynus tables defy explanation. [From MS Dublin Chester Beatty 3673, fols. 86v-87r, courtesy
of the Chester Beatty Library.]
These tables, extant in MSS Berlin Ahlwardt 5753 (incomplete), Dublin CB 3673, Escorial
r. 924,7, Cairo DM 108, Cairo Azhar falak 4382, Cairo MM 137, Gotha A 1410, Cairo DM
53, Istanbul Nuruosmaniye 2903 and 2925, Cairo MM 64, Cairo DM 690, DM 616, DM 739,
DM 786, DM 1101, DM 1224 and Cairo K 4044, are very accurately computed. Tables for
h = 30 and 35 are found in the kim Zj, MSS Leiden Or. 143, pp. 389-390, and Oxford
Hunt. 331, fols. 124v-127v.2 On the related tables of a(,h) in MS Cairo DM 1108,9 see 5.2.1
below.
It seems that the entries in Ibn Ynus tables for altitudes above 80 were either not complete
or were not contained in any manuscripts that were available in the 14th century. Additional
entries for altitudes 81 to 89 were added by some anonymous incompetent who was oblivious
of the fact that the maximum solar altitude in Cairo is about 831/2. Thus, for example, MS
2 On Ibn Ynus azimuth tables see also King, Ibn Ynus, III.22.2 (kim Zj); and idem, Astronomical
Timekeeping in Medieval Cairo, pp. 362-364 (Kitb al-Samt). The tables for altitude 30 in the two sources
are reproduced ibid., pp. 354-355.
104 part i, chapter five
Fig. 5.1.1b: Ibn Ynus table of the azimuth as a function of the solar altitude at the equinoxes and solstices.
On the left is an auxiliary table for sundial construction, of the kind for which another detailed survey would
be worthwhile. [From Cairo MM 57, fols. 1v-2r, courtesy of the Egyptian National Library.]
Gotha A 1410 contains entries for h running from 1 to 89, MS Dublin CB 3673 entries to
88 (see Fig. 5.1.1a), MS Cairo MM 137 entries to 86. The entries for 81 to 83 were
recomputed properly by Ibn al-Rshid (2.1.5), and were incorporated in al-Bakhniqs edition
of the Cairo corpus (2.1.1). Thus, for example, MS Cairo Azhar falak 4382 contains a corrected
set of azimuth tables with entries to 83 for h, and in MSS Cairo DM 108 and Cairo DM 53
these corrected entries are tabulated alongside the corresponding values of T(h,) and t(h,).
Further information on these tables is given in II-4.4 and II-5.2.
Ibn Ynus also tabulated the function a(h) at the equinoxes for h = 1, 2, ... , 60 (= )
with additional values for the solstices. This table is extant in MSS Cairo Azhar falak 4382,
fol. 71r, and Cairo MM 57, fol. 1v see Fig. 5.1.1b. MS Vatican Borg. ar. 217,2 (fols. 6r-
7r) of al-Kha}s auxiliary tables (6.15.1) contains a table of this same function for the
equinoxes with minor variants from Ibn Ynus table. Furthermore, al-Kha} reiterates Ibn
Ynus remark in the kim Zj to the effect that such a table also displays d().3 Notice that
if = 0 the formula for a(h) is (cf. F15):
a(h) = arc Sin { Tan h Tan / R } ,
3 Cf. King, Ibn Ynus, III.15.1b, and idem, Astronomical Timekeeping in Medieval Cairo, pp. 389-390.
tables of solar azimuth 105
Fig. 5.1.2: The last sub-tables in the corpus for Damascus, Fig. 5.2.1: An extract from the anonymous solar
in which al-alabs values for the azimuth are tabulated azimuth tables for Cairo serving solar longitudes
alongside al-Khalls values of the time since sunrise and Aries 4-6 / Virgo 24-24. [From MS Cairo
time to midday. [From MS Cairo MM 71, fols. 65v-66r, DM 1108,9, courtesy of the Egyptian National
courtesy of the Egyptian National Library.] Library.]
5.2 Tables of solar azimuth as a function of solar longitude and solar altitude, for a specific
latitude
5.3 Tables of solar azimuth as a function of meridian altitude and instantaneous altitude
The simplest method for calculating a(H,h) involves first finding the Sine of the altitude in
the prime vertical (4.8) corresponding to each value of H, thus (cf. F13):
Sin h0(H) = R Sin ( H - ) / Sin .
Then the azimuth is given by (cf. F16):
a(H,h) = arc Sin { [ Sin h - Sin h0(H) ] Tan / Cos h } .
Note that a(H,h) is independent of . Only one Islamic table of this function has come to my
attention.
5.4 Tables of solar azimuth as a function of time and solar longitude, for a specific latitude
The solar azimuth tables noted in this section were probably compiled from the corresponding
solar altitude tables listed in 4.1. The following list is not exhaustive.
5.5 Tables of solar azimuth at the equinoxes as a function of time, for a specific latitude
Tables of solar azimuth at the equinoxes are useful in sundial construction. The azimuth is
related to the time since sunrise by the simple formula:
a(T,) = arc Tan { Sin Tan T / R } .
5 See n. 1:22.
tables of solar azimuth 109
The solar rising amplitude (sa{at al-mashriq in Arabic) is determined by the equivalent
expressions (cf. F4 and F14):
() = arc Sin { R sin () / Cos } = arc Sin { Sin Sin / Cos }
= arc Sin { Sin [max ()] Sin / R } .
Several tables of () for specific latitudes are preserved, occasionally accompanying tables
of Sin () or Sin () (6.1 and 8.3).
6 The existence of this work is noted in Brockelmann, GAL, I, p. 625. On al-Kawsh and his zj see King,
Astronomy in Yemen, no. 7.
7 Published in King & Gingerich, Astronomical Observations from 13th-Century Egypt.
8 On al-Maghrib see Suter, MAA, no. 376; and Kennedy, Zj Survey, nos. 41 and 108. Another zj for
Maragha by al-Maghrib is preserved in Cairo: see Cairo ENL Catalogue, I, sub MM 188, etc.
110 part i, chapter five
= 30;0 (Cairo) and = 23;35. The values are given to two digits for each 5 of and are
intended to serve the construction of the markings on a horizontal dial (an orthogonal projection
of the celestial sphere onto the plane of the horizon). They are not based on the corresponding
values of Ibn Ynus (5.6.2). Yet another table based on the same parameters but now with
values to two digits for each 1 of is found in MS Berlin Ahlwardt 5773/5776,1 (Mq.
98,16+18), fol. 25v, of an Egyptian zj (5.6.1).
Only one Islamic table for finding the rising amplitude of non-circumpolar stars has come to
light. The underlying formula is (cf. F14):
( ) = arc Sin { R Sin / Cos } .
Note that al-Khalls auxiliary function f() (9.5) displays Sin (,) for each 1 of both
arguments.
In Almagest VI.12 Ptolemy presented a table of the maximum solar rising amplitude for the
seven climates. The values for each zodiacal sign are given in degrees and minutes and are
arranged in concentric circles for each climate, the so-called horizon diagrams.9 The
underlying value of is 23;51,20. I have not consulted any of the Arabic versions of the
Almagest in which this diagram is surely reproduced, but in three sets of Islamic sources we
find similar tables.
9 See Ptolemy, Almagest, transl. Gerald Toomer, p. 320; and Neugebauer, HAMA, I, pp. 37-39. On the climates
see n. 1:11.
112 part i, chapter five
Fig. 5.8.1: A table of solar rising amplitudes in the Ptolemaic tradition. Values are given for each pair of zodiacal
signs with the same amplitude (and declination), radially for the first climate out to the seventh. The (false)
impression given is of directions around the horizon, not least because the 16 winds are displayed in the inner
ring. [From MS Berlin Ahlwardt 5751, p. 215, courtesy of the Deutsche Staatsbibliothek (Preuischer Kultur-
besitz).]
11 On Abu l-Waf} (Suter, MAA, no. 167) see Sezgin, GAS, VI, pp. 222-224, and the references there cited,
also the article by A. P. Youschkevitch in DSB.
114 part i, chapter six
CHAPTER 6
The computation of time from solar altitude (cf. F9-12) is considerably facilitated if the
function:
B() = Cos () Cos / R
is tabulated first. This function B is called in Arabic al-al al-mulaq, the absolute base, or
simply al-al, the base. Tables of the function:
G() = R2 / [Cos () Cos ]
would be yet more useful, but Muslim astronomers generally prefered to use B(). To compute
B() it is necessary either to compute Cos () or Sin H(). Given a table of Sin H(), B()
can be found using (cf. F8):
B() = 1/2 { Sin H() + Sin H(*) }
This is clearly valid since:
H() = + () and H(*) = - ()
and (in modern notation):
cos cos = 1/2 { sin ( + ) + sin ( - ) }.
See also 2.1.1.
Another important function of Islamic timekeeping is:
C() = Sin () Sin / R ,
called in Arabic bu{d al-qur, literally the distance of the diameter. It measures the height
of the day-circle centre above the horizon. To compute C() one can either use Sin () or
B() together with Sin H() since:
C() = B() - Sin H(*) .
Given tables of B() or G() and C() or Sin H(), the determination of the hour-angle reduces
to an application of the equivalent formulae:
Vers t(h,) = [Sin H() - Sin h] R/B() = [Sin H() - Sin h] G()
Cos t(h,) = [Sin h - C()] R/B() = [Sin H() - C()] G() .
If we now define the function:
H(H,h) = Sin H - Sin h ,
called in Arabic fal al-jaybayn, the difference between the two sines, the first formula for
the hour-angle reduces to:
Vers t = H R/B = H G .
Likewise if we define the function:
b(h,) = Sin h - C() ,
called in Arabic al-al al-mu{addal, the modified base, the second formula for the hour-angle
reduces to:
tables of auxiliary functions for timekeeping 115
Cos t = b R/B = b G .
Similarly the functions:
B() = Cos Cos / R , C() = Sin Sin / R and G() = R2 / [Cos Cos ]
are useful in computing time from stellar altitudes.
I now present a description of various Islamic tables of Sin , Cos and Sin H (6.1. to 6.3),
B, C, G and related functions (6.4 to 6.11), b and H (6.12 and 6.13), and the inverse
trigonometric functions necessary to find the hour-angle once one has used the functions B,
C and G or Sin H (6.14 and 6.15). Finally I discuss certain star catalogues which present
numerical information for each star to facilitate reckoning time by night (6.16).
numerous other spherical astronomical tables for Damascus attributed to al-Khall (2.1.4). This
particular table, which is based on parameter = 23;31, is also contained in MS Damascus
hiriyya 9233, p. 156, of the main Damascus corpus for timekeeping, there introduced in the
name of al-anw (2.2.5). See also 6.3.2 and 8.3.2 and II-11.10 and 11.13.
To compute Cos () for given the easiest procedure is to compute () first and then use
a Sine table to find Cos (). In the Islamic sources we find tables of both Cos () and Cos
(). The former can be used to compute B() for a particular latitude (6.4) and the latter is
of limited use in certain operations in spherical astronomy, such as the conversion from ecliptic
to equatorial coordinates.7 The following tables of Cos () or Cos () are known.
6 On Riwn Efend (I-6.1.3) see Dorn, Drei arabische Instrumente, pp. 32-34; Brockelmann, GAL, II, p.
47l, and SII, p. 487, Azzawi, History of Astronomy in Iraq, pp. 319-320; Mayer, Islamic Astrolabists, pp. 81-
82, Kennedy, Zj Survey, no. X209; and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 377-384,
no. 246. See also II-7.10.
7 Cf. Jensen, Ab Nars Table of Minutes, pp. 6-7 and 14.
tables of auxiliary functions for timekeeping 117
Fig. 6.2.1: abashs table of the cosine of the declination of the complement of the solar longitude, together
with a table of the absolute base for latitude 29;30 (6.4.5), and tables of the auxiliary azimuth function for
various latitudes in Iran (8.1.2). [From MS Paris BNF supp. pers. 1488, fol. 207v, courtesy of the Bibliothque
Nationale de France.]
and is attributed to x-kh-s-r-w, where x denotes a carrier which could serve as a b}, t} or
th}, nn or y}, the combination corrupted from -b-sh for abash! see Fig. 6.2.1.
6.3 Tables of the sine of the meridian altitude, for a specific latitude
Tables of H() for a particular latitude are standard in zjes. The underlying formula is (cf.
F1):
H() = + () .
Tables of Sin H() (Arabic, jayb al-ghya) are less common and can be used either to compute
the function B() (6.4) or to find the hour-angle t(h,) using the formula (cf. F9):
Vers t(h,) = [ Sin H() - Sin h ] R / B() .
6.4 Tables of the absolute base for timekeeping by the sun, for a specific latitude
Several sources listed below contain tables of B() for particular latitudes. The function is
invariably called al-al al-mulaq or simply al-al. Note that in al-Khalls first set of auxiliary
tables (9.4 and also 9.7 and 9.11) the function 1/2B(,) is tabulated for each degree of both
arguments.
8 On Shhmardn Rz see Storey, PL, II:1, p. 45, no. 81; Sezgin, GAS, VI, p. 246; and Cairo ENL Survey,
no. B91.
9 On al-Nasaw (Suter, MAA, no. 214) see Sezgin, GAS, V, pp. 345-348, VI, pp. 245-246, and VII, pp. 182,
402, and 410-411, also III, p. 311, and the article by A. S. Saidan in DSB.
120 part i, chapter six
Fig. 6.4.1: This table of the auxiliary function called the base (al) or later the absolute base (al-al al-
mulaq), computed in the 10th century by a scholar well versed in mathematical astronomy as well as pure
mathematics, rather than by a scholar specializing in astronomical timekeeping, shows that Muslim interest in
practical astronomy started early. [From MS Cairo FF 11, fol. 78v, courtesy of the Egyptian National Library.]
and is referred to as al in the text. The values are computed to three digits for each degree
of with first differences (taful) to two digits. The latitude 35;34,40 is mentioned in the
accompanying text, and is attested for Rayy, being associated with the observations of al-
Khujand in the year 994.10 However, the text further states that the Cosine of this latitude
is 48;47,52, whereas the accurate value is 48;47,58,31. In fact, the value in the text was derived
10 On the derivation of this value of the latitude of Rayy see Schirmer, Studien zur Astronomie der Araber,
pp. 63-79.
tables of auxiliary functions for timekeeping 121
by linear interpolation (!) between the entries for 35 and 36 given to three (or perhaps four)
significant digits. If we, not unreasonably, assume that this erroneous value of the Cosine was
actually used to compile the table, Benno van Dalen has shown that the values used for Cos
() were extremely accurate for = 23;35,44 plus or minus 1 second. This unfortunately does
not correspond to the obliquity associated with the first derivation of the latitude 35;34,40,
namely, 23;32,21. We note in passing that 23;35,43,57,41 is the arc Sine of 24;1: this may
somehow be relevant to explaining the makeup of this table.
Other spherical astronomical tables in the Rawa, for example, those for () (fol. 59v of
the Cairo manuscript), are computed for = 36, a popular medieval value for Rayy and also
for the 4th climate.11
Kha} points out the fact that the latter can be computed from the former using the relation
noted in 6.0. Indeed the errors in the table of B() can be explained by the fact that the values
of Sin H() were derived from H() by linear interpolation in an accompanying table of Sin
which displays the function to three digits for each integral value of .
In MS Berlin Ahlwardt 5710, fol. 24r, of the prayer-tables of Riwn Efend (6.1.3) the
same table of B() occurs with a few additional copyists mistakes, alongside a more accurately
computed table of C(). Note that C() is easier to compute than B() for = 30 (see further
6.10.1).
The purpose of the tables described below is to facilitate the marking of B() on the instrument
called the sine quadrant (Arabic, al-rub{ al-mujayyab), a quadrant fitted with a trigonometric
grid that was popular amongst Muslim astronomers from the 9th to the 19th century.13 Com-
plicated trigonometric problems can be solved with the sine quadrant without any calculation
whatsoever. The grid consists of two orthogonal sets of equi-spaced lines drawn parallel to
the axes of the grid and dividing each axis into 60 parts. The circumference is divided into
90 equal degree intervals. A thread attached to the centre of the quadrant bears a movable
marker (Arabic, mur or mr). On the use of the instrument to compute the hour-angle from
solar or stellar altitudes 6.16.4 below.
Fig. 6.5.1a-b: Diagrams illustrating the use of al-Khalls tables for marking the absolute base on a sine quadrant.
13 On the sine quadrant, see, for example, Schmalzl, Geschichte des Quadranten; Wrschmidt, Gedosi ber
den Quadranten; and King, al-Khwrizm, pp. 28-29. The relatively complicated operation of finding the qibla
with a sine quadrant is discussed in detail in King, al-Khalls Qibla Table, pp. 111-115.
126 part i, chapter six
The first function b1() is intended to be used with a sine quadrant (radius R = 60) marked
with two special semi-circles (Arabic, d}irat al-tajyb) for finding Sines, drawn with the two
axes of the quadrant as diameters. In Fig. 6.5.1a if the thread OC cuts off the arc AC equal
to b1() then OD equals Sin b1(), that is, B(). Thus the marker can be moved to D and the
quadrant is ready to be used for computations of time from solar altitude which involve a
division by B().
The second function b2() is intended to be used with a Sine quadrant bearing the usual
grid and thread, but no marker. The use of this instrument is described by al-Khall in MS
Cairo MM 201, fols. 11v-12r, which also contains the table of b2() (fol. 12r). His treatise on
this quadrant is of considerable interest but I here restrict attention to the use of the table of
b2(). In Fig. 6.5.1b the thread OC is moved to the position such that the arc AC equals b2().
To divide any quantity, say x, by B(), simply measure EF equal x and then, because AD equals
Tan b2(), that is, B(), OE equals Rx / B(). Such a simple operation is clearly extremely
useful in timekeeping problems.
The existence of a table of B() for the latitude of Mecca has been noted in 6.4.4. Since the
argument runs only as far as = the table was clearly intended to be used for timekeeping
by the sun. No other tables of B() for particular latitudes which would be suitable for
timekeeping by the stars have been located. See, however, 6.16.4 below. Note also that al-fs
second auxiliary function (9.10) actually defines B(,) for each degree of both arguments,
and that al-Mridns second auxiliary function M2 (9.6) can be used to find B(,) with
facility, namely, by simply using and as arguments.
6.7 Tables of the reciprocal of the absolute base and related functions for timekeeping by the
sun
It is rather curious that we find tables of the function B() rather than its reciprocal G() in
the Islamic sources. As yet I have not come across any tables of G specifically intended for
timekeeping by the sun see, however, 6.8. The following tables of functions related to G()
have been located.
tables of auxiliary functions for timekeeping 127
14 On the Muala Zj see Kennedy, Zj Survey, no. 47; Cairo ENL Survey, no. C12; King, Lunar Equation
Table, p. 135; and idem, Astronomy of the Mamluks, pp. 535-536. See also Fig. I-9.8a.
128 part i, chapter six
6.8 Tables of the reciprocal of the absolute base for timekeeping by the sun and stars
The reader will by now have gained the impression that the Muslim astronomers tabulated
virtually every function that was of any use in computing the hour-angle from solar or stellar
altitudes. Yet I have not come across any Islamic tables of the function:
G = Vers D / Sin H = R2 / [ Cos Cos ] ,
either for a specific latitude or for all latitudes, such as might have been used to facilitate the
computation of a aylasn table of T(H,h) (2.3) or a universal table of T(H,h,D) (2.6). Note
that using such a table of G(H) for a particular latitude or G(H,D) for all latitudes, along with
a table of the function:
H(H,h) = Sin H - Sin h ,
such as compiled by Ibn al-Mushrif (6.13.1), the computation of the hour-angle reduces to (cf.
F9):
Vers t = H G .
On the other hand, two tables of the function G() have come to light and are described below.
The first was compiled for Cairo, where, however, as far as we know, no aylasn tables were
prepared. The second was compiled for Tunis and may have been used to compute a set of
tables for reckoning time by stellar altitudes. Note also that al-Waf}s first auxiliary function
(9.9) defines G(,) for each degree of and some 13 different values of .
Fig. 6.8.2: The table of the reciprocal of the absolute base in the Tunis corpus, together with a table of the sine
of the declination labelled as if it were an auxiliary azimuth function (see 7.1.5). [From MS Berlin Ahlwardt
5724, fols. 42v-43r, courtesy of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
130 part i, chapter six
The Muslim astronomers did not generally exploit the standard trigonometric functions other
than the sine and (co-)tangent.15 However, a table of the cosecant function is found in one
version of the Zj of the early-9th-century astronomer abash and others are found in the sets
of tables used by the Muslim astronomers for marking curves on astrolabes and quadrants (see
further 9.2). Also, a few simple tables of the secant function have recently come to light but
they are not referred to by the standard, if uncommon, Arabic term qur al-ill, hypotenuse
of the (horizontal) shadow (i.e., the length of the hypotenuse of the triangle formed by a
vertical gnomon and its shadow on a horizontal plane). Rather, they are considered as auxiliary
tables for timekeeping. See also 8.4 for tables of the secant function intended for azimuth
calculations.
I denote the secant function as found in the Arabic sources by G1, where:
G1() = R / Cos .
It can be used to find the auxiliary function G (6.7 and 6.8) using the relation:
G(,) = G1() G1() .
Thus, for example, in the text of the Zj of Ibn al-Bann} (6.9.4) the formula outlined for finding
t from h, and is equivalent to (cf. F9):
Vers t(h,,) = ( Sin H - Sin h ) G1() G1() .
Note that each of the four tables of G1 listed below contains the same significant error in the
entry for = 60. (Compare the Cosecant table of abash discussed in 9.1.) All of these tables
merit a new investigation.
Fig. 6.9.3: A table of the reciprocal of the absolute base (far left) amidst some anonymous tables for Cairo.
Two other tables display the tangent of the declination and the tangent of the solar declination (both omitted
from 7.1!). Another displays for each degree of solar longitude the difference between the lengths of daylight
in Cairo and Mecca, which should be investigated in the light of the acompanying treatise. [From MS Cairo
MM 204,7, fols. 80v-81r, courtesy of the Egyptian National Library.]
15
On Islamic trigonometric tables in general see n. 1:23.
16
Mills, Azarquiel, pp. 72-237. More recent studies on Ibn al-Zarqlluh include Toomer, Solar Theory of
al-Zarql, p. 331; and Julio Sams and Eduardo Mills, Ibn al-Bann}, Ibn Isq and Ibn al-Zarqlluhs Solar
Theory, in Sams, Studies, X. See also the article al-Zarql by Juan Vernet in DSB.
tables of auxiliary functions for timekeeping 131
The main part of the table is as follows. (The errors in the third digit are indicated in square
brackets.)
3 1;0, 5a [0] 27 1; 7,21 [+1] 51 1;35,27 [+7]
6 1;0,20 [0] 30 1; 9;17 [0] 54 1;42, 2 [-3]
9 1;0,45 [0] 33 1;11,32 [0] 57 1;50,10 [0]
12 1;1,21 [+1] 36 1;14,10 [0] 60 2; 1, 2 [!!!]b
15 1;2, 5 [-2] 39 1;17,11 [-1] 63 2;12c,10d [0]
18 1;3, 6 [+1] 42 1;20,44 [0] 66 2;27,31 [0]
21 1;4,16 [0] 45 1;24,51 [0] 69 2;47,27 [+1]
24 1;5;40 [-1] 48 1;29,40 [0] 72 3;14,10 [0]
a Text: 30, Mills: 5. b See below. c Text: 17, Mills: 12. d Text and Mills: 2
The remaining entries in the table in the Munich manuscript, compared with Mills entries
and the accurate values of Sec , are as follows:
Text Mills Accurate
75 3;51,50 75 3;51;50 75 3;51;49
78 4;48,21 78 4;48,21 78 4;48,35
81 5;23,33(!) 81 6;23,33 81 6;23,33
84 6;14, 1(!) 84 9;34, 1 84 9;34, 0
87 7; 6,30(!) 87 19; 6,30 87 19; 6,25
88 8;18,42(!) 90(!) 29,18,42 88 28;39,11
89 9;17,42(!) - (!) 56;17,40 89 57;17,56
90 -; 0,47(!!) 90 infinite
The last few entries in the Munich manuscript are thus badly garbled. Of considerable interest
is the entry for = 60, namely 2;1,2 rather than the accurate 2;0,0. It is difficult to account
for this error in view of the fact that most of the other entries are fairly carefully computed.
The tables of G1() discussed in 6.9.3-5 have the entry 2;1 for = 60.
17 In the 1980s all one that was known about Ibn Isq was what was recorded in Suter, MAA, no. 356, and
King, Astronomy in Yemen, p. 83. His zj is not listed in Kennedy, Zj Survey. The Zabid manuscrip of the
zj of the Yemen astronomer al-Daylam (4.3.4) contains some planetary tables due to Ibn Isq. MS Milan
Ambrosiana C82 of the perpetual almanac of Zacuto (4.3.7) contains some spherical astronomical tables due
to him. On the Hyderabad manuscript see King, Astronomy in the Maghrib, p. 32; Sams, Maghrib Zjes,
p. 93; and, more especially, Mestres, Hyderabad MS of the Zj of Ibn Isq, and idem, Zj of Ibn Isq. In
the new edition of Mestres (ibid., pp. 275 and 279-281), a value 2;0 occurs at argument 61, also in the
recomputation! The values in the table for 59-61 are 1;57, 1;1 (sic) and 2;5.
132 part i, chapter six
6.10 Tables of the altitude of the solar day-circle centre and related functions
18 On Ibn al-Bann} see Suter, MAA, no. 399; and Cairo ENL Survey, no. F23; as well as Renaud, Ibn al-
Bann}. His zj, not listed in Kennedy, Zj Survey, is entitled Minhj al-lib and is extant in several
manuscripts. The text of the introduction to the zj and a Spanish translation were published in Vernet, Ibn al-
Bann}, and the planetary tables have recently been studied in Sams & Mills, Ibn al-Bann} on Planetary
Longitudes.
19 On Ibn al-Raqqm see Suter, MAA, nos. 388 and 417; and Cairo ENL Survey, no. F22. His two zjes are
not listed in Kennedy, Zj Survey, but see now idem, Ibn al-Raqqms Tables. See also Carandell, Ibn al-
Raqqm sobre los cuadrantes solares.
tables of auxiliary functions for timekeeping 133
as that of al-Kha} (6.4.2), but there is no corresponding table of C() in MS Vatican Borg.
ar. 217,2 of al-Kha}s set. Note that for = 30:
C() = 1/2 Sin () .
6.11 Tables of the height of the centre of the arc of visibility for timekeeping by the stars
No tables of C() for particular latitudes have been located in the manuscript sources. (See,
however, 6.16.4 below where C is tabulated for specific stars.) Note that both al-fs first
tables of auxiliary functions for timekeeping 135
auxiliary function (9.10) and al-Mridns second auxiliary function M2 (9.6) define C(,)
for each degree of both arguments.
The function:
b(h,) = Sin h - C() ,
called al-al al-mu{addal, the modified base, is useful in timekeeping because the hour-angle
t(h,) is defined in terms of b by the simple relation (6.0):
Cos t(h,) = R b(h,) / B() = b(h,) G() .
In only one set of hour-angle tables has the compiler left us the set of tables of b(h,) which
he used to compute the hour-angle.
6.13 Tables of the difference between the sines of the meridian altitude and the instantaneous
altitude
The function:
H(H,h) = Sin H - Sin h
is useful in timekeeping because the hour-angle t(H,h) is defined in terms of H by the simple
relation (6.0 and 6.8):
Vers t(H,h) = H(H,h) G(H) .
Only one table of this function has come to light.
Tables of inverse trigonometric functions are comparatively rare in the Islamic sources. al-
Marrkush (4.2.4) tabulated the inverse Sine and Cotangent (base 12) (I.121-124 and 168-
169), and a later Egyptian copy of al-Khalls Universal Table, MS Cairo MM 43 (9.5),
contains tables of the inverse Sine and Tangent (base 12). Again, a 14th-century copy of the
main Cairo corpus, MS Dublin CB 3673, fol. 8r (2.1.1), and also MS London BL Or. 3624,
136 part i, chapter six
fol. 169v, of the Mukhtr Zj of Abu l-{Uql (2.1.2) contain tables of the inverse Sine function,
the latter probably taken from an earlier source. Tables of inverse spherical astronomical
functions are likewise rare. MS Cairo MM 43, mentioned above, contains a table of the function
() with values to two digits for each degree of , and MS Cairo MM 58, fol. 7r, contains
a similar table with argument increment 0;15.
In timekeeping the most useful tables of inverse trigonometric functions would be for the
Cosine and Versed Sine. No tables of the former have come to my attention, but notice that
al-Khalls auxiliary function K and al-fs function S (9.5 and 9.10) perform the final
operation for finding t(h,D). Several tables of the inversed Versed Sine have been located, and
the auxiliary functions V and V tabulated by al-Kha}, al-Khall and Ibn al-Mushrif (6.15.1,
9.4 and 9.8) likewise perform the final operation for finding t(h,H).
Fig. 6.14.3: Tables of the inverse Sine and Versed Sine functions and a table of the maximum half daylight
for all latitudes (based on obliquity 23;35) appended at the end of a copy of al-Khalls universal auxiliary
tables (see Fig. I-9.5a-b), of which the last few sub-tables for his third function are shown on the right. From
MS Istanbul Hamidiye 1453, fols. 265v-266r, courtesy of the Sleymaniye Library, Istanbul.]
I now describe a table for finding the hour-angle which has the absolute base B as one argument.
This is devised for a particular latitude. Note that the auxiliary tables of al-Khall for
timekeeping by the sun (9.4 and also 9.7 and 9.11) have B = 1/2 B as one argument but serve
138 part i, chapter six
Fig. 6.15.1: An extract from the auxiliary tables of al-Kha}, showing the surviving part of the table of V . [From
MS Vatican Borg. ar. 217,2, courtesy of the Biblioteca Apostolica Vaticana.]
all latitudes. Likewise the later auxiliary tables of Ibn al-Mushrif and al-f (9.8 and 9.10)
have B as one argument and serve all latitudes.
al-Kha}, an individual whose name is new to the history of Islamic science. His father was
a Mamluk prince who flourished around 1450.20 (The other tables in this manuscript are by
al-Waf}: see 9.9.) The main function tabulated by al-Kha} is discussed below. We first
consider the smaller tables of the standard functions which accompany the main set. Those
having solar longitude as argument have the same format as the tables for timekeeping in the
main Cairo corpus (2.1.1 and 5.1.1) and the underlying parameters are the same, namely,
= 30;0 and = 23;35.
(a) A table of Sin to three digits for each degree of .
(b) A table of Sin H() to three digits (6.3.1).
(c) A table of B() to three digits (6.4.2).
(d) A table of the solar altitude at the beginning of the afternoon prayer ha(). The entries
are not the same as those in the main corpus of tables for timekeeping compiled for
Cairo (II-4.9).
(e) A table of the solar azimuth as a function of the altitude, computed for the equinoxes.
Values of a(h) are given to two digits for each degree of h from 1 to 60 (= ) and
but for minor variants the entries are the same as those of Ibn Ynus (5.1.1).
(f) A table of a function f() which I am unable to identify. The table bears no title and
is not referred to in the extant fragment of the instructions. The equinoctial value of
f is noted in the margin. The following sample entries from the table may eventually
facilitate the identification of the function tabulated:
0 30 60 90 180 210 240 270
f() 54,45 58;24 59;49 59;59* 54;45 48;23 42;28 39;57
* read 60;0 (?)
At first sight the entries for > 0 resemble the values of Cos (), but closer
investigation reveals that this is not the function tabulated.
The entries for Sin H() and B() are generally in error by several digits in the third place.
Notes at the side of the tables indicate how to derive B() from Sin H(), and indeed the errors
in these tables can be explained by the fact that the values of Sin H were found by linear
interpolation in the Sine table. al-Kha} also points out that the table of a(h) displays d()
(5.1.1), adding that Ibn Ynus had compiled azimuth tables for the latitude of Cairo. Otherwise
there are no references to any of his predecessors, but unfortunately as already noted the
instructions to his tables in the Vatican manuscript are incomplete.
The main set of tables is likewise incomplete. The function tabulated is:
V(x,y) = arc Vers { R y / x },
and values are given to two digits for the domains:
x = 47;37, 47;57, ... , 51;57 and y = 1, 2, ... , Y(x),
where Y(x) is a certain maximum defined below. The instructions for finding t(h,) indicate
that one should first find B() and:
H(h,) = Sin H() - Sin h
using tables (a), (b) and (c) above. Then these arguments are to be entered in the main table,
which is called jadwal fal al-d}ir, hour-angle table: see Fig. 6.15.1. The result is indeed
the hour-angle, since:
t(h,) = V { B(), H(h,) } .
The argument x runs in intervals of 0;20 between the limits of B() for Cairo and the argument
y runs in unit intervals up to Y(x), which is the greatest integer less than the value of Sin H
corresponding to the value of underlying the argument x. The table originally contained about
825 entries, but in the Vatican manuscript only the page for y 31 is to be found now. The
entries are rather accurately computed.
Being devised for latitude 30 al-Kha}s tables are less useful than al-Khalls auxiliary
tables for timekeeping by the sun (9.4) which serve all latitudes. The possibility that al-Khall
knew of al-Kha}s tables can be ruled out because the Damascene astronomer preceded the
Cairene astronomer by close to a century. Nevertheless, the existence of al-Kha}s tables
implies that there were no tables of t(h,) and/or T(h,) currently available for Cairo (see,
however, 2.1.1). Ibn al-Mushrifs auxiliary tables (9.8) are an extension of the kind of tables
compiled by al-Kha}, for timekeeping by both the sun and the stars and intended to serve
all latitudes.
Most zjes contain a catalogue displaying the coordinates of prominent stars for a particular
epoch, either in the ecliptic or equatorial system or both.23 A smaller number of zjes and tables
21 On al-Jdar see Suter, MAA, no. 424a, Renaud, Additions Suter, no. 424a, and Cairo ENL Survey,
no. F26. The anonymous commentary has been previously attributed to Ibn al-abbk (Suter, MAA, no. 435
and Cairo ENL Survey, no. F28), which cannot be correct because he is mentioned in the text.
22 For a survey of trepidation in the Islamic sources see King & Sams, Islamic Astronomical Handbooks
and Tables, Section 3.7. See also the text to n. II-12:20.
23 See n. 1:44.
tables of auxiliary functions for timekeeping 141
for timekeeping give additional numerical information about particular stars.24 Thus, for
example, al-Battn in his Zj, after his main catalogue which gives the coordinates of 533 stars,
tabulates the meridian altitude, half-arc of visibility of five stars, as well as the longitudes of
the points of the ecliptic which rise, culminate and set with each star, all computed for the
latitude of Raqqa (36).25 The two Yemeni astronomers Abu l-{Uql (ca. 1300) and Ibn al-
M-s-r-b (?) (ca. 1325)26 tabulated the same functions for 30 stars, computed respectively for
the latitude of Taiz (13;40) and Zabid (14;40) (in both the Berlin and Sanaa manuscripts).
Our present concern is with a category of star catalogue which displays information particularly
useful for reckoning time by the stars, of which the following have come to my attention.
24 See Girke, Die frhesten islamischen Sternkataloge (cited in n. 1:44). A prime example of such tables
are those presented by abash al-sib for his instrument for reckoning time by night for any latitude (XIIb-
12), on which the various quantities are registered graphically: see Charette & Schmidl, abashs Universal
Plate, pp. 138-146.
25 Cf. Kennedy, Zj Survey, p. 155.
26 King, Astronomy in Yemen, no. 12. Perhaps this is Ibn al-Mushrif, who cannot, however, be identical with
the 15th-century Cairene astronomer with the same name (see I-9.8 and II-6.15).
142 part i, chapter six
Fig. 6.16.1: Some auxiliary tables for timekeeping by the stars, apparently for the latitude of Qandahar. [From
MS Berlin Ahlwardt 5751, pp. 186-187, courtesy of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
The underlying latitude is 33;25 (Baghdad), which is used elsewhere in the Berlin
manuscript, as in the table of oblique ascensions. By virtue of the epoch 304 whether Hijra
(?) [= 916/17] or Yazdigird (?) [= 936] the star catalogue cannot be due to abash himself.
The star catalogue in MS Istanbul Yeni Cami 784,2, of abashs Zj (see fol. 192r) is simply
that of the Mumtaan Zj, compiled for the year 214 H [= 829/30] (6.16.3).27
27 Debarnot, Zj of abash al-sib, p. 57. These tables are edited in an unpublished paper by Dorothea
Girke (see n. 1:44). See now Charette & Schmidl. abashs Universal Plate, p. 139, n. 113.
28 There are no geographical tables in the Escorial manuscript. According to Kennedy & Kennedy, Islamic
Geographical Coordinates, pp. 55-56, only the zjes of al-Maghrib (Kennedy, Zj Survey, nos. 41 and 108
see note 5:8) out of all the numerous Islamic sources investigated have this value for Baghdad. See also King,
Earliest Muslim Geodetic Measurements (cited in n. 2:26), pp. 221 and 225-227.
29
See Delphin, Lastronomie au Maroc.
144 part i, chapter six
Fig. 6.16.3: The star-table in a recension of the Mumtaan Zj, showing various functions useful for timekeeping
in addition to the basic ecliptic coordinates of the stars and their declinations. [From MS Escorial r. 927, fols.
96v-97r, courtesy of the Biblioteca de El Escorial, taken from the 1986 Frankfurt facsimile.]
Cos t = R [ Sin h - C ] / B .
On the other side of the instrument there is an alidade and altitude scale for making such
measurements of stellar altitude, as well as a table purporting to display the rising times of
the signs of the zodiac for each degree of latitude from 1 to 30 (the entries do not make
much sense to me).
tables of auxiliary functions for calculating the length of daylight 145
CHAPTER 7
of oblique ascensions are also contained in the later zjes of al-s, al-Ksh and Ulugh Beg.4
Ibn Ynus entries are computed to degrees and minutes, al-ss and al-Kshs to seconds,
and Ulugh Begs to thirds.
7.1 Tables of the Tangent of the declination for calculating the equation of half-daylight for any
latitude and right ascensions
As noted in 7.0, the function Tan () is of use in computing either d() or (). Most of the
tables of Tan () listed below have been mentioned in previous studies. Note that tables of
the auxiliary functions Sin d() and Sin (), which are easily derived from tables of Tan (),
also occur in the Islamic sources (see 7.1.2 and 7.1.5; and 7.2, 9.2 and 9.3, respectively). See
also Fig. I-6.9.3 for some other tables of Tan and Tan (), which by an oversight have not
been included in this section.
and that he used the value for from the Handy Tables, namely 23;51 and it seems fairly
certain that the above-mentioned tables are ultimately due to al-Khwrizm.11 The factor 150/
12 essentially converts the product of two tangents (the second to base 12) into a sine (to base
150). See also 7.1.9 on what is probably another such table from 9th-century Baghdad.
Zj of al-Majrit published in Suter, al-Khwrizm, pp. 169-170, which is to base 60, is not original to al-
Khwrizms Zj. On these tables see also King, al-Khwrizm, pp. 2 and 34, n. 12.
11 In MS Berlin Ahlwardt 5793 (Landberg 56), fols. 93v-95v, appended to a unique copy of the treatise on
the construction and use of the astrolabe by al-Khwrizm (see King, al-Khwrizm, p. 23), there are some
tables which may also be by him. Of particular interest is a table of normed right ascensions () with values
to 3 digits for each 3 of from 273 to 360. This may be al-Khwarizms original table, for as shown in
Neugebauer & Schmidt, op. cit., p. 224, the values in the table of () in al-Majrs Zj do not correspond
to values of () which underlie certain calculations in a text relating to the use of al-Khwrizms table of
e(). Neugebauer and Schmidt noted three values of () in this text as follows:
0;54,53,40 ( = 1), 1;49,33,20 ( = 2), and 27;53 ( = 30).
The corresponding values which I derive (using linear interpolation for = 1 and 2) from the table of ()
are 0;54,54, 1;49,48, and 27;53,26, so that this table of () is probably related to al-Khwrizm.
12 See further King, Ibn Ynus, III.13.3d, on this table.
148 part i, chapter seven
13 The existence of this table was noted in Kennedy, Zj Survey, p. 156. There is more work to be done
on the two zjes of Kshyr (see n. 3:4); in particular, an investigation of the tables relating to spherical astronomy
in the various available manuscripts might establish the location of his activities. In the DSB article on him it
is stated, after a medieval source, that he was active in Baghdad, but somehow I doubt this.
14 On this work, not listed in Kennedy, Zj Survey, see Zimmermann, Dustr al-Munajjimn.
tables of auxiliary functions for calculating the length of daylight 149
(p. 174). The entry for = 90 in the table of e() is 2;10,46 (accurately, 2;10,45). The same
table with minor variants occurs in MS Istanbul Kandilli 249 of the Zj of the early-14th-century
Tunisian astronomer Ibn al-Raqqm (6.9.4), whose declination table, however, is based on the
distinctive value 23;32,40 for . Here the function is called nisab juyb al-falt li-{amal al-
mali{, the ratios of the Sines (i.e., Sines and Cosines) of the excesses (?) for finding
ascensions.
A related table in the Almanac of Ibn al-Zarqlluh (p. 226) is entitled jadwal anf ful
uhrt al-kawkib, table of the half excesses of visibility of stars, and displays values of
the function:
e() = 1/12 Tan60
to three digits for each 3 of (and also = 88 and 89). It is of interest that the entry for
45 is 4;59,59 rather than 5;0,0: this suggests that the table was not compiled directly from
a tangent or cotangent table. (See also 6.9.1.)
Another table of the same function e() displaying values to three digits for each degree
of argument from 1 to 60, and also having the entry 4;59,59 for argument 45, is contained
in MS Berlin Ahlwardt 5724, fol. 43r, of an anonymous corpus of tables for Tunis (2.3.5).
In the Berlin copy the table is entitled jadwal al-ikhtilf al-ufuq, table of variation due to
the horizon and the function is labelled nisbat jayb al-bu{d, the ratio of the Sine of the
declination see Fig. 6.8.2.
The advantage to be gained from dividing Tan60 () or Tan60 by 12 is that using only
tables of Cotangents to base 12 and Sines to base 60 one can compute d() or d() for any
latitude with facility. The formulae are:
d(,) = arc Sin60 { Tan60 () / 12 Cot12 }
d(,) = arc Sin60 { Tan60 / 12 Cot12 } .
Paris BNF ar. 2486, fol. 235v, there is a table of the function Sin d(), perhaps based on the
above-mentioned table of e() and computed for = 33;25 (Baghdad).
17 On al-Fris and the Muaffar Zj see Suter, MAA, nos. 349 and 349N; Kennedy, Zj Survey, no. 54,
and King, Astronomy in Yemen, no. 6. An 1822 study Lee, Astronomical Tables of Al Farsi, is still most useful.
See also n. 7:26 below and II-2.2 and 12.0.
18 On al-Fahhd see Kennedy, Zj Survey, nos. 23 and 84, inter alia, and, most recently, Pingree,
Astronomical Works of Gregory Chioniades.
19 See Neugebauer, Studies in Byzantine Astronomy, p. 36b (entries for fols. 43lr-433r).
tables of auxiliary functions for calculating the length of daylight 151
set of al-Khalls hour-angle tables for Damascus (2.1.4); and in MS Cairo TR 275, p. 73,
copied in 858 H [= 1454], of the zj entitled al-Lum{a by the early-15th-century Egyptian
astronomer al-Kawm al-Rsh20 (but not in other copies of this zj); and in the manuscript in
a private collection in Sanaa of the astronomical miscellany of the 14th-century Yemeni ruler
al-Suln al-Afal (2.1.2). A significant entry in all of these sources is 1;50,41 for = 60:
al-Marrkush has the accurate value 1;50,48.
20 On al-Kawm al-Rsh (Suter, MAA, no. 428) see Cairo ENL Survey, no. D41.
152 part i, chapter seven
is tabulated to four significant decimal digits for each degree of . This table and accompanying
one of () correspond closely to recomputation with the parameter = 23;28,54. Taqi l-
Dns Sine and Tangent tables in MS Istanbul Esat Efendi 1976 are also given decimally, but
to base 10 rather than base 1.21
7.2 Tables of auxiliary functions for calculating the equation of half daylight for any latitude
21 See n. 1:22.
22 Nallino, al-Battn, II, pp. 58-59; also Neugebauer, HAMA, II, pp. 980-982.
23 See, for example, Stahlman, Handy Tables, p. 264.
tables of auxiliary functions for calculating the length of daylight 153
The entries in the tables are rather accurately computed, but, unlike many of the other
standard spherical astronomical tables in the Zj, are not taken from the kim Zj of Ibn Ynus
(7.1.3). Indeed it would appear that Abu l-{Uql took these auxiliary tables from the same
source from which he took the table of Indian Sines, the identity of which is not yet
established. The attribution of these auxiliary tables to Ibn Ynus is put in question by the
fact that the values of max d() do not generally correspond to his values for = 90 in his
tables of oblique ascensions, and also, for example, by the fact that the auxiliary tables have
Sin [max d(45)] (= Tan ) = 23;11, whereas Ibn Ynus value for Tan is 23;11,33,23 (
23;12). The formula for d(,) noted in 7.2 was used by Ibn Ynus, but it was also used by
certain of his predecessors such as al-Battn and anyway is equivalent to the method proposed
by Ptolemy in the Almagest.24
24 Cf. King, Ibn Ynus, III.14.1d; Nallino, al-Battn, I, p. 188; Neugebauer, al-Khwrizm, p. 51; and
Almagest, II.7 (pp. 92-94).
25 On al-Khzin and the Sanjar Zj see Kennedy, Zj Survey, no. 27, and the article by Robert E. Hall
in DSB. The existence of these tables is noted by Kennedy on p. 159. More recent studies include King, Mecca-
Centred World-Maps, pp. 71-75 and 564-585, and Pingree, Editing the Zj al-Sanjar, pp. 105-113.
26 al-Fris lists some 28 zjes known to him (see Lee, Notice (cited in n. 7:17), and also Kennedy, Zj
Survey, no. 54), and the Sanjar is not one of them.
27 Neugebauer, Studies in Byzantine Astronomy, p. 30 (Appendix 12).
154 part i, chapter seven
7.3 Tables of proportional parts of the tangent of the local latitude, for calculating the equation
of half-daylight
28 Kennedy & Kennedy, Islamic Geographical Coordinates, p. 473 (sub UTT). See also n. II-10:34.
tables of auxiliary functions for calculating the length of daylight 155
7.4 Tables of auxiliary functions for calculating the arc of daylight approximately
In this section I describe a table for computing the arc of daylight for any latitude, underlying
which is the assumption that d(,) varies linearly with each of and . It is worth mentioning
that in various Byzantine manuscripts of the Handy Tables there is a table of a function (),
tabulated to one digit for each degree of , with which the equation of daylight is given
approximately using:
d(,) () max d() .
William Stahlman has suggested that this table is a Byzantine addition to the Handy Tables.29
By virtue of the fact that:
() Sin () /R
and that tables of Sin () /R (to two digits) were used by certain Muslim astronomers to
determine d(,) (7.2), it may be that this particular table is ultimately of Islamic provenance.
On the other hand, if the table is due to Theon then it may have provided the inspiration for
the more elaborate Islamic tables.
29
See Stahlman, Handy Tables, pp. 265 and 117-122, for table and commentary; also Neugebauer, HAMA,
II, pp. 980-982.
156 part i, chapter eight
CHAPTER 8
As noted above (5.1), the standard Islamic method for finding the solar azimuth from the solar
altitude involves the use of two auxiliary functions (cf. F15). The first of these is called in
Arabic ikhtilf al-ufq, literally difference of the horizon or iat al-samt, azimuth com-
ponent, and is defined by:
k(h,) = Sin h Sin / Cos .
For lack of simple, meaningful English equivalent, I shall refer to this function as the auxiliary
azimuth function. The second auxiliary function is called in Arabic jayb sa{at al-mashriq or
simply jayb al-sa{a, which means Sine of the rising amplitude. This is determined by:
L(,) = Sin (,) = R Sin / Cos .
With these two functions the azimuth is defined by:
a(h,,) = arc Cos { R [ k(h,) - L(,) ] / Cos h } .
In 8.1 and 8.3 I describe the tables of k(h) and L(,) for specific latitudes which have been
located in the Islamic sources, and in 8.5 al-Khalls tables of both functions for each latitude.
The quantity:
m(h,,) = k(h,) - L(,)
is usually called ta{dl al-samt, equation of the azimuth, in late medieval Arabic. At least
one Muslim astronomer compiled a table of the secant function G1(h) (see already 6.9)
specifically to facilitate azimuth calculations. With this the determination of the azimuth
reduces to:
a(h,,) = arc Sin { m(h,,) G1(h) }.
I discuss this secant table in 8.4. al-Khall further tabulated a double-argument auxiliary
function with the equation of the azimuth as one argument and the solar altitude as the other
(8.5.1*): the function gives the azimuth without further ado.
Tables of k(h) for particular latitudes are found in the following sources. Note that al-Khalls
auxiliary function g (9.5) defines k for each degree of both h and , thus:
g(h) = Sin h Sin / Cos = k(h,) .
respectively. Values are given to three digits for each degree of argument and are extremely
accurately computed. Ibn Ynus doubtless used his table for Cairo to compile his extensive
azimuth tables (5.1.1).
8.1.2 Sanjar al-Kaml: Shiraz, Rayy, and two other cities in Iran
In MS Paris BNF supp. pers. 1488, fol. 208r, of the Ashraf Zj (2.3.1) there are tables of k(h)
for the latitudes of four cities in Iran see Fig. 6.2.1. They bear the title jadwal al-sumt, table
of azimuths, but the tabulated function is referred to as al, base. Values are given to two
digits for each degree of h and the underlying latitudes are stated to be:
29;30, 30;22, 33;18 and 36;0 .
The first latitude is that of Shiraz and is used in other tables in the Zj (see, for example, 6.4.5),
but the second and third values do not occur in the Ashraf geographical tables, or in any other
known medieval geographical tables,2 and so the tables of k(h) may have been taken from an
earlier source. The fourth value is standard for Rayy and the 4th climate.3
8.2 Tables of proportional parts of the tangent of the local latitude, for calculating the auxiliary
azimuth function
2 In other words, they are not found in Kennedy & Kennedy, Islamic Geographical Coordinates. For the
geographical table in the Ashraf Zj see now King, Mecca-Centred World-Maps, pp. 74-75 and 564-585.
3 See Kennedy & Kennedy, Islamic Geographical Coordinates, p. 284, and King, Geography of Astrolabes,
pp. 6-8.
158 part i, chapter eight
8.3 Tables of the Sine of the solar rising amplitude, for a specific latitude
On tables of the secant function intended to facilitate the computation of the hour-angle see
6.9 above.
4 The table for Baghdad is published already in Schoy, Gnomonik der Araber, p. 81.
tables of auxiliary functions for azimuth calculations 159
Fig. 8.4.2: The very special auxiliary function, actually just the secant, for calculating the rising amplitude, found
in the anonymous Sfax treatise. That such a table just pops up in such a work points to the fact that a large
proportion of our potential sources are lost for all time. [From MS Cairo K 7584,1, fol. 55r, courtesy of the
Egyptian National Library.]
al-samt li-l-shams wa-l-kawkib, table for solar and stellar azimuth and there is no reference
to the use of the function G1 in timekeeping calculations.
5 See n. 1:48.
160 part i, chapter eight
8.5 Auxiliary tables for finding the solar azimuth for any latitude
Only in January, 2001, did the following tables come to my attention. I mention them here
only in passing: a more detailed account is in 9.4* and II-10.3*.
8.5.1* al-Khall
MS Bursa Haraiolu 1177,4 (fols. 72r-90r), copied ca. 1450, is a unique copy of a set of
tables by Shams al-Dn al-Khall, not identified as the compiler, which contain sub-tables for
the following functions:
L() = Sin () and k(h)
for each 1 of from 1 to 49, and each 1 of the arguments and h. The value of underlying
the first table is about 23;30, possibly 23;31. With these the instructions describe how to
determine the ta{dl al-samt. The third sub-table displays the function:
K(x,y) = arc Cos { x R / Cos h }
for each unit value of x and each 1 of h. These universal auxiliary tables represent the ultimate
solution to the problem of determining the solar azimuth a(h,,). See further 9.4*, where they
are considered in the light of al-Khalls other tables.
tables of universal auxiliary functions 161
CHAPTER 9
In the 9th century abash compiled a set of tables of functions which had no specific
astronomical significance, but which were so conceived that combinations of them would lead
to the solution of certain problems in spherical astronomy. In the 10th century al-Nayrz and
Ab Nar compiled other sets of such functions. These three sets of tables contain a few hundred
entries and they are not ideally suited for use in the solution of problems in timekeeping.
However, in the early 14th century, the Cairo astronomer compiled a monumental table for
timekeeping, with close to half a million entries, and proposed its use as a universal auxiliary
table for solving all of the standard problems of spherical astronomy, including the determi-
nation of the qibla and the coversion of ecliptic and equatorial coordinates. Such a table, before
the age of printing, was doomed from the outset, and it is small wonder that it survives only
in a single copy in the hand of its author. In the mid 14th century, the Damascus astronomer
al-Khall compiled his more subtle, and eventually more useful, universal auxiliary tables.
These contain over 13,000 entries and were likewise devised to solve the standard problems
of timekeeping for all latitudes, with no calculation beyond addition, subtraction and inter-
polation.
In the sequel I briefly describe the auxiliary tables of abash, al-Nayrz, Ab Nar and
al-Khall (9.1, 9.2, 9.3 and 9.5), all of which are known from previous studies, and I also present
a description of several other sets of auxiliary tables which have come to light during recent
research. Most of these, like al-Khalls tables, were intended to be used for problems in
timekeeping. Most of the functions displayed in these tables are extensions of the auxiliary
functions discussed in Chs. 6 and 8 to serve more than one latitude. Only the auxiliary tables
of al-Khall, Ibn al-Mushrif and al-Waf} (9.5, 9.8 and 9.9) take advantage of the fact that
the problems of determining the hour-angle and azimuth from the altitude, declination and
latitude, are mathematically equivalent.
abash, called al-sib, meaning the astronomer rather than the calculator, lived for over
a hundred years and was most active in Baghdad, Samarra and Damascus in the first half of
the 9th century.1 He was the author of several zjes, one of which survives in a later recension
1 On abash (Suter, MAA, no. 22) see Kennedy, Zj Survey, nos. 15 and 16; Sezgin, GAS, V, pp. 275-
276, and VI, pp. 173-175; King, Mecca-Centred World-Maps, pp. 40-41 and 345-349; and also nn. 2-4 below.
162 part i, chapter nine
Fig. 9.1a-b: The auxiliary tables of abash as they occur in a Mamluk zj. On the right of the first double page
is a table of the tangent of the solar declination (7.1.9). [From MS Paris BNF ar. 2520, fols. 69v-71r, courtesy
of the Bibliothque Nationale de France.]
tables of universal auxiliary functions 163
in MS Istanbul Yeni Cami 784, fols. 69v-229r, from the 13th century. Another anonymous
recension of a different zj by abash is contained in MS Berlin Ahlwardt 5750 (Wetzstein
90), copied ca. 1300. The first source has been studied in detail by Marie-Thrse Debarnot.2
The Istanbul manuscript contains two related sets of auxiliary tables, both apparently
compiled by abash himself and both entitled Jadwal al-taqwm, Rectification Table.3 The
two sets of tables in the Istanbul manuscript have been analyzed in detail in a valuable study
by Rida A. K. Irani, alas unpublished.4 The main difference between the two sets is that one
is based on = 23;35 and the other on = 23;33. Ibn Ynus (2.1.1) noted with disapproval
that in his Mumtaan Zj called al-Qnn abash used = 23;35 in the solar declination
tables and = 23;33 in the auxiliary tables.5
A set of abashs tables for = 23;35 is also contained in a treatise preserved in MS
Bankipore 2468,8 (fols. 50v-66r), copied 631 H [= 1233/34], which consists of a discussion
by Ab Nar (9.3) of abashs tables and procedures.6 Yet other sources exist for the study
of abashs tables.7 For example, there are various related auxiliary tables in MS Berlin
Ahlwardt 5750 of the recension of abashs Zj (9.2), and another copy of the auxiliary tables
has been located in MS Paris BNF ar. 2520 of a recension of the Egyptian Muala Zj see
6.7.1 and Fig. 9.1a-b. Again, a table of one of abashs functions occurs in MS Paris BNF
supp. pers. 1488 of the Persian Ashraf Zj see 6.2.1 and Fig. 6.2.1.
abashs auxiliary functions are the following (R = 60):8
F1() = 2() = arc Tan { Tan Sin / R }
F2() = Cos () = Cos { arc Sin [ Cos Sin / R ] }
F3() = R Cos / F2() = R Cos / Cos ()
F4a() = Tan Sin / R
F4b() = Tan .
They are of use in solving particular problems of spherical astronomy, though not those directly
relating to timekeeping. Values are given to three digits for each degree of argument and I
have not investigated their accuracy. Maybe before I retire I shall organize a seminar to prepare
a publishable version of abashs work.
MS Leiden Or. 468 (280 fols., copied ca. 1400), contains the introduction and first of five
maqlas of a work entitled Kanz al-yawqt fi st{b al-mawqt and compiled by an anony-
mous Egyptian author ca. 1350.9 The work consists mainly of material culled from the Muala
Zj and the Kitb al-Mabdi} wa-l-ghyt of al-Marrkush (II:6.6-7), both compiled in Cairo
in the late 13th century. Besides various spherical astronomical tables also found in al-Marr-
kushs treatise, the Kanz al-yawqt contains a set of auxiliary tables not found in any other
known sources. The tables (fol. 147v-150r) are entitled jadwal al-nisab, table of ratios and
display values of four functions for each half-degree of argument from 0;30 to 90 to three
sexagesimal digits. The entries in the tables are garbled but the tabulated functions are recog-
nizable as sexagesimal multiples of:
cos () , cos / cos () , tan tan and tan .
I have not compared the entries in this Egyptian source with those found in MS Paris ar. 2520
of the Egyptian Muala Zj mentioned above.
In MS Berlin Ahlwardt 5750 of the anonymous recension of one of the zjes of abash (9.1)
there are two sets of auxiliary tables, the first (fols. 82r-84v) entitled al-jadwil al-jmi{a, the
universal tables, and the second (fols. 85r-87v) entitled jadwal al-nisab, table of ratios.
Various operations with the first set are outlined in the text of the Zj. These tables have not
been discussed before in any detail,10 and they deserve a more thorough investigation than is
possible here. Indeed our knowledge of early Islamic mathematical astronomy in general will
be greatly improved when each of MSS Istanbul Yeni Cami 784 (fols. 69v-229r) and Berlin
Ahlwardt 5750 of abashs Zj, as well as MS Escorial r. 927 of the anonymous recension
of the Mumtaan Zj, have been subjected to a critical comparative analysis.11
The first set is clearly related to al-jadwal al-jmi{, the universal table, of the late-9th-
century astronomer al-Nayrz.12 Ab Nar (9.3) mentions that al-Nayrz incorporated
abashs jadwal al-taqwm into his Zj and added some more functions of his own.13 This zj
is alas no longer extant. The first set of auxiliary tables in the Berlin manuscript displays ten
functions, with entries computed to three digits for each degree of argument. Five of these are
simply as follows (R = 60):
Sin and Vers ,
(), Sin () and Cos () ( = 23;35) .
The remainder are labelled abashs second, third and fourth (functions), and al-Nayrzs
first and second (functions). The first three are indeed abashs functions F2 (computed for
= 23;33), F3 and F4a (computed for = 23;35), and the remaining two are simply F2 and
F3 (also for = 23;35), to base 150 rather than 60.
This curious assortment of tables is followed immediately by the second set displaying
another ten functions. The entries are also computed to three digits for each degree of argument,
and the functions tabulated are the following (R = 60):
g1() = Cos () [= F2()]
g2() = Sin () [= g10()]
g3() = ()
g4() = Tan Tan / R
g5() = Tan
g6() = ()
g7() = ()
g8() = R / Sin
g9() = Tan / R
g10() = Sin () [= g2()]
Note the following points. First, the tables of g1, g2, g3, g4, g6, g7 and g10 are based on =
23;35 but the entries in the table of g10 do not correspond precisely to those in the table of
g2. Second, the table of g1 differs from both of abashs tables of F2. Third, the value of Tan
underlying the table of g4 is 26;11,34 (the accurate value is 26;11,33). (See 7.1 on other values
of this parameter.) Fourth, g4() is tabulated only up to = 60. For the remaining arguments
up to 90 a different function, which I am unable to interpret, is tabulated. Sample entries are:
61 0;47,15 . 85 4;49,22
62 0;49,13 . 86 6;14,34
63 0;51,53 75 1;37,45 87 8;19,20
64 0;53,41 . 88 12;31,30
65 0;56,11 . 89 26;30, 0
. 80 2;33,33 90 50; 0, 0
. .
70 1;11,18 .
Fifth, the function f8 is simply the modern cosecant: tables of this function were also used in
Islamic instrument construction.14
Here again we have at first sight a rather motley set of auxiliary tables. But we do well
not to forget that we are dealing with tables compiled almost over a thousand years ago. The
anonymous compiler has included a function equivalent to abashs F2 and Ab Nars f5 as
well as others equivalent to Ab Nars f2 and f3. I have not investigated the accuracy of either
set of auxiliary tables in the Berlin manuscript.
14 The cosecant table is published in Schoy, Beitrge zur arabischen Trigonometrie, p. 15, from defective
photographs (Schoy was able to publish only 60 entries rather than the 90 that are found in the manuscript).
For an example of a cosecant table for use in marking curves on astrolabes and quadrants see the Ottoman
source MS Princeton UL Yahuda 373, fol. 74r. Here the cosecant and cotangent are tabulated side by side to
three sexagesimal digits for each degree of arc: on one possible application of such a pair of tables see King,
al-Mridns Universal Quadrant, App. A.
166 part i, chapter nine
Ab Nar ibn {Irq15 worked in Khwrizm ca. 1000 and was a teacher of the illustrious al-
Brn.16 He was the author of numerous works of considerable merit, two of which concern
the present study. The first of these is a commentary on the auxiliary tables of abash and
al-Nayrz (9.1-2), and the second is a treatise written to al-Brn in which he introduces a
set of auxiliary tables of his own. Both of these texts were published in Hyderabad in 1948
using MS Bankipore 2468,14 (fols. 86v-93r), copied 631 H [= 1233/34], and Ab Nars
auxiliary tables have been analyzed in a valuable study by Claus Jensen.17 MS Oxford Bodleian
Thurston 3, fols. 111r-114r, copied 675 H [= 1276/77] is another copy of Ab Nars treatise
on his own auxiliary tables see Fig. 9.3.
Fig. 9.3: The end of Ab Nars treatise on his auxiliary tables and the first tables. [From MS Oxford Bodleian
Thurston 3, fols. 113v-114r, courtesy of the Bodleian Library.]
15 On Ab Nar (Suter, MAA, no. 186) see the article Manr ibn {Al ibn {Irq by Julio Sams in DSB,
and Sezgin, GAS, V, pp. 338-341, and VI, pp. 242-245, and more especially Sams, Estudios sobre Ab Nar.
16 On al-Brn see the splendid article by Ted Kennedy in DSB, and also II-2.2.
17 See Jensen, Ab Nars Table of Minutes.
tables of universal auxiliary functions 167
The underlying trigonometric functions used by Ab Nar are to base 1 rather than 60, but
values are, of course, given sexagesimally. This explains the title Jadwal al-daq}iq, Table
of Minutes (i.e., sexagesimal fractions), which he gave to his auxiliary tables. The functions
tabulated are the following:
f1() = sin / cos
f2() = tan tan
f3() = tan () / tan
f4() = sin
f5() = cos () .
Ab Nar used Ptolemys value 23;51,20 for rather than the value 23;35 which was generally
accepted in his time. His tables contain 225 entries, rather carelessly computed to four
sexagesimal digits.18
Ab Nars f4() is simply the sine function, and the other functions f1(), f2() and f3()
are respectively:
sin [max ()] , sin [max d()] and sin () .
Tables of max (), max d() and its Sine, and Sin () occur in several other Islamic sources
(7.2 and 9.2). Note that Ab Nars f5 is equivalent to abashs F2.
Ab Nars functions are not ideally suited to solving the problems of timekeeping. For
example, to find T(h,,) he suggests a method which involves first finding () and sin d(,).
Thus we begin by using the relations:
() = f4-1 { f4() f4() } and sin d(,) = f2() f3()
and then apply:
T(h,,) = f4-1 (sin d) + f4-1 { [ f4(h) sin d + f4(h) ] / f4( + ) - sin d } .
It is not difficult to show that this last formula is equivalent to F11.
Ab Nar does not describe the computation of the corresponding azimuth using his tables.
The procedure would be clumsy, although the value of sin () is conveniently given by:
sin () = f1() f4() .
Najm al-Dn al-Mir (I-2.6.1) intended his monumental triple argument table of a function
F(x,y,z) to be used as a universal auxiliary table. A commentary on the use of this table for
this purpose is in MS Dublin 102,1, copied ca. 1325, and has been studied in detail by Franois
Charette.19 I refrain here from a detailed commentary, not least in order to encourage the reader
to consult Charettes study, which is to be regarded as a supplement to the present work. Suffice
it to say that Najm al-Dn describes not only the use of the table to find the time since rising
and the azimuth, and special cases thereof such as the determination of the length of daylight
and rising amplitude, but also to find the qibla using the accurate formula and to covert between
ecliptic and equatorial coordinates. Also, that Najm al-Dn shows as much originality in his
18 On the accuracy of these tables see King, Al-Khalls Auxiliary Tables, p. 107, n. 7.
19 Charette, Najm al-Dns Monumental Table.
168
part i, chapter nine
Figs. 9.4a-b: Two extracts from al-Khalls universal auxiliary tables for timekeeping by the sun. The first (a) shows a sub-table for the (half)
al function for latitudes 31, 32, 33, 33;30 (Damascus), and 34, with the solar longitude entered vertically. The second (b) shows a sub-
table for the hour-angle as a function of the same al function entered horizontally, here with arguments 29 and 30, and the difference of the
Sines of the meridian and instantaneous altitudes entered vertically. [From MS Dublin CB 4091, fols. 163r and 166r, courtesy of the Chester
Beatty Library.]
tables of universal auxiliary functions 169
instructions as in his tirelessness at compiling his enormous table. However, all this seems to
have taken its toll on his mind.
In the unique source MS Dublin CB 4091, copied 833 H [= 1429], there is a set of auxiliary
tables by the Damascus astronomer al-Khall (2.1.4) for finding t(h,,), which is quite
different from his Universal Table (9.5). The principal functions tabulated are the following:
B(,) = Cos () Cos / 2R = 1/2 B(,) ( = 23;31)
for the domains:
= 1, 2, ... , 49 as well as 33;30 (Damascus) and = 1, 2, ... , 90 ,
and:
V(x,y) = arc Vers { R y / 2x }
for the domains:
x = 30, 29, ... , 19 and y = 0;10, 0;20, ... , 60;0 .
These two sets of tables see Figs. 9.4a-b contain respectively 4,500 and 4,680 entries, rather
accurately computed to two digits. Horizontal differences are shown in the second set. The
first function is called al-al, the base (cf. 6.0) but a marginal note at the beginning of the
table points out that it is in fact nif al-al, half the base. The second function is called fal
al-d}ir fq min al-al wa-min fal jayb al-ghya {an jayb al-irtif{, the hour-angle for all
latitudes as a function of the base and the excess of the Sine of the meridian altitude over the
Sine of the (instantaneous) altitude.
al-Khalls tables are complete with instructions and are preceded by two small tables of
the function:
() = 90 + () ( = 23;31)
and the Sine function, with entries to two digits for each degree of argument. To find t(h,,)
al-Khall proposes first using the table of () to find H(,), thus:
H(,) = () - ,
and then using the Sine table to find Sin H and Sin h in order to establish their difference:
H(h,,) = Sin H(,) - Sin h (> 0) .
Then, using the two main sets of auxiliary tables one should find B(,), with which:
t(h,,) = V { B(,) , H(h,,) } .
al-Khalls tables are clearly an extension of those of al-Kha} (6.15.1), which were devised
for a specific latitude, but the two sets may have been compiled independently. Although only
one copy of these original auxiliary tables of al-Khall has come to light, as opposed to the
several extant copies of his Universal Table, it appears that they became known in both Tunis
and Cairo: see further 9.7 and 9.11 and also II-10.11.
9.4* al-Khalls universal auxiliary tables for finding the solar azimuth
The following tables are mentioned by al-Khall is the introduction to his main universal
auxiliary tables (9.5) but the unique surviving copy came to my attention only in January, 2001.
170
part i, chapter nine
Fig. 9.4*a-b: Extracts from al-Khalls tables of f' and g for latitudes 33 and 33;30 (a) and his table of K for arguments 16, 17, ... , 20 (b).
See also Fig. II-10.3*a-b. [From MS Bursa Haraiolu 1177,4, fols. 80v-81r and 86v-87r, courtesy of the Genel Ktphanesi.]
tables of universal auxiliary functions 171
I have mentioned them briefly in 8.5.1*, and they will be treated again in the context of al-
Khalls other tables in II-10.3*.
MS Bursa Haraiolu 1177,4 (fols. 72r-90r), copied ca. 1450, is a unique copy of a set
of tables entitled Jadwal al-samt li-kulli rtif{ f {ur al-aqlm al-sab{a, table of the azimuth
for all altitudes in the latitudes of the seven climates (i.e., for all latitudes).20 No compiler is
mentioned, but it can only be Shams al-Dn al-Khall (9.4 and 9.5). The set contain sub-tables
for the following functions:
f'() = R Sin () / Cos and g(h) = Sin h Tan / R
for the domains:
and h = 1, 2, ... , 90 ,
= 1, 2, ... , 48, as well as 33;30 (Damascus) ,
and:
K(x,h) = arc Sin { R x / Cos h }
for the domains:
x = 1, 2, ... , 59 and h = 0, 1, ... , n(x) ,
where n(x) is the largest integer such that x Cos n(x). The functions f' and g are labelled
jayb sa{at al-mashriq and iat al-samt by al-Khall, and it is clear that they represent Sin
() and k(h) see 5.0 and 8.5. The value of underlying the first table is about 23;30, possibly
23;31.
The instructions describe how to determine the ta{dl al-samt using the first two tables, and
then how to find the azimuth from the third table. These universal auxiliary tables represent
the ultimate solution to the problem of determining the solar azimuth a(h,,). It is clear that
all al-Khall had to do to compile his splendid tables described in the next section was to replace
the table of f'() with a new one of f() and rewrite the instructions!
After compiling his two sets of auxiliary tables for timekeeping by the sun (9.4) and for finding
the solar azimuth (9.4*), al-Khall prepared another set which can be used for solving all of
the standard problems of spherical astronomy for any latitude.21 These tables are appropriately
called al-Jadwal al-fq, The Universal Table. They are preserved in several sources,
including MSS Paris BNF ar. 2558, Berlin Ahlwardt 5754/5/6 (Wetzstein 1138) and 5739
(Wetzstein 1144), Escorial r. 931,8, fols. 171r-211v, Istanbul Ayasofya 2590, Istanbul Hami-
diye 1453 (fols. 232v-266v), Istanbul Serez 1914, Oxford Marsh 95 (Uri 961), Cairo MM 43,
Cairo MM 98, Cairo DM 758, and Princeton Yahuda 861,2. See also II-10.7.
The functions tabulated by al-Khall are:
f() = R Sin / Cos and g() = Sin Tan / R
for the domains:
= 1, 2, ... , 90 ,
= 1, 2, ... , 55 as well as 21;30 (Mecca) and 33;30 (Damascus) ,
20 I owe the reference to this manuscript to ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 805.
My thanks are due to Dr. Sonja Brentjes of Berlin for showing me a microfilm thereof.
21 See King, al-Khalls Auxiliary Tables, for the first analysis.
172 part i, chapter nine
Fig. 9.5a-b: The sub-tables of al-Khalls first and second functions for latitude 20 (a) and his third function
for arguments 3-4 (b). See also Fig. I-6.14.3. [From MS Paris BNF ar. 2558, courtesy of the Bibliothque
Nationale de France.]
and:
K(x,y) = arc Cos { R x / Cos y }
for the domains:
x = 1, 2, ... , 59 and y = 0, 1, ... , n(x) ,
where n(x) is the largest integer such that x Cos n(x). al-Khall calls f and g al-mafz
al-awwal and al-thn, the first and second functions, and he refers to K by the same
expression used for the argument x, namely jayb al-tartb, the auxiliary Sine. His tables
contain over 13,000 entries, rather accurately computed see Fig. 9.5a-b. The values for the
second and third function were, of course, simply taken over from his universal auxiliary tables
for finding the solar azimuth (9.4*).
To find D(,), t(h,,) and a(h,,) using his tables al-Khall outlines the following rules:
D(,) = 180 - K { g() , } (>0) or K { g() , } (<0)
tables of universal auxiliary functions 173
The late-14th-century astronomer Jaml al-Dn al-Mridn, who appears to have worked in both
Damascus and Cairo,23 compiled an auxiliary table which is extant in MS Paris BNF ar. 2525,1
(fols. 1v-16v), copied ca. 1450, and also in the much later copy MS Cairo K 4026. al-Mridn
called his table al-shabaka, which means net or grid and I have analyzed it in detail in
a previous publication.24 The shabaka displays values of three trivial functions, arranged in
a single 90 90 table containing 8,100 entries. No indication is given in the instructions that
more than one function is tabulated and in the sequel I use M(x,y) to denote all three functions,
which are in fact:
M1(x,y) = R y / x for 0 x 59, y < x
M2(x,y) = Sin x Sin y / R for 0 x 90, x < y 90
M3(x,y) = 24 y / x for 60 x 90, y < x
The function M3 is intended to be used for calculating conjunctions and does not concern the
present study.
To derive the hour-angle using his table al-Mridn suggests finding the three quantities:
B(,) = M(,) ( > ) or M(,) ( < )
C(,) = M(,) ( < ) or M(,) ( > )
b(h,,) = Sin h m C(,) (as \ 0) ,
22 I am indebted to the late Prof. M. Buhairi of the American University of Beirut for showing me this
manuscript from his personal collection.
23 On al-Mridn (Suter, MAA, no. 421) see Cairo ENL Survey, no. C47; and King, al-Mridns Universal
Quadrant, p. 219, n. 2.
24 Ibid., pp. 231-240. The term shabaka is not to be confused with the same term used for a table displaying
the solar longitude over a four-year period: see n. 9:26.
174 part i, chapter nine
which he calls respectively al-al al-mulaq, jayb irtif{ qur al-madr and al-al al-mu{addal
(see further 6.0). With these t is defined by:
Cos t(h,,) = M{ B(,) , b(h,,) } .
To find the azimuth a(h,,) one of al-Mridns methods involves first finding the arc:
= arc Sin (Tan )
and then the three quantities:
k(h,) = M(,h) ( < h) or M(h,) ( > h)
Sin (,) = M(Cos ,Sin )
m(h,,) = k(h,) m Sin (,) (as \ 0)
which he calls respectively iat al-samt; jayb al-sa{a and ta{dl al-samt. al-Mridn also
suggests the following alternative method. First form the three quantities:
p(h,) = M(h,) ( > h) or M(,h) ( < h)
q(h,,) = p(h,) - Sin
r(h,) = M(,) ( > h) or M(,) ( < h)
and then:
Sin a(h,,) = M{ r(h,), q(h,,) } .
MS Cairo DM 689, copied ca. 1600, contains an extensive set of auxiliary tables copied in
an elegant Maghribi hand.25 The tables conclude with a star catalogue dated 801 H [= 1398]
and they appear to have been compiled in Tunis. The title folio, instructions and first few tables
are missing from the manuscript, which begins with the last page of a set of tables of the
shabaka variety26 displaying the solar longitude for each day of a period of four Syrian years
(see Fig. 1.2d). The tables for timekeeping follow this see the extracts in Figs. 9.7a-b.
The main functions tabulated are the same as those in al-Khalls auxiliary tables for
timekeeping by the sun (9.4), namely B(,) and V(x,y). In fact these Tunisian tables are
merely an extension of al-Khalls tables and the corresponding entries in both sets are the
same. The Tunisian tables of B are computed for each degree of from 1 to 48 and also
21;40 (Mecca) and various non-integral latitudes between 30 and 40 some of which were
intended to serve specific localities in Ifriqiyya and the Maghrib and perhaps also Sicily. al-
Khall had a separate table for latitude 33;30 (Damascus) and the Tunisian set has separate
tables for latitudes:
30;30, 31;30, 32;30, 33;30, 33;40, 34;30, 35;30,
36;30, 36;40, 37;10, 37;30, 38;30, 39;30 .
The Tunisian tables of V(x,y) are simply those of al-Khall rearranged so that the horizontal
argument is increasing.
In MS Cairo DM 689 the tables of B are preceded by a set of tables of the function Sin
H(), called jayb al-ghya, computed to two digits for each degree of and the latitudes of
Fig. 9.7a-b: (a) The sub-tables for the al B() in the Tunisian corpus for latitudes 36;30, 36;40, 37 and 37;10.
(b) This extract shows the table of the hour-angle V'(x,y) for arguments x = B from 19 to 26, and y = H' =
Sin H - Sin h from 55;10 to 60. [From MS Cairo DM 689, fols. 26v-27r and 47v-48r, courtesy of the Egyptian
National Library.]
Mecca (21;40) and Medina (25), as well as each of the latitudes between 30 and 38 for
which B is tabulated. These tables were probably used by the anonymous Tunisian astronomer
to compile his tables of B, since:
B(,) = 1/2 { Sin H(,) + Sin H(,*) } .
The underlying value of is 23;35, whereas al-Khall used 23;31. However, the change in
hardly affects the values of B() given to two digits. The Tunisian tables also contain a table
of Sin H() for = 0, which is simply Cos (). See further 6.2.5 and 6.3.3, and II-13.4.
MS Cairo MM 241 is an apparently unique copy from ca. 1450 of a set of auxiliary tables
by Ab Bakr ibn Ism{l, known as Ibn al-Mushrif (see already 6.13.1, 7.3.1 and 8.2.1).27 The
work is entitled Nr al-adhq li-ma{rifat a{ml al-falak f s}ir al-fq, The Light of the
27 On Ibn al-Mushrif see Cairo ENL Survey, no. C43. He was still active in 848 H [= 1444/45]: see II-5.7,
so he cannot be identical with the individual with the same name mentioned in King, Astronomy in Yemen, no.
12, and 6.16.
176 part i, chapter nine
tables of universal auxiliary functions 177
Intelligent for Performing Astronomical Calculations for all Latitudes, and appears to have
been compiled in Cairo in 1435. The title-folio of this manuscript is shown in Fig. 9.8a. The
three main functions tabulated are:
B(,) = Cos Cos / R
for the domains:
= 0, 1, ... , 89 and = 0, 1, ... , 90 ;
H(H,h) = Sin H - Sin h
for the domains:
H = 1, 2, ... , 90 and h = 1, 2, ... , H - 1 ;
and:
V(x,y) = arc Vers { R y / x }
for the domains:
x = 1, 2, ... , 60 and y = 0;15, 0;30, 0;45, 1, 2, ... , 60 .
These three functions are called al-al, the base, fal m bayn al-jaybayn, the difference
between the two Sines, and fal al-d}ir idh kn al-al kadh daraja wa-ikhrj al-samt, which
means hour-angle or azimuth for base so-and-so. An extract is shown in Fig. 9.8b. The tables
are complete with instructions and to find the hour-angle t(h,,) one simply finds B(,) and
H(H,h) and uses these as arguments in the third table, thus:
t(h,,) = V { B(,) , H(H,h) } .
The method for finding the azimuth a(h,,) is to find B(,h) and:
H { ( + h ), } = Sin ( + h ) - Sin
using the first two tables, and then the azimuth (measured from the north point) is given by:
a(h,,) = V{ B(,h), H{ ( + h ), } } .
Various other minor functions are tabulated by Ibn al-Mushrif including () for = 23;35
and B() for = 30 (6.4.3) and the inverse Versed Sine function (6.14), as well as a table
for finding the equation of half daylight or the auxiliary azimuth function (7.3.1 and 8.2.1).
One small table, which is not referred to in the instructions and whose purpose eludes me, is
entitled jadwal tawl juyb al-tartb, Table for converting the auxiliary Sines. The argument
is al-{ar, latitude, and the values of the function f() are labelled al-ta{dl, equation or
interpolation factor. The tabulated function displays discontinuities between = 30 and 31
and between = 31 and 32. Sample entries are:
Figs. 9.8a-b: (a) The title-folio of the unique copy of Ibn al-Mushrifs tables. The owner was the copyist Abu
l-Yumn Muammad ibn Muammad ibn Muammad ibn {Arab (Cairo ENL Catalogue, I, pp. 714 and 761).
There is a note in a different hand listing the planetary apogees for epoch 740 H [= 1339/40] from the Muala
Zj, the most popular zj in medieval Egypt. A further note in the lower left corner is in the hand of Ibn Abi
l-Fat al-f, the leading astronomer in Egypt ca. 1500: it deals with what one might call advanced
timekeeping. (b) The last part of the table of the second function and the beginning of the table of the third
function in Ibn al-Mushrifs set. By the nature of the functions both tables are of the aylasn variety. [From
MS Cairo MM 241, fols. 1r and 11v-12r, courtesy of the Egyptian National Library.]
178 part i, chapter nine
f() . 30 ; 0,35,24 .
1 ; 0, 0, 0 . 31 1; 0,37,11 .
2 ; 0, 1, 27 20 ; 0,20,36 32 1; 1,16,48 40 1; 7,49,50
3 ; 0, 2, 21 . 33 1; 1,24,24 .
. . . .
. 29 ; 0,33,37 35 1; 3,25,19 45 1;13,29, 4
10 ; 0, 9, 16
The tabulated function is certainly neither Sec nor arc Vers (x) / x nor Vers .
MS Cairo DM 512, copied ca. 1080 H [ 1670], contains an extract from Ibn al-Mushrifs
tables by Sdn al-Bashtak.28 The author presents only the tables of V(x,y), with his own
introduction. MS Cairo MM 209,1 (fols. 1r-37r) contains tables of normed right ascension to
three digits for each minute of ecliptic longitude attributed to Ab Bakr ibn al-Mushrif,
copied in 873 H [= 1468/69] from a copy by the author dated 848 H [= 1444/45].
al-Waf} was an Egyptian astronomer who died ca. 1470.29 MSS Vatican Borg. ar. 217,1 (fols.
1v-5v), copied ca. 1500, and Istanbul Nuruosmaniye 2921,2 (fols. 22r-26v) contain a small
set of auxiliary tables attributed to him entitled Kifyat al-waqt li-ma{rifat al-d}ir wa-falihi
wa-l-samt, which might be rendered All you need to find the time, hour-angle and azimuth.
The first function tabulated is:
G(,) = R3 / [ Cos Cos ]
with values to three digits for horizontal arguments (latitudes):
= 12 (Hadramawt, Aden), 15 (Zabid, Sanaa), 18 (Ethiopia), 21 (Aswan, Jidda),
24 (Qift, Yathrib (= Medina)), 27 (Akhmim, Yenbo), 30 (Fayyum, Aqaba),
32 (Jerusalem, Hebron), 33;30 (Damascus, Tripoli), 36 (Mardin, Tunis),
39 (Malatya), 42 (Edirne) and 45 (Constantinople)
and vertical argument:
= 0, 1, ... , 89 ,
and the second function is simply:
V(x) = arc Vers(x) ,
with values to degrees and minutes for:
x = 0;5, 0;10, ... , 0;30, 0;40, ... , 2;0, 2;15, ... , 3;0, 3;20, ... , 4;0, 4;30, ... ,
16, 17, 18, ... , 120 .
The two tables of these functions are labelled jadwal al-ia and jadwal fal al-d}ir, that
is argument table and hour-angle table. These titles generally refer to tables giving the
times of prayer and tables in which the hour-angle is given as a function of solar altitude,
respectively. An extract is shown in Fig. 9.9a-b.
Fig. 9.9a-b: Extracts from the auxiliary tables of al-Waf}. They are followed by part of the incomplete auxiliary
tables of al-Kha} (6.15.1). [From MS Vatican Borg. ar. 217,1, courtesy of the Biblioteca Apostolica Vaticana.]
al-Waf}s tables, which contain 1,320 entries, are rather accurately computed and are
complete with instructions. To find t from h, and one first computes:
H(h,,) = Sin H(,) - Sin h = Sin ( + ) - Sin h
and then uses:
t(h,,) = V { H(h,,) G(,) } .
Similarly, to find a(h,,) one computes:
H(,h,) = Sin ( + h ) - Sin
and then uses:
a(h,,) = V { H(,h,) G(,h) } ,
where a is measured from the north point. al-Waf}s use of the functions G and H to determine
t or a with equal facility is ingenious and it fully exploits the equivalence of the underlying
mathematical problems. He also demonstrates how to compute the duration of twilight and the
direction of Mecca with his tables.
180 part i, chapter nine
Ibn Abi l-Fat al-f was an Egyptian astronomer who died ca. 1495.30 He prepared a
recension of the early-15th-century Zj of Ulugh Beg of Samarqand, adapting the planetary
tables for the longitudes of Cairo and he was also the author of several short treatises on
quadrants and sundials. His auxiliary tables are contained in the apparently unique source MS
Oxford Seld. Supp. 101 (Uri 1040), copied in 918 H [= 1513], and they are entitled Kitb al-
Jawhir f ma{rifat al-samt wa-fal al-d}ir, Book of Jewels for Finding the Azimuth and the
Hour-angle.
The first and second functions tabulated by al-f are the following:
C(,) = Sin Sin / R and B(,) = Cos Cos / R ,
computed to two digits for the domains:
= 0, 1, ... , 89, = 0, 1, ... , 90 and also = 33;30 and 23;35 (= ) .
The third is:
S(x,y) = arc Cos { R x / y } ,
computed likewise to two digits for the domains:
x = 0;30, 1;0, ... , 59;30 and y = [x] + 1, [x] + 2, ... , 60 .
The first two functions are called ufq and al respectively, that is, horizon and base (see
further 6.0). The third function is referred to by the term applied to the argument x, namely,
al-muqanara, almucantar, for what reason is not clear.
al-fs tables, which contain about 15,000 entries, are more sophisticated than those of
al-Mridn (9.6) but are hardly an improvement over the Universal Table of al-Khall (9.5).
In fact, it seems that they owe their inspiration to both al-Mridns set and al-Khalls first
set of auxiliary tables (9.4): compare al-fs function C with al-Mridns M1 and his
functions B and S with al-Khalls B and V. The link with al-Khall is also apparent from
the fact that al-f, working in Cairo, tabulates B and C for = 33;30, which serves Damascus.
Likewise it may be that al-f noticed that al-Mridns function M1 would be more useful
if one could feed in the non-integral value of as one argument in order to facilitate the
computation of (): his tabulating C for = 23;35 satisfies this need. al-f clearly had
not learned much from the investigations of the Damascus and Samarqand astronomers on the
secular change of : he preferred to use the standard value, 23;35. He begins his work with
a table for finding the solar longitude for the period 877-1004 H [= 1472-1596], based on Ibn
Ynus solar parameters derived five centuries previously. al-fs auxiliary tables can be used
for computations involving the stars as well as the sun, whereas al-Khalls first set of auxiliary
tables are for timekeeping by the sun only. al-f notes that his two functions B and C are
related by:
B(,) = C(,)
but states that he preferred to tabulate both functions rather than only one. To find the hour-
angle t(h,,) his instructions prescribe the simple rule:
30 On Ibn Abi l-Fat al-f see Suter, MAA, no. 447; Kennedy, Zj Survey, no. 37; Cairo ENL Survey,
no. C98; and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 116-126, no. 58 (where this work is not
mentioned).
tables of universal auxiliary functions 181
9.11 Two sets of anonymous Egyptian auxiliary tables for timekeeping by the sun
MS Istanbul S. Esad Efendi Medresesi 119,2 (fols. 82v-97v), copied ca. 1700, contains a set
of auxiliary tables bound at the end of a set of prayer-tables for Cairo by Riwn Efend (6.1.3).
The same auxiliary tables occur in MS Cairo DM 644,1 from ca. 1700, where they are followed
by a star catalogue dated 1061 H [= 1651]. In both sources the work is anonymous. The title
is Fat al-Karm al-Bq f ma{rifat al-d}ir wa-falihi fq, The Victory of God, the Noble
and Eternal, for Finding the Time since Sunrise and the Hour-angle for all Latitudes.31
The main functions tabulated are:
B(,) = 1/2 B(,) = 1/2 Cos () Cos / R
for the domains:
= 1, 2, ... , 90 and = 1, 2, ... , 27, 28, 30, 32, 33;30, 36, 38, ... , 48 ,
and:
V(x,y) = arc Vers { R y / 2x }
for the domains:
x = 19, 20, ... , 30 and y = 0;5, 0;10, ... , 60;0 .
The first auxiliary function is labelled al-al, (half) the base, and the arguments in the table
of the second function are called al-al al-mu{addal, the modified base, and fal al-jaybayn,
the difference between the two Sines. The entries in the tables, which are given to two digits,
were lifted from al-Khalls auxiliary tables for timekeeping by the sun (9.4). Note that in al-
Khalls set B is tabulated for some 50 values of and the increment in the argument y for
the tables of V is 0;10 rather than 0;5. Also, in the later Egyptian set no horizontal differences
are given in the tables of V but the simplified format makes them slightly easier to use.
MS Cairo MM 203,3 (fols. 153r-160v), copied ca. 1700, contains an anonymous set of tables,
probably of Egyptian provenance, entitled Daq}iq al-raq}iq f ma{rifat fal al-d}ir li-s}ir
al-fq, which means something like Nice Numbers for Finding the Hour-angle for any
Latitude. The set consists of simple tables of the inverse Versed Sine (6.14.2) and Sine
functions and solar declination (based on = 23;35), followed by a table of a function:
G3(,)
tabulated to one significant digit for the domains:
= 1, 2, ... , 60 and = 1, 2, ... , 24 .
31 A third copy in Diyarbakr, also anonymous, is mentioned in ~hsanolu et al., Ottoman Astronomical
Literature, II, p. 743.
182 part i, chapter nine
Fig. 9.11: An extract from some anonymous Egyptian auxiliary tables, serving latitudes 31-50. [From MS Cairo
MM 203, fols. 159v-160r, courtesy of the Egyptian National Library.]
See the extract in Fig. 9.11. The methods outlined in the instructions for finding D and t by
means of this function are equivalent to the formulae:
Vers D(H,) = Sin H + Sin H G3(,) / R ,
Vers t(h,H,) = [ Sin H - Sin h ] + [ Sin H - Sin h ] G3(,) .
It is clear that for these formulae to be valid we must have:
G3 = R3 / [ Cos Cos ] - R ,
which is confirmed by inspection of the table. The inspiration for tabulating this function was
probably the little table of G3 for the latitude of Cairo due to al-Marrkush (6.7.2).
It is not inappropriate that I should conclude this study with a brief description of the late
Ottoman set of auxiliary tables for timekeeping which shows the influence of Western
mathematics on traditional Islamic astronomy. MS Istanbul Kandilli 226 is a copy of the
auxiliary tables of Sa{d (or Sayyid ?) Beg Zde for computing the prayer-times in Turkish
time for any latitude.32 The main tables are based on the principle that to find the hour-angle
in equinoctial hours and minutes corresponding to a particular prayer time ti one uses three
functions:
Hi(H), G(,) and V(x) ,
thus:
ti = V{ Hi(H) + G(,) }.
Likewise, to find the hour-angle corresponding to a particular solar altitude one uses other tables
to find a function H(H,h), and then uses:
ti = V{ Hi(H,h) + G(,) } .
These functions are the following (compare the auxiliary tables of al-Waf} discussed in 9.9
above !):
H(H,h) = log [Sin H - Sin h] = log H(H,h) ,
G(,) = log { R2 / [ Cos () Cos ] } = log G(,),
V(x) = 1/15 arc Vers (alog x).
Both the trigonometric functions and the logarithms are to base 60 and are written
sexagesimally! Further details are given in II-14.16.
MS Istanbul UL T4203 contains another set of auxiliary tables for computing the hour-angle,
also using logarithms but not the sexagesimal system. The tables were computed by an
individual identified simply as Jaml,33 and I have not investigated their underlying structure.
33 Ditto.
184 part i, chapter ten
CHAPTER 10
ber das Wie und Warum der Benutzung dieser Tabellen knnen wir nur spekulieren.
D. A. King, in Schweinfurt 1993 Exhibition Catalogue, p. 352.
I have not looked systematically for tables for timekeeping in medieval European manuscripts
or early printed works; however, I am confident that these are far more widespread than has
hitherto been thought. I mention here just a few examples from the published literature and
from European manuscripts which have come to my attention by chance. One feature of these
tables is clear: medieval European astronomers liked to tabulate the solar altitude as a function
of time. The reason is not completely clear, although it had to do with the regulation of clocks.1
Muslim astronomers, on the other hand, preferred tabulating something that one wants to know,
namely, the time, as a function of something that one can measure, namely, solar or stellar
altitude.
We begin with the Toledan Tables, a motley corpus of tables derived mainly from the Zjes
of al-Khwrizm and al-Battn, which has been surveyed by Gerald Toomer, with a detailed
new study by Fritz Pedersen.2 In this corpus, as well as in a 15th-century English manuscript,
there is a table of the function e() = 1/12 150 tan () (7.1.1), labelled tabula differencie
ascensionum universe terre.3 A table of h(T,H) for all latitudes with entries tabulated to two
digits for each 0;15 of T (up to 6sdh) and each 0;30 of h from 10 to 90, with a total of
over 3,500 entries (4.3), in a 14th-century English astronomical manuscript of both the Toledan
and Alphonsine Tables, has also been noted by Toomer.4 Universal tables of T(H,h) and h(H,T)
based on the standard approximate formula are also found in two Byzantine manuscripts.5
The Alphonsine Tables generally replaced the Toledan Tables and were used in different
recensions all over Europe until the 16th century. An isolated table displaying the function
H(T,) for a particular latitude (3.0) is found in an early-15th-century manuscript of the
Alphonsine Tables written in Germany: it is entitled Tabel der ofsteygenden Czeychn and
the entries are given in zodiacal signs and degrees for each hour and each degree of solar
longitude.6
1 Geoffrey Chaucer referred to them as tables ... for the governaunce of a clokke: see North, Chaucers
Universe, p. 87, and also Eisner, ed., Kalendarium of Nicholas of Lynn, pp. 29ff.
2 See Toomer, Toledan Tables, and now Pedersen, The Toledan Tables.
3 Ibid., p. 33, and Neugebauer & Schmidt, Hindu Astronomy at Newminster, p. 226. See also North, Richard
of Wallingford, II, pp. 12-14, and idem, Horoscopes and History, p. 14. See now Pedersen, Toledan Tables,
III, pp. 986-991.
4 Toomer, Toledan Tables, p. 155. See now Pedersen, op. cit., III, pp. 1134-1138.
5 Jones, Byzantine Astronomical Manual, p. 174.
6 I have examined MS Gotha Forschungsbibliothek Membr. I. 109, apparently dating from 1428, in which
the tables occur on fols. 22r-24v.
european tables for timekeeping 185
Fig. 10.1a: Extracts from the table of solar altitude at the hours in the treatise on astrology attributed to Enrique
de Villena. The format of the table is dictated by the changes in meridian altitude of 1. [From Ctedra & Sams,
Astrologa de Enrique de Villena, pp. 171 and 173.]
A table of hi(H) for the equatorial hours and each degree of H was computed for the latitude
of Baeza (38) by pseudo-Enrique de Villena ca. 1430: see Fig. 10.1a.7 The entries display
error patterns which could be taken as implying that they were based on the approximate
formula; however, the table may have been compiled using an instrument such as an astrolabe,
for even the values at the equinoxes contain errors of as much as 2. A less extensive table
of this kind, with entries for each 5 of h, is found in a medieval Hebrew manuscript.8
Tables of hi() ( = 30) and/or the associated vertical or horiontal shadows abound in
the medieval European sources, but few have been published, and even fewer investigated.9
There is, for example, a table of solar altitudes hi for latitude 48 in the treatise on the astrolabe
attributed to Hermannus Contractus (1013-1054).10 The author states that he derived the entries,
which are given to degrees, using an astrolabe and they do indeed correspond quite well to
computation with the exact formula. The available version of the table of vertical shadows zi()
7 See Ctedra & Sams, Astrologa de Enrique de Villena, pp. 171-176, and the commentary on pp. 56-57.
8 MS Munich heb. 343, which was brought to my attention by Professor Bernard Goldstein in the 1970s.
9 Zinner, Astronomische Instrumente, pp. 159 and 50-51, also lists tables for Oxford, London, Rome, Venice,
Nuremberg, Augsburg and Regensburg, which I have not investigated. See North, Richard of Wallingford, I,
p. 18, on a table for Oxford by John Maudith (14th century), and idem, Chaucer, p. 114, on various other tables
of this kind.
10 On Hermannus see the article by Claudia Kren in DSB. For the table see Gunther, Early Science in Oxford,
II, p. 419 (but it is not contained in the treatise translated in Joseph Drecker, Hermannus Contractus ber das
Astrolab, Isis 16 (1931), pp. 200-219).
186 part i, chapter ten
= Tan12 hi() ( = 30) for latitude 47;46 (Vienna) compiled by John of Gmunden (fl. ca.
1430) for marking the hour-curves on a cylindrical sundial, is full of errors.11
In the Kalendarium of Nicholas of Lynn (1386) there are tables of the solar altitude hi and
shadows z (base 6!) for the seasonal hours on each day of the year for the latitude of Oxford,
taken as 51;50.12 Petrus de Dacia, canon in Roskilde, Denmark, ca. 1300, presented tables
of meridian altitudes H and lengths of day and night 2D and 2N for each day of the year,
computed for latitude 48;50, serving Paris, and similar to others by William of St Cloud in
his Kalendarium edited in 1296.13 In one manuscript of the treatise on astrolabe construction
by Jean Fusoris (Paris ca. 1400) there is a table, also for Paris, of the solar altitude hi() at
each equinoctial hour for each degree of and a separate table for each 15 days of the year
specifically for constructing a cylindrical sundial.14
This activity was clearly continued by Renaissance astronomers, but there is precious little
documentation as yet. A manuscript in the Stadtsarchiv in Schweinfurt contains a mixture of
tables in the hands of two 16th-century German astronomers, Johannes Hommel and Johannes
Praetorius.15 The former contributed two tables for a latitude of 51 (location unspecified),
displaying of h(T,) and a(T,) for each equinoctial hour and each 1 of . The latter contributed
tables of h(T,) for latitudes 50 (locality unspecified) and 49;26 (Nuremberg is specifically
mentioned), for = 3 and 5, respectively, as well as a set of tables for coordinate conversion
for latitude 50 (copied in Cracow) after the model of the tables appended to Ptolemys
Analemma.16 Also in the hand of Hommel we find a table of h(,) for each 1 of from 39
to 63 and each 1 of . No other such universal table is known from the medieval and
Renaissance European sources.
In Europe, between the late 18th and the early 20th century, there appeared a series of
extensive tables for timekeeping for purposes of navigation, of which I present scant details.17
Already in 1770 the French astronomer Csar-Franois Cassini (= Cassini III) prepared in his
Almanach perpetuel pour trouver lheure par tous les degrs de hauteur du soleil a skeleton
set of tables for latitudes between 34 and 51. Joseph-Jrme Lalande indicated the usefulness
of such tables in the second edition of his Trait de lastronomie (1771), and referred to the
tables computed by Cassini. We can take a closer look at the tables published in Leipzig in
1791 by one Friedrich Christoph Mller. These display the two times of day in the morning
and afternoon for each day of the year when the sun has a given altitude. Entries are given
11 On John of Gmunden see the article by Kurt Vogel in DSB. The table is published in Claudia Kren, The
Travellers Dial, p. 431.
12 Eisner, Kalendarium of Nicholas of Lynn, pp. 68-69 (Jan.), 74-75 (Feb.), etc., and the commentary on
p. 26. On the latitude see ibid., p. 17. See also North, Chaucer, pp. 115-116, and on the problems of the edition
see ibid., p. 115, n. 54. These tables include values for twilight: see also Pedersen, Toledan Tables, III, pp. 1132-
1133.
13 Pedersen, F. S., ed., Astronomical Works of Petrus de Dacia and Petrus de S. Audomoro, I, pp. 39 and
320-322.
14 Poulle, Fusoris, pp. 184-185 and 186. On this manuscript see also King, Ciphers of the Monks, p. 155,
and for more on Fusoris see ibid., pp. 397-398.
15 On the manuscript (AvS Ha 21) see Schweinfurt 1993 Exhibition Catalogue, pp. 351-353, no. 169.
16 See n. 1:40.
17 The information in this paragraph is taken from Cotter, Nautical Tables. Alas the details of the individual
tables are not standardized. The subject could use a new study.
european tables for timekeeping 187
Figs. 10.1b: The title-page and an extract from the tables of F. C. Mller, 1791, showing time as a function
of solar altitude throughout the year for a range of latitudes in Germany. The corpus bears the title tables of
solar altitudes, which is mathematically inexact, and curiously corresponds precisely to the Turkish expression
irtif{ jadwali (irtif cedveli) for tables of the same kind. Mllers introduction makes no mention of con-
temporaneous Turkish tables, and the West was to remain innocent of any knowledge of them for almost another
two centuries. [From Mller, Tables des hauteurs du soleil.]
to hours and minutes for each degree of altitude, and there are tables for each degree of latitude
from 47 to 54, intended to serve all of Germany and the countries surrounding it. See Fig.
10.1b for an extract from the French edition, apparently published simultaneously with the
original German version.18 In his preface the author complained about the difficulties of
18 It is a pleasure to thank Dr. iva Vesel of Paris for sending me a copy of the French edition of these
tables, listed in the bibliography as Mller, Tables des hauteurs du soleil.
188 part i, chapter ten
compiling the tables and finding a publisher, and mentioned that he was considering compiling
similar tables for the solar azimuth.
In his Abrg de navigation (1793) Lalande presented tables of t(h,,) for each value of
h between 2 and 48, each degree of between 0 and 24, and each degree of from 0
to 60. The tables displayed the first differences and an auxiliary table was provided to facilitate
interpolation. The first English tables of this kind were published in 1827 by Thomas Lynn:
they were not widely used except by officers of the Honourable East India Company, to whom
Lynn was engaged as an examiner in navigation. The Time-Azimuth-Tables of the naval officer
John Burdwood published from 1852 onwards became most popular on British vessels. Percy
Davis of the Royal Naval Office published in 1902 a set of tables of t(h,,) with values for
each degree of each argument. In 1913 a set of tables of stellar altitudes and azimuths, for
declinations between -30 and +30, and specifically for latitude 55, was published in Berlin
for the German Navy; within the next few years companion volumes were published for
latitudes 50 and 70. (It is not clear what the advantage of such tables would have had beyond
enabling a vessel to be steered along one of three parallels of latitude.) A set of tables published
by the United States military authorities in 1919 displayed altitude and azimuth in parallel
columns (no further information available). In fact, various astronomers from the late 19th
century to the mid 20th century compiled tables of the hour-angle and the solar azimuth, both
as functions of solar altitude and declination, or solar altitude as a function of hour-angle and
declination, all for extensive ranges of latitudes.
Although previous studies of Islamic auxiliary tables have failed to draw attention to European
tables of the same kind, the tradition of compiling auxiliary tables for solving problems of
spherical trigonometry inevitably relating mainly to spherical astronomy was continued
in Europe.19
Regiomontanus table for solving spherical triangles in his Tabula primi mobilis prepared
in 1463 displays values of:
x(a,b) = arc sin { sin a sin b }
for each degree of both arguments. Values are given in degrees, minutes and seconds, and are
generally accurate in the last digit. First horizontal and vertical differences are also tabulated.
As Anton von Braunmhl has shown, linear interpolation applied horizontally and vertically
yields results in which the seconds are not at all to be trusted (his example shows an error
of about 30). This table is useful for solving right-angled spherical triangles with any of the
rules involving only sines and cosines.
19 The information in this section is summarized from Cantor, Geschichte der Mathematik, II, pp. 273-275;
von Braunmhl, Geschichte der Trigonometrie, I, pp. 122-124 and 231-236, II, p. 231; Wolf, Handbuch der
Astronomie, I, p. 438; and Fletcher et al., Index of Mathematical Tables, esp. I, pp. 224-227 (altitude tables,
azimuth tables and hour-angle tables). The tables of F. C. Mller are not mentioned in any of these works. As
far as I know, the Renaissance tables have never been studied in any depth. Glowatzki & Gttsche, Die Tafeln
des Regiomontanus, deals only with standard trigonometric tables, but not only those of Regiomontanus.
european tables for timekeeping 189
It is not necessary to assume a knowledge of the spherical cosine formula when an equivalent
solution occurs in an ancient or medieval text. Thus, for example, it is often asserted that al-
Battn (ca. 910) knew the formula and used it to derive the hour-angle from an instantaneous
celestial altitude.20 This is simply not the case. al-Battn merely used an Indian procedure
for deriving the hour-angle by projection methods. His formula can, of course, be shown to
be equivalent to an application of the spherical cosine rule. As far as we know, no Muslim
scholar wrote on the cosine formula per se, and we cannot even assume that even Najm al-
Dn al-Mir or al-Khall knew of it in its generality.21 All that can be asserted is that both
of them and many other medieval Muslim astrnomers were fully aware of the mathematical
equivalence of the two major problems of spherical astronomy (determination of the hour-angle
and the azimuth from an instantaneous celestial altitude) which in modern terms can most
readily be solved by an application of the formula. As far as the available sources indicate,
most of the astronomers who produced the tables I have described preferred projection methods
to spherical trigonometry anyway.
There is no evidence that Regiomontanus or any of the later European astronomers were
familiar with the Islamic auxiliary tables. Nor did the orientalists at Oxford in the 16th and
17th centuries, for all their interest in the Islamic astronomical texts and tables preserved in
Oxford, encounter any Islamic tables for timekeeping.22 Nevertheless, the auxiliary tables of
the Muslim astronomers, particularly the splendid universal tables of Najm al-Dn al-Mir and
al-Khall, as well as some of the more impressive of tables of other Muslim astronomers from
the 9th to the 16th century, would surely have been of interest to any serious European
astronomer, navigator or naval officer from the 18th to the 20th century. For during that time,
numerous sets of tables of trigonometric functions with two or three arguments were
compiled.23
Certainly the tables of the Muslim astronomers merit appropriate consideration in any new
history of astronomy or history of trigonometry. But most people would think that these
histories have already been written, and it is unlikely that newly-discovered historical sources,
such as a few hundred medieval tables, will change old attitudes.
20 Similar confusion about the use of the cosine formula for deriving the azimuth from the altitude is found
in Sezgin, GAS, V, pp. 261 and 288, and VI, p. 185 (based on earlier secondary sources).
21 Similar confusion relates to the attribution of the so-called prosthaphaeresis formula of trigonometry to
Ibn Ynus: see the remarks at the end of 2.1.1.
22 See Mercier, English Orientalists and Islamic Astronomy. As I have shown in a study listed as Ibn
Ynus and the Pendulum, the misconception that Ibn Ynus discovered the principle of the pendulum, an
association now widely spread on the Internet, actually stems from Edward Bernard in 17th-century Oxford.
23
Fletcher et al., Index of Mathematical Tables, esp. I, pp. 169-195 (tables of trigonometric functions and
miscellaneous functions associated with the circle and sphere), 207-229 (trigonometric functions of two or three
arguments).
european tables for timekeeping 191
Part II
A survey of tables
for regulating
the times of prayer
192 part i, chapter ten
To Bernard Goldstein
194 part i, chapter ten
This study is dedicated to the best teacher and thesis advisor a graduate student could ever
have wished for. Bernard Goldstein was just that, a man whose generosity, humour and breadth
of scholarship I have always admired and felt fortunate to have experienced and enjoyed. It
was he who taught me how to read Arabic scientific manuscripts and what to do with the
numbers one finds in them: in his seminar we read the Arabic version of Ptolemys Planetary
Hypotheses, in which he had been the first to discover Ptolemys inter(n)esting spheres.
One day Bernie told me to get hold of microfilms of the Leiden, Oxford and Paris
manuscripts of the Zj of Ibn Ynus and redo Delambre: this would serve as a doctoral thesis.
But I should also take a look at a certain Dublin manuscript: it had been catalogued by A.
J. Arberry as another copy of the Zj. The Dublin manuscript turned out to be unrelated to the
Zj, rather it was a copy of the Cairo corpus of tables for timekeeping, and the reader will
imagine my surprise and delight as I drooled over the microfilm for the first time. When I
returned to Yale in 1971 after a year in Beirut with Ted Kennedy, I showed Bernie not only
the first draft of my thesis on Ibn Ynus, but also my analysis of the corpus of tables for
timekeeping attributed to him in the Dublin manuscript. Bernie corrected the first draft of my
thesis over a weekend. Should I include the timekeeping stuff in the thesis?, I asked. No,
he replied, fully aware of the syndrome of the graduate student who cannot finish a thesis.
He was right, of course, and as soon as the thesis was submitted, I turned back to the time-
keeping. Thus Bernie is directly responsible for the genesis of this study, albeit unintentionally.
(Now this study is published, and the thesis not.)
When my wife and I went off to Cairo in 1972, Bernie gave me a compass and a book entitled
How to Survive in the Outdoors, both of which I have always cherished. I came out of Egypt
after seven fat years rather than the formidable forty years which others are said to have
taken. I used Bernies compass off and on, not to get back to the U.S. but to measure orientations
of various medieval mosques; thus he also unintentionally inspired another major area of my
research (see VIIa). In 2001 he further gave me a very useful little Eskimo man, carved out
of bone, and with a flat head: this one should turn upside down and address when one has
problems with colleagues.
At the special session of the Annual Meeting of the (U.S.) History of Science Society held
in Bernies honour in Pittsburg in November, 1999, four colleagues presented state-of-the-art
reports on their joint research with him, in ancient Greek, medieval European and Renaissance
astronomy, and, last but not least, things Keplerian. There were no presentations on Islamic
or Hebrew astronomy, which was a pity because Bernies contributions to both have been of
the same order of magnitude. However, the Festschrift dedicated to him in 2003 contains
contributions relating to each of these six widely-different various areas of the history of
astronomy and is appropriately entitled Astronomy and Astrology from the Babylonians to
Kepler. In each of these areas Bernie Goldstein has made ground-breaking discoveries, and
not only because he too had the best teacher.
196 part ii, table of contents
TABLE OF CONTENTS
CHAPTER 1
ON ISLAMIC PRAYER-TABLES
But now Islam has appeared in the eastern and western parts of the world and has spread
between Andalusia in the West and parts of China and central India in the East, and between
Abyssinia and Nubia in the south and the Turks and Slavs in the North. It has united all
the different nations in one bond of love, a handiwork which can be made by God only.
al-Brn (Afghanistan, ca. 1025), Tadd, p. 225, transl. by J. Ali, p. 190-191.
The historian of the Islamic exact sciences is frequently confronted with an embarras de
richesse hundreds of manuscript sources which have never been studied in modern times.
E. S. Kennedy, Islamic Mathematical Geography (1996), p. 185.
1.0 Introduction
The times of the five daily prayers of Islam are astronomically defined and for a given locality
vary throughout the year. Given a clear sky and unrestricted view of the local horizon, it is
possible to regulate the daytime prayers using a gnomon and the nighttime prayers by
observation of twilight phenomena. The legal scholars of medieval Islam, who spoke with
authority, advocated the methods of folk astronomy, devoid of any mathematics beyond simple
arithmetic, in their discussions of this subject (see III). However, the astronomers of medieval
Islam developed a new branch of science out of the religious obligation to pray at specific times,
which they called {ilm al-mqt, the science of astronomical timekeeping. In the early centuries
of Islamic astronomy, this topic was treated almost as in passing in zjes and treatises on
instruments. But in Egypt in the 13th century a new professional class emerged, astronomers
who specialized in this discipline and who were associated with a particular religious institu-
tion. These men were called muwaqqits, literally, timekeepers; astronomers who also worked
in this discipline independently bore the epithet mqt (see V). The writings of these men
consist mainly of treatises on the determination of the prayer-times either by direct calculation
or with the aid of analogue computers such as the astrolabe and astrolabic quadrant or
calculating devices such as the trigonometric quadrant. But they also produced an enormous
amount of tables of one sort or another as aids to their task. Muslim astronomers were also
concerned with the determination of the qibla, that is the direction of Mecca, which the Muslim
should face in prayer, and also the determination of the visibility of the lunar crescent, and
hence the regulation of the lunar Muslim calendar with its holy months of fasting and religious
festivals.
We have already seen the practical aids to their work which the astronomers and muwaqqits
had at their disposal in the form of tables for timekeeping by the sun and stars (I). The most
tangible and colourful evidence of the activity of the muwaqqits and their role in Islamic
societies is to be found in the mathematical tables which they prepared displaying the prayer-
times throughout the year for a particular locality. It is clearly convenient to have the prayer-
202 part ii, chapter one
times tabulated, either for each degree of solar longitude or for each day of the year, particularly
in localities where clouds are frequent or where the local horizon is obscured by mountains
or tall buildings. However, such tables must be used in conjunction with some kind of
instrument for measuring the passage of time, be it an astrolabe, a quadrant or a sundial, or,
from the 17th/18th/19th century onwards (depending on the location), a clock. The times of the
muezzins call to prayer in the modern Islamic world are still regulated by tables, compiled
by modern methods, and displayed in almanacs, calendars and pocket-diaries (see V-13).
The purpose of this study is to present a survey of the medieval prayer-tables which have
been located in recent years in various libraries around the world.1 The vast majority have never
been studied previously, indeed, they were not known to exist. These tables are of interest not
only by virtue of their mathematical sophistication and accuracy and for the light they cast
on medieval computational techniques, but also for the limited but significant information they
reveal on one aspect of devotional life in medieval Islam. I discuss mainly tables that were
used in Egypt and Syria, since these were the centres of {ilm al-mqt in medieval times. How-
ever, I have also included descriptions of all medieval prayer-tables from other areas known
to me in 1975 (and in 1999 I know of no others). I make no claim to have exhausted the available
material even for the study of Egyptian and Syrian prayer-tables, although I suspect that any
fresh tables from these two areas that may come to light will not differ greatly from those
described in the sequel. However, new Egyptian and Syrian manuscripts may clarify the picture
which I present of the transmission of prayer-tables in the medieval period and help to establish
the identity of those who compiled the tables, where this is still in doubt. The reader will observe
that many of the tables that I discuss exist in what appear to be unique manuscripts. However,
the main corpuses of tables for Cairo and Damascus and also Istanbul, exist in dozen
manuscripts.
In an early publication I analyzed in some detail the corpus of tables for timekeeping that
was used in medieval Cairo.2 In each of the seven manuscripts that formed the basis of that
analysis the tables are attributed to the 10th-century astronomer Ibn Ynus. Fresh manuscript
material enables me now to present new information on the main Cairo corpus (Chs. 4 and
5), showing that some of the tables were in fact not computed by Ibn Ynus but by later
Egyptian astronomers. There are still many questions surrounding our knowledge of astronomy
in Egypt before about 1400, particularly with regard to the activities of such individuals as
Najm al-Dn al-Mir and al-Marrkush (Ch. 6). Other later Egyptian tables for timekeeping
(Chs. 7-8) can generally be shown to be closely related to the main Cairo corpus. The
1 I use the terms medieval to cover the entire period from the 8th to the 19th century: see n. I-1:1. I use
the expression prayer-table to denote any set of spherical astronomical tables displaying functions relating
to twilight or the afternoon prayer. Many medieval prayer-tables also contain tables of standard spherical
astronomical functions: these I have discussed. Many zjes (see n. I-1:3) contain similar tables of spherical
astronomical functions: these I have generally not discussed. In this study I also treat tables for finding the local
qibla by means of the sun, but not tables displaying the qibla for a whole range of latitudes and longitudes,
except for that of al-Khall (10.8), the only such table in a corpus of tables for timekeeeping.
2 See n. 4:1 below.
on medieval islamic prayer-tables 203
The definition of the intervals during the five daily prayers should be performed have been
traditionally traced to certain verses of the Qur}n and various adth attributed to the Prophet
Muammad. The adth dealing with the prayer-times have been subjected to a preliminary
analysis by Arent J. Wensinck,5 and some of the various definitions found in the later
astronomical sources have been discussed in a valuable study by Eilhard Wiedemann and
Fig. 1.1a: The Islamic day, showing the times for the beginnings of the five prayers at sunset, nightfall, daybreak,
midday and mid-afternoon.
Fig. 1.1b: The standard definition of the beginning of the interval for the {ar is the time when the shadow increase
equals the gnomon-length. In some circles, notably amongst the anafs, the end of the permitted interval is
defined when the shadow increase is double that amount. In Andalus and Maghrib practice, the interval for
the uhr prayer begins when the shadow increase is one-quarter of the gnomon-length.
on medieval islamic prayer-tables 205
Joseph Frank.6 In the 1970s Ted Kennedy published a translation and commentary of al-
Brns book On Shadows, probably the most important single work on timekeeping compiled
in the Middle Ages, albeit without tables.7
The definitions which were used by the Muslim astronomers, and which underlie most of
the tables discussed in this study, are as follows: The Muslim day begins at sunset, and the
interval during which the first prayer (al-maghrib) is to be performed lasts from sunset to
nightfall. The interval for the second prayer (al-{ish}) begins at nightfall and lasts until
daybreak. The third prayer (al-fajr) is to be performed during the interval between daybreak
and sunrise. The permitted time for the fourth prayer (al-uhr) begins when the sun has crossed
the meridian and ends when the interval for the fifth prayer (al-{ar) begins; namely, when
the shadow of an object equals its meridian shadow increased by the length of the object. The
interval for the fifth prayer may last until the shadow increases again by the length of the object
or until sunset. In the medieval period there were occasional variations to this practice, such
as, for example, the definition of the uhr in al-Andalus and the Maghrib, whereby the prayer
began not at midday but when the shadow increase over its midday minimum was one-quarter
of the length of the gnomon. Also in some communities there was a prayer at mid-morning
called the u. Indeed there is sufficient material available to explain the origin of the curious
definitions for the uhr and {ar, which are not specifically mentioned either in the Qur}n
or the adth. I have investigated the origin of these definitions in IV.
The determination of the prayer-times either by observation or by computation was a
relatively simple problem for a competent medieval astronomer. The midday and afternoon
prayer-times can be determined with a facility using a gnomon. The corresponding solar altitude
can be readily computed so that one could also use an astrolabe or a quadrant to know when
the time for prayer had arrived. To determine the times from sunrise to midday, and from
midday to the beginning of the afternoon prayer (when the sun has a predetermined altitude),
are standard problems of medieval spherical astronomy. The times of the prayers at nightfall
and daybreak can be likewise be determined either by observation or by calculation. To compute
the duration of twilight one must assume that the transitions between twilight and total darkness,
that is, daybreak and nightfall, occur when the sun has a certain angle of depression below
the horizon. The determination of the duration of twilight is then a simple extension of the
problem of finding the time to sunrise or since sunset when the sun has a particular altitude,
which is again a standard problem of medieval spherical astronomy. Thus from an astronomical
or mathematical point of view the determination of the prayer-times is a straightforward
application of certain standard procedures in spherical astronomy.
Underlying several of the tables amongst those I shall be describing is a value for the qibla
or local direction of Mecca. the reader is referred to VIIa-c for an introduction to this topic.
In early Islam the times of prayer were regulated by observation. In the case of the daytime
prayers, the uhr and {ar, a gnomon erected in the mosque could be used to determine the
times with facility. Such a gnomon exists to this day in the beautiful Mosque of Janad in the
Yemen (see Fig III-1.2b); the mosque dates from the 7th century, but I do know when this
gnomon, a (concrete?) block of rectangular cross-section about 2 metres high and about 15
cm thick, was erected. In the case of the nighttime prayers, the moment of sunset can generally
be ascertained from the vantage of a minaret, and the twilight phenomena are likewise readily
observable.
In early Islam, time in general was reckoned either in seasonal hours of the day or night
or with respect to the times of prayers. The time of day could be calculated from shadow length
by simple formulae ultimately of Indian origin, as recorded in the treatises of al-Brn and
Ibn Raq (2.2-3). The Bedouin of the Arabian peninsula have possessed, since pre-Islamic
times, a detailed astronomical folklore relating the changing night sky to meteorological and
agricultural patterns; in particular, timekeeping by night was effected by observing the lunar
mansions.8 The time of night in early Islam could also be measured, for example, by a simple
candle-clocks: one such clock described in the medieval sources consists of a set of twelve
oil lamps which are filled in such a way that one goes out at each hour of the night.9
From the 9th century onwards the times of the uhr and {ar were sometimes regulated by
sundials.10 Also, individual prayer-times could be determined using special curves marked on
astrolabes and quadrants.11 However, the passion of the early Muslim astronomers for the
compilation of tables found a ready outlet in astronomical timekeeping. Already in the mid
10th century {Al ibn Amjr in Baghdad compiled two extensive tables displaying the time
since sunrise as a function of instantaneous solar altitude and meridian altitude. One of these
was based on an accurate formula and computed for the latitude of Baghdad, and the other
was based on an approximate formula and was intended to serve all latitudes (3.2). Ibn Amjur
and various other Muslim astronomers, even from the 9th century, likewise tabulated certain
functions relating to the times of prayer, such as the solar altitude at the beginning of the {ar
as a function of its meridian altitude, and the duration of twilight assuming that daybreak and
8 See, for example, the articles Anw} and Layl and Nahr by Charles Pellat and Manzil by Paul
Kunitzsch in EI2, and now Varisco, Islamic Folk Astronomy.
9 This clock is described in Kennedy & Ukashah, The Chandelier Clock of Ibn Ynis. The author is the
{Irq craftsman Ynus al-Asurlb, not the celebrated Egyptian astronomer: see King, Ibn Ynus, I.2.10 and
III.15.1c, and idem, Review of Hill, al-Jazar, n. 2 on p. 286.
10 On Islamic gnomonics see Schoy, Gnomonik der Araber, and, more recently, the survey in the article
Mizwala in EI2, repr. in King, Studies, C-VIII, and also X-7. Many new insights are to be gained from Charette,
Mamluk Instrumentation.
11 On the astrolabe in Islam see the articles Asurlb by Willy Hartner in EI , also listed under Hartner,
2
Astrolabe. On the quadrant see the ground-breaking study Schmalzl, Geschichte des Quadranten, and my
article Rub{ in EI2. Many new insights are to be gained from Charette, op. cit. Several Islamic instruments
displaying the curves for the various times of prayer are illustrated in Gunther, Astrolabes, I, and Michel, Trait
de lastrolabe. See further X-4.
on medieval islamic prayer-tables 207
nightfall occur when the sun is at a fixed arc of depression below the horizon (3.1 and 3.3).
Later some astronomers in Iran followed the tradition of Ibn Amjr in compiling tables of
the time since sunrise for the latitudes of Maragha and Shiraz, inter alia (3.8, 3.11 and 3.12).
The corpus of tables for Cairo associated with Ibn Ynus (Ch. 4) marks a new trend in
astronomical timekeeping in Islam. The major part of the corpus, which contains some 200
tables and over 30,000 entries, consists of tables for reckoning time and solar azimuth from
solar altitude. Other tables relate to the beginning of the afternoon prayer, the duration of
morning and evening twilight, and the time when the sun is in the azimuth of Mecca. Numerous
simple spherical astronomical functions are also tabulated, such as the solar meridian altitude
and the rising amplitude. All of the functions are displayed for each integral degree of solar
longitude, and the times are usually given in equatorial degrees and minutes. Not all the tables
in this corpus were computed by Ibn Ynus (Ch. 5), but the corpus set the pattern for most
subsequent prayer-tables in Egypt and Syria. In certain tables prepared by later Egyptian
muwaqqits (Chs. 7-8) we find the values of the functions relating to the prayer-times in the
Cairo corpus tabulated for each minute of solar longitude, with values derived from these by
interpolation. The extensive tables for regulating time by the sun and stars for all latitudes
compiled by Najm al-Dn al-Mir in the early 14th century (6.2) are rather different in character
from those of the Cairo corpus, but they were not widely used. In another set of tables,
apparently dating from the 15th century, we find prayer-tables like those in the Cairo corpus
computed for each degree of terrestrial latitude (8.1). The Cairo corpus itself was used in Egypt
as late as the 19th century alongside some of its later manifestations.
In Syria, tables for timekeeping were computed independently by three 14th-century
astronomers of exceptional competence, namely, al-Mizz, Ibn al-Shir and al-Khall (Chs.
9-10). al-Mizz, who studied astronomy in Cairo and then moved to Damascus, computed a
set of hour-angle tables and prayer-tables for his city which were virtually identical in their
conception to those of the Cairo corpus. Ibn al-Shir computed some prayer-tables for an
unspecified locality near Damascus, but this was his colleague al-Khall who was responsible
for the extensive corpus of tables for Damascus that was used in Syria until the beginning of
the 20th century. The main corpuses of tables that were used in the Yemen and in Tunis also
owe their inspiration to the Egyptian and Syrian traditions. The achievements of the muwaqqits
in Egypt and Syria after about 1500 are unimpressive, but this is partly explained by the fact
that the tables which were their major concern had already been computed by competent
astronomers in earlier centuries.
In the earliest Islamic prayer-tables, the functions relating to the times of prayer are tabulated
with the solar longitude as argument. In later tables of this kind, the argument was the date
in a calendar based on the solar year, such as the Syrian or Coptic calendars. Additional
calendrical tables could enable the user to find the corresponding date in a solar-based calendar
to use in such prayer-tables. Tables displaying the prayer-times for a given Muslim year were
also compiled.
In medieval prayer-tables it was usually the permitted intervals for the prayers that were
tabulated, these being expressed in equatorial degrees and minutes. In Ottoman times it
generally became the practise to tabulate the times of the prayers in equinoctical hours and
208 part ii, chapter one
minutes, according to the Ottoman convention that sunset is 12 oclock.12 The rationale for
this system derives from the fact that the Muslim day begins at sunset and also that the Muslim
astronomers generally used equinoctial hours rather than seasonal hours. However, this system
is hardly convenient when used with mechanical clocks, and when Western clocks were
introduced in the Ottoman Empire they were at first generally used with the Turkish time
system, the hands of the clock being adjusted to 12 oclock every evening at sunset. This custom
still prevails in the Arabian peninsula, although Yemeni friends of mine have claimed, and
not during a qt session, that their watches register 12 oclock at sunset throughout the year.
Modern prayer-tables are in a sense an extension of the Ottoman-type tables, now displaying
the times in Western time: see some examples in Fig. V-13.1. Also, since they are intended
to be used with clocks, the time given for midday usually incorporates the equation of time.13
In the text notation in the form (1.2.3) stands for (see Section 1.2.3). Occasionally in Part
II there are also cross-references to Part I in the form (I-1.2.3), and vice versa.
In the mathematical analysis of tables the notation f(x,y) is used for a function f tabulated
so that x is the horizontal argument and y is the vertical argument. I use the following
mathematical notation:
Latin
a azimuth, measured from the prime vertical
a earliest permitted time for the afternoon prayer (Ottoman convention)
-a relates to the earliest time for the afternoon prayer
b time for the end of the afternoon prayer (Ottoman convention)
-b relates to the end of the interval for the afternoon prayer
b1,2 auxiliary functions related to B
B the absolute base, an auxiliary function of Islamic timekeeping (see I-6.0)
C the altitude of the day-circle centre, an auxiliary function of Islamic timekeep-
ing (see I-6.0)
d half-excess of daylight, or equation of half daylight (d = D-90)
(uncial d) time of the u
- relates to the time of the u (only in Maghribi and Ottoman sources)
D semi diurnal arc
e the auxiliary function tan tan
G an auxiliary function for timekeeping related to sec sec
h altitude
h0 solar altitude in the prime vertical
ha solar altitude at the beginning of the interval for the afternoon prayer
hb solar altitude at the end of the interval for the afternoon prayer
hq solar altitude in the azimuth of the qibla
12 On this see Wrschmidt, Osmanische Zeitrechnung, and the text to n. I-1:36.
13 On this notion in medieval astronomy see the article Ta{dl al-zamn in EI2.
on medieval islamic prayer-tables 209
ta/b refers to a time between the beginning and end of the {ar (see 14.10)
tb time from midday to the end of the interval for the afternoon prayer (see also
b )
tq hour-angle when the sun is in the azimuth of the qibla
tq* hour-angle when the sun is in the direction perpendicular to the qibla
tz hour-angle at the beginning of the midday prayer (see hz)
-t relates to the ta}hb (Maghribi sources only)
T time since sunrise or remaining until sunset (measured from horizon)
Ta time from the beginning of the interval for the afternoon prayer to sunset
Tq time from sunrise to the moment when the sun is in the azimuth of the qibla
z horizonal shadow of a vertical gnomon (base given in parentheses as subscript)
za shadow at the beginning of the afternoon prayer
z vertical shadow
-z relates to the uhr
Z horizontal shadow at midday
Greek
right ascension
normed right ascension ( = - 90)
a oblique ascension of the ascendant at the {ar
r oblique ascension of the ascendant at daybreak
s oblique ascension of the ascendant at nightfall
ascension of the ascendant at the time of the salm
oblique ascension for latitude
solar declination
* solar declination augmented by 90 (useful since H = *-)
2 second declination (see F6 in I-1.3)
D correction to D for horizontal refraction
z increase of the shadow over its midday minimum
rising times of the zodiacal signs
obliquity of the ecliptic
time of the prayer at the {d (Ottoman convention)
independent variable
solar longitude or ecliptic longitude
elongation of sun from nearer equinox
* longitude of point of ecliptic opposite point with longitude (* = + 180)
H longitude of the ascendant (horoscopus)
M longitude of upper midheaven
time from sunset to the salm
- relates to the time of the salm
time from sunset to the time of the afy in Raman
a time before the beginning of the afternoon prayer
z time from the beginning of the midday prayer (see hz) to the beginning of the
afternoon prayer
on medieval islamic prayer-tables 211
longitude and local latitude , is symmetrical with respect to the solstices and so the
equinoctial value and 180 values suffice to define it for each integral degree of solar longitude.
This also holds for functions defining the permitted intervals for the prayers. In virtually all
the tables consulted, full advantage is taken of the symmetry of the functions tabulated. I denote
the longitude argument for such functions by ( \ 0), where measures the solar elongation
from nearer equinox. Generally early tables begin with entries for Aries and Virgo and later
ones begin with entries for Capricorn and Sagittarius.
Some of the Muslim astrononomers whose writings and tables we shall be considering used
approximate formulae, as a result of which computation is much facilitated. I mention just two
of these here, since they recur frequently.
First, an Indian arithmetical formula for finding the time in seasonal hours (T < 6) from
the increase in the shadow length z for a gnomon length n is equivalent to the following:
T 6 n / ( z + n ) .
Note that when z = 0 then T = 6, and when z then T = 0. For intermediate hours the
results are respectable to a first approximation. As we shall see, this formula underlies the
standard definitions for the {ar as well as the Andalus and Maghrib definitions of the uhr
(see further IV-2.4). See 2.3, 2.4, 2.5, 2.9, and 9.3 (Ibn al-Shir!) for the formula in the present
study, and also III and IV.
Second, an approximation based on trigonometric rather than simple arithmetical considera-
tions. The time in seasonal day-hours T (< 6) is given in terms of the instantaneous solar altitude
h and the meridian altitude H thus:
T 1/15 arc Sin { R Sin h / Sin H } .
Here T = 0 when h = 0 and T = 6 when h = H. The formula is in fact accurate when the
sun is at the equinoxes. For intermediate solar longitudes and latitudes between 15 and 40
this formula too gives results that are reasonable actually more reasonable than one would
expect to a first approximation. We find this formula in numerous of the sources studied
here, from the 9th century onwards: see I-2.5 and 4.3, II-2.4, 2.5, 2.9, 3.2 and 3.3, and further
XI. It underlies the universal horary quadrant that was a standard feature of astrolabes, both
Islamic and European, although those who used it in Europe were apparently aware that the
formula is no longer reliable at European latitudes.
In addition, various approximate rules for quick calculation of various spherical astronomical
functions are proposed by certain Muslim astronomers: see 2.4, 2.4a (Andalus source), 2.5,
2,9, 3.1 (early {Abbsid source!), 3.11, 6.1, 6.3, 6.12, and 9.3 (Ibn al-Shir!). It seems that
those who proposed such rules looked at tables of a function f(), noted the changes between
the values at the equinoxes f0 and solstices, and then determined the factors m and n (+
or -) in simple fractions which would best fit in a simple relation of the kind:
f() = f0 + m () for > 0 and f0 + n () for < 0 .
Sometimes other variables, such as d or were used instead of . I am not aware of any tables
based on such approximations.
on medieval islamic prayer-tables 213
Given the algebraic expression defining a particular function and the underlying parameters
it is possible to compile extensive tables of the function with an electronic computer (see
already I-1.5). The use of the computer in the analysis of medieval Islamic tables was pioneered
by Ted Kennedy, and is now being pursued by Benno van Dalen, Glen Van Brummelen and
Franois Charette.20 Once the underlying latitude is established, it is often useful to consult
the lists of coordinates taken from Islamic sources prepared by Ted and Mary Helen Kennedy.21
In 1970-71 I recomputed the main Cairo corpus in the Computer Centre at the American
University of Beirut and in 1971-72 recomputed similar sets for other latitudes at the Yale
University Computer Center. At the Computer Centre of the American University in Cairo,
programmes were prepared in 1972-73 to recompute prayer-tables of each of the various types
described in the sequel for any latitude. A given set for a particular latitude can be recomputed
in a matter of seconds once the underlying parameters have been established.22
The tables discussed in the present study contain a total of several million entries. Some
of them merit publication in toto, others are of little scientific or historical interest. In this study
I present sample entries only from tables that I consider to be of singular interest for one reason
or another. The values agree with recomputation except where the error in the second digit
is shown in square brackets, computed according to the convention:
error = (value in text) - (recomputed value) .
Rather than give sample entries, I have generally considered it more important to identify the
structure and underlying parameters of all the tables, to comment on their accuracy, as well
as to present illustrations of as many tables as possible.23
20 See n. I-1:24.
21 See n. I-1:4.
22 Computer time was made available by the Department of Mathematics, The American University of Beirut;
the Department of Near Eastern Languages and Literatures, Yale University; and the Smithsonian Institution,
respectively.
23 Ideally, sample entries should be given for tables forming part of a clearly defined corpus of prayer-tables
for a given locality. (A model for such an endeavour for standard astronomical tables is Toomer, Toledan
Tables.) In most cases such entries (say for each 30) are adequate to assess the accuracy of the tables and
to identify related material. In King, Astronomical Timekeeping in Medieval Cairo, pp. 380-384, I present
sample entries from the tables in the Cairo corpus, and in an unpublished study Shams al-Dn al-Khall and
the Culmination of the Islamic Science of Astronomical Timekeeping, yet more extensive samples from the
Damascus corpus. The latter work is unpublished not least because the prospective publisher, the Third World
Academy of Sciences, some years ago understandably baulked at the prospect of printing masses of tables as
well as photos of the most important manuscripts (including MSS Paris ar. 2558 and Dublin CB 4091). The
text, without the tables, is now incorporated in Ch. 10 of the present study.
214 part ii, chapter two
CHAPTER 2
An enormous amount of Arabic scientific literature deals with astronomical timekeeping. Also,
the topic of spherical astronomy is dealt with in the majority of Islamic zjes. In order to display
the importance of this literature for the documentation of the history of the development of
mathematical methods, and the history of astronomical inquiry into such phenomena as twilight
and refraction at the horizon, I have selected a few of the more historically interesting of the
available treatises and have attempted to survey their contents. I have deliberately included
one treatise in the folk astronomical tradition (2.3). There are others, also of widely varying
levels of competence on the part of their original authors, to which the interested reader may
have recourse: Kshyr ibn Labbn on spherical trigonometry;1 Ab Nars treatises on
spherical astronomy;2 al-Brns treatise on spherical trigonometry applied to problems of
spherical astronomy;3 al-Kshs treatment of spherical astronomy in his Khqn Zj;4 and Ab
Miqra{s treatment of timekeeping by folk astronomical methods.5 Also the simple techniques
for reckoning time of night by the lunar mansions have been studied recently by Miquel
Forcada.6 In addition there are several accounts in the published literature of the solution of
such problems using astronomical instruments, such as al-Khwrizm on the astrolabe;7 abash
on the celestial globe;8 Ibn al-Zarqlluh on the universal plate (afa shakkziyya);9 Ibn al-
Shir on an astronomical compendium (multi-purpose instrument) of his own design;10 and
Jaml al-Mridn on the universal quadrant.11 See further X. As this book goes to press, I
am gratified that Julio Sams has published an account of a medieval Maghrib treatise
preserved in MSS London B.L. 411 and Cairo K 4311 which deals critically with parameters
for twilight (and also trepidation) as discussed by various earlier authors.12
2.1 A poem on the times of prayer attributed to the Imm al-Shfi{ and Ibn Ynus
In MS Cairo DM 181, fols. 46v-48r, copied ca. 1800, there is a short poem on the prayer-
times in awl metre attributed to Ibn Ynus (4.1): see Fig. III-9.7b. The Egyptian astronomer
was renowned as a poet and several of his poems are preserved in various anthologies and
biographical lexicons.13 However, in MSS Berlin Ahlwardt 5820 (Wetzstein 175), fol. 65r, and
5700 (Landberg 953), fol. 11r, copied in 944 H [= 1537/38] and 1231 H [= 1816], respectively,
a similar poem is attributed to the famous legal scholar al-Shfi{ (d. 819/20). This poem is
also attributed to al-Shfi{ in various later Egyptian works on timekeeping, such as those of
{Abd al-Ramn al-Tjr and Ysuf Kilrj (8.8). al-Shfi{ was renowned for his poetry, but
this particular poem is not included in a modern anthology of his poetical works.14 The poem
in both Berlin copies is extremely corrupt and is clearly an improvisation of the original poem
by someone who was not very gifted in either astronomy or poetry. I present a prose translation
of the poem from MS Cairo DM 181, incorporating a few corrections necessary to the sense
of the poem, some of which are confirmed by the readings in the second Berlin manuscript
(Ahlwardt 5700):15
Knowing the prayer-times is a prescribed duty for discerning Muslims. This is
summarized in the Qur}n, my friend, and was explained by Amad [i.e., the Prophet
Muammad, referred to as Amad in the Qur}n], the most outstanding of men.
Perform the midday prayer whenever you observe the shadow starting to increase. Add
a length (of the gnomon) to the shadow: this gives you (the shadow) at the time of
the afternoon prayer. At sunset get up and perform the evening prayer: this is the only
permissible time. Perform the night prayer when, looking at the sky, you see the upper
part of the evening twilight fade away and disappear. As for the end of this prayer
time, if you wish you can wait until one-third of the night; indeed, it is better that
you wait. But do not wait until whiteness appears (on the eastern horizon): it will last
for a time in the sky. Bear in mind that there are two stages of daybreak according
to our doctrine: distinguish between them carefully you are the one who decides
this. The first daybreak looks like a wolfs tail rising in the sky: this is the false dawn.
The later one is the true dawn: you see it illuminate the sky like a fire. The end of
this prayer-time is sunrise: at that moment the best time for prayer is over. There is
no virtue in a person who is neglectful of the prayer-times and he has no knowledge
of Him who is to be worshipped.
Ibn Ynus is the more likely author because in the time of al-Shfi{ the standard definitions
mentioned here were not yet definitively formulated (IV-3).
13 On Ibn Ynus poetry see King, Ibn Ynus, I.1, and the references there cited.
14 On al-Shfi{s poetry see Sezgin, GAS, III, p. 647.
15 The text of the poem is published in my contribution to the Aleppo 1976 Symposium Proceedings, I (Arabic
section), pp. 393-394, but is not included in the English version Astronomical Timekeeping in Medieval Syria,
repr. in King, Studies, A-X.
216 part ii, chapter two
al-Brn was the most outstanding scientist of the Islamic Middle Ages.16 He was a prolific
writer and did original work in fields as varied as ethnography, geology and astronomy. Our
present concern is with his Ifrd al-maql f amr al-ill, Unique Treatise on Matters Relating
to Shadows, in which he devotes a whole chapter to the opinions of the religious leaders about
the times of prayer and the obligation to determine them properly. The Arabic text of this work
was published in Hyderabad in 1948 and E. S. Kennedy published a translation and commentary
in 1976.17 al-Brns treatise is a mine of information concerning the prayer-times and time-
keeping in general. However, al-Brn does not make a single reference to tables for regulating
the prayer-times, although in his treatise on the construction of the astrolabe entitled al-Ist{b
he does mention aylasn tables for time-keeping.18 Thus this study is in a sense a supplement
to Kennedys study of al-Brns Shadows.
MS Berlin Ahlwardt 5664 (Landberg 108), achieved ca. 1350, is an apparently unique copy
of a treatise on folk astronomy by an individual named Ab {Abdallh Muammad ibn Raq
ibn {Abd al-Karm, otherwise unknown to the literature.19 The treatise was compiled in Mecca
in the early 11th century and it contains information of a very simple kind on the calendar,
the lunar mansions and timekeeping. Some of the material on the latter topic is said to be taken
from the large book of Ab {Al {Araqa, the muezzin of the mosque of {Amr (al-jmi{ al-
{atq) in Fus. The treatise is an important source for our knowledge of the practices of the
muezzins of early Islam, before sophisticated mathematical methods were applied to time-
keeping. A detailed study of this and two other Rasulid Yemeni treatises on the same subjects
by Ibrhm ibn {Al al-Aba20 and Muammad ibn Ab Bakr al-Fris21 is currently being
conducted by Petra Schmidl.22
Ibn Raq begins his treatise by asserting that the times of prayer should be determined by
observation with ones own eyes (al-{iyn wa-l-raad), not by using an astrolabe or any of
this astronomy nonsense (Arabic, tanjm: the word usually means astrology). He states that
he will follow the adth about the Archangel Gabriel recorded in the canonical collections
of Muslim and al-Bukhr (IV-1.2), rather than the writings of those take their learning from
the infidels and the Sindhind. (By these he means the Greek and Indian traditions in Islamic
astronomy, respectively.)
With regard to the uhr prayer Ibn Raq recommends the use of an Indian sundial (al-
rukhma al-mabniya (?) al-hindiyya) with which one can watch for the moment when the
16 On al-Brn see the splendid article by E. S. Kennedy in DSB. His work on timekeeping is listed as al-
Brn, Shadows.
17 See n. 1:7.
18 See n. I-1:53.
19 On Ibn Raq and his works see King, Astronomy in Yemen, no. 2, and also III-3.2 and IV-1.5.
20 On al-Aba see King, Astronomy in Yemen, no. 5; Cairo ENL Survey, no. E5; and IV-1.5, etc.
21 On al-Fris see n. I-7:17, and also III-4.2 and IV-1.5.
22 See Schmidl, Islamische Volksastronomische Abhandlungen, and also eadem, Qibla und Winde.
selected medieval arabic treatises on timekeeping 217
shadow cast by the sun begins to increase.23 Ibn Raq records that {Araqa had reported that
Amad al-Rz, the muezzin at the Sacred Mosque in Mecca, had observed the sun over a period
of seven years and had found that the shadow length for the Coptic months, expressed in terms
of the height of a man (qma), were as follows (beginning with the 10th month):
X: 1/3 1/5 XI: 0 a XII: 1/
6
I: 1/3 + 1/2 1/6 b II: 1/2 III: 2/3 + 1/
4
IV: 1 + 1/6 1/8 V: 1- 1/2 1/8 VI: 1/3 + 1/
4
VII: 1/4 + 1/8 VIII: 1/6 IX: 0
a Arabic: l shay} b text has: 1/33 + 1/2 = 1/6
These fractions, written in words in the text, can be simplied thus:
1/
18 0 1/
6
5/
12
1/
2
11/
12 11/48 15/16 7/
12
3/
8
1/
6 0
Ibn Raq states that the beginning of the {ar prayer should be when the shadow is 61/2
feet (qadam, pl. aqdm) more than the midday shadow, which indicates that he favored
61/2 units for the length of the gnomon. He records the following maximum shadow lengths
reported by Ab {Al {Araqa:
Mecca 7 - 1 /5 (= 64/5)
Medina 7 + 1/2
Sirrayn (in Yemen) 7 - 1 /3 (= 62/3)
Alexandria, Baghdad, Crete 10
Jerusalem 9 + 2/3
Cairo-Fus (Mir) 9
Ifrqiyya (i.e., Tunis), Rayy 11
Qum 13
Kufa 8 + 1 /2
Armenia 6 (read 16?)
{Araqa also measured the shadow at Sanaa on Kayhak 25 (year unspecified), in the company
of the adth scholar Amad ibn li, and found it to be 41/2 feet.
In a section on measuring time of day approximately from shadow lengths Ibn Raq presents
a method for finding the seasonal hours from the excess of the instantaneous shadow over the
midday shadow, which he asserts is valid for the latitudes of al-{Irq, Syria, the Maghrib and
Egypt. When these excesses are:
39, 191/2, 61/2, 31/4, 1+
according to Ibn Raq, the time is:
1, 2, 3, 4, 5 or 11, 10, 9, 8, 7
seasonal hours after sunrise. The figure for 5 and 7 seasonal hours is expressed as qadam rji,
which means just more than one foot. No explanation is given for these figures. Consider,
however, the following approximation for the time T in seasonal day-hours as a function of
the excess z of the instantaneous shadow of a gnomon length n over the midday shadow,
namely:
T 6 n / [ n + z ] .
23 It is not completely clear what is meant here. The name Indian circle (al-d}ira al-hindiyya) is usually
given in the Arabic sources to a simple device and procedure a gnomon and a a geometric construction based
on observations with it for determining the local meridian. On this see Kennedy, al-Brns Shadows, II, pp.
80-90, and, most recently, King, Mecca-Centred World-Maps, pp. 93 and 94, and also n. 24 on p. 134.
218 part ii, chapter two
This formula, which is found in several Islamic sources (1.5), is ultimately of Indian origin
(III-1.3 and IV-2.4). If we substitute T = 3, 4 and 5sdh, we obtain respectively:
x = 61/2 , 31/4 and 13/10 ,
as in the text. Further, if x is large compared with n and we use the cruder approximation:
T 6n/x,
we obtain for T = 1sdh and 2sdh Ibn Raqs values:
x = 39 and 191/2 .
Thus each of the five values is explained. Ibn Raq adds a note that the 9th hour marks the
beginning of the {ar prayer according to the fourth caliph {Al ibn Ab lib (III-2.2 and IV-
2.2).
Here we have, perhaps, the explanation of the origin of the definition of the {ar which was
not stated in the adth but which was adopted already in the 8th century and has persisted
to this day. If we postulate that the {ar divides the period from midday to sunset into two
equal intervals, then, according to the approximate Indian formula quoted above, the length
of the shadow at the {ar is equal to the midday shadow increased by the length of the gnomon.
If this is indeed the origin of the definition of the {ar, Ibn Raq has underestimated the debt
of Islam to at least the Indian tradition in early Islamic astronomy. Note also that the shadow
length at the beginning of the 10th hour, according to the Indian formula though not according
to Ibn Raq, is in excess of the midday shadow by twice the length of the gnomon. This is
Ab anfas definition of the {ar (1.1 and IV-3).
Ibn Raq states that the evening prayer (al-maghrib) should begin when redness appears
in the western sky after sunset. He distinguishes between the true and false dawn for the
beginning of the morning prayer (al-fajr).
Our author quotes from a work of Muhammad ibn Surqa al-{@mir (d. 1019) information
about the determination of the qibla in various parts of the Muslim world. The following extract
is typical of the information contained in an extensive corpus of material dealing with what
I call sacred geography and which I have surveyed elsewhere.24 It is traditional knowledge,
independent of the scientific tradition of mathematical geography and cartography in medieval
Islam.
The people of Alexandria and Egypt up to Kairouan and Sus (al-Ss al-aq) and
the Black Sea and all places in the same direction (with regard to Mecca) should face
the same direction as (when one is standing in front of the Ka{ba) between its western
corner to the draindripe (at the middle of the north-western wall). In any of these
places, if one should stand with the three main stars of the Great Bear (bant na{sh)
at ones right shoulder when they are setting and at ones left shoulder when they are
rising, with the celestial pole behind ones back and the west wind (al-dabr) on ones
right, then one is facing the direction of the Ka{ba.
A large part of Ibn Raqs treatise consists of simple information about each of the lunar
mansions throughout the year, such as is found in several later Yemeni works (12). The
following is an extract:
24This kind of qibla determination is surveyed in my unpublished study The Sacred Geography of Islam,
summarized in the article Makka. iv. As centre of the world in EI2, repr. in King, Studies, C-X. See also
Schmidl, Islamische Volksastronomische Abhandlungen (n. 2:22).
selected medieval arabic treatises on timekeeping 219
The lunar mansion al-Ikll rises at daybreak on Htr 17, which is Tishrn II 13,
November 13 and @dharmh 10. Morning twilight (?: al-ill) is when the next mansion
al-Qalb (is rising over the horizon) and the sun is in (the next mansion) al-Shawla.
The beginning of the night is when al-Haq{a is rising. The mansions culminating at
dawn and nightfall are al-Jabha and al-Mu}akhkhar. This mansion is (completely)
visible on the 12th day of its rising and contains three bright stars which (form) the
crown of the Scorpion. The midday shadow at Mecca may God exalt her is six
and one third feet.
MS Istanbul Haci Mahmud Efendi 5713, penned in the 12th century, is a unique copy of three
Egyptian astronomical treatises of considerable historical interest. These are: (a) a treatise on
timekeeping by an individual called Ibn al-ammm (fols. 1r-10v), (b) a treatise on the
astrolabic quadrant by the same author (fols. 10v-25v), and (c) a treatise on the trigonometric
quadrant by Fat al-Dn al-Qays (fols. 25v-34v). The treatise on the astrolabic quadrant
predates the earliest previously-known Arabic treatises on quadrants (by al-Mizz) by at least
a century.25 I have no information on either of the two authors.26
Our present concern is with the first of these treatises, which contains various approximate
rules for the standard functions of timekeeping, such as those given by the later Egyptian
astronomers Najm al-Dn al-Mir, al-Bakhniq and al-Fawns (3.3, 3.11 and 6.12).
Ibn al-ammm first records the following approximate changes in and for each 30
of measured from the equinox:
: 12 8 4 and : 21 24 30 34 36 35 ,
and the following midday shadows for the zodiacal signs:
7 4 2 1;21 2 4 7 10 14 16;16 14 10
To find the time T in seasonal hours since sunrise from the instantaneous and midday shadows
z and Z he prescribes the rule:
T = 72 / [ ( z + n ) - Z ] (n = 12) .
To find the shadow z1 and z2 at the (second optimum?) time for the midday prayer (ill ikhtiyr
al-uhr) and the end of the midday prayer (ill khirihi), the author prescribes:
z1 = Z + 1/4 n and z2 = Z + n (n = 12) .
Likewise the shadows z3 and z4 at the beginning and end of the afternoon prayer (ill awwal
waqt al-{ar and ill khirihi) are given as:
z3 = z2 + and z4 = Z + 2n ,
where is referred to as adn ziyda, the last (perceptible) increase. I know of no other
Egyptian treatise in which the times of the midday and afternoon prayers are defined in this
way (IV-4).
In the course of his discussion of the determination of the standard functions of timekeeping
Ibn al-ammm presents several approximate rules. As an example I quote the rule for finding
ta and Ta:
To find the time between the midday and afternoon prayers and between the afternoon
prayer and sunset, add one-half of one-sixth of the solar declination to 52 for the
northern signs and subtract one-quarter of the declination and one-sixth of the
declination for the southern signs. The result will be the time between the midday and
afternoon prayers. Subtract this from half the diurnal arc and the remainder will be
the time between the afternoon prayer and sunset. God knows best.
In all there are five such approximate rules:
(a) ha 32;30 + 2/5 ( \ 0)
(b) hq 47 + + /5 2 ( \ 0)
(c) ta 52 + 1/2 1/6 ( > 0)
ta 52 - /4 - /6
1 1 ( < 0)
(d) s 18;30 + 1/8 ( > 0)
s 18;30 + 1/2 1/8 || ( < 0)
(e) r 23;30 + 1/5 ( > 0)
r 23;30 + 1/2 1/6 ( < 0)
Rule (b) corresponds slightly more closely to the paramater q = 52 than to 53 (though see
below) and rules (d) and (e) imply that Ibn al-ammm accepted the parameters 20 and 16
for morning and evening twilight, which is confirmed on fols. 22r and 22v of this Istanbul
copy of his treatise on the astrolabic quadrant. I consider it likely that Ibn al-ammm derived
these simple rules by inspecting some tables of these five functions. Those in the main Cairo
corpus give the following values for the equinoxes and the solstices:
ha hq hq ta s r
q=52 q=53 hs=-16 hr=-20
EQ (=1) 32;33 47;23 46;43
51;54 18;35 23;18
SS 41;58 80; 2 79;48
53;48 22;13 28;28
WS 23; 0 13;39 12;32
41;53 20; 4 24;54
This having been said, it should be pointed out that Ibn al-ammm predates al-Maqs and
so was more likely to have access to earlier tables that may not have been preserved for us.
See, for example, 5.3.
Our author concludes his treatise by describing some simple operations with an astrolabe,
including how to use the alidade to indicate the direction of the qibla, stated as 37 south of
east (q = 53).
The Andalus astronomer Ibn Bo (d. 1316) is best known for his ingenious universal plate,
which became a feature of many later Maghribi and even a few Eastern Islamic astrolabes.27
27 See Calvo, Ibn Bo and his Universal Plate, where the definitions are discussed on pp. 78-80.
selected medieval arabic treatises on timekeeping 221
In his treatise on this plate he advocates approximate rules for the times of the uhr and {ar,
as follows:
hz ( H - 10 ) - 1/10 ( H - 30 ) for H > 30 and hz H - 10 for H < 30
ha 1/2 H + 1/10 ( 80 - H ) for H < 80 and ha 1/2 H for H > 80
The second rule is curious because for the latitude of, say, Cordova, the solar altitude does
not reach 80. I have not encountered these elsewhere, although certain Egyptian treatises (2.9
and 6.1) have a similar rule for ha with underlying latitude 30.
MS Istanbul Hamidiye 1453, fols. 219r-228v and 228v-230, copied in Damascus in 869 H [=
1464/65], contains two related treatises on timekeeping, the second being attributed to Najm
al-Dn al-Mir, who worked in Cairo in the early 14th century (see I-2.6.1 and I-9.3*, and also
II-6.5 on his extensive tables for timekeeping). The first treatise is entitled Tarr al-maqla
f ma{rifat al-awqt bi-ghayr la, and the second Ikhtir al-maqla ... , meaning respectively
Redaction of and Abridgement of of a Treatise on Determining the Times of Prayer
without Instruments. A third treatise on timekeeping by Najm al-Dn al-Mir is preserved
in MS Milan Ambrosiana 227a (C49), fols. 85v-97r, and represents a much more sophisticated
level of spherical astronomy. Neither of the first two treatises is related to the third, since they
both deal only with approximate methods, and not always the same ones at that. However,
I see no problem in accepting Najm al-Dn as the author of all three.
In the first treatise Najm al-Dn first records the following approximate changes in , and
for each 30 of measured from the equinox:
: 12 - 8 - 4
: 28 - 30 - 32
: 21 - 24 - 30 - 35 - 35 - 35
To find the shadow lengths (base 12) corresponding to a particular altitude he considers the
four cases:
(a) h < 27 , (b) 27 < h < 45 , (c) 45 < h < 63 and (d) h > 63
stating that the corresponding shadow lengths are respectively:
(a) h / 41/2 , (b) (h-27)/3 + 6 , (c) (63-h)/3 + 6 , and (d) (90-h) / 41/2 ,
where the first two are vertical (manks) and the last two horizontal (mabs). He also notes
that the horizontal and vertical shadows z and z are related by:
z z = 144 (= 122) .
Najm al-Dn then discusses the altitudes corresponding to (vertical) shadow lengths in feet
(aqdm) and mentions the values:
z 1 2 3 4 5 6 6;40
h 9 17 24 31 37 42 45
For timekeeping by the sun he notes that the length of half-daylight in equinoctial hours D
is given approximately by:
D 11/60 ,
a formula which underlies a simple table in MS Paris BNF ar. 2513 of the contemporary
222 part ii, chapter two
Muala Zj (6.6 and I-7.4.1), and then prescribes the following rules for finding the time in
seasonal hours since sunrise, T, from the instantaneous and midday shadows, z and Z:
T 40 / [ (6;40 + z) - Z ] (shadows in feet)
T 40 / [ (12 + z) - Z ] (shadows in digits)
The author also presents the following approximate rules for finding the duration of morning
and evening twilight r and s in seasonal hours when the solar longitude is :
r 72 / [ (33+12) - Z* ] and s 72 / [ (42+12) - Z* ] ,
where Z* is the horizontal midday shadow (n = 12) cast by the sun when it has longitude *
= + 180. Note that Cot12 20 33 and Cot12 16 42, so that these rules are simply extensions
of the above rule for reckoning time by day.
In the second treatise Najm al-Din records the following approximate changes in , and
for each 30 of measured from the equinox:
: 12 - 8 - 4 : 28 - 30 - 32 and : 21 - 24 - 30 - 34 - 36 - 35
and the following midday shadows (n = 12) for the zodiacal signs beginning with Aries:
7 4 2 1;20 2 4 7 10 14 16;15 14 10 .
Note that the solar meridian altitude in Cairo at the equinox and solstices are 60, 83;35, 36;25
and the corresponding shadows are 6;56, 1;21, and 16;16.
The remainder of the treatise deals with the determination of the standard functions of
timekeeping. Najm al-Dn mentions the following approximate formulae:
(a) ha 32;32 2/5 || ( \ 0)
(b) d 1/2 + 1/4
(c) ta 52 + 1/3 1/5 d ( > 0)
ta 52 - 1/3 |d| ( < 0)
(d) + /6 1
so the rule is a fair approximation. (No tables of aa() are known to have been prepared for
Cairo, but the function is discussed in several later treatises on sundial theory.)
In passing I note the existence of a related anonymous treatise preserved in MS Escoral
r. 961,3 (fols. 13v-16v), copied in 863 H [= 1459]. This consists of a preface and 21 very
short chapters, each dealing with one of the standard functions of timekeeping. The
approximations for d, , s and r given in this source are the same as those of Najm al-Dn,
and the other approximations are clearly related to his:
(a) ha 32 + 2/5
(b) ta 52 + 1/3 1/5 d
(c) hq 47 + 1/6 + 1/4
(d) h0 2 ( < 15)
h0 30 + /4 (2 - 30)
1 ( > 15)
(e) aa 21 + 1/8 .
The work attributed to Najm al-Dn al-Mir in MS Milan Ambrosiana 227a (C49), fols.
85v-97r, is entitled al-Risla al-isbiyya fi l-a{ml al-fqiyya, which means Treatise on
Computations in Spherical Astronomy for all Latitudes. This short treatise of 31 chapters
contains material of considerably greater sophistication than his two other treatises. All of the
theoretical procedures advocated are exact, and are expressed in the terse technical Arabic
standard in later Egyptian and Syrian works on astronomy. I shall refer to the section on
reckoning time from celestial altitude elsewhere (5.2), and here restrict attention to the only
historical reference in the work, which occurs in Ch. 11 on the determination of twilight. Najm
al-Dn advocates the parameter 18 for both morning and evening phenomena but states that
if one is calculating for Cairo in particular according to the opinions of the celebrated Ibn
Ynus, one should use 20 and 16. This reference confirms that Ibn Ynus wrote more on
twilight than has been preserved in MS Leiden Or. 143 of his kim Zj (4.5 and 5.1) and
indicates that the Egyptian tradition of using the parameters 20 and 16 which we shall
encounter frequently was much older than the 13th century.
{Abd al-{Azz ibn Amad al-Drn was an Egyptian dervish who lived from about 1215 to 1297.
He was the author of numerous works, one of which deals with astronomical timekeeping.28
This particular compilation is a poem with intermittent prose passages and simple tables and
diagrams (see Fig. 2.6a). I have examined MS Istanbul Hamidiye 1453, fols. 85v-102v, copied
in Edirne in 869 H [=1464/65], and MS Istanbul Ayasofya 2711 of this poem, which is entitled
al-Yawqt f {ilm al-mawqt, Gems for Timekeeping.
al-Drns discussion of calendars, lunar mansions and the motions of the sun and moon
do not directly concern the present study. He notes the following midday shadows for the twelve
signs beginning with Cancer, which appear to be based on n = 6;40:
2/
3 1 2 31/2 5 1/4 + 1/8 71/4 81/8 7 1/ 4 ...
28 On al-Drn see Brockelmann, GAL, I, pp. 588-589, and SI, pp. 810-811; and Cairo ENL Survey, no. C14.
224 part ii, chapter two
Figs. 2.6a-b: Illustrations of the bright stars ({ayyqt) in each of the 28 lunar mansions (a), and a curious table
purporting to show the solar altitude for each integral shadow length (to base 6;40!) (b). [From MS Istanbul
Hamidiye 1453, fols. 87v-88r and 97v-98r, courtesy of the Sleymaniye Library, Istanbul.]
selected medieval arabic treatises on timekeeping 225
He also notes the daily change in shadow length for each sign. In one of the two manuscripts
there is a table displaying the midday shadows for each of the Coptic year. al-Drns values
for the solar meridian altitude and for the lengths of daylight in equatorial degrees are:
84 80 72 60 48 40 36 40 ...
210 205 195 180 165 155 150 155 ...
His values for the change in oblique ascensions for each 30 of ecliptic longitude are:
21 24 30 34 36 35 ..
A table of solar altitude corresponding to particular horizontal shadow is also presented (see
Fig. 2.6b). The first six entries:
z 1 2 3 4 5 6
h 81 73 66 59 53 48
correspond to those in the first treatise of Najm al-Dn al-Mir (2.5); the remaining entries
are somewhat confused. The reader should bear in mind that they are written in abjad notation
in the manuscripts:
z (Hamidiye) 6;47 7;50 8;48 11;15 14; 6 22;30 44;60
z (Ayasofya) 6;80 7;50 8;48 11;15 14;100 22;30 44;60
h 45 42 37 31 24 17 9
It is difficult to explain these numbers. However, it is clear that z(45) was intended to be 6;40
and the mm representing 40 has been miscopied as mm-zy 47 and (by a further stretch of
the imagination) f} 80. The values of h in the second part of the table are the complements
of those in the first part, and for these complementary values we should have shadows z and
z such that:
z z = (6;40)2 = ( 62/3 )2 = 400 / 9 .
In order to complete the table one must divide this product by 1, 2, 3, 4, 5 and 6 to obtain
the shadow lengths corresponding to 9, 17, 24, 31 and 42. The values that one obtains
this way are:
z 7;24 8;53 11; 7 14;49 22;13 44;27
h 42 37 31 24 7 9
I suggest that whoever compiled the table in al-Drns treatise was incapable of expressing
the quotient 400/9 as a sexagesimal and wrote z(9) = 44;60. This explains the values z(17)
= 22;30 and z(31) = 11;15. To find z(24) our author divided 44;60 by 3 and somehow derived
14;100 (the qf for 100 was confused for ww, 6 in the Ayasofya copy) and hence 7;50 for
z(42). To find z(37) he divided 44;60 by 5 and was content to a settle for 84/5 = 8;48 as the
quotient. The above explanation is somewhat far-fetched, but how else can one explain the
absurd entries in the table?
Ab {Ali al-Marrkush was the author of a substantial summa of spherical astronomy and
instruments entitled Jmi{ al-mabdi} wa-l-ghyt. Part of this work was translated by Jean-
Jacques Sdillot (pre) in 1834-35 as Trait des instruments astronomiques des Arabes,
although in fact the part that he translated deals only with spherical astronomy and sundials.
226 part ii, chapter two
Timekeeping
I have discussed elsewhere (I-6.7.2) al-Marrkushs small auxiliary table for timekeeping
operations (I.264). It is rather curious that he does not tabulate the functions:
B() = Cos () Cos / R or B() = Cos Cos / R ,
which he calls the al, base, and uses freely throughout his discussion of the determination
of time from solar and stellar altitudes (I.258ff.). I have also discussed (I-7.1.7) al-Marrkushs
auxiliary tables for finding the half-excess of daylight (I.209). The method he expounds is
repeated by al-Bakhniq (see 5.6, note (k) to the translation).
al-Marrkush advocates the parameters 20 and 16 for morning and evening twilight
(I.295-296). He states that observations which he had made in various latitudes between 20
and 45 led him to the conclusion that these parameters, or what differs from them by one
degree, by which he surely means 19 and 17, were the most suitable.
32 On Amn al-Dn al-Abhar (not to be confused with the well-known Athr al-Dn al-Abhar) see Suter, MAA,
nos. 369 and 393 (confused).
33 On Abu l-Fid} see Suter, MAA, no. 392, and the articles in DSB by Juan Vernet and in EI by H. A. R.
2
Gibb.
34 Gotha Catalogue, pp. 64-65.
35 See the article Shakkziyya in EI , and X-5.2.
2
selected medieval arabic treatises on timekeeping 229
2.10 Sib al-Mridn on the differences of opinion of the religious scholars and the muwaqqits
Sib al-Mridn was one of the leading astronomers in Cairo towards the end of the 15th century;
he was a prolific author, mainly of treatises on spherical astronomy and instrumentation.37 One
of his works (see also 5.8 on another) exists in an apparently unique copy, MS Princeton Garret
Hitti 531 (1960). This is a discussion of the validity of the times of prayer determined by the
muwaqqits, and it is disappointing for the paucity of information which it presents on our
subject. Nevertheless I present a free translation of the entire treatise.
Questions and answers by the scholar Sib al-Mridn concerning the determination
of the prayer-times, and whether or not the opinion of the astronomers agrees with
the opinion of the legal scholars, and other similar problems.
36 See n. I-1:32.
37 On Sib al-Mridn see Suter, MAA, no. 445; Cairo ENL Survey, no. C97; and ~hsanolu, et al., Ottoman
Astronomical Literature, II, p. 1014 (index).
230 part ii, chapter two
In the Name of God, the Merciful and Compassionate. The lawyers, who are the imms
of religion and the scholars of the Muslims may God grant them all success for their
obedience have written in their books, and commented thereon with their legal
opinions, concerning those daily phenomena which can actually be seen, such as:
sunset, when someone who is fasting can break the fast; the disappearance of the red
twilight glow, when the time for the evening prayer begins; and the first visibility of
the lunar crescent, when the fast of Raman begins.
They also dealt with similar problems like the determination of midday, on which the
determination of all the prayer-times depends, noting that when the sun is in the east
the shadow it casts will be in the west, and that as the suns altitude increases the
length of this shadow decreases, and stops decreasing when the sun culminates; and
that as soon as the sun starts to sink from the meridian, the midday shadow starts to
increase, and this phenomenon marks the beginning of the time of the midday prayer.
If the sanctity of this prayer-time is neglected, and this increase has not taken place,
the prayer is not correct. Now it has become apparent that most muezzins for a long
time have not been taking heed of this. They say: We take heed of what the
professional time-keepers have done, and of what they have prescribed in their treatises
on quadrants and astrolabes; this is what we go by. What they mean is that they
consider the actual time of midday as the beginning of the prayer-time and do not wait
to observe the increase in shadow length prescribed by the legal scholars. It is clear
that the opinion of the muezzins is less correct than that of the legal scholars, and
it is the latter opinion which should be used as a basis for the determination of the
prayer-time.
2.11 A late Syrian treatise on fifteen times of day with religious significance
MS Aleppo Awqf 970, copied ca. 1850, contains a short treatise of a few folios compiled
in Aleppo and deals with 15 times of day with religious significance. This treatise is of
considerable interest for late Ottoman practice (IV-5.4 and 6.1). I have come across no other
treatise in which these 15 times are astronomically defined. The times, which include an
indication of when the performance of the prayer is no longer approved (karha), are as follows:
(1) the uhr;
(2)-(5) the first and second {ar and the times of karha for both {ars;
(6)-(7) sunset and the time of karha for the maghrib prayer;
(8)-(9) the first and second {ish}s;
(10) the imsk;
(11)-(12) the false dawn and the true dawn;
(13) sunrise;
(14) the small awa; and
(15) the large awa.
The author states that the time of karha for the {ar is when the solar altitude is 41/2. The
time of the karha for the maghrib is when the stars become visible (waqt itibk al-nujm),
selected medieval arabic treatises on timekeeping 231
namely, when sun is 10 below the horizon (h = -10). The first and second {ish} occur at
the disappearance of the red and the white twilight glow, respectively, that is, when h = -17
and -19. The false dawn occurs when h = -20, altered in the text to -21, and the true dawn
when h = -19. The imsk is 20 minutes before the true dawn. The small awa, the time before
which prayer is forbidden and the time for the prayer at the {d and the u, is when h =
+41/2. The large awa, which is the time after which the statement of intention to fast is
not valid and after which the u prayer should not (be performed), is halfway between the
true dawn and sunset, when the sun is at the mid-point of the legal arc of daylight (qaws
al-nahr al-shar{).
The treatise includes some remarks about the effect of refraction at the horizon and concludes
with some tables for facilitating the determination of the hour-angle at these times: see further
11.15.
232 part ii, chapter three
CHAPTER 3
On the other hand, creative interest in astronomy continued in Iran and Central Asia, both
under the Timurids and later under the Safavids.5 But the tables for timekeeping that we find
are also few and far between. Nevertheless, there is every hope that such materials will be
uncovered when the rich manuscript collections in Iran are exploited for their scientific
contents.
The anonymous tables in MS Berlin Ahlwardt 5793 (Landberg 56), fols. 93v-95v, copied ca.
1450, are appended to al-Farghns treatise on the construction of the astrolabe (fols. 1r-77r)6
and two treatises on the construction and use of the astrolabe by al-Khwrizm (fols. 77v-93r).7
In view of the nature of these tables, I deem likely that they were compiled in the time of al-
Farghn and al-Khwrizm, namely the early 9th century, and the most probable author is al-
Khwrizm himself.
Firstly, there is a table (fol. 93v) of normed right ascensions, (), with entries to three
digits for = 273, 276, ... , 360, accurately computed but for a few scribal errors for =
23;33, the first Mumtaan value of this parameter.8 The second table (fol. 94r see Fig. 3.1)
displays the shadows to base 12 at the beginning of the midday prayer, and at the beginning
and end of the afternoon prayer, that is:
zz, za and zb .
Entries are given to one digit for each 6 of ( \ 0). The latitude is stated as 13, which
is a scribal error for 33. The values in the table correspond roughly but not precisely to the
standard definitions of later Islamic practice:
zz(12) = Z(12) , za(12) = Z(12) + 12 and zb(12) = Z(12) + 24 .
In particular, it seems that the values of za and zb were computed using functions of H different
from these. The table merits detailed investigation.
In al-Khwrizms treatise on the astrolabe the following definitions are given for the solar
altitude at the uhr and {ar (fol. 89v of the Berlin manuscript):
hz H - 7 and ha 1/2 H + 1/10 .
5 See the survey articles Kennedy, Seljuq and Mongol Science, and Timurid Science, and Winter,
Safavid Science, and on Safavid astronomy see more recently King, Mecca-Centred World-Maps, pp. 128-
138. The standard Western bio-bibliographical reference work on Persian manuscripts is Storey, PL, II:1,
although much can now be added from Iranian reference works, such as those of @gh Buzurg, al-Dhar{a, and
Monzawi, Persian Manuscripts.
6 On al-Farghn see the article in DSB by A. I. Sabra, and Sezgin, GAS, VI, pp. 149-151. On his astrolabe
tables see King & Sams, Islamic Astronomical Handbooks and Tables, Section 4.11. His astrolabe treatise
is currently being edited by Richard Lorch (Munich).
7 On al-Khwrizm (I-7.1.1) see the article in DSB by G. Toomer; Sezgin, GAS, VI, pp. 140-143; and on
various Abbasid works attributed to him, but not all by him, King, al-Khwrizm. One of his treatises, on
the use of the astrolabe, is translated in Frank, al-Khwrizm ber das Astrolab. On the other, on the
construction of the astrolabe, see King, al-Khwrizm, pp. 23-27. An edition of both treatises with English
translation and commentary has been prepared by Franois Charette and Petra Schmidl.
8 See I-5.6.1 and 9.1.
234 part ii, chapter three
Fig. 3.1: al-Khwrizms prayer-tables for Baghdad. Note the sporadic use of Hindu-Arabic numerals in the table:
this is most unusual in tables copied in the Mamluk period. It may indicate that al-Khwrizm used such notation
in his original table. From MS Berlin Ahlwardt 5793 fol. 94r, courtesy of the Deutsche Staatsbibliothek
(Preuischer Kulturbesitz).]
It remains to explain these formula, the first of which is not attested in any other source known
to me, and the second of which is similar to that used by later astronomers in both Egypt and
Syria (see, for example, 2.9).
al-Khwrizms definition of the uhr is intended to correspond to a time one seasonal hour
after midday as defined by shadow lengths using the Indian formula:
T { 6 n / (z + n) } sdh ,
discussed above (1.4): see further IV-4.5.
The second of al-Khwrizms formulae is more a careful approximate solution to the
condition z = n. If we add n = 12 to the midday shadows for Baghdad noted above we obtain:
30;29 19;48 13;54
{ir@q and iranian tables for timekeeping 235
9 On Ibn Amjr (n. I-2:29) see Sezgin, GAS, VI, pp. 177-178, where the table in MS Paris BNF ar. 2486
is listed.
236 part ii, chapter three
2486, fols. 239v-255r, of the Zj of al-Baghdd (copied by the author in 1285), where it is
expressly attributed to the earlier scholar and called al-zj al-aylasn (fol. 239r), the word zj
here apparently meaning a large table rather than an astronomical handbook.
In MS Paris BNF supp. pers. 1488, copied ca. 1500, of the early-14th-century Persian Ashraf
Zj,10 there is a similar table computed for an unspecified locality, actually Shiraz (3.10),
preceded by yet another aylasn table (fols. 201v-204v), of which it is stated: this table was
compiled by Abu l-Qsim al-Mjr in his zj called al-Zj al-aylasn. The title reads istikhraj
al-s{t al-zamniyya min irtif{ al-waqt wa-nif al-nahr, for finding the temporal hours from
the instantaneous and meridian altitudes. The function tabulated is T(H,h) in seasonal hours
and minutes for degree of both arguments such that 1 h H 90. The table is in fact intended
to display the time since rising for any latitude and the tabulated function is based on the
approximate formula:
T(H,h) arc Sin { R Sin h / Sin H } sdh .
See further I-2.5.1 on the table and II-1.4 on the formula, also XI.
Thus Ibn Amjrs work contained at least two extensive tables, one for Baghdad and the
other for all latitudes. I suspect that several of the other spherical astronomical tables in the
Zj of al-Baghdd (3.2) are also due to Ibn Amjr. We are most fortunate to have these
aylasn tables of his, since the other five zjes compiled by him and his sons have vanished
almost without trace. The little evidence we have about the Amjr family ranks them close
to the top of the list of Muslim astronomers.
The unique MS Paris BNF ar. 2486, copied 684 H [= 1285], of the Zj of al-Baghdd (3.2)
contains numerous spherical astronomical tables, but I here restrict comments to those found
on fols. 117r-122v. Several functions are tabulated to four digits for the parameters = 33;25
(Baghdad) and = 23;35, namely:
d, 2Dh, and H .
These are followed by a set of prayer-tables see Fig. 3.3 for the same parameters, with
values given to one or two digits for each day of the Syrian year. The corresponding position
of the sun (mawi{ al-shams) in the zodiacal signs (al-awl) is given (Tishrn I l corresponds
to Libra 16) and the functions tabulated are as follows:
(1) Cot7 H is given to two digits.
(2) A function is labelled ul{ al-fajr, daybreak, is actually the number of seasonal night-
hours and minutes from sunset to daybreak (see below). To find the duration of evening
twilight one should subtract the values from 12 (see below).
(3) Values of functions H, ha and hb are given to one digit.
al-Baghdds twilight table is based on the parameter 17. The function tabulated is:
r() = 12snh - s()
237
238 part ii, chapter three
MS Paris BNF ar. 2514 contains an anonymous set of tables for timekeeping, copied or maybe
compiled in 612 H [= 1215]. The work consists of two parts, for timekeeping by day and night,
respectively. The first set of tables display the time since rising in seasonal hours as a function
of meridian altitude and instantaneous altitude. The underlying formula is approximate and the
entries can be used for all latitudes; however, for meridian altitudes such that - H
+ , the time since rising for the sun is given in equinoctial hours for the latitude 33;30,
which could serve Baghdad or Damascus. For each value of H the value of ha is also given.
On this table see further I-2.5.2.
The remainder of the tables (fols. 28v-48v) display the longitude of the horoscopus and its
oblique ascensions as a function of the altitudes of various stars. See further I-3.2.2. On some
similar tables for latitude 30;25 (Qandahar) see also I-3.2.1.
A pair of tables for the duration of morning and evening twilight was computed by an individual
named {Al ibn {Abdallh al-Qyin who lived in the city of Qayin in Eastern Iran, probably
ca. 1000.11 These tables are preserved in the unique MS Bankipore 2468,23 (fols. 114v-115r),
copied in 632 H [= 1234/35], from which the text was published in Hyderabad in 1947; they
have been studied by Marie-Louise Davidian and Ted Kennedy.12
al-Qyin states in his introduction that he has tabulated the functions r(H) and s(H + 180),
where H() is the longitude of the horoscopus at daybreak in Qayin when the solar longitude
is . He also states that the latitude of Qayin is 33;55, a value not known from other sources,13
and that he is using the parameter 17 for both morning and evening twilight. In the introduction
it is noted that:
r(H) = s(H + 180) = s(H*) ,
and this relation also holds for the tabulated functions, values of which are given in degrees
and minutes for each 5 of argument. However, the tabulated functions do not correspond to
the accurate values of these functions for these parameters. Kennedy and Davidian demonstrated
that the entries corresponded more closely to the lengths of twilight as a function of solar
longitude, but then we are confronted with the fact that if is considered constant for a given
day then r() = s().
To compute the function r(H) using an accurate formula we first determine the arc of the
ecliptic which rises during morning twilight. The simplest formula is suggested by al-Qyin,
namely:
= arc Sin { r Sin 17/ Cos }
where (H) is the function known as {ar iqlm al-ru}ya, the latitude of visible climate, which
measures the angle between the ecliptic and the horizon.14 With this r(H) is given by:
r(H) = (H + ) - (H) .
No tables of () or () specifically for latitude 33;55 are attested in the known manuscript
sources.
Like Kennedy and Davidian I am unable to explain al-Qyins table. I have considered the
possibility that he used a table for the latitude of visible climate which was computed for
the latitudes of one of the climates rather than for Qayin, and also that the quantity was
computed by an approximate formula. But neither of these possibilities seems to account for
the peculiarities of al-Qyins functions.
Abu l-Yusr al-Bazdaw was a anaf legal scholar active in Samarqand in the 11th century.15
He wrote a treatise on the qibla in early Islamic Transoxania which is extant in a unique
manuscript preserved at one time in Sohag, from which MS Cairo B 19385 was copied in 1355
H [= 1936].16 al-Bazdaw promises a table displaying the solar altitude in the azimuth of the
qibla hq which will serve Samarqand, Bukhara and Nasw. The table is not contained in the
Cairo manuscript, but al-Bazdaw mentions three values from it which are the following:
EQ: 40 SS: 66 W: 11 .
For = 40, which al-Bazdaw gives for Samarqand, these correspond roughly to q = 45, which
is a happy compromise between the directions of due west, used for the qibla by the anafs
in Samarqand, and due south, used there by the Shfi{s.17
14 On this concept see Kennedy, Zj Survey, p. 145b. See also nn. 3:29 and 7:15.
15 On al-Bazdaw see Brockelmann, GAL, I, p. 460, and SI, pp. 637-638, and Sezgin, GAS, I, pp. 412-428.
16 Text, translation, and commentary are contained in King, al-Bazdaw on the Qibla in Tranoxania. That
study is based on the Cairo manuscript because the original in Sohag had disappeared when I tried to locate
it in the 1970s.
17 Op. cit., and King, Mecca-Centred World-Maps, p. 125.
240 part ii, chapter three
The anonymous zj preserved in MS Paris BNF ar. 5968, copied ca. 1250, and known as the
Dustr al-munajjimn is based mainly on the zjes of al-Battn, Kushyr, Ab Ja{far al-Khzin
and al-Brn.18 There are two tables for morning and evening twilight in this zj (fols. 187v-
188), which are based on the approximate formulae noted in 3.3. The tables are anonymous,
and someone has written a non-commital li-ba{ihim, by one of the astronomers, by each
of the titles see Fig. 3.7. The underlying parameter is found by inspection to be 16, although
the parameters mentioned in the accompanying text are firstly 18 and then 17 (fols. 187r).
Corresponding entries in the two tables add up to twelve seasonal night-hours: the table for
evening twilight measures the duration of twilight and the other measures the time from sunset
Fig. 3.7: The tables for the duration of twilight in the the Dustr al-munajjimn. The format is the same as that
of the tables in the Cairo corpus, and it may be that there is Egyptian influence behind these Iranian tables.
[From MS Paris BNF ar. 5968. fols. 187v-188r, courtesy of the Bibliothque Nationale de France.]
18This zj is not listed in Kennedy, Zj Survey, but the introduction has been studied in Zimmermann,
Dustr al-munajjimn (cited in n. I-7:14). See also Sezgin, GAS, VI, pp. 63-64.
{ir@q and iranian tables for timekeeping 241
until daybreak. The underlying latitude is as stated, namely, = 36;21. This is the value used
in several Islamic sources for the Ism{l stronghold of Alamut (and also Nishapur), although
it was probably derived by calculation as the latitude of the 4th climate rather than derived from
any observation.19 Elsewhere in the Dustr (fols. 161r-163r) there are tables of the functions:
H, d, D, 2Dh and ,
with values to two digits for parameters = 36;21 and = 23;35. See further XI-5.2.
3.8 The I lkhn aylasn table for Maragha
The lkhn Zj of Nar al-Dn al-s and his collaborators at the Observatory in Maragha
ca. 126020 contains a aylasn table displaying the times since sunrise as a function of the solar
meridian altitude and instantaneous altitude for = 37;20 (Maragha). The table is not contained
in all of the numerous available copies of the Zj since its use is restricted to one latitude, but
it is contained, for example, in MS Florence Medici 269, fols. 150r-152v, and Oxford Hunt.
143, fols. 155v-160v, of the Zj. See further I-2.3.2.
Muyi l-Dn al-Maghrib was one of the astronomers who assisted Nar al-Dn al-s at the
Observatory in Maragha (I-4.3.1* and 5.6.3).21 His own Zj purports to be more representative
of the work done at Maragha than al-ss lkhn Zj, but neither Zj has been fully studied
yet: a brief survey has been made of the lkhn Zj, but the Zj of Muyi l-Din al-Maghrib
has hardly been studied at all. Whilst the lkhn Zj exists in several copies, only three copies
are known of al-Maghribs Zj, namely, MSS Dublin CB 3665, Meshed Shrine Library
332(103), and Medina Aref Hikmet mqt 1.
Certain features of al-Maghribs Zj indicate his debt to Ibn Ynus (5.1). First, his Sine
tables are based on values lifted from the kim Zj. (Likewise, his Tangent table is based
on values lifted from al-Brns Qnn.) Second, other than Ibn Ynus he is the only known
compiler of a zj who tabulates both the solar rising amplitude and the solar altitude in the
azimuth of the qibla (MSS Dublin CB 3665, fols. 96v-97r and Meshed Shrine Library 332(103),
fols. 99r-99v).
al-Maghrib tabulated () to three digits for each degree of , and the latitude of Maragha,
taken as 37;20,30 (I-5.6.3). Ibn Ynus had also tabulated the function to three digits for the
latitude of Cairo (5.1). al-Maghribs table of hq() gives values to two digits for each degree
of , as did Ibn Ynus table for Cairo (see also 5.1). At the head of al-Maghribs table of
hq() in both sources it is stated that the underlying parameter is q = 16;59. In al-Maghribs
19 Kennedy & Kennedy, Islamic Geographical Coordinates, pp. 15 and 245-246; also King, Geography of
Astrolabes, pp. 6-9.
20 On al-s and the lkhn Zj see n. I-2:31. The existence of the lkhn aylasn table was noted in
Kennedy, Zj Survey, p. 161.
21 On Muyi l-Dn al-Maghrib see n. 5:8.
242 part ii, chapter three
geographical tables (MSS Dublin CB 3665, fols. 57v-59r, and Meshed Shrine Library 332(103),
fols. 61r-61v) we find the coordinates:
Mecca : 21;40 L: 77;0
Maragha 37;20,(30) 82;0
Accurately computed for these coordinates q is 16;49. However, in MS Istanbul Ayasofya
2984, fols. 85-86v, of the early-14th-century Persian Zj of al-Wbiknaw,22 the calculation is
presented on which the value 16;59 is based. The value of q which underlies the table of hq(),
however, appears to be 16;49 rather than 16;59 and the value used for is 23;30. al-Maghrib
has the three values: 51;27, 75;38, 27;15 for hq at the equinoxes and summer and winter
solstices. Recomputation with the parameter 16;49 reproduces precisely these values.
Recomputation with the parameter 16;59 yields 51;25, 75;37, 27;13. al-Maghribs table
of hq differs in format from Ibn Ynus table in the Hkim Zj. The latter is arranged in 12
columns of 30 entries for each degree of , so that no advantage is taken of the symmetry of
the function. In the main Cairo corpus hq() is displayed in six columns of 30 entries and no
value is given for the equinoxes. In al-Maghribs table the function is displayed in six columns
of 30 entries beginning with Aries, but the vertical argument runs from 0 to 29 instead of Ibn
Ynus 1 to 30. At the bottom of the third column (for Gemini) and the sixth column (for
Sagittarius) an extra entry is added for argument 30. Thus values are given for the equinoxes
and both solstices. al-Maghribs entries are rather accurately computed.
The Medina manuscript, but not the other two copies of the Zj, contains a table displaying
the solar altitude at each seasonal hour for each degree of solar meridian altitude. This table
is based on an approximate formula see further I-4.3.1*.
MS Paris BNF supp. pers. 1488, copied ca. 1550, is an unique copy of the Ashraf Zj, compiled
in 702 H [= 1302/03] by the Iranian astronomer Sanjar al-Kaml also known as Sayf al-
Munajjim (3.2). It contains two aylasn tables, the first being the universal table of Ibn Amjr
(3.2) and the second being computed for = 29;30 (Shiraz). Various other tables in the Ashraf
Zj relate to spherical astronomy and have been discussed in I-2.3.3, etc. see Fig. I-6.2.1
for some of these. However, there are no tables for regulating the times of prayer.
MS Leiden Or. 199, fols. 21v-27v, contains an anonymous aylasn table of the function t(H,h)
for latitude 36. This table, which contains many errors difficult to attribute to copyists, has
been studied by Bernard R. Goldstein.23 See further I-2.3.4 and, on the structure, XI-4.2
(illustrated).
22 Kennedy, Zj Survey, no. 35 (not in Storey, PL, II:1); and King, Mecca-Centred World-Maps, n. 84 on
p. 158.
23 See Goldstein, Medieval Table for Reckoning Time, the first modern study of a medieval Islamic table
for timekeeping.
{ir@q and iranian tables for timekeeping 243
MS Cambridge Add. 3527 (date?) contains a set of anonymous timekeeping tables for Baghdad,
bound in some confusion. The underlying latitude is = 33;25 (Baghdad). The functions
tabulated for each degree of argument are (fols. 120v-121r), 2Dh (fol. 121v), and (fols.
158v-159). In view of the format of the tables, which have the main vertical argument running
from 0 to 29 rather than the more sensible 1 to 30, I estimate that they are rather late, perhaps
from the 16th century. There is also a table for converting hours and minutes to equatorial
degrees for each minute from 1m to 15h59m (fols. 122v, 157r-158r) and another displaying the
astrological houses for latitude 33;25.24 Two other tables in the set, each copied in different
hands, are spurious, namely:
(1) a table displaying the astrological houses for latitude 32 (fols. 166v-172r); and
(2) a table displaying the time of daybreak for Baghdad according to the Ottoman convention
for each day of the Persian year (fol. 165v) see the next section.
The anomalous table in MS Cambridge Add. 3527 (3.12), fol. 165v, displays the duration of
twilight (s{t ub wa-shafaq) for the latitude of Baghdad, stated to be 33;25. Values are
given in equinoctial hours and minutes for each day of the Persian year starting with the month
of Farwadn and ending with the five intercalary days. The maximum value of 1;41h corres-
ponds nicely to a solar depression of 17 and the minimum of 1;17h equally well to 16, so
the table invites further investigation.
MS Cairo FF 14, copied ca. 1700, is the only copy known to me of an astronomical work
entitled Tuhfa-yi Sulaymani compiled in 1078 H [= 1667/68] by Muammad Zamn ibn Sharaf
al-Dn usayn al-Mashhad, well-known as a maker of astronomical instruments.25 The treatise
contains some tables of various spherical astronomical functions (fols. 103r-105v) such as:
(), (), and ()
based on the parameter = 37;0 (Meshed), a standard value.26 There is also a table displaying
the time of culmination of the star al-kaff al-khab (= Cas) as a function of solar longitude
(fol. 42v) see further I-2.8.2. Two prayer-tables concern the present study. They are entitled
jadwal-i irtif{-i aftb dar samt-i qibla-yi Mashad-i muqaddas bi-taqwim-i shams birikrand
(fol. 102v) see Fig. 3.14 and jadwal-i s{t m bayn ul{-i ub wa shams wa ghurb-
i shams wa maghb-i shafaq (fol. 111r) and display the functions:
hq() and rh() or sh() .
Fig. 3.14: Muammad Zamns table of the solar altitude in the azimuth of the qibla at Meshed, followed by
the first and second tables in a set for each of the zodiacal signs for the equalisation of the astrological houses.
[From MS Cairo FF 14, fols. 102v-103r, courtesy of the Egyptian National Library.]
The format is the same as that of the twilight tables in the Cairo corpus (4.10). The underlying
latitude here is stated on other tables to be 37. The qibla used by Muammad Zamn for
Meshed in the gazetteers engraved on one of his astrolabes is 45;23 (accurately for his
coordinates 45;9); on another he has 55;28(!), and on another no value at all (!).27
MS Istanbul Topkap B 411, an encyclopedia compiled ca. 1413 for Isakandar Suln ibn {Umar
Shaykh of Isfahan, contains an extensive section on astronomy with numerous tables and
illustrated constellation figures.28 Amidst the tables is a set of prayer-tables for Isfahan. Values
of the following functions are displayed to two digits for each degree of ( \ 0):
d, , d, D, , 2Dh, H, Z(12), Z(7), ha, za(12), za(7), tah, ta, 12/2D, hb, zb(12), (2N-r)h and rh .
The underlying parameters are:
= 32;25 and = 23;30 .
Although no locality is specified, elswhere in this compilation there are tables of the function
called {ar-i iqlm-i ru}ya, latitude of visible climate29 for:
= 29;30 (Shiraz) and 32;25 (Isfahan).
The times defined by ha and hb are called awwal al-{ar and ajz}-yi ikhtiyrt, the latter
suggesting an astrological association but simply an unfortunate error for khir al-ikhtiyr, the
end of the most favourable time. The names of all the functions tabulated are given in Arabic
although the remainder of the astronomical section is in Persian.
27 Ibid., p. 504.
28 See ibid., pp. 143-145, on the world-map in this manuscript.
29 See n. 3:14.
{ir@q and iranian tables for timekeeping 245
The celebrated Iranian astrolabist {Abd al-A}imma, who worked in Isfahan about 1700 and is
best known for his elegant astrolabes,30 engraved a set of shadow tables on a sundial and qibla-
indicator which he made and which is now preserved in the Time Museum, Rockford, Illinois.
Another unsigned instrument of the same kind is in the Museum of the History of Science,
Oxford: see Fig. 3.16.31 The instrument is fitted with a pin-gnomon at the centre a set of
Fig. 3.16: The prayer-tables on the compass cum qibla-indicator attributable to {Abd al-A}imma. [Photo courtesy
of Museum of the History of Science, Oxford.]
30 On {Abd al-A}imma see Mayer, Islamic Astrolabists, pp. 23-26, and supplement, p. 293; Gingerich & King
& Saliba, The {Abd al-A}imma Astrolabe Forgeries; and also King, Mecca-Centred World-Maps, pp. 175-177.
31 On these two instruments see ibid., pp. 118-121. On a fake see Khalili Collection Catalogue, II, p. 274
(no. 162).
246 part ii, chapter three
quadrants of a circle radii 1, 2, ... , 40 where 12 is the length of the gnomon. These curves
can be used to regulate the uhr and {ar using the tables, which display values of the functions:
zz (= Z), za and Ta (in hours)
to two sexagimals for each zodiacal sign. The underlying latitude is = 32 (Isfahan).32 Given
the paucity of Safavid manuscript sources on timekeeping these tables are of particular
historical interest. It is unlikely that {Abd al-A}imma computed them himself, for, as far as
we know, neither he nor any of his colleagues astronomers and instrument-makers from
17th-century Isfahan were not into compiling tables.33
I recall seeing a set of such tables in a manuscript at the Oriental Institute in Tashkent. My
notes thereon have disappeared.
32 On the qibla values for various cities also engraved on these instruments see ibid., pp. 118-121, 177, 518-
519, and 545.
33 Ibid., pp. 128-138.
the main cairo corpus 247
CHAPTER 4
The Cairo corpus exists in a number of a manuscript sources, usually containing tables from
three main categories:
(1) tables of solar azimuth as a function of solar altitude and longitude;
(2) tables of time as a function of solar altitude and longitude;
(3) tables of standard spherical astronomical functions of solar longitude and various
other functions for regulating the prayer-times.
All of these tables are computed for the latitude of Cairo taken as 30;0 and obliquity 23;35.
No two manuscripts have been found yet containing precisely the same tables.
In the first category the function tabulated is a(h,), and the tables contain over 10,000
entries. In the second the functions tabulated are either t(h,) or T(h,) or both. These are closely
related and the computation of one given the other is trivial, since their sum is D(). Each set
contains over 10,000 entries. The third category consists of a total of about 30 tables, each
containing 90 or 180 entries. Most of these can be derived very easily from basic tables of
() (as is the case with, say, H, whence za and ha can be found), or d() (as is the case with,
say, D, 2N, , 1/, ), or directly from the tables of T(h,) (as in the case with r and s). Likewise,
given ha or hq, it is not difficult to find the corresponding hour-angles and time since sunrise,
ta and Ta, and tq and Tq, using the tables of t(h,) and T(h,) or D().
Most of the sources which I used in my first study of the Cairo corpus1 mention the 10th-
century astronomer Ibn Ynus as the compiler, but several sources identified thereafter include
instructions by the the late-13th-century astronomer al-Maqs and others contain some notes
by the mid-14th-century astronomer al-Bakhniq. In numerous late manuscripts in which only
the tables of standard spherical astronomical functions and the prayer-tables are found, these
tables are attributed to the 16th-century astronomers al-Ldhiq and al-Ikh (7.8).
The location in the manuscripts of the various categories of main tables, as well as the authors
to whom they are attributed, is shown below (see already I-2.1.1). (The tables in MS Cairo
TR 354 falsely attributed to Ibn Ynus are computed for Alexandria see 8.5.) The notation
T/t means that T and t are tabulated on facing pages, and the notation T/t/a means that T, t
and a are tabulated together, with triplets of values for each pair of arguments. Otherwise the
functions listed are tabulated separately. An asterisk indicates that tables are preceded by an
introduction on their use prepared by al-Maqs (5.4), and a double asterisk indicates that this
introduction is accompanied by some notes by al-Bakhniq describing the way in which he
had rearranged the tables (see further 5.6).
1 See King, Astronomical Timekeeping in Medieval Cairo, cited in n. I-2:2. That study was based mainly
on MSS Dublin CB 3673 and Cairo DM 108, also Berlin Ahlwardt 5753, Cairo Azhar falak 4382, Cairo TR
191, and Escorial r. 924,7.
248 part ii, chapter four
Having already analyzed the main Cairo corpus from a mathematical point of view in my
previous study, I propose to consider each of the tables afresh (4.3 to 4.9), paying particular
attention to the question of authorship. I begin with a few brief remarks (in 4.1) about each
of the individuals known to have contributed to the corpus. I then (in 4.2) present brief
descriptions of several new manuscripts of the corpus to supplement those in my original
analysis.
My main purpose in 4.3-9 is to describe the tables in two manuscripts of the corpus, MS
Dublin CB 3673, copied in the 14th century, and MS Cairo DM 108, copied at the beginning
of the 19th century, in the light of the several dozen related manuscripts that I have located.
In 4.3 I discuss the tables of standard functions for spherical astronomy. In 4.4 I discuss
the tables of a(h,) and show that they were indeed compiled by Ibn Ynus. Certain manuscripts
the main cairo corpus 249
of the azimuth tables contain some nonsensical values added by an incompetent anonymous
and in others these have been corrected by the early-14th-century astronomer Ibn al-Rashd.
In 4.5 I discuss the tables of t(h,) and T(h,) and present the evidence which makes it clear
that these tables in their present form were not compiled by Ibn Ynus. Three centuries after
his time, al-Maqs claimed to have prepared the tables of T(h,), but I suspect that he had at
his disposal some tables for timekeeping by Ibn Ynus which have not survived in their original
form. In the 14th century Ibn al-Kattn prepared a set of tables of t(h,) using al-Maqss tables
of T(h,). In 4.6-11 I attempt to cast some light on the highly complicated problem of the
authorship of the minor tables in the corpus. Most of the mathematically-significant tables were
either computed by Ibn Ynus or owe their inspiration to him. The development of the main
Cairo corpus from the 10th to the 14th century is then discussed in Ch. 5.
4.1.2 al-Maqs
Shihb al-Dn Amad ibn {Umar al-Maqs was one of the leading astronomers in Cairo about
1275.5 In my previous study of the Cairo corpus, on the basis of my examination of the set
of tables in MS Cairo DM 53, introduced in al-Maqss name, I suggested that he had merely
plagiarized some of Ibn Ynus tables. I have since been able to consult other such sets, for
example, MSS Gotha A1402, Istanbul Nuruosmaniye 2925 and Cairo DM 616, and al-Maqss
Fig. 4.1.1: A miniature showing a rather angry-looking Caliph al-kim apparently giving Ibn Ynus a hard
time. The astronomer is holding his zj, marked with Hindu-Arabic numerals on the spine. Actually the kim
Zj would have filled several volumes of this size. [From MS Istanbul Topkap Revan Kku 1638 (400 folios)
of al-Qnn fi l-duny, an astrological work of the prolific Ibn Zunbul (~hsanolu et al., Ottoman Astronomical
Literature, I, pp. 183-184, no. 84); photo courtesy of the Topkap Library.]
contribution is reconsidered in 5.4 below. I have also studied the only other known work
attributed to al-Maqs, a treatise on sundial theory extant in MSS Cairo Azhar falak 5528,
Dublin CB 4090 and Cairo DM 103. This consists mainly of tables, which are to be used for
marking shadow traces and curves for the seasonal hours on sundials erected in various planes.6
There are over 100 tables each containing about 30 entries and this illustrates that al-Maqs
was not afraid of extensive calculation. These tables are related to the more accurate but less
extensive ones of his contemporary Ab {Al al-Marrkush (6.7) although neither individual
refers to the other. The treatment of the standard problems of spherical astronomy in al-Maqss
introduction to his sundial tables is not impressive, and it is of interest that in this introduction
al-Maqs used some material from the kim Zj without mentioning his illustrious predecessor.
6 On al-Maqss sundial tables see King, Islamic Astronomical Tables, pp. 51-53, and idem, Astronomy
of the Mamluks, pp. 547-548, and also I-4.1.3 and X-7.1.
the main cairo corpus 251
4.1.5 al-Bakhniq
Shams al-Dn Amad ibn Muammad al-Bakhniq worked in Cairo and also in the Yemen
in the 14th century.9 He was at some time associated with the Azhar Mosque since he is also
called al-Azhar. MS Istanbul Topkap A3343 of the summa of al-Marrkush (2.7 and 6.7)
bears his mark of ownership dated 776 H [= 1374/75]. However, he must have been active
for rather a long time because one of his works was compiled for a Yemeni wazr and can
be dated ca. 1325. This particular compilation, extant only in MS Dublin CB 4092, was an
extension of the tables of the early 9th-century Baghdad astronomer al-Farghn for constructing
the various markings on astrolabe plates.10 al-Farghns table enabled the user to construct
7 On Ibn al-Rashd (I-2.1.5), whose full name was Shams al-Dn Ab {Abdallh Muammad ibn Burhn
al-Dn Ibrhm al-Rashd, see n. I-2:18.
8 On Ibn al-Kattn see n. I-2:6.
9 On al-Bakhniq see n. I-2:5.
10 On al-Farghn and his astrolabe tables see the article in DSB by A. I. Sabra and King, Islamic
252 part ii, chapter four
the altitude and azimuth circles on plates for each degree of latitude from 15 to 50 and are
easily derived from a single very useful little table of an auxiliary trigonometric function; al-
Bakhniq simply completed the set for each degree of latitude from 0 to 90. Another work
by al-Bakhniq is a treatise on the prayer-times in which he outlines a set of approximations
for the standard functions of timekeeping: I have summarized this in 2.9.
al-Bakhniq was responsible for rearranging the tables in the Cairo corpus so that triplets
of values (T,t,a) are displayed side by side as functions of (h,). I assess his other contributions
to the corpus in 5.6.
The manuscripts in which the main tables of the Cairo corpus are to be found date from the
14th to the 19th century.11 No two copies contain the same tables. No single manuscript contains
reliable information on the identity of the compiler of the individual tables. In my previous
analysis I relied on MSS Berlin Ahlwardt 5753, Cairo TR 191, Dublin CB 3673, Escorial r.
924,7, Cairo DM 108, Cairo Azhar falak 4382 and Cairo MM 137; all of these happened to
be preserved in collections which had been catalogued and all happened to be attributed to
Ibn Ynus, which explains how I located them first and why I had no problem attributing all
of the tables to him. I also noted the existence of MS Cairo DM 53, in which part of the corpus
is attributed to al-Maqs, of al-Bakhniqs notes in MS Cairo DM 108, and of various copies
of al-Ldhiqs prayer-tables, but I assumed that each of these authors had merely plagiarized
tables of Ibn Ynus. The new sources listed below help explain more precisely the development
of the corpus between the 10th and the 14th centuries and although my doubts about the integrity
of al-Maqs, al-Bakhniq and al-Ldhiq still appears to be correct (5.4, 5.6 and 7.8), one
should bear in mind that in the world of the Cairo astronomers in the 13th and 14th centuries
it must have been difficult to simply plagiarize tables.12 The discovery of other early copies
of the corpus might well cast new light on its development and on the identity of the compilers
of some of the tables, where this is still in doubt.
MS Berlin Ahlwardt 5753 (Landberg 574):13
This is a disordered fragment of the corpus copied ca. 1400. The title on fol. 1r reads Kitb
fhi l-Samt li-bn Ynus al-sib, Azimuth Tables by Ibn Ynus, the astronomer (see Fig.
4.2a). There are several notices of possession on the title folio, one dated 840 H [= 1436].
The tables on fols. 1v-20r are in considerable confusion. Each is headed only by the degree
of solar altitude. For altitudes 1 through 19 (fols. 1v-10v) the function tabulated is a(h,);
Astronomical Tables, pp. 53-55; idem & Sams, Islamic Astronomical Handbooks and Tables, pp. 91-92;
and X-4.2.
11 On the Cairo manuscripts see Cairo ENL Catalogue (in Arabic) I, in which they are listed by accession
numbers, and II, especially under 3.1.2 (azimuth tables of Ibn Ynus), 3.1.2 (tables of al-Maqs), and 3.1.10
(edition of al-Bakhniq); as well as Cairo ENL Survey (in English), no. B59/3.1.1 (Ibn Ynus), C15/3.1.2 (al-
Maqs), and C28 (al-Bakhniq). For Istanbul manuscripts some catalogues have been prepared after I worked
on the manuscripts, but I have not seen them. For manuscripts from other catalogued collections references to
the catalogues are given ad loc.
12 See also nn. 4:31 and 10:2.
13 Berlin Catalogue, pp. 206-207.
the main cairo corpus 253
Fig. 4.2a: The title folio of the Berlin manuscript, Fig. 4.2b: The title-folio of the Dublin manuscript. The
in which the azimuth tables are attributed to Ibn decorated title, illegible in this photograph, announces
Ynus. The tables that follow are, however, a only the azimuth tables of Ibn Ynus, and the double
disordered mixture of tables of a (which correspond tables of the time since sunrise and time before midday
to the title), then t and T as functions of h and . which follow are in fact in a different hand. Compare
See also Figs. II-5a-b. This was the only manuscript Figs. I-2.1.1a and I-5.1.1a. This did not stop me in the
of the corpus known to Carl Schoy when he worked early 1970s from blithely assuming that the whole lot
on Ibn Ynus in the early years of the 20th century. were by Ibn Ynus. I had ordered a microfilm of the
[From MS Berlin Ahlwardt 5753, fol. 1r, courtesy manuscript because Bernard Goldstein had told me it
of the Deutsche Staats-bibliothek, Berlin.] had been catalogued as a new copy of the kim
Zj. [From MS Dublin CB 3673, fol. 121r, courtesy of
the Chester Beatty Library.]
for 8 and 9 (fols. 11r-11v) the function is t(h,); and for 52 to 83 (fols. 12v-20r) we are
dealing with T(h,)! On fol. 20v there is a table of ta(), which was published by Carl Schoy,14
and on fol. 21r, a table of 1/(). An additional folio at the end of the manuscript contains
an illegible title and some instructions on a calculation in spherical astronomy.
MS Cairo TR 191:
This is a copy of part of the corpus written in 1215 H [= 1800/01] by {Abd al-Bri} al-{Ashmw
(also the copyist of MS Cairo DM 108 see below). The title reads Jadwil fal al-d}ir min
qibal al-irtif{ isb ... Ibn Ynus, and most of the manuscript consists of a complete set of
tables of t(h,). These are followed by tables of the functions:
H, s, n, Tq, za(12), r, 2N, Z(12), tq, , , , s and r .
The twilight tables are based on parameters 19 and 17, the qibla tables on q = 53 and the
salm is assumed to be 1 before daybreak. The manuscript concludes with an undated cata-
logue of the equatorial coordinates of 60 stars, copied by al-{Ashmw in 1217 H [= 1802/
03].
MS Dublin CB 3673:15
This manuscript (annoyingly foliated 121r-1v as for a Latin manuscript) is a virtually complete
copy of the Cairo corpus, incorrectly catalogued as the kim Zj. The title on fol. 121r reads:
Kitb al-Samt li-bn Ynus, but this refers only to the first part of the tables: see Fig. 4.2b.
This folio also contains some notices of possession, a floriated ughr signature, and the date
of copying, 772 H [= 1371]. The tables are copied in the distinctive elegant hand of Muhammad
ibn Muhammad ibn Muammad ibn {Abd al-Qaw al-Qurash known as Ibn al-Kattn see
the notes to MS Cairo MM 72 below. This part of the manuscript contains a complete set of
tables of a(h,), with additional entries for altitudes 84 to 88 (fols. 121v-96r) (see Fig. 5.1.1a),
and these are followed by tables of
hq (q = 52), h0, d and hv .
(The entries in the table of hv decrease to 0;41 at = 270. If Ibn al-Rashd compiled this
table, why is it presented here together with the uncorrected azimuth tables that al-Bakhniq
states Ibn al-Rashd had corrected?)
The remaining part of this manuscript contains tables of T(h,) and t(h,) (fols. 82v-15r)
and various other functions for timekeeping and the prayers, also copied in an elegant hand,
but not that of Ibn al-Kattn: see Figs. 2.1.1a and II-4.7a. This part of the manuscript has
no separate title page but begins with the instructions on the use of the tables (fols. 83r-83v),
in which the name of al-Maqs has been altered to that of Ibn Ynus. The minor functions
tabulated (fols. 15v-3r) are:
ta, d, H, hq (q=53), Ta, , , h0, D, ta, , r (hr=20), s (hs=16),
ha=60, ha=30, tq (q=53), arc Sin (x), r (hr=19), s (hs=17), 2N, n, 2N, r (hr=20),
d, H, , , h0, 1/
Fol. 82r has been filled with odd items such as notices of possession, verses of poetry, and
a list of Graeco-Coptic numerals, given with their Arabic abjad equivalents.16 The manuscript
ends with some astrological writings in a different hand (fols. 2r-2v).
MS Escorial r. 924,7:17
This is an incomplete copy of Ibn Ynus azimuth tables dating from the 14th (?) century. The
title on fol. 30r reads: Kitb al-Samt li-bn Ynus wa-Kitb al-ill mall daqqa daqqa,
Azimuth Tables of Ibn Ynus, and Cotangent Tables computed for each Minute of Argument.
Below this is the curious note: f {ilm al-rumz li-fiqh al-majma{, on the science of secret signs
for the under-standing of the community (?). Fols. 30v-63v contain azimuth tables for solar
15 Dublin CB Arabic Catalogue, p. 73. The date of copying is misread as 672 H.
16 On these see, for example, Ritter, Griechisch-koptische Ziffern. See also n. 4:21.
17 Escorial Catalogue, pp. 32-33.
the main cairo corpus 255
altitudes from 1 to 78. The remainder of the azimuth tables, and the Cotangent tables, are
missing from the manuscript.
MS Cairo DM 108:
This is a fairly complete copy of the main Cairo corpus as edited by al-Bakhniq. It was copied
by {Abd al-Bri} (ibn) Nar al-{Ashmawi (fol. 74v) in 1218 H [= 1803] (see also MS Cairo
TR 191 above) (see Fig. II-4.11a). The title folio gives the title as Kitb Ghyat al-intif{ f
ma{rifat al-d}ir wa-falihi wa-l-samt min qibal al-irtif{ and identifies the author as Ibn
Ynus. The manuscript contains al-Maqss introduction followed by al-Bakhniqs notes (see
Fig. II-5.6c), a complete set of tables of the functions (T,t,a) (see Fig. II-5.6b), followed by
various prayer-tables (see below), as well as some calendrical tables tables not found in the
other main copies of the corpus. The latter part of al-Maqss introductions, al-Bakhniqs
notes, and the first five folios of tables are in the older and more elegant hand of the copyist
of MS Cairo DM 777 (see below). The parameters for twilight in this fragment are 20 and
16.
Following the main set of tables are some calendrical tables for tables for finding the date
in the Hijra calendar for the years 1521-1680 in the Coptic calendar [= 1219-1383 H], and
tables of the following functions:
H, D, ta, ha, Ta, hq, tq, hq*, 2N, n, s, r, , D, , 2Dh, , and tq* .
The twilight tables are based on parameters 19 and 17, the qibla tables on q = 53 and q*
= 37, and the salm is taken to be 1 before daybreak. See Fig. II-4.11a.
MS Cairo Azhar falak 4382:
This is a carefully-copied manuscript of tables of t(h,) (fols. 1v-35r) and a(h,) (fols. 36v-
70v), and also of a(h) for certain altitudes at the equinoxes and solstices (fol. 71r) and ()
together with h0() (fol. 71v). All of the tables are in the same 14th-century hand. The hour-
angle tables are preceded by a page on which the title Kitb Fal al-d}ir is mentioned but
no author is identified. The azimuth tables, which contain entries for arguments h = 1 to 83,
are preceded by a page on which the title Kitb al-Samt is mentioned and Ibn Ynus named
as author.
MS Cairo MM 137:
This manuscript, which was copied about 1500, contains a complete set of Ibn Ynus azimuth
tables bearing the title Kitb al-Samt. The altitude argument runs beyond 83 to 86 (4.4).
Information about the solar altitude for which the azimuth changes direction is given in the
margins of the sub-tables.
MS Dublin CB 4078 (new):
This is a complete set of carefully-copied tables of T(h,) and t(h,) for Cairo dating from ca.
1450. The title appears to read Kitb al-D}ir wa-faluhu, and there is no mention of the
compiler or date of copying.
MS Gotha A 1410 (new):18
This consists mainly of a complete set of Ibn Ynus azimuth tables with entries for altitudes
up to 89. The copyist was {Al ibn Muammad ibn {Al al-Dalm19 and the date of completion
841 H [= 1437/38]. An undated notice of possession names the owner as Jall al-Dn
Muammad al-Raml, muwaqqit at the the Umayyad Mosque (in Damascus). There are
additional tables of hq (q = 53) and hv (with entry 0;41 for = 270). The remaining tables
were originally from a different manuscript and consist of an incomplete anonymous set of
double-argument lunar equation tables, attributed elsewhere to Ibn Ynus.20
MS Cairo MM 184 (new):
This contains tables (fols. 7v-10r, copied ca. 1700) of the following functions:
hq, , hv, B and C ,
as well as a table of the difference between sunrise and sunset at Mecca and Cairo, such as
is contained in MS Cairo MM 204. The table of hq is based on q = 53 and the entry for hv
for = 270 is 0;41.
MS Cairo DM 190,2 (new):
The tables (fols. 25r-33r) listed below were copied ca. 1550 and begin with a page of notes
on the values of D for each sign (max. 62, min. 32), the (Coptic) numbers for the warrqn,21
and a statement that the sun moves 18 miles on its sphere each minute. The tables display the
following functions:
D, ta, ha, Ta, s, n, r, , h0, s, , and .
MS Cairo DM 954,2:
This manuscript, copied ca. 1475, contains an incomplete set of tables from the corpus (fols.
2r-10v) with no accompanying title-page. The functions tabulated are:
tq, hq*, tq*, , , hv, Z(12), ha=60, , za(12), ha=30,
, (D+s), , d, H, 2N, and hq .
The tables relating to the qibla are based on q = 53 and the table of hv has 0;41 at = 270.
MS Cairo DM 53 (new):
This contains a complete set of tables of T(h,), t(h,) and a(h,) with a triplet of entries for
each pair of arguments. The colophon indicates that the copyist was the Azhar muwaqqit {Abd
al-Fatt ibn Muammad al-Dulaj and that the manuscript was completed in the Nile Valley
town of Jirj al-{ara (Girga) in 1117 H [= 1705]. The first folio bears the title Kitb al-D}ir
wa-faluhu wa-l-samt and the tables are preceded by al-Maqss introduction and al-
Bakhniqs notes. In the introduction the parameters advocated for twilight are 19 and 17.
At the end of the manuscript are some absurd instructions on the use of the tables for latitudes
other than 30.
MS Cairo DM 45:
This contains a collection of spherical astronomical tables in different hands and in considerable
disorder. The various parts of the manuscript date from about 1700. The first function tabulated,
as indicated on the title folio, is the oblique ascensions of the ascendant at the beginning of
19 On al-Dalm and other manuscripts copied by him see Cairo ENL Catalogue, I, p. 704.
20 See King, Lunar Equation Table (n. 4:4), p. 132.
21 See already n. 4:16.
the main cairo corpus 257
the afternoon prayer according to the new observation, i.e. based on obliquity 23;30 (7.1).
The tables which follow are for finding the entry of the sun into each zodiacal sign and are
due to Riwn Efend, as stated (7.10). The next tables display the functions (T,t,a) for =
30, a triplet of entries being given for each pair of arguments (h,) but only for altitude
arguments 21 to 57. These are followed by tables of H for = 30 ( = 23;35), as well
as for eight different values of between 32 and 45 ( = 23;30). Of particular interest is
an odd folio of tables (fol. 27) of an unspecified function f(h,) tabulated for arguments h =
24 and 25 and each degree of solar longitude: the function is in fact t(h,) for = 32 (6.7).
The manuscript also contains an incomplete sexagesimal multiplication table.
MS Cairo MM 204:
This manuscript, copied in 1052 H [= 1642/43], contains an unusual set of tables displaying
the following functions (fols. 80v-81r and 86r):
Tan5 , Tan5 (), G(), ha(H) and Cot12 h
for each degree of argument. The table of G is labelled table for finding the hour-angle with
ease: see further I-6.9.3 (illustrated). Another table in this set displays the difference between
sunrise or sunset at Mecca and Cairo for each degree of , such as is also found in MS Cairo
MM 184 (see above).
MS Cairo MM 58,2:
This was copied 1450 by Ibn Abi l-Fat al-f and contains tables (fols. 2v-12v) of the
following functions:
(=31, Alexandria), , , s, n, ha,
() ( = 0;15), hv, ha(H), and 2Dh ( = 0;6) ,
as well as some astrological tables. The table of hv has 0;41 at = 270.
MS Gotha A1402 (new):21a
This manuscript (date ?) bears the title Risla fi l-falak li-l-f, Treatise on astronomy by
al-f (that is, al-Maqs), and contains a complete set of tables of T(h,) preceded by an
introduction introduced in the name of al-Maqs (written al-Maqdis, as if he or his forbears
were from Jerusalem, a curious mistake for a Cairene copyist, who must have known the suburb
of al-Maqs) and followed by an incomplete and disordered set of various other tables for
timekeeping. These additional tables are numbered and are bound in the manuscript now as
follows:
1: ta; 2: D; 3: Tq; 4: 1/; 5: hq; 23: Tan ;
24: Tan60 ; 9/10: d and ; 11: hq (= no. 5); 21: Vers ; and 22: Sin .
It is clear that there were originally some two dozen additional tables. The first four appear
to be original to al-Maqss set (5.4). All of the qibla tables are based on q = 53. The table
of () with values to three digits for each degree of is based on that of Ibn Ynus in the
kim Zj. The table of Tan (), also giving values to three digits, is not identical to that
of al-Marrkush, but the entry 26;11,41 for = 90 is used by al-Maqs in his treatise on
sundials.22 The three trigonometric tables give values to three digits for each degree of arc.
In the introduction the parameters advocated for twilight are 20 and 16.
21a Gotha Catalogue, III, pp. 56-57.
22 See n. 4:6.
258 part ii, chapter four
MS Cairo MM 64:
A beautifully-executed copy of Ibn Ynus azimuth tables, dating from about 1400. The title
folio is missing, and there is no colophon. The manuscript begins with the last page of a set
of tables displaying the solar longitude for each day of a four-year period in the Coptic calendar.
For each month of the Coptic year four columns display the solar longitude in signs, degrees
and minutes for each of the four years, labelled sana rub{, sana nif, sana nif wa-rub{, sana
kabs. Only the tables for Ba}na, Abb and Misr are contained in the manuscript now. The
entry for Ba}na 1 (first year) is Gemini 12;16. The altitude argument in the azimuth tables
runs up to 89 and there is a marginal note which attempts to rationalize this (4.4). For each
altitude above 36 the meridian altitude corresponding to the last longitude argument for which
there is an entry for the azimuth is given to three digits and is called ghyat irtif{ htayn al-
darajatayn ru}s al-jadwal, the meridian altitude for these two solar longitudes (whose signs
are given) at the head of the table.
MS Cairo DM 444:
This manuscript contains a complete set of tables of T(h,) copied in 858 H [= 1454] in the
elegant hand of Abu l-Yumn Muammad ibn Muammad ibn {Arab al-Shabb (or al-
Shayyib).25 It bears a notice of possession (fol. 3r) by the Cairo muwaqqit {Abd al-Ramn
al-ln26 dated 1046 H [= 1636/37]. The title is given as Kitb Ghyat al-intif{ f ma{rifat
al-d}ir min al-falak min qibal al-irtif{ li-{ar lm shaml li-l-Maqs, and the tables are
preceded by al-Maqss introduction (with parameters 20 and 16 advocated for twilight). al-
Maqss four additional tables are not contained in this manuscript; indeed the tables of T
conclude with the words: tamma wa-kamala wa-l-llh al-amd, it is completed and fulfilled
and praise be to God. The tables of T(h,) end on the recto side of the last folio and the verso
is blank except for a notice of sale dated 997 H [= 1588/89]. The tables for 17 and 19 bear
instructions for finding the duration of twilight. See Figs. I-2.1.1b-c and II-5.6a.
MS Cairo DM 690:
This consists mainly of two incomplete sets of tables of (T,t,a) bound in disorder. The
handwriting in both cases is neat and can be dated to about 1500. The last folio of tables (for
altitude arguments 40-42) is copied in the later hand of asan al-ablw ca. 1700.27 There
is no title folio. The tables for altitudes 19 and 17 bear instructions for finding the duration
of morning and evening twilight.
MS Cairo DM 777:
This contains an incomplete set of tables of T(h,) and t(h,) on facing pages, for values of
h from 2 to 46 and then 76 to 83. There is no title folio and no compiler is mentioned.
The tables are copied in the elegant hand of the copyist of MS Cairo MM 241 of Ibn al-
Mushrifs auxiliary tables (6.9) and can thus be dated ca. 1500; see also the notes on MS Cairo
DM 616 below.
25 On other manuscripts copied by Abu l-Yumn see Cairo ENL Catalogue, I, p. 714.
26 On {Abd al-Ramn al-ln see Cairo ENL Survey, no. D16, and on his library see V-10.
27 Close to 20 scientific manuscripts copied by him ca. 1700 are listed in Cairo ENL Catalogue, I, p. 688
(also n. 4:29). See also Cairo ENL Survey, no. D56, on some tables for constructing sundials apparently computed
by him, as well as ~hsanolu et al., Ottoman Astronomical Literature, I, p. 347, no. 212.
260 part ii, chapter four
MS Cairo DM 616:
This consists of four parts, the first two forming a complete set of tables of (T,t,a) preceded
by al-Maqss introduction and al-Bakhniqs notes. The title page and tables for altitude
arguments 7 and above are written in the same hand and dated (10)98 H [= 1686/87]. The
title is given as Kitb Kifyat al-waqt f ma{rifat al-d}ir wa-falihi wa-l-samt and is attributed
to Abu l-{Abbs Amad ibn {Umar al-f, that is, al-Maqs. Part of the introduction and the
tables for altitude arguments 1 to 7 (fols. 2r-9v) are in an older hand datable to about 1500,
which is in fact the same hand as that of MSS Cairo MM 241 and Cairo DM 777 (see above).
The third part of the manuscript, copied about 1600], consists of a set of prayer-tables
displaying values of the functions:
H, D, ha, ta, Ta, d, n, 2N, , , s and r ,
side by side for each degree of starting with = 1. The fourth part of the manuscript contains
some tables for finding the longitude of the horoscopus and upper midheaven from the oblique
ascensions of the horoscopus.
MS Cairo DM 651,5 (new):
This source, copied ca. 1700, contains (fols. 59r-64v) a set of tables of various functions from
the corpus tabulated together, namely:
H, D and N, 2D and 2N, ta and Ta, s and r, ha and za(12) ,
as well as tables of and and a table of geographical coordinates.
MS Cairo DM 739:
This is in the hand of the copyist of the title folio of MS Cairo DM 616 (see above) and can
hence be dated to about 1685. It bears the title Kitb Kifyat al-waqt f ma{rifat al-d}ir wa-
falihi wa-l-samt, attributed to Abu l-{Abbas Amad ibn {Umar al-f, that is, al-Maqs. It
contains al-Maqss introduction followed by al-Bakhniqs notes and a complete set of tables
of the functions (T,t,a). These tables end on the recto side of the last folio (fol. 64). On the
verso there are displayed values of a(h,) only, for altitude arguments 84 to 88, and values
of a(h) at the solstices.
MS Cairo DM 776, fols. 1v and 10r-43r:
This contains a complete set of tables of T(h,), apparently copied in 981 H [= 1573/74]. The
title folio, which is in the hand of an owner Muhammad Sinr / Sannr (?) al-Amad,28 gives
this date of copying and the title as al-D}ir li-rtif{ al-shams. No compiler is mentioned and
there is no introduction. The tables for arguments 16 and 20 are marked as suitable for twilight
by the original copyist, and are marked b--l for baal(a), is no good, in a different hand.
The tables for 17 and 19 are marked in another different hand as suitable for twilight
according to the new observations, a phrase which usually refers to the Samarqand observa-
tions of Ulugh Beg, but here may refer to the opinion of the Cairo editor of the Samarqand
zj, namely, Ibn Abi l-Fat al-f.
28 He is to be identified with the Muammad Sinr / Sannr who was a muwaqqit at the Amad (from Sayyid
Amad Badaw) Mosque in Tanta, on whom see Cairo ENL Survey, no. D108; and also ~hsanolu et al., Ottoman
Astronomical Literature, II, pp. 625-626, no. 460.
the main cairo corpus 261
MS Cairo DM 778:
This was copied by asan ibn {Al al-Ghamr and appears to date from ca. 1500. It contains
a disordered incomplete set of hour-angle tables followed by various prayer-tables. The title
is given as Kitb Fal al-d}ir and no compiler is mentioned. Following the main set there
are tables of:
D, ta, Ta, s, ha, r and .
The twilight tables are based on parameters 19 and 17 and the salm is assumed to be 1
before daybreak. These prayer-tables are copied in the same hand as the hour-angle tables, but
the colophon occurs at the end of the latter.
MS Cairo DM 786:
An incomplete copy of the tables of (T,t,a) for altitude arguments 45 to 83. The manuscript
bears neither title nor colophon but is copied in the distinctive and untidy hand of asan al-
Tablw29 and can thus be dated ca. 1700.
MS Cairo DM 1101:
A late copy of Ibn Ynus azimuth tables, perhaps copied ca. 1785, and entitled Jadwal al-
samt al-mall daqqa daqqa. Entries are given for altitude arguments up to 89.
MS Cairo DM 1108,9 (fols. 35v-59r):
An unique copy, dated 1053 H [= 1643/44], of a set of azimuth tables for Cairo arranged so
that for each degree of solar longitude the solar altitude is entered vertically. The altitude
arguments are arranged in columns of up to 30 entries, and the meridian altitude is given to
three digits below the last integral altitude argument for certain degrees of solar longitude. The
last southern azimuth in each solar longitude sub-table is written in red ink and the
corresponding solar altitude argument is labelled darajat al-intiql.
MS Cairo DM 1109:
A complete set of tables of (T,t,a) copied in 1128 H [= 1716] by asan ibn {Abdallh,
nicknamed al-Qi al-miskn, perhaps meaning Poor Pussy. There is no introduction and the
title al-D}ir wa-faluhu wa-l-samt isb Ibn Ynus al-Mir li-{ar Mir is a later addition.
On fol. 1r there is a list of pilgrim stations on the road from Cairo to Mecca with latitudes
and qiblas (see 8.4 and 10.9).
MS Cairo DM 1224:
An incomplete copy of the tables of the functions (T,t,a) bearing the tile Jadwil al-d}ir wa-
fal al-d}ir wa-l-samt min qibal al-irtif{ and attributed to Ibn Ynus. There are some simple
instruments on the use of the tables but no other introduction. The manuscript was copied by
{Abd al-Bri} al-{Ashmw (see MSS Cairo TR 191 and DM 108) and can hence be dated ca.
1800. The tables end with those for altitude argument 51.
MS Cairo K 4044:
This copy of the main tables in the corpus is dated 1308 H [= 1890/91]. The work is divided
into two parts, the first being entitled Kitb shar {al fal al-d}ir li-l shaykh Abi l-{Abbs
29 See n. 4:27.
262 part ii, chapter four
al-Maqs. This refers to al-Maqss introduction which precedes a complete set of tables of
T(h,), not t(h,) as one might have expected from the title. The second part of the manuscript
is written in a different hand but also dated 1308 H. It is entitled Kitb Mall al-samt li-bn
Ynus li-{ar lm shaml and contains a set of azimuth tables for altitude arguments up to 83.
The last folio of the manuscript contains tables of the solar azimuth at the equinoxes and
solstices and the functions () and h0().
MS Princeton Yahuda 861,1:
See 8.1.
MS Cairo J 367,5, fols. 35r-40r (new):
Copied ca. 1450, this contains part of an introduction and part of a solar longitude table,
followed by tables of the functions:
r, H, hq, h0, , d and
from the corpus. The work bears the title Natjat al-afkr f a{ml al-layl wa-l-nahr and is
attributed to the 15th-century Egyptian astronomer {Izz al-Dn al-Waf}. However, the prayer
tables are not mentioned in the introduction (see further 5.7). The same introduction and solar
tables, followed by a much more extensive set of prayer-tables for different latitudes, is
contained in MS Princeton Yahuda 861,1 (8.1).
Some copies of the corpus contain a number of spherical astronomical functions of the kind
standard in Islamic zjes. The notes which follow give a clearer picture of those tables than
was possible in my original analysis.
(b) )
Half excess of daylight, d(
The tables of d() to two digits in the Chester Beatty manuscript are not based on the
corresponding table in the kim Zj, which has entries to three digits: about one-sixth of the
entries in the table in the former source are in error by 1 in the second digit. See further (d)
below on the related tables of D().
(c) )
Meridian altitude, H(
The errors in a table of H() will reflect those in the underlying values of () since H =
+ . Furthermore, the values for > 0 and < 0 can be compared to distinguish computational
errors from copyists errors. Now the table of H in MS Dublin CB 3673, fol. 14v, contains
only one error of the former kind: for = 15 we find H = 65;58, and for = 195 we find
the main cairo corpus 263
H = 54;2, which means that (15) was taken as 5;58 instead of the accurate value 5;57.
If this table were based on Ibn Ynus declination table in the kim Zj the entries would
all be correct, since of his entries for integral values of only seven are in error by only 1
in the third sexagesimal digit. However, it is possible that Ibn Ynus accurately-computed
entry 5;56,37 for (15) was misread as 5;57,37 and rounded to 5;58. In MS Dublin CB
3673, fols. 4r and 11r, however, it is clearly 5;56,37.
I do not think that al-Maqs could have computed () accurately to two digits. al-
Marrkushs declination table (computed to two digits) contains some nine errors of -1 in the
second digit, but it seems that he compiled this table by truncating values to three digits found
in an earlier {Irq source.
(e) )
Nocturnal arc, 2N(
The tables of 2N() in MS Dublin CB 3673, fols. 5v and 6v, are based on different tables of
D(). The values of 2N on fol. 6v in the Dublin manuscript are fairly closely related to those
of D on fol. 12r. The values of 2N on fol. 5v are based on a set of values of D or d which
is significantly different from those on fol. 12r and also from the more accurate set of values
of d in MS Paris BNF ar. 2513, fol. 53v, of the Muala Zj. The table of length of darkness
n() (see 4.10) on fol. 6r of the Dublin copy is based on the table of 2N() on fol. 6v.
end of the book, but in the Cairo copy there are four tables following the one of l/(). In
MS Berlin Ahlwardt 5753, fols. 20v and 21r, however, immediately following an incomplete
set of tables of T(h,), there are others of ta() and 1/() which conclude the work.
(j) ) for = 30
Oblique ascensions for Cairo, (
The tables of () to two digits in the Hartford and Paris copies of al-Ldhiqs prayer-tables
contain numerous errors, which fact immediately distinguishes them from the corresponding
table in the kim Zj, since it contains but two small errors. In fact, these tables are al-
Marrkushs. On the other hand, the table of () in the Natja attributed to al-Waf} (8.1),
which contains a few copyists errors, appears to have been taken from the kim Zj.
(k) )
Solar altitude in the prime vertical, h0(
The kim Zj contains a table of h0() carefully computed to three sexagimal digits. In MS
Dublin CB 3673, fol. 85v, of the corpus contains this same table with numerous copyists errors,
but it is two other tables of h0() to two digits in the Chester Beatty manuscript, also occurring
in other copies of the corpus, which merit comment. Firstly note that:
h0() = arc Sin { R Sin () / Sin } = arc Sin { 2 Sin () } for = 30 .
Since also:
Sin = Sin Sin / R ,
we have:
h0 = for = 30 .
the main cairo corpus 265
Now in the table of h0() in MS Dublin CB 3673, fol. 3v, the entry for = 30 is 24;34,
which is clearly absurd. The entry for = 29 has been miscopied as 23;49 instead of
22;49,(32): the 24 in 24;34 appears to have been fixed up and the 34 can be explained
as a copyists error for 35. Nevertheless, the entry for = 31 is 24;21. (Even in the table
to three digits in MS Dublin CB 3673, fol. 85v, the entry for 30 is miscopied as 23;45,0.)
These same errors occur in the Hartford and Paris copies of al-Ldhiqs prayer-tables (7.8),
but in MS Princeton Yahuda 861,1 of the Natja attributed to al-Waf} (8.1) they have been
corrected. In the table of h0() on fol. 12v of the Dublin manuscript, the entries in the third
column have been further garbled, apparently by someone who believed that the maximum
value of h0 was 53;46 rather than the accurate value 53;9. Again, in the prayer-tables of al-
Minf (7.1) most of the standard functions are recomputed for = 23;30: his table of h0 is
an exception, since he simply plagiarized an earlier source. The entries in his table for =
30 and 90 are again 24;34 and 53;9. Finally, in the prayer-tables of al-Fawns (7.5), the
values of h0 corresponding to = 30 and 90 are respectively 23;47 and 53;21, both
inaccurate.
(l) )
Solar rising azimuth, (
This function is defined by:
() = arc Sin { R Sin () / Cos } ,
and so for = 30 is slightly more difficult to compute than h0() (see (k) above). In the kim
Zj Ibn Ynus computes () carefully to thirds. His table occurs again in MS Dublin CB 3673,
fol. 84r, and those on fols. 12v and 3v of the same manuscript, computed to two digits, are
slightly less accurate than can be obtained from his values by rounding. Nevertheless the errors
are much smaller than those in the associated tables of h0.
() ( = 30) al-Marrkush
arc Sin (x) anonymous
The tables of h0() and () to three digits were appended to Ibn Ynus Kitb al-Samt (5.2).
The table of l/() was appended to al-Maqss tables of time since sunrise (5.4). al-Bakhniq
appears to have compiled the table of 2Dh() and added those of () and () ( = 30)
to his edition of the corpus (5.6). The identity of the person who compiled the remaining tables
remains obscure.
These, called Kitb al-Samt when copied separately, display the solar azimuth a(h,) measured
from the prime vertical, for each integral degree of solar altitude h up to the maximum 83
and each degree of solar longitude ( \ 0).
In each of MS Berlin Ahlwardt 5753, Dublin CB 3673, Escorial r. 924,7, Cairo Azhar falak
4382, Cairo MM 137 and Gotha A 1410, the title Kitb al-Samt is given and Ibn Ynus is
named as the compiler (see Figs. 4.2a-b). al-Bakhniq and al-Kha} (5.6 and 6.8) also say
that Ibn Ynus compiled the azimuth tables. I see no reason to doubt this. In the kim Zj
there are tables of a() for altitudes 30 and 35 with values for each degree of in 12 columns
of 30 entries (5.1): no reason is given for choosing these two altitudes. The two tables in the
Zj are slightly more accurately computed than the corresponding tables in the Kitb al-Samt.30
It seems that Ibn Ynus azimuth tables for altitudes 80 to 83 were either not completed
or were lost at some stage in their transmission. Some anonymous with a very limited
knowledge of spherical astronomy computed a set for 80 to 89 (extant in MS Gotha A 1410;
MSS Dublin CB 3673 and Cairo MM 137 have entries up to 88 and 86: see Fig. I-5.1.1a).
A corrected set for 80 to 83 was compiled by Ibn al-Rashd (4.1.3) and these azimuth tables
were used by al-Bakhniq (MSS Cairo DM 108 and DM 53) and were also copied separately
(MS Cairo Azhar falak 4382). I must withdraw my previous statement that al-Maqs
plagiarized Ibn Ynus azimuth tables,31 since now it is clear that al-Bakhniq put together
the tables of T, t and a, and the reference to azimuth tables in the introduction to MS Dublin
CB 3673 is probably not original to al-Maqss introduction. Ibn Ynus tabulated certain other
functions along with the azimuth see 5.2.
In the margin of the tables in some copies of the corpus, e.g., MSS Dublin CB 3673 and
Escorial r. 924,7, for altitudes 1 to 53 it is stated for which solar longitudes the azimuth
changes from north to south for that particular altitude. In MS Berlin Ahlwardt 5753 the change
in direction is more crudely indicated. In the margin of the tables for altitude 54 in the Chester
Beatty manuscript we find the comment that there are no more azimuths in the north for this
altitude and those above it. When, for certain longitudes, the sun does not attain a given
altitude, the entries in the table are left blank. For altitudes of about 60 and above, the entries
for more than one altitude can be tabulated on one folio. In the Escorial manuscript, for a given
30 Both are edited in King, Astronomical Timekeeping in Medieval Cairo, pp. 354-355.
31 Ibid., p. 364. See also the text to nn. 4:12 and 10:2.
the main cairo corpus 267
altitude, the meridian altitude is given for the longitude whose azimuth is closest to 90. Thus,
in the tables for altitude 37, the last significant entry in the right-hand side of the table is
an azimuth of 87;3 for Sagittarius 17 and Capricorn 13. The next and final entry is given
as 0;0, which does not mean due east or west, but that the sun does not attain the altitude
37 for Sagittarius 18 or Capricorn 12. The meridian altitude is accurately stated as 37;3,22
for the solar longitudes having azimuth 87;3. The purpose of giving this detail is not fully
clear to me. However, in the unique MS Cairo DM 1108,9, copied in 1053 H [= 1643/44],
the solar azimuth is displayed as a function a(,h), with values a(h) for each value of , and
the values of H() are given for each . It seems unlikely that Ibn Ynus original tables were
in this form, not least because he tabulated a(h,) for two altitudes in the kim Zj (5.1).
By virtue of the format of the tables no values are shown for the equinoxes. However, in
some copies of the corpus, such as MS Cairo Azhar falak 4382, fol. 71r, there is a table
displaying a(h) at the equinoxes for each degree of h from 1 to 60 (= ) and also for each
0;5 of h between 36;5 and 36;25 (= -) at the winter solstice and again for each 0;5 of
h between 83;5 and 83;30 and also for 83;33 and 83;35 (= +) at the summer solstice.
al-Bakhniq (5.6) states that this table was computed by Ibn al-Rashd (4.1.3).
These, often called Kitb al-D}ir and Kitb Fal al-d}ir when tabulated separately, display
T(h,) and t(h,) for each integral degree of h up to the maximum 83 and each degree of
( \ 0). There are no such tables in the kim Zj. In MSS Dublin CB 3673, Dublin CB
4078, Leipzig 817 and Cairo DM 777, the tables of both functions for a given altitude generally
occupy two facing pages of the manuscript. In MSS Cairo TR 191 and Cairo Azhar falak 4382
only t is presented, and in MS Gotha A1402 only T. In MS Berlin Ahlwardt 5753 the functions
are tabulated separately. In MSS Cairo DM 108, Cairo DM 53, Istanbul Nuruosmaniye 2903
and 2925, Cairo DM 690 and DM 616, triplets of values (T,t,a) are given for each pair of
arguments (h,), where a is the solar azimuth. See Figs. I-2.1.1a-b, 2.1.1c, II-5-4b and 5.6a.
These tables of T and t are less accurately computed than the azimuth tables. In general
the entries are less accurate for higher altitudes, as a result of the nature of the tabulated
functions. When the three functions are tabulated side by side, for altitudes over 80 the number
of longitude arguments for which entries are given for T and t differs slightly from the number
for which entries are given for a: see Fig. 5.6b.
The values of T and t are related by the values of D() tabulated elsewhere in the corpus
(4.3d). al-Maqs (5.4) claims to have computed the tables of T(h,) himself. Each of al-Karak,
Ibn al-Kattn and al-Bakhniq (9.4, 5.5 and 5.6) attribute these tables to him. Ibn al-Kattn
claims to have computed the tables of t(h,) from al-Maqss tables of T(h,) and al-Bakhniq
claims to have done the same, adding that his entries may differ in some cases from previous
calculations.
I am now of the opinion that al-Maqs did compile the tables of T(h,) and Ibn al-Kattn
compiled the tables of t(h,). The fact remains that in order to tabulate T(h,) one must first
compute t(h,): I suspect that al-Maqs used an earlier set of tables of either t(h,) or T(h,)
268 part ii, chapter four
compiled by Ibn Ynus which are no longer extant in their original form and were not available
to astronomers after al-Maqs: see further 5.2 and 5.3.
Two tables of the function h(a,) for a = 30 and 60 and each degree of ( \ 0) occur
in MSS Dublin CB 3673, Cairo DM 954,2 of the Cairo corpus. The second table alone is also
found in MS Cairo DM 153,2, of the corpus, and the first table alone is found in MS Cairo
DM 620 of al-Qaymars tables for the afternoon prayer (7.3). The functions are labelled al-
irtif{ idh kna l-samt lm/sn. In MS Cairo DM 153 Ibn Ynus is specifically identified
as the compiler and indeed the same tables occur along with others based on different azimuths
in the kim Zj (5.1), where Ibn Ynus describes the way in which they can be used to
determine the meridian. The tables in each of the above-mentioned copies of the corpus are
arranged symmetrically in the same way as the azimuth and time tables, whereas those in the
Zj give values of h for each degree of in 12 columns of 30 entries.
The Cairo corpus also contains tables of solar altitude in the azimuths of the qibla and the
rising sun at the winter solstice: see 4.7 and 4.8.
4.7 Tables of the solar altitude and the time of day when the sun is in the azimuth of the qibla
The tables in the corpus relating to the qibla are based on the assumption that the qibla at Cairo-
Fus is 53 E of S. This value is rounded from the 53;0,6 and 53;0,17 derived by Ibn Ynus
in his Zj, and it was used at least by the astronomers for the qibla in Cairo throughout the
medieval period. (Other directions were used by different interest groups, such as the azimuth
of the rising sun at the winter solstice: see 4.8.) Ibn Ynus discussion of the qibla in his Zj
concludes with a table of the solar altitude in the azimuth of the qibla, hq(), based on the
parameter q = 52, computed some time ago. The value 53 is based on a different value
of the longitude difference between Cairo and Mecca, but no further information is provided
about the way in which this longitude difference was measured or what coordinates were used
to derive the value q = 52.
The table of hq() in MS Dublin CB 3673, fol. 86v, of the corpus is based on the parameter
q = 52 and is arranged with values for each degree of , that is, in 12 columns of 30 entries.
There are numerous corrupt readings, as can be seen by checking entries symmetrically-placed
with respect to the solstices or by comparing the table with the more accurate one in the kim
Zj. This table is not contained in any other copies of the corpus, although al-Marrkush (6.7)
tabulated hq() for each 30 of and six of the seven values, which are based on q = 52,
are the same as the corresponding entries in the kim Zj. The Dublin manuscript (fol. 14r),
and several other copies of the corpus contain a table of hq() for q = 53, symmetrically
arranged like the azimuth and time tables, but less accurately computed: see Fig. 4.7a.
A number of sources, including the Dublin manuscript, contain a table of tq(), the hour-
angle when the sun is in the direction of Mecca, also symmetrically arranged. This is based
on q = 53 and is not carefully computed. Using the values of solar altitudes for this azimuth
the main cairo corpus 269
Fig. 4.7a: The table of the hour-angle when the sun is in the azimuth of the qibla at Cairo. On the left is a table
for the duration of evening twilight, based on a value of 17 for the angle of solar depression below the horizon.
[From MS Dublin CB 3673, courtesy of the Chester Beatty Library.]
Fig. 4.7b: The tables for the solar altitude and hour-angle when the sun is perpendicular to the qibla. [From
an unidentified Cairo manuscript, courtesy of the Egyptian National Library.]
as given, for example, on fol. 14r of the Dublin manuscript, together with the tables of t(h,)
in the corpus, I derive more accurate values of tq than those in this table of tq(). A table of
Tq(), measuring the interval from sunrise to the moment when the sun is in the azimuth of
the qibla, is found only in MS Dublin CB 3673, and the values in it are easily derived from
the tables of tq() and D().
In a few sources, including MS Cairo DM 108, there are tables for the solar altitude hq*()
and the corresponding hour-angle tq*() when the sun is in the direction perpendicular to the
qibla ({al tarb{ al-qibla). These tables are based on q = 53 and are less carefully computed
than those of hq and tq: see Fig. 4.7b.
The table of hq (q = 52), which occurs in only one source, is certainly due to Ibn Ynus.
The tables of hq and hq* (q = 53) are also due to him, if my interpretation of al-Bakhniqs
remarks (5.6) is correct. The table of Tq appeared already in al-Maqss set compiled in the
270 part ii, chapter four
late 13th century (see 5.4) and was probably based on an earlier table of tq. The table of tq*
appears to be a later addition.
One copy of the corpus, MS Gotha A1411, contains a table displaying the qibla as a function
of the latitude and longitude differences from Mecca. Values of q(,L) are given for each
degree of each argument from 1 up to 20 and are based on the approximate formula:
Tan q = R Sin L / Sin .
The table is attributed to Ysuf ibn al-Damr, an individual as yet unidentified.32 However,
the same table exists in several other manuscripts, all of Egyptian provenance, and in one of
these the compiler is named as Ysuf al-Damr and in another as the 9th-century Baghdad
astronomer al-Khwrizm. I have discussed this table and related approximate methods for
determining the qibla in a separate study.33
Several copies of the corpus contain a table displaying the solar altitude, hv, in the azimuth
of the ventilator (al-bdahanj), for each 1 of ( \ 0). The table is based on an azimuth
of 27;30 (S of E), that is, the azimuth of the rising sun at the winter solstice. It is understood
that the back of the base of the ventilator should be aligned in this direction and that the front
should be exposed to the winds from the north east. See Figs. VIIb-3.1 and 4.1.
We know from medieval travel accounts that ventilators were a common feature on the roofs
of the multi-storey houses of al-Qhira and Fus; those few which survive in Cairo are late
Ottoman constructions. The table of hv() in the Cairo corpus, as well as other information
on ventilators in medieval Egyptian astronomical sources, inform us that in Mamluk and early
Ottoman times it was the custom to erect ventilators in this astronomically-defined direction.
But the qibla of the first Muslim settlers in Egypt was towards winter sunrise, so that, for
example, the Mosque of {Amr in Fus faces this direction. In addition, the entire medieval
city of al-Qhira is, by hydrographical convenience and human design accompanied with a
good deal of luck, aligned in this direction. The ventilators were simply aligned with the
roughly-orthogonal street-plan of the medieval city, and they were in fact open to a direction
which is not entirely appropriate for gathering favourable winds.
The entries in the table of hv(), which are given to two digits, are not as accurately computed
as those in Ibn Ynus tables of h(a,) (4.4). Many of the entries are in error by about -0;5,
which would suggest that the value used for he(a) (4.6 and 5.2) was in error by this amount.
In certain sources, e.g., MSS Cairo MM 58 and Istanbul Nuruosmaniye 2925, the entries in
the sixth column decrease to 0;41 rather than to 0;0 as one would expect, since the sun rises
in this azimuth when = 270. We can be certain that the correction to 0;41 represents an
attempt to account for the displacement of the visible horizon from the true horizon as a result
of refraction (see further 4.11). Yet not all later astronomers included this adjustment: the
32 In 2001 a set of tables of double-argument planetary (but no lunar) equation tables compiled by him was
located in MS Istanbul Selim Aa 728,2, fols. 38r-181r, copied in the 14th century. These merit detailed
investigation.
33 King, Earliest Qibla Methods, pp. 133-141, 110-112, and especially p. 112.
the main cairo corpus 271
The tables in the corpus relating to the afternoon prayer are based on the standard definition
that the permitted time for the prayer commences when the gnomon shadow exceeds its midday
minimum by the length of the gnomon, and ends at sunset. The most commonly-tabulated
functions are:
ha(), ta() and Ta() ,
labelled respectively irtif{ awwal waqt al-{ar, al-d}ir bayna l-uhr wa-l-{ar and m bayn
al-{ar wa-l-ghurb and representing the solar altitude and the hour-angle at the beginning
of the prescribed interval and the time remaining until sunset: see Fig. 5.4b. In MS Cairo TR
191 only there are tables of:
za(60)() and za(12)()
instead, which measure the shadow at the time for the prayer. Clearly, at least one of these
quantities, easily determined from the corresponding shadow at midday, Z, by addition of one
gnomon length, would need to be calculated in the process of deriving ha.
al-Maqs claims to have computed the table of ta() himself (5.4). The entries in the table
for ha() in the corpus differ slighly from those in the corresponding tables in the Muala
Zj and in the auxiliary tables of al-Kha} (6.6 and 6.8).
The tables of r() and s(), measuring the duration of morning and evening twilight, are labelled
respectively iat al-fajr wa-huwa min ul{ al-fajr il ul{ al-shams and iat al-shafaq wa-
huwa min al-ghurb li-maghb al-shafaq al-amar. These titles indicate only that morning
twilight lasts from daybreak to sunrise and evening twilight from sunset to the disappearance
of the red twilight glow. The underlying values of the solar depression are 19 for morning
twilight and 17 for evening twilight and the two tables are derived from the tables of T(h,)
in the corpus (4.3) using:
r() = T(19,*) and s() = T(17,*) .
272 part ii, chapter four
MS Dublin CB 3673 contains these tables see Fig. 4.7a and also another pair based on
the parameters 20 and 16, both marked a{f, weak, in a later hand, suggesting that someone
objected to these parameters.
Ibn Ynus discussed the determination of twilight in the kim Zj34 but did not present
any tables for twilight. An interesting question that cannot yet be answered definitively, is
whether he himself computed any tables for twilight, using either the parameter 18 mentioned
in the kim Zj or the parameters 20 and 16 attributed to him by Najm al-Dn al-Mir
(2.5) three centuries after his time. al-Maqs (5.4) did not include any twilight tables in his
set but advocated the parameters 20 and 16 in his introduction. Ibn al-Kattn (5.5) included
twilight tables based on parameters 19 and 16 in his set and also additional tables for morning
twilight based on 20, which parameter he says was used by recent scholars (presumably he
was referring to al-Marrkush and al-Maqs), and 18, which he says was used by earlier
scholars (presumably referring to Ibn Ynus). The parameters 19 and 17 were favoured by
most later muwaqqits in Egypt and Syria from the 15th century onwards. Various tables for
twilight purporting to be adjusted for refraction at the horizon occur in late Egyptian prayer-
tables in the tradition of al-Minf (7.1).
The table in the corpus displaying the duration of darkness, n(), labelled jawf al-layl, was
compiled using the simple relation:
n() = 2N() - s() - r() (hr = -19, hs = -17).
As indicated by the title of the table in MS Dublin CB 3673, darkness lasts from the beginning
of the prayer at nightfall to the true dawn (wa-huwa m bayn al-{ish} al-khira wa-ul{
al-fajr al-diq).
The tables of r(), s() and n() are the main twilight tables in the corpus, but in various
late copies such as MSS Cairo DM 108 and Princeton Yahuda 861,1 (8.1), as well as in the
prayer-tables of al-Ldhiq (7.8), other tables derived from these are presented either to
facilitate timekeeping relative to the times of nightfall and daybreak or to serve the
determination of various twilight-related religious institutions. These I now describe.
Firstly, tables of s(), labelled matli{ al-shafaq, the (oblique) ascensions of (the
ascendant at) nightfall, occur in various late Egyptian prayer-tables, notably those of al-
Ldhiq, and are based on the relation:
s() = (H) = (*) + s() ,
where H denotes the ascendant at nightfall and the parameter for s is 17. Tables of r(),
labelled mali{ al-fajr, the (oblique) ascensions of (the ascendant at) daybreak, are less
common in late Egyptian prayer-tables, the function () (see below) generally being
preferred. The underlying relation is:
r() = (H) = () - r() ,
where H denotes the ascendant at daybreak and the parameter for r is 19. I have no clues
as to the identity of the individual who first compiled these tables. However, the earliest
reference to the two functions in the sources known to me is in the treatise by Najm al-Dn
al-Mir on spherical astronomy (MS Milan Ambrosiana 227a (C49), Ch. 24: see 2.5), written
ca. 1325. He prescribes procedures that are equivalent to the above formulae.
A group of tables occuring in MS Cairo DM 108 of the corpus, as well as in the prayer-
tables of al-Ldhiq, is associated with two activities called salm and afy. It was the custom
in medieval Egypt to light candles (qandl) on the minarets of mosques during the nights of
Raman: the term afy refers to the extinction of these candles 20 minutes before daybreak,
which may have been the signal to the faithful that the days fasting should begin. (In Egypt
today the beginning of the fast is called the imsk and is also 20 minutes before daybreak.)
The salm was a special call of the muezzin, a few minutes before daybreak, invoking blessings
on the Prophet Muammad. It is clear that the muwaqqits in medieval Cairo were not of one
mind concerning the precise time of the salm. From the sources investigated below it appears
that the salm was variously timed to begin 1, 1;30 or 2 (that is, 4, 6 or 8 minutes) before
daybreak: see Fig. 4.10 for a table based on a parameter of 1.
The table in MS Cairo DM 108 and in most copies of al-Ldhiqs prayer-tables which is
entitled time from sunset to the salm, two (equatorial) degrees before daybreak is based
on the relation:
() = 2N() - r() - 2 (hr = -19)
In MS Paris ar. 2553 of al-Ldhiqs tables the corresponding entries are 1 more, so that the
salm is assumed to be at 1 before daybreak. In a later anonymous set of this function preserved
in MS Cairo DM 57, computed to three digits for each 0;6 of by linear interpolation (7.2),
the entries are based on those in MS Paris ar. 2553. Whilst this means that the salm is 1
before daybreak, the title of the tables states that it is 2 before. On the other hand, a note
by the title states that the shaykh who compiled the tables was in error because the time-
difference is actually 1 not 2.
A related table is found only in al-Ldhiqs prayer-tables. It is entitled (oblique) ascensions
of (the ascendant at the time of) the salm, two (equatorial) degrees before daybreak. The
title also states that if one subtracts three degrees from the entries the result will be the
ascensions of the afy. An entry is given for each degree of in 12 columns of 30 entries.
If H is the ascendant at daybreak, then:
() = ( H) = () - r() - 2 (hr = -19) .
In MS Istanbul Reislkttap Mustafa Efendi 582 the table of is based on the assumption
that the salm is 1 before daybreak. In later related tables of this function in MSS Cairo DM
157, datable ca. 850, and DM 158, datable ca. 1600, values are given for each 0;3 or 0;6
of , respectively, generated by linear interpolation. In these two sources the entries are larger
than those in the Hartford copy by 0;30 and 1, respectively.
Tables of the function (), al-d}ir li-l-afy f Raman {al anna baynahu wa-bayn al-
fajr h} daraj, time from sunset to the afy in Raman, 5 before daybreak, occur only in
MSS Princeton Yahuda 861,1 and Cairo DM 108 amongst the sources investigated. In the
former the function tabulated is:
() = 2N() - r() - 5 (hr = -19) ,
as one might expect. In the latter the function tabulated is in fact:
() = 2N() - r() - 2D() (hr = -19)
where the values of D, a correction for refraction at the horizon, are those attributed elsewhere
to Ibn Ynus (4.11). Note that this formula includes a correction for both horizons. This table
274 part ii, chapter four
Fig. 4.10: On the right is a table displaying the time from sunset until the moment for extinguishing the candles
in the lanterns on the minarets (al-afy) in Raman, taken as 5, that is, 20 minutes, before daybreak. On the
left is a table displaying the corresponding time to the benediction on the Prophet (al-salm), taken as 1, that
is, 4 minutes, before daybreak. The entries in the second table are, of course, 4 more than the corresponding
entries in the first. Both tables are adjusted using the difference minutes for the time of sunset over the visible
(mar{) horizon. [From an unidentified Cairo manuscript of the Cairo corpus, courtesy of the Egyptian National
Library.]
is the only one in MS Cairo DM 108 which is adjusted, silently at that, for refraction at the
horizon.35
Tables of the function (), labelled mali{ al-afy, the (oblique) ascensions (of the
ascendant at the time) of the afy, occur in MS Cairo DM 153 and Princeton Yahuda 861,1
of the sources investigated, as well as in MS Istanbul Reislkttap Mustafa Efendi 582 of al-
Ldhiqs prayer-tables. Now the table of in MS Princeton Yahuda 861,1 is not based on
the table of () in MS Cairo DM 108. Rather, the tabulated function is:
() = () - r() - 5 (hr = -19) ,
as one would expect. The table in this Istanbul manuscript is in fact based on:
() = () - r() - 5 - D() (hr = -19) ,
35 In Astronomical Timekeeping in Medieval Cairo, p. 371, I noted that the entries would correspond to
the relation one would expect (namely, that used in the Princeton manuscript) only if the value used for hr was
about -20;15. The structure of this anomalous table is now explained.
the main cairo corpus 275
using the values of D attributed elsewhere to Ibn Ynus. This formula includes a correction
D only for the western horizon, but there is no indication of any correction at all in the title
of the table. Furthermore, the accompanying table of () in the same manuscript is not
corrected for horizontal refraction.
I have not found any references to functions relating to the salm and afy in the Egyptian
astronomical literature. In MS Princeton Yahuda 861,1 (8.1 and 9.5) there is a table giving
the time from sunset to the salm for Jerusalem, but the underlying time is daybreak rather
than a few minutes before. Shihb al-Dn al-alab, writing in Damascus, stated that daybreak
(al-fajr) is the time of the prayer-call (al-dhn) in Raman and of the salm (written s-l-
m) in the other months of the year (11.2). In MS Damascus hiriyya 7564 (11.11) there are
some cryptic notes about the time for the salm as practiced in Aleppo.
In MS Cairo DM 108 of the corpus, dating from the early 19th century, there is a table entitled
daq}iq al-ikhtilf, literally, minutes of difference, here denoted by D(): see Fig. 4.11a.
The function is tabulated to two sexagesimal digits for each degree of . In some manuscripts
of al-Ldhiqs prayer-tables (7.8), such as MS Cairo DM 190, values of the same function
to one digit are given for each zodiacal sign, along with the remark that it measures the time
in minutes (daq}iq) of an equatorial degree taken by the sun to move between the true horizon
(al-ufq al-aqq) and the visible horizon (al-ufq al-mar}).
This table in MS Cairo DM 108 also occurs in the prayer-tables of the early-16th-century
astronomer Muammad al-Minf (MS Cairo DM 107, but not MS Cairo DM 177). al-Minf
modified each of the tables associated with horizon phenomena in the prayer-tables of the Cairo
corpus to account for the effect of refraction at the horizon (7.1). The same table occurs in
MS Alexandria 4441J of the prayer-tables of the 18th-century muwaqqit Ibn Ab Rya (7.9).
Although al-Minfs table of D() is the earliest table of the function in any of the known
sources, the underlying theory, at least that part of it which is in reasonable conformity with
physical reality, appears, as I shall now attempt to show, to date from the 10th century.
The entries in the tables of D() in MSS Cairo DM 108 and DM 107, and Alexandria 4441J
were calculated by linear interpolation between the values at the equinoxes and the solstices.
The significant entries, in minutes of an equatorial degree, are the following:
EQ: 47 SS: 62 WS: 32 .
The attribution of these values to Ibn Ynus is claimed by al-Minf. In the introduction to
his prayer-tables (e.g., MS Cairo DM 107, fol. 7v), he states that Ibn Ynus had found the
difference minutes to be 47 at the equinoxes, increasing by 5 for each sign up to the summer
solstice and decreasing by 5 for each sign to the winter solstice. We may presume that al-Minf
had access to a work by Ibn Ynus that is no longer extant, for none of his known works contains
any mention of refraction. In an earlier, 15th-century Egyptian source, these same values are
associated with the 14th-century Damascus astronomer Ibn al-Shir (5.7), but this association
is less likely.
Now only part of the theory described by al-Minf and reflected in the tables makes sense.
276 part ii, chapter four
Fig. 4.11a: A table for the time taken by the sun to move
between the true and visible horizons. [From MS Cairo
DM 108, fol. 74v, courtesy of the Egyptian National
Library.]
It seems reasonable to assume that Ibn Ynus would have considered the visible horizon to
be at a certain arc of depression below the true horizon. If this is so, the above equinoctial
and solstitial values are easily shown to be mutually inconsistent. On the other hand, if the
three values are the results of observations made at the equinoxes and solstices, then only the
equinoctial value is reasonable. But it seems much more likely that the solstitial values were
simply computed from the equinoctial value, which was possibly derived by observation. In
any case, it is worthwhile to investigate the three values further.
Consider Fig. 4.11b, which represents a portion of the western sky about the west-point
in plane cross-section. The sun sets over the horizon at an angle . AB represents the true
horizon and AB the visible horizon at an arc of depression h below the true horizon. The
problem is to find to a first approximation how long the sun takes to move between these two
horizons. XX represents the path of the centre of the sun at the equinoxes and YY and ZZ
its path at the summer and winter solstices, respectively. Clearly:
XX = YY = ZZ = h sec .
the main cairo corpus 277
Now the time taken by the sun to move from X to X is measured by the corresponding arc
of the celestial equator. We assume, to a first approximation, that the difference minutes at
the equinoxes, D0, equal h sec . Since Ibn Ynus, according to al-Minf, took these as
0;47 for latitude 30, we have:
D0 = h sec 30 = 0;47 ,
whence:
h 0;40 = 2/3 .
This, then, is an approximate value of the vertical displacement of the horizon due to refraction
corresponding to the value for D0 attributed to Ibn Ynus. Note, however, that the value 0;41
is used in the table of hv (4.8) to represent the apparent altitude of the sun when its centre
is on the true horizon. The modern value for this displacement is 035.36
The solstitial values attributed to Ibn Ynus are, in the light of modern knowledge, less
reasonable. A distance of 0;47 on either of the solstitial day-circles corresponds approximately
to an arc of:
0;47 sec = 0;51
on the celestial equator and it is clear that the time taken by the sun to move between the two
horizons is the same at both solstices. This time is some 0;4 more than the corresponding
time at the equinoxes, according to the above calculation. Note too that, if we assume a value
36 For a modern discussion see Smart, Spherical Astronomy (n. 1:37), pp. 69-70, etc.
278 part ii, chapter four
of 0;40 for h, the corresponding times D can be found from tables of T(h,) in the same
way as the duration of twilight (4.10). Thus:
D() = T(2/3,*) 2/3 T(1,*) .
From the table of T(h,) for 1 (see Fig. 4.11c) we can read off the time taken by the sun
to move 1 below the horizon, namely:
EQ: 70 SS/WS: 79 ,
from which it follows that the time taken by the sun to move 2/3 below the horizon is
approximately:
EQ: 47 SS/WS: 53 .
Since the variation in D is relatively small, it is quite reasonable to interpolate linearly between
the equinoctial and solstitial values of D for intermediate solar longitudes.
Ibn Ynus was a master of spherical astronomy, and I deem it unlikely that he would make
such an elementary error as to give radically different values of D at the solstices. Rather,
I suspect that he originally stated that the correction was 47 minutes at the equinoxes and
increased about 5 minutes to maxima at the solstices, which would be virtually correct. The
assertion of al-Minf that Ibn Ynus had stated that the correction was 47 minutes at the
equinoxes and changed by 5 minutes for each sign with a maximum at the summer solstice
and a minimum at the winter solstice is then simply garbled. My hypothesis has no substantia-
tion in the manuscripts examined thus far and my investigations of the various Arabic manus-
cript sources dealing with refraction at the horizon contribute little to our understanding of the
problem as treated by Ibn Ynus and (mis)construed by his successors.
The theory of refraction at the horizon associated with Ibn Ynus as interpreted by al-Minf
was adopted by the latters son {Abd al-Qdir al-Minf to serve all latitudes (8.2), using the
approximation:
D(,) / 30 D(30,) .
Four sources have been identified in which values based on this formula are presented, namely,
for Mecca, Crete, Damascus and Istanbul (6.10, 8.8, 11.13 and 14.9). No Muslim astronomer
is known to have noted that the underlying equinoctial and solstitial values were mutually
inconsistent.
The instructions accompanying the table in MS Dublin CB 3673 and the slightly modified
version in MS Cairo DM 108 contain virtually no information about the way in which the tables
were intended to be used in practice, nor do they reveal how the tables were computed. In
the Dublin copy the title Kitb Ghyat al-intif{ ... , which may be (and has been) rendered
Very Useful Tables ... , is given to the whole corpus. Also al-Kattn called his tables Kitb
Nihyat al-intif{ ... , which may be rendered Extremely Useful Tables ... . Otherwise the
various copies of the tables of T(h,), t(h,) and a(h,) bear the titles Kitb al-D}ir, Kitb
Fal al-d}ir and Kitb al-Samt, or various combinations, or they bear no titles at all.
The introduction begins with pious statements about God as the Creator of the Universe.
The Qur}nic verse X.5/5:
the main cairo corpus 279
(God) made the sun as a shining light and made the moon a luminary and measured
out its mansions
is continued thus (fol. 83r:5):
for perfecting prognostications about what He orders and what He effects.
Then the verse XXXVII.6/5:
(God) adorned the lower heavens with the stars
is continued (MS CB 3673, fol. 83r:5-6):
and the times are known by what rises and sets.
Since, states the author, astronomical timekeeping ({ilm al-mawqt) was ordained by God, he
thought he would prepare these table to be the limit of usefulness (ghyat al-intif{). The
instructions proper are translated in full below, using MSS fols. 83r-83v of the Dublin
manuscript and fols. 2v-3r of the Cairo one. The paragraph numbers are my own. 9 does not
occur in copies of the introduction that are not followed by the azimuth tables. 10a appears
to be original to al-Maqss introduction, whereas 10b, whose highly sophisticated contents
merit special commentary, was probably inserted by someone more competent that al-Maqs.
The instructions read:
1 To determine the time since sunrise up to any instant you wish, enter the solar
longitude in the table of the instantaneous altitude and you will find the required time,
if the solar altitude is in the east. If it is in west, subtract the value in the table from
the diurnal arc and the remainder will be the time up to that instant. God may He
be exalted knows best. Example: Suppose the sun is in 21 of Taurus and the
instantaneous altitude is 30. We find 35;38a in the table: this is the time from sunrise
to the instant in question. If the altitude is in the west, subtract 35;38 from the diurnal
arc, which is 201;46b: the remainder is 166;8, which is the required time.
2 To determine how much of the day has passed and how many hours of daylight
are left, divide the time since sunrise by the number of degrees corresponding to a
seasonal day-hour for the day in question, or by 15 if you want equinoctial hours. The
result will be the number of hours of daylight that have passed. Subtract the seasonal
hours from 12 and the equinoctial hours from the number of hours of daylight in
question: the remainder will be the number of hours of daylight left. Example: Suppose
the sun is in 11 of Aries and the instantaneous altitude is 45 in the east. We find
the time since sunrise for this altitude as before: it is 53;19c. We divide this by the
number of degrees corresponding to a seasonal day-hour, which is 15;24d, and the
result is [3]e;27. These are the hours of daylight that have passed; subtract them from
12 and the remainder, which is the number of hours of daylight left, is 8;33. If we
divide the time since sunrise by 12, the quotient is [3;33]f and this is the number of
equinoctial hours that have elapsed. We subtract these from the number of equinoctial
hours of daylight on the day in question, which is 12;20, and the remainder is 8;47.
This is the number of hours of daylight remaining. I have included a table at the end
of the book in which the entry corresponding to the solar longitude is multiplied by
the time since sunrise to give the seasonal day-hours passed since sunrise.
3 To determine (the altitude of the sun for) the seasonal and equinoctial hours, enter
the solar longitude in the tables of the time since sunrise for each altitude and turn
280 part ii, chapter four
through table after table until you find a time as large as the degrees corresponding
to the seasonal or equinoctial hours. Interpolate if necessary and the result with be
the required altitude. Example: Suppose the sun is in the last degree of the sign of
Gemini and we want the altitude at the end of the first seasonal hour. We enter the
solar longitude in the tables, turning through table after table, and find opposite the
longitude in the table for (altitude argument) 13 a time of 16;24g. We subtract this
from the number of degrees in one seasonal hour, which is 17;26h, and the remainder
is 1;2. We divide this by the difference between the time [in the tables for altitude
13 and for altitude 14, which difference is 1;15. The ratio of one to the other is
0;50, which we add to the first altitude]i and so the required altitude is 13;50. If we
want the altitude after one equatorial hour, we enter the solar longitude in the tables
and we find opposite it 13;56. We subtract this from 15 and the remainder is 1;4,
which we divide by the difference between the time for altitude 11 and that for 12,
which is 1;15. The ratio of the one to the other is 0;51, which we add to the first
altitude, obtaining 11;51 for the required altitude.
4 To determine the ascendant at the eastern horizon, add the time since sunrise to
the oblique ascensions of the solar longitude and the result will be (the oblique
ascensions of) the ascendant at that time. Example: Suppose the sun is in 18 of Scorpio
and the instantaneous altitude is 25. We take the time since sunrise for this altitude
and find it to be 33;20j. We add this to the oblique ascensions of the sun, namely,
235;52k, and the sum is 269;12. We take the longitude corresponding to this: the
result is 16;18, which is the degree of Sagittarius that is rising.
5 To determine the hour-angle at the beginning of the afternoon prayer, you should
know that there is a table for this included at the end of the book. You enter the solar
longitude as argument and read off the required value.
6 To determine the semi diurnal arc, you enter the solar longitude in the appropriate
table and read the required value.
7 To determine the duration of morning twilight, which is the remainder of the
nocturnal arc to sunrise, you enter the longitude directly opposite the solar longitude
on the ecliptic in the table of the time since sunrise for altitude 19 and the value you
find will be the required time.
8 To determine the time from sunrise to the moment when the sun is in the azimuth
of the qibla, enter the solar longitude in the appropriate table and read off the required
value.
9 To determine the azimuth, proceed as with the tables of time. The direction of the
azimuth is south if the sun is in the southern zodiacal signs, or if it is the northern
signs and the altitude is greater than the altitude in the prime vertical; otherwise, it
is north. God knows best.
10a Numerous interesting facts about the operations of timekeeping by day and night
will be revealed to anyone who examines these tables closely. However, in this book
we have chosen to keep the discussion short.
10b Certain interesting facts will be revealed to anyone who examines these tables
closely. For example, if the solar altitude equals the declination, then both the azimuth
the main cairo corpus 281
and the time since sunrise are equal to the complement of the other, Further, if you
take the declination as altitude and you know the complement of the time since sunrise
for the solar longitude, then, if the latitude is 30, it will be the half excess of daylight.
Numerous other facts and points of interest relating to night and day also become
evident. God may He be exalted knows best.
_______________
aAs on fol. 54v. b As on fol. 12r. c As on fol. 39v. d Fol. 13r has the more accurate value 15;25.
eText: 4 (sic). f Text: 4;13 (sic). g As on fol. 71v. h As on fol. 13r. i [ ] missing in the MS. j As
on fol. 59v. k This is accurate, and is the value given in the kim Zj. MS Hartford TS 621, p. 21,
has 235;51.
These instructions (with the exception, perhaps, of 10b) were written by al-Maqs, not by Ibn
Ynus as I stated in my first study of the Cairo corpus.37 They were intended to precede his
tables of T(h,) discussed in 5.4. This explains why there is no mention here of the hour-angle
t(h,) and most of the minor functions tabulated in the corpus. Nevertheless the style differs
markedly from that of al-Maqss introduction to his tables for sundial construction.38
In the new source MS Gotha A1402 the instructions are in al-Maqss name and are followed
by his tables of T(h,). In MS Cairo DM 53 they also bear al-Maqss name and are followed
by al-Bakhniqs arrangement of the tables of (T,t,a). In MS Cairo DM 108 the introduction
ends with al-Maqss name. However, in MS Dublin CB 3673, which was the major source
of my first study of the Cairo corpus, the instructions are introduced in the name of Ibn Ynus.
It is now apparent that the earlier scholars name was inserted by the copyist: it is preceded
by a badly-written phrase with Arabic letters improperly pointed, which I originally read as
anwa} al-ns and, with considerable stretching of the imagination, interpreted as meaning
something like most proficient of astrologers. However, the phrase must be a corruption of
Abu l-{Abbs, the kunya of al-Maqs. The copyist of the Chester Beatty manuscript was of
the opinion that Ibn Ynus deserved more credit that al-Maqs for the tables which he was
copying: that copy does indeed contain more tables by Ibn Ynus than any other known source.
He omitted al-Maqss name and inserted that of Ibn Ynus, not realizing, perhaps, that Abu
l-{Abbs was al-Maqss kunya rather than that of Ibn Ynus. This copyist also included the
remarks of 10b above, which imply that the author had before him tables of both a(h,) and
T(h,). I do not think that either al-Maqs or al-Bakhniq would have noticed these subtleties.
Rather, they are reminiscent of Ibn Ynus remark in Ch. 15 of the kim Zj (MS Leiden
Or. 143, p. 335:17-20):39
If for a particular locality there are no (tables of) oblique ascensions available but
there is a table of the azimuth computed for the first point of Aries giving values for
each degree of solar altitude up to 24, then enter the solar declination as argument
in this table and the corresponding value in the table will be the half excess of
daylight.
In al-Maqss introduction the parameters advocated for morning and evening twilight were
probably originally 20 and 16. This is the case in MSS Cairo DM 108 and Gotha A1402,
but in MS Dublin CB 3673 only the value for morning twilight is mentioned, and in MS Cairo
DM 53 the parameters 19 and 17 are proposed. On the other hand, MS Dublin CB 3673 is
the only copy of the corpus containing twilight tables based on parameters 20 and 16
(although tables based on the more common parameters 19 and 17 are also included in this
manuscript). See further 4.10.
In the introduction to the tables in the Chester Beatty manuscript separate tables of:
(), ta(), D() and Tq()
are specifically mentioned. In the introduction in MS Cairo DM 108 only separate tables of:
(), D() and Tq()
are specifically mentioned. In both sources there is mention of the tables of a(h,), but this
seems to be a later addition to al-Maqss original introduction. In 5.4 I discuss al-Maqss
introduction and tables further.
the development of the main cairo corpus 283
CHAPTER 5
It is not yet possible to give a completely clear picture of the development of the main Cairo
corpus described in Ch. 4. New manuscripts of tables from the corpus could add considerably
to our knowledge and it is not too much to hope that some will turn up in future investigations
of uncataloged manuscript collections.
The celebrated Fatimid astronomer Ibn Ynus (I-2.1.1 and II-4.1.1) tabulated numerous
spherical astronomical functions in his kim Zj.1 The title-folio of the Leiden manuscript
is shown in Fig. 5.1. It is written in the hand of Ibn Abi l-Fat al-f, and bears a notice
of ownership by Taqi l-Dn ibn Ma{rf, indicating that it once formed part of his library at
the short-lived observatory in Istanbul.2 No other zj contains so many tables of this kind; Ibn
Ynus thus began a trend which was to distinguish the later traditions of astronomical
timekeeping in Egypt and Syria. The tables in the Zj which have values computed to minutes,
seconds or thirds [here indicated by [1], [2] and [3]) for each 1 of (or h, where relevant),
occasionally for each 0;10 (here indicated by *). The underlying parameters, where relevant,
are:
= 30;0 and = 23;35 .
The tables display the following functions:
()* [2]; () [2]; () [1]; Sin () [3]; Tan () [3];
() [1] (for each 1 of from 1 to 48);
d() [3] and Sin d() [3];
() [2] and Sin () [2]; h 0() [2]; k(h) [2];
a() [1] for two different altitudes; h() [1] for ten different azimuths;
and h q() [1] (q = 52).
Of particular importance for the development of the main Cairo corpus was the inclusion of
tables of a(h,) for two values of h (I-5.1.1), and h(h,) for ten values of a (I-4.7.1); such tables
are not known from earlier sources. The tables of:
1 On Ibn Ynus see n. I-2:3. King, Ibn Ynus, deals specifically with the spherical astronomy in the kim
Zj.
2 See also King, Astronomy in Fatimid Egypt, p. 499, and also nver, Istanbul Observatory (in Turkish),
figs. 41-42, for the marks of ownership on other manuscripts from Taqi l-Dns library.
284 part ii, chapter five
Fig. 5.1: The title folio of the kim Zj, containing biographical information on the author in the hand of the
late-15th-century Cairene astronomer Ibn Abi l-Fat al-f (I-9.10). On the right is a notice of possession of
the Istanbul astronomer Taqi l-Dn (see Fig. I-1.0a). [From MS Leiden Or. 143, fol. 1r, courtesy of the
Universiteitsbibliotheek.]
There are also tables of Sin () [3] and k(h) [2] for = 33;25, serving Baghdad and perhaps
an indication of the source of part of his inspiration.
Ibn Ynus compiled a work called Kitb al-Samt containing a complete set of tables of a(h,).
Subsequently the entries for h 80 were lost and they were recomputed by an anonymous
incompetent who included tables for 80 h 89 even though the maximum solar altitude
at Cairo is only about 831/2. In the 14th century the tables for 80 h 83 were recomputed
more carefully by Ibn al-Rashd (4.4). al-Bakhniq used the better set for his edition of the
corpus (5.6), extant in MSS Cairo DM 108, Cairo DM 53, Istanbul Nuruosmaniye 2903 and
2925, Cairo DM 690 and Cairo DM 616, but the uncorrected tables also survive in MSS Dublin
CB 3673, Cairo MM 137 and Gotha A 1410 of the Kitb al-Samt. In these three sources, entries
are given up to 88, 86 and 89, respectively: see Fig. I-5.1.1a. The corrected tables of the
Kitb al-Samt were also copied separately, as in MS Cairo Azhar falak 4382.
Some copies of the Kitb al-Samt have tables of hv, h q (q = 52 or 53), hq* (q = 53) and
a(h) ( = 0 and 90) appended, as well as and h0. MS Dublin CB 3673 contains a set of
uncorrected tables of a(h,) as well as of hv, hq (q = 52), and h0. MS Gotha A 1410 contains
the uncorrected azimuth tables and also tables of hq (q = 53) and hv. MS Cairo Azhar falak
4382 contains the azimuth tables and a(h) ( = 0 and 90). MSS Berlin Ahlwardt 5753 and
Escorial r. 924,7 of the azimuth tables are incomplete. The tables of hq (q = 52), and and
h0 (both to three digits), were taken from the kim Zj. The rather carelessly computed table
of hv usually has the entry 0;41 for the winter solstice rather than zero, to correct for refraction
at the horizon. The remark in MS Istanbul Nuruosmaniye 2925 that the table of hv (which has
the entry 0;41) is due to Ibn al-Rashd may mean either that he computed the table or that
he introduced the modification for the winter solstice. If I interpret al-Bakhniq correctly as
asserting that the table of hq* (q = 53) was due to Ibn Ynus, then it is reasonable to suppose
that Ibn Ynus also computed hq (q = 53), although these tables for q = 53 are less accurate
than his table for q = 52.
I conjecture that Ibn Ynus also compiled a small set of prayer-tables for the latitude of Cairo-
Fus, rather like those for Mecca preserved in MS Cairo MM 68 (6.10), and containing tables
of such functions as:
H, D, ha, ta, r and s .
Perhaps it was in such a work that he proposed the equinoctial and solstitial values for the
correction D due to refraction at the horizon. MS Cairo MM 68 contains a set of these
corrections for Mecca.
al-Bakhniq (5.6) appears to state that at the end of a copy of Ibn Ynus azimuth tables
used by him there had been added (uyafa) tables of the functions:
286 part ii, chapter five
The tables for timekeeping called Kitb al-D}ir compiled by al-Maqs (4.5) consisted of a
complete set of tables of T(h,) and four additional tables of:
D(), ta(), Tq() (q = 53) and 1/() .
The tables were preceded by an introduction in which each of these functions was specifically
mentioned (4.12). MS Gotha A1402 (4.1) is the closest manuscript to al-Maqss original set.
It contains his introduction and a complete set of tables of T(h,) followed by the four smaller
tables in the order noted above. The parameters advocated for twilight in the introduction are
20 and 16. The other tables which follow the one of l/() are to be regarded as later additions
to al-Maqss set. In MS Berlin Ahlwardt 5753 (4.1) there are tables of ta() and l/()
immediately following a set of tables of T(h,): see Figs 5.4a-b. However, in all other sources
investigated T(h,) is tabulated either together with t(h,), or together with t(h,) and a(h,)
in al-Bakhniqs arrangement of the corpus.
4 Ibid., III.15.3d.
5 See n. I-1:22.
288 part ii, chapter five
Fig. 5.4a-b: (a) The tables of time since sunrise for values of the solar altitude from 77 to 83 as found in
the Berlin manuscript, and (b) the tables of ta() and 1/() which immediately follow. These were the only
tables of the corpus known to Carl Schoy and mentioned in his Gnomonik der Araber (1923), p. 53. [From
MS Berlin Ahlwardt 5753, fols. 19v-20r and 20v-21r, courtesy of the Deutsche Staatsbibliothaek, Berlin.]
the development of the main cairo corpus 289
al-Maqs uses the phrase ra}aytu an uarrir, I thought I would compile / edit / prepare
when introducing his tables. He makes no reference to any earlier tables of the same kind. In
the introduction to his sundial tables he uses the verb asaba, to compute, and mentions
various of his predecessors and contemporaries, with the notable exception of al-Marrkush
who compiled similar tables (2.7). It seems to me probable that al-Maqs compiled his tables
of T(h,) by using his own table of D() and some earlier hour-angle tables. I suspect also
that his table of ta() is not original and that his table of Tq() was derived from an earlier
table of tq(): I think Ibn Ynus was compiler of both the tables of ta() and tq(). Finally,
al-Maqss table of l/(), which is rather carelessly computed, was probably compiled from
rounded values of D().
MS Istanbul Kl Ali Paa 684 is a unique copy of an extensive set of tables compiled by
Ibn al-Kattn (4.1.4). The manuscript was executed in Cairo in the authors own elegant hand
and is dated 768 H [= 1366]. MS Cairo MM 72, preserved in Cairo, contains a set of solar
altitude tables copied in 747 H [= 1346/47] by Ibn al-Kattn but not necessarily computed
by him (6.9, illustrated).
The Kl Ali Paa manuscript is of considerable importance for the present study. The work
it contains is entitled Kitb Nihyat al-intif{ f ma{rifat fal al-d}ir min al-falak min qibal
al-irtif{, Extremely Useful Book for Finding the Hour-angle from the (Solar) Altitude, and
was patronized by Burhn al-Dn Ibrhm al-Battann.6 In his introduction to the tables Ibn
al-Kattn states that when he saw the tables of time since sunrise (jadwal al-d}ir min al-falak)
of al-Maqs, he got upset about the fact that the format of the tables caused the upward
arguments to run only from 0 to 29 and he wanted to make both vertical arguments run to
30. He gives an example for = 360 and h = 60 and computes T = 90, remarking that
al-Maqs gives the value 85;1, and there is a difference of 4;59. Ibn al-Kattns navet is
well illustrated by the above: al-Maqss value 85;1, the first entry in his tables of T(h,) for
h = 60, is for = 1 rather than the equinox. Ibn al-Kattn goes on to say that most of the
astronomical works he knew of dealt with the hour-angle rather than the time since sunrise
and this fact was what inspired him to prepare the present work. The major part of the tables
which follow consists of a complete set of tables of the hour-angle t(h,) derived from al-
Maqss tables of T(h,). The entries are the same as those in the hour-angle tables in the main
Cairo corpus (4.5) except that there are 31 rows of entries on each page rather than 30, both
arguments running from 0 to 30. Thus Ibn al-Kattn has contrived to include the values for
the equinoxes in the hour-angle tables: of course, the other entries in the top row merely
duplicate the corresponding entries in the bottom row.
The remainder of the tables of Ibn al-Kattn consists mainly of a set of prayer-tables, having
basically the same entries as the Cairo corpus, but Ibn al-Kattns distinctive format. The
functions which he tabulates are:
Of prime importance for our investigations is a note by al-Bakhniq (5.1) which he added
to al-Maqss introduction (5.4) when preparing his own edition of the corpus. The
distinguishing feature and virtually sole innovation of this edition is that the values of the time
since sunrise, hour-angle and azimuth are tabulated together, a triplet of entries (T,t,a) being
given for each pair of arguments (h,): see Fig. 5.6a-b. al-Bakhniqs note is contained, for
example, in MSS Cairo DM 108, Cairo DM 53, Istanbul Nuruosmaniye 2925, Cairo DM 616
and Cairo DM 739. The version in MS Cairo DM 53 is incomplete and rather corrupt. I present
an annotated translation of the version in MS Cairo DM 108 (see Fig. 5.6c). (The texts in the
other manuscripts offer no variants of consequence.) The most important sections, which are
omitted in MS Cairo DM 53, are in italics in the translation.
The wretched slave of God Amad ibn Muammad ibn Amad al-Azhar known as
al-Bakhniq said: I have arranged this book so that under each zodiacal sign there
are three tables. The first of these gives the time since sunrise by al-Maqs may God
have mercy on him and the second gives the hour-angle. The values of this latter
function may differ in some cases from previous calculations:a I calculated them by
using the corresponding values (of the time since sunrise and half diurnal arc).b The
the development of the main cairo corpus 291
third function is the azimuth for the altitude in question computed by the Shaykh
Shams al-Dn Ab {Abdallh Muammad ibn Ibrhm ibn {Abdallh al-Rashd may
God have mercy upon himc correcting the errors introduced by repeated copying
of the azimuth tables of Abu l-asand ibn Ynus may God have mercy on him.
He e had preceded these with two tables on a single page, one giving the solar rising
azimuth and the other the solar altitude in the prime vertical. (He also included) tables
of the solar altitude in the azimuth of the ventilator and in the azimuths of the qibla
(S.E.) and the perpendicular direction (S.W.), on the basis that its azimuth (measured
from the prime vertical) was 37, and an additional complementary f table for the
azimuth at the equinoxes and solstices. To these there was added g at the end of the
book the tables which al-Maqs had at the end of his book, h as well as the hour-angle
when the sun is in the azimuth of the qibla, the solar meridian altitude, the solar
292 part ii, chapter five
Fig. 5.6c: The text of al-Bakhniqs remarks on the corpus. [From MS Cairo DM 108, fols. 3v-4r, courtesy
of the Egyptian National Library.]
altitude at the beginning of the afternoon prayer and its Cotangent, the time from
midday to the beginning of the afternoon prayer and the time remaining till sunset,
and also the number of equatorial degrees in the seasonal day-hours, and the duration
of morning and evening twilight, (all) as functions of solar longitude.
I arranged these tables in a satisfactory fashion,i and added to them others giving
the normed right ascensions, regular right ascensions and oblique ascensions for
latitude 30, as well as tables of Cotangents to base 60, 12 and 7. j If you enter these
with the altitude you obtain the horizontal shadow of a vertical gnomon, and if you
enter with the complement you obtain the vertical shadow of a horizontal gnomon.
If you take one-half of the Cotangents to base 12 you obtain Cotangents to base 6,
and if you take one-half of one-sixth of the entries you get the function which, with
the declination as argument, is multiplied by the same function, that is the Cotangent,
of the equinoctial meridian altitude or the Tangent of its complement (to find the Sine
of of the half excess of daylight).k (I also added) a table of the number of equatorial
hours of daylight for latitude 30.
the development of the main cairo corpus 293
If you enter the solar longitude under the appropriate solar altitude in the tables of
the time since sunrise and hour-angle you get for eastern altitudes the time since sunrise
and the time till midday, and for western altitudes the time till sunset and the time
since midday. The sum of the time since sunrise and the hour-angle is the semi diurnal
arc. If you add the time from midday to the semi diurnal arc you get the time to sunset
for eastern altitudes and the time since sunrise for western altitudes. As for the azimuth,
it will be northern if (the solar longitude) on the given page is before the longitude
at which the azimuth changes direction (from north to south), and it will be southern
if the solar longitude is after this particular longitude. For altitudes in the east the
azimuth moves from north to south and for altitudes in the west the opposite is the
case. But God knows better.
In this book (I have also included) the universal table which I computed for finding
the time: this is for obtaining the (hour-angle) [text: solar longitude] according to the
admirable method of Ibn Ynus. I know of no easier method, and he had a treatise
(on the subject) which was unrivalled.l I also computed a star catalogue, in which
the declinations, right ascensions, and degrees of transit can be used in all localities,
but the arcs of visibility, oblique ascensions of their risings and settings, as well as
their meridian altitudes, are for the latitude 30 only. The text (explaining their use)
is at the end of this book.m God knows the truth best and, knowing all our works, is
our best Judge.
Notes:
(a) This sentence implies that al-Bakhniq had seen some tables of t(h,) computed for
Cairo but preferred to prepare a new set.
(b) The expression bi-l-tabuq means using the relation:
T(h,) = D() - t(h,)
to compute the entries.
(c) The text has them both rather than him, and the phrase should probably occur after
the name of Ibn Ynus, and refer to both him and al-Rashd. On the other hand, it
may refer to al-Rashd and his father, but his grandfathers name is also mentioned.
(d) The text has incorrectly Abu l-usayn.
(e) This could refer to either Ibn Ynus or Ibn al-Rashd the latter seems the more likely.
(f) The Arabic term jmi{ seems to indicate that this supplementary table together with
the main azimuth tables display the solar azimuth for all longitudes.
(g) I read the Arabic as uayfa, but since the subject is plural the word should be uyafat.
Note that al-Bakhniq did not use the word afa, which would refer to Ibn al-Rashd.
(h) These are the tables of the functions D, ta, Tq and 1/ see below.
(i) This seems a reasonable rendering of rattabtuh tartban. The word {ajb generally
means wonderful / admirable / strange.
(j) Called sittn, iba{ and qadam, sexagesimal, digit, and foot, respectively. These
expressions are standard.
(k) This procedure is equivalent to the formula:
Sin d = 1/12 Tan60 Cot12 ,
294 part ii, chapter five
being the declination of a star or the sun, and being the co-latitude. See I-7.0 and
7.1.
(l) The text has darajat al-shams solar longitude rather than hour-angle, a curious
error. al-Bakhniq may have been referring to the function:
G() = R2 / { Cos () Cos }
discussed by Ibn Ynus in Ch. 15 of the kim Zj (see also I-6.7.1).7 The term risla
means treatise or method in medieval scientific Arabic: it probably refers to the
passage in the kim Zj.
(m) al-Bakhniqs star catalogue and notes thereon have not been identified in any of the
known manuscripts of the corpus.
Thus it appears that al-Bakhniq rearranged the tables of Ibn Ynus and al-Maqs as follows.
Taking Ibn Ynus tables of a(h,) corrected by Ibn al-Rashd (5.2), al-Maqss tables of T(h,)
(5.3), and Ibn al-Kattns tables of t(h,) (5.4), he then tabulated the values (T,t,a) as triplets
for each pair of arguments (h,). This arrangement of the tables is found for example, in MSS
Cairo DM 108, Cairo DM 53, Istanbul Nuruosmaniye 2925, Cairo DM 616, and Cairo DM
739. If I understand al-Bakhniq properly he is saying that Ibn al-Rashd added the tables
of:
, h0, hv, hq, hq* and a(h) for = 0, 90, 270
before the azimuth tables, and that after the azimuth tables he had put the tables which al-
Maqs had at the end of his book, namely, tables of:
D, ta, Tq (q = 53), and 1/ ,
as well as the tables of:
tq, H, ha and za, ta and Ta, 2Dh, r and s .
But if Ibn al-Rashd had indeed added all these tables, why is he not generally acknowledged
as the author of all the minor tables in the corpus? His name is specifically mentioned in the
tables of the corpus only in relation to the two tables of hv() and a(h) at the equinoxes and
solstices.
To al-Maqss introduction al-Bakhniq added the note translated above, describing what
he had done to the tables. The five manuscripts listed above contain this modified introduction
as well as the tables of (T,t,a), but MS Cairo DM 108 also contains a variety of other tables.
In al-Bakhniqs note in that copy (DM 108), but not in MS Cairo DM 53, he states that he
included tables of , and ( = 30). These are not contained in MS Cairo DM 108, but
in later sets of prayer-tables for Cairo we find, for example, Ibn Ynus tables of (MS Cairo
MM 43) and ( = 30) (MS Princeton Yahuda 861,1), and al-Marrkushs tables of (
= 30) (MSS Hartford TS 621 and Paris BNF ar. 2507). al-Bakhniq further asserts that he
included tables of the Cotangent function to bases 60, 12, and 7: these are likewise not contained
in MS Cairo DM 108. He also mentions Cotangents to base 6 (cf. 2.9).8 Finally, al-Bakhniq
states that he himself added a table of to the corpus (extant in MS Cairo DM 108) as well
as a star catalogue and a table for finding the hour-angle (if my correction hour-angle for
solar longitude is correct), which may have been some kind of auxiliary table (both not extant
in the known sources).
MS Cairo J 367,5, fols. 35r-40r, copied about 1450, contains a short introduction and a solar
longitude table followed by tables of the functions:
r, H, hq, h0, , d, and
from the corpus. The work bears the title Natjat al-afkr f a{ml al-layl wa-l-nahr and is
attributed to the 15th-century Egyptian astronomer al-Waf} (1-9.9). The solar table (fol. 37v)
is not complete and the prayer-tables which follow (fols. 38r-40r) are in the same hand but
are not mentioned in al-Waf}s introduction. If the title, which mentions timekeeping by day
and night, is original, it seems probable that the prayer-tables were included in the work by
al-Waf} himself. If it is indeed the case then it was al-Waf} who set the trend followed
by later Egyptian astronomers such as al-Ldhiq, al-Ikh, Ibn Ab Rya and Riwn Efend
(7.8-10), of presenting updated tables for calendar conversion and finding the solar longitude
for a given date, and including selected tables from the already available corpus.
MS Cairo DM 157,2, copied ca. 1450, contains a set of tables of the function (), the
oblique ascension at the time of the salm, computed for each six minutes of argument with
entries given to minutes. These tables are specifically attributed to al-Waf}. A similar set of
tables for the time of nightfall is attributed in MS Cairo DM 1218, copied ca. 1500, to
Muammad al-Ghazw, a muwaqqit in the Muqassam (?) Mosque in Cairo, whose name is
new to the modern literature. MS Cairo MM 209,1 contains a set of normed right ascensions,
() ( = 23;35), with values for each minute of argument attributed to Ibn al-Mushrif (I-
9.8 and II-6.15), copied in 873 H [= 1468/69] by Ibn Abi l-Fat al-f (6.15) from a copy
by the author dated 848 H [= 1444/45]. Apart from the earlier ascension tables of Baylak al-
Qipjq (6.4), all other Islamic tables of this kind, which are a particularly Egyptian
phenomenon, and in which values for each degree of solar longitude are stretched to each six,
three, or even one minute of argument, date from the period of Ibrhm ibn Qyitby and {Abd
al-Qdir al-Minf, namely the late 16th and early 17th centuries. See further 7.2.
MSS Cairo MM 181,2 (fols. 4v-5r, copied 1030 H [= 1620/21], TR 303,2 (fol. 33v, copied
1266 H [= 1849/50]), and MM 155,2 (fol. 23v, ca. 1700, defective) are copies of a short treatise
(nubdha) attributed to al-Waf} (at least in the first source) concerning the daq}iq al-ikhtilf,
that is, the the time taken by the sun to cross between the visible and true horizons (4.11).
The first two sources are displayed in Fig. 5.7. In the treatise it is stated that the values:
EQ: 47 SS: 62 WS: 32
were derived by the Damascus astronomer Ibn al-Shar (9.3). This is curious for the following
reasons:
(1) the Minfs, father and son, attributed these values to Ibn Ynus (7.1 and 8.2);
(2) Ibn al-Shar is known to have written that he had observed the half arcs of visibility
of the sun and stars and found that they were greater than the arcs found by calculation
by an amount exceeding 2/3 (9.3), but is not otherwise known to have written more
296 part ii, chapter five
Fig. 5.7b: The same notes abridged and without
the attribution to al-Waf}, but note that in both
these Cairo manuscripts the theory is attributed to
the illustrious Ibn al-Shir. [From MS Cairo TR
303,2 (fol. 33v), courtesy of the Egyptian Na-
tional Library.]
the development of the main cairo corpus 297
Sib al-Mridn (2.10), in his treatise on sexagesimal arithmetic entitled Raq}iq al-aq}iq
f isb al-daraj wa-l-daq}iq, The Finest Truths concerning Sexagesimal Arithmetic, dis-
cusses interpolation9 in tables for timekeeping. The relevant passage (I have used MS Gotha
A1390, fol. 33v) translates as follows:
The conclusion of this work deals with interpolation (ta{dl m bayn al-sarayn). The
tables of the semi diurnal arc, the half excess of daylight, solar rising amplitude,
ascensions, and the like, are generally computed for each integral degree of argument
although they may also be computed for each minute. Certain tables are generally
computes for each minute, such as the (Co)tangent, Sine and Versed Sine. Entries in
these are usually given for argument differences of a minute. Likewise the tables of
the inverse (Co)tangent, Sine and Versed Sine also have entries for each minute of
argument. In tables of this sort there is no need to interpolate... .
It is of interest that he mentions tables of spherical astronomical functions and trigonometric
functions computed for each minute of argument: see below.
Sib al-Mridn then gives examples of how to use linear interpolation in a table of, say,
D() for non-integral . He also shows how to interpolate in double-argument tables, quoting
an example from his teacher, that is, the celebrated Ibn al-Majd (1359-1447).10 To find the
solar azimuth when the altitude is 19;24 and the solar longitude is 53;20 he interpolates in
9 See n. I-1:21.
10 On Ibn al-Majd see Suter, MAA, no. 432; Cairo ENL Survey, no. C62; and also King & Kennedy, Ibn
al-Majds Tables, pp. 48-49.
298 part ii, chapter five
the azimuth tables for altitudes 19 and 20 as follows. The relevant entries which he cites
(all accurate) are:
h = 19 20
= 53 11; 2 10;31
54 11;20 10;49
(as in Ibn Ynus tables), and our author computes what he calls al-ta{dl al-awwal and al-
thn, the first and second interpolations:
a(19, 53;20) = 11;8 and a(20, 53;20) = 10;37 ,
from which:
a(19;24, 53;20) = 10;56 .
He states that the accurate value of a is 10;55,36, which is just one second too high, and calls
this kind of interpolation ta{dl al-tajyb. He makes no mention of the author of the tables.
Sib al-Mridns remark that certain tables for timekeeping had entries for each minute
of argument (solar longitude) is confirmed by the tables of Ibn al-Mushrif, Ibrhm ibn
Qyitby and al-Waf} (5.7). The Sine and Cotangent tables attributed to Ibn Ynus give values
also to each minute, but to five sexagesimal digits.11 By the 15th century such tables were
standard in the major centres of astronomy in the Islamic world, the most popular ones being
the Sine and Cotangent tables from the Zj of Ulugh Beg.12 However, there was also an
independent Egyptian tradition of tables of the standard trigonometric functions with entries
to three digits for each minute of argument, of which several manuscripts survive: see Fig.
I-1.0e for an example.13
CHAPTER 6
Virtually no Egyptian astronomical works survive from the period between Ibn Ynus and the
middle of the 13th century. (This accounts, for example, for the importance of the treatises by
Ibn al-ammm discussed in 2.4.) For the history of astronomy in Egypt during that period
we have little more than biographical references to various individuals, and al-Maqrzs
account of the observatory that was built in Cairo in the 12th century.1 From late 13th-century
Cairo we have the A-Z on spherical astronomy and instruments by Ab {Al al-Marrkush,
the treatise on sundial construction by al-Maqs, and a zj called al-Zj al-Muala. al-
Marrkushs treatise enjoyed considerable popularity in later centuries in Egypt, Syria, and
Turkey. I discuss his spherical astronomical tables in 6.7. In his encyclopedia Irshd al-qid,
the early-14th-century scholar Ibn al-Akfn states that the two books on mqt that were used
(in Egypt) in his time were a short work Naf}is al-yawqt and the longer work Jmi{ al-
mabdi} wa-l-ghyt of Ab {Al al-Marrkush.2 He does not mention the author of the first
work, but it may be that the Durar al-yawqt described in 3.2 is identical with, or at least
related to the Naf}is al-yawqt. Ibn al-Akfn also mentions that the closest astronomical
observations to his time were those conducted under Hulagu by Nar al-Dn al-s and his
associates, and that in Egypt the yearly ephemerides were prepared from a zj which had been
compiled from a number of other zjes and which was called the Muala Zj. In 6.6 I discuss
the spherical astronomical tables found in the two extant versions of this work. Various
Egyptian astronomers compiled auxiliary tables for facilitating the solution of problems in
spherical astronomy. The historical context of the monumental table of Najm al-Dn al-Mir
(6.5) is still something of a mystery, even though Franois Charette has produced a detailed
study of the table itself and gathered all the known material relating to its author. It seems
that the tables of al-Kha{ (6.8) do not, after all, and as one might have expected, precede
those of al-Khall of Damascus (10.3). al-Khalls tables were also used in Egypt (6.14). All
of the other Egyptian auxiliary tables postdate those of al-Khall (6.15). Egyptian astronomers,
in particular Ibn al-Rashd, compiled timekeeping tables for the latitudes of Mecca and
Jerusalem (6.11 and 6.12).
ibn Amad ibn {Al ibn al-asan al-Ifahn and two other shorter works in poetry and
calligraphy.3 A treatise on astronomical timekeeping, written in a different hand but datable
to the 14th century, has been bound in the manuscript (fols. 94v-130r) after the astrological
work. It is entitled Kitb al-Durar wa-l-yawqt f {ilm al-raad wa-l-mawqt, Book of
Pearls and Sapphires on Astronomical Observations and Timekeeping. On the title pages (fols.
94v-95r) the authors name has been partially deleted, but the following letters are still legible:
Ab A... ... ibn ...d ibn Amad, al-.... al-... (perhaps, al-Shfi{ al-madhhab, referring to his
adherance to the Shfi{ legal school) al-Mir al-wan (the Egyptian). This name, however,
does not correspond exactly with the authors name which appears on fol. 97v, namely: [Abu]
l-{Abbs Amad ibn Ab {Abdallh Muammad ibn Amad. The copying of the manuscript
was completed in 734 H [= 1334], and the name of the copyist was Abu l-asan {Al ibn
(?) Muammad ibn (?) Bar (?). The original work can be dated to the early 13th century (see
below).
The treatise is a complete muwaqqits manual for timekeeping by the lunar mansions and
regulating the prayer-times approximately (see also V-5). It may be that Ibn al-Akfn, writing
in the early 14th century, was referring to this work when he stated that the book on simple
timekeeping which was currently most widely used in Egypt was called Naf}is al-yawqt
(6.0). The treatise begins with a set of prayer-tables for Cairo (fols. 98r-104r) of a kind not
attested in any other known Egyptian source: see Fig. 6.1. For each day of the Coptic year
and corresponding date in the Syrian calendar the lunar mansion rising at daybreak is indicated
and following functions are tabulated:
darajat al-shams
H ghyat al-irtif{
Z(12) ill al-zawl
ha irtif{ al-{ar
2D qaws al-nahr
2N qaws al-layl
hq samt al-qibla
zq(12) (= Cot12 hq) ill al-samt
r iat al-fajr.
The values of H, 2D and 2N are given in degrees and minutes, and the remaining functions
are displayed to the nearest degree or digit. From the values of H the parameters are found
to be:
= 30 (Cairo) and = 23;35 .
All of the functions are approximate. The values given for H for each 30 of starting from
the winter solstice are:
36;25 40;8 48;4 60;0 70;36 79;56 83;35 ,
but starting from the summer solstice are:
83;35 80;0 72;6 60;0 48;4 40;20 36;25 .
The values of vary in each of the four quadrants of the ecliptic! In the text the author states
3 On the astrological work see Ullmann, Die Natur- und Geheimwissenschaften im Islam, pp. 344-345; Sezgin,
GAS, VII, pp. 24-25; and the detailed study of the art-historical aspects in Carboni, Kitb al-bulhn.
other early egyptian tables for timekeeping 301
Fig. 6.1: The prayer-tables for the months of Kayhak and ba in the Kitb al-Durar wa-l-yawqt. [From MS
Oxford Bodley Or. 133, fols. 99v-100r, courtesy of the Bodleian Library.]
that varies by 12, 8, and 4 for each 30 of measured from the equinox, but he has not
used these figures in his tables. The scheme underlying the table of 2D, which is outlined
elsewhere in the text, is that the function increases by 0;30 for each degree of Aries, 0;20
for each degree of Taurus, and 0;10 for each degree of Gemini, etc. This explains, for example,
the extremal values 150;0 and 210;0 (accurately, 150;48 and 209;12). The function Z(12)
is made to assume values:
16 13 10 7 4 2 1;21
for each 30 of solar longitude starting from the winter solstice (entries less than 2 are given
to two digits). The underlying scheme is clearly very crude, although the (accurate) value 1;21
for the summer solstice was probably copied from another source. The accurately computed
shadow lengths are:
16;16 14;26 10;38 6;56 4;0 2;03 1;21
The values of ha for each 30 of starting at the winter solstice are:
21 25 26 31 37 40 42 ,
compared with the accurate values:
23 24 28 33 37 40 42 .
In the text the author advocates the approximate rule (cf. 2.9):
ha 1/2 H + 1/10 (83 - H) ,
which gives values:
23 24 27 32 37 40 42 .
The tabulated function hq assumes values:
13 19 35 45 58 71 80
for each 30 of starting at the winter solstice. The corresponding values in Ibn Ynus table
of hq (q = 52) and the other table of hq in the Cairo corpus (q = 53) round to:
302 part ii, chapter six
14 19 31 47 63 75 80
13 18 30 46 62 75 80 ,
respectively. Using the approximate formulae for hq found in the treatises of Najm al-Dn al-
Mir and al-Bakhniq (2.5 and 2.9), namely:
hq 47 + 1;25 and hq 47 + 1;24 :
we obtain values:
14 18 31 47 63 76 80 .
Again, using linear interpolation between the extremal values 13 and 80, we obtain:
13 24 35 47 58 69 80 .
Thus the tabulated values correspond to none of these approximations.
No other known Egyptian prayer-tables display the function zq(12) (= Cot12 hq), using which
one can determine the qibla from the gnomon shadow.
The tabulated values of r are:
22 21 21 24 25 25 27
for each 30 of starting at the winter solstice. It is difficult to explain these values. For
parameters 18, 19, and 20, the corresponding accurate values are respectively:
22 22 21 21 22 24 25,
24 23 22 22 23 26 27
25 24 23 23 25 27 28 .
Note that the compiler realized that r displays a secondary maximum at the winter solstice.
Most of the remainder of the treatise and the other tables are for timekeeping by the lunar
mansions. Some of the tables resemble those found in a later Yemeni almanac (12.5). Thus
there are tables displaying the names of the lunar mansions rising, culminating and setting at
nightfall, and at one-third, one-half, two-thirds and three-quarters of the night, as well as at
daybreak, for each 13 days of the Coptic year. Other tables display the 13th divisions of the
lunar mansions rising at daybreak and the solar longitude for each day of the Coptic year (the
entry in the latter for Tt 1 is Virgo 14). The longitudes of 23 prominent stars are given; these
correspond to the early 13th century (e.g. the longitude of Aldebaran is 2s 29;37), enabling
us to date the compilation of the treatise to the early 13th century.
MS Damascus hiriyya 10732, fols. 3r-3v, copied ca. 1600, contains some simple prayer-
tables for an unspecified latitude with values of the functions:
D, ta, Ta, s, r, n and 2N
to one degree for each zodiacal sign. The tables occur on a single folio that may be related
to the text that immediately precedes and follows, namely, simple material on the lunar
mansions and shadow lengths attributed to the late-12th-/early-13th-century legal scholar and
theologian of Jerusalem, Ibn Qudma.4 Whether or not this is the case, the values in the table
can be derived from those in the main Cairo corpus.
4 On Ibn Qudma see the article Ibn ^udma in EI by George Makdisi. On the shadow scheme associated
2
with him see III-8.1.
other early egyptian tables for timekeeping 303
MS Cairo DM 187 (74 fols., copied ca. 1850) is the only known copy of an anonymous almanac
based on the Coptic calendar of the same kind as the Calendar of Cordova,5 and accompanied
by various simple calendrical and astrological tables. This almanac merits detailed investigation.
I am unable to ascertain the date of its compilation, but have noted a list of planetary apogees
amongst the tables, as follows:
Saturn Jupiter Mars Sun (and Venus) Mercury
245173 125 83 202
These correspond very closely to those in the 9th-century Abbasid Mumtaan Zj, which are
as follows for the year 199 Yazdigird [= 831]:6
244;30 172;32 128;33 82;39 201;0
On the other hand, some prayer-tables contained in this work see Fig. 6.3 do not, I think,
predate the 14th century. These display values of the following functions for each five days
of the Coptic month:
D (al-uhr), ta (al-{ar), Ta (al-maghrib), s (al-{ish}), n (jawf al-layl), r (al-fajr)
Values are given in degrees and also in hours (misspelled sa{t !) and minutes, and correspond
to those in the main Cairo corpus. The twilight tables are based on 19 and 17, which are
not attested in Egyptian sources before the 14th century.
At the end of the almanac there are various notes, including one expounding some ap-
proximate formulae for timekeeping, of which the first three are not attested elsewhere. These
are as follows:
ta 4/7 D , s 1/8 (2N) , r 1/7 (2N) and ha 1/2 H + 1/2 1/6 [90 - 1/2 H] .
Fig. 6.3: The part of the prayer-tables in an anonymous Egyptian almanac serving the months of Baramht and
Barmda. [From MS Cairo DM 187, courtesy of the Egyptian National Library.]
5 This is published with translation in Charles Pellat, Calendrier de Cordoue. On the shadow schemes in
that work and also those in MS Cairo DM 187 see III-5.1 and 9.7a.
6 On the Mumtaan Zj see n. I-4:7. The apogees given are extracted from Caussin, Table Hakmite (cited
in n. I-2:3), pp. 234 ff.
304 part ii, chapter six
MS Cairo MM 41, copied in 1191 H [= 1777/78], contains a set of tables of the right ascensions
for each minute of argument computed to three digits, attributed to the well-known scholar
Baylak ibn {Abdallh al-Qipjq7 and stated to have been compiled in 674 H [= 1274/75]. The
manuscript contains 41 folios of tables. There is an introduction by Sayf (?) al-Dn Ibn al-
Mushrif (I-9.10).
The tables of either of the two functions t(h,) or T(h,) which were compiled for particular
latitudes each contain over 10,000 entries. To compile a table displaying either function for
each degree of latitude would be an enormous undertaking. Nevertheless, around the year 1325
the Egyptian astronomer Najm al-Dn al-Mir (2.5) compiled a table displaying for any latitude
the time since rising of either the sun or any non-circular star as a function of the solar or
stellar altitude. This remarkable table is the largest known table from the entire medieval period
and contains over 440,000 entries.
MSS Cairo MM 132 and Oxford Marsh 676 (Uri 944 = 995) are two halves of a unique
copy of Najm al-Dns tables, copied ca. 1325 probably by the author. The arguments which
one feeds into the main table to find the time since rising of any celestial body for any latitude
are H, h, and D: for details see I-2.6.1. The table can also be used for solving all problems
of spherical astronomy for any latitude: for details see I-9.3*. Other tables are included to
facilitate the determination of H and D for any latitude, starting from the solar longitude
or stellar declination . Solar tables based on Ibn Ynus parameters, and are a star catalogue
giving the equatorial coordinates of 367 stars, are also provided.
The existence of Najm al-Dns tables raised certain questions. What, if any, was his
relationship with al-Maqs and al-Marrkush? What, again if any, was his relationship with
the Muala Zj ? What were his sources? One surely does not just sit down to compile a
table containing over close to half a million entries which works for all latitudes, without having
seen smaller tables for specific latitudes. If Najm al-Dn preceded al-Maqs, as I believe he
did, then there were not yet any tables for T(h,) available for Cairo. No tables of T(H,h), the
so-called aylasn zjes, are known to have been compiled for Cairo, although they had been
used by astronomers elsewhere for centuries. I suspect that Najm al-Dns inspiration for his
main table might have come from some tables of Ibn Ynus which are no longer extant.
Suppose, for example, that Ibn Ynus had compiled a table of:
G(H,D) = Vers D / Sin H ,
for each degree of both arguments. The values of T(H,h,D) can then be generated using the
simple formulae:
7 On Baylak see the article by Martin Plessner in DSB, and Cairo ENL Survey, no. C9. The unique manuscript
of a treatise on mechanical clocks by Riwn al-Khurasn (MS Istanbul Kprl 949 see Cairo ENL Survey,
no. C1) was copied by Baylak in 658 H [= 1260]; this treatise is now published see Hill, Studies, p. xix, n.
4.
other early egyptian tables for timekeeping 305
The Muala (Technical) Zj appears to have been the most popular Zj from the 13th to
the 15th centuries (6.0). The 17th-century Turkish bibliographer jj Khalfa attributes it to
an individual named Muammad ibn Muammad al-Friq,9 on whom I have no further
information, and he notes that it is based on the Zj of Ibn Ynus. Two recensions of this work
are known to me, MSS Paris BNF ar. 2513 from the 13th century and Paris ar. 2520 from the
14th century, and both are incorrectly attributed to Ibn Ynus on their title folios. In the former,
the zj is entitled Kitb al-Ta{lm f wa{ al-taqwm, Book of Instruction on Finding Planetary
Positions, and in the latter Kitb al-Zj al-Muala. The two copies are by no means identical,
but their relationship to each other is firmly established by the similarity of the introductory
text and the planetary tables, if not the spherical astronomical tables.
The planetary tables in the two Paris manuscripts are a motley collection lifted from various
sources, including the Zj of Ibn al-A{lam,10 the Zjes of Ibn Ynus, the Zj of al-Kammd,11
and a zj whose name is variously written as Shh, Shw, and Shm, and which may be the
Shh Zj of um al-Dn Slr.12 There are also tables in both manuscripts taken from
Abbasid sources such as Mumtaan Zj (solar eclipse table, MSS ar. 2513, fol. 48r, and ar.
2520, fols. 58v-59r),13 the Zj of abash (auxiliary tables for spherical astronomy, MS ar. 2520,
fols. 70r-71r),14 and another early zj, perhaps of Andalus origin (lunar crescent visibility table
for the seven climates based on Indian theory, MS ar. 2513, fol. 71v).15 MS ar. 2520 seems
to be closer to the original than MS ar. 2513. In the former the mean motion tables run from
630 H [= 1232] and in the latter from 750 H [= 1349]. In MS ar. 2513, fol. 11r (but not in
MS ar. 2520) there is a numerical example of the determination of the solar longitude for the
year 682 H [= 1283]. I suspect that further investigation of these and related sources will reveal
a connection with the scholar Ibn al-Lubd, who lived in the mid 13th century and is known
to have moved from Damascus to Cairo.16 He is also known to have compiled two zjes, one
of which was based on the Shh Zj. At this stage of the research it is perhaps not amiss to
mention an anonymous Yemeni zj preserved in MS Paris BNF ar. 2523, copied in the late
14th century, which was compiled in Taiz in 775 H [= 1374] and contains planetary tables from
the kim, Shh and Muntaal (?) Zjes.17 Thus the combination of kim and Shh tables
is attested in both Egypt and the Yemen, but we also know that the Yemeni astronomers of
the 14th century were familiar with the Muala Zj.18
The potential value of each of these manuscripts for furthering our knowledge of Islamic
astronomy is obvious, and all of them deserve detailed study. My present purpose is to briefly
note certain of the spherical astronomical tables which occur in the two copies of the Muala
Zj and relate to our subject. Most of them are not found in any other Egyptian sources currently
known to me.
Geographical coordinates
The latitude of Cairo is taken as 30;0 in the spherical astronomical tables in the Muala
Zj. However, MSS Paris ar. 2513 and 2520 contain two different sets of geographical tables,
the former also with values of q,19 as is shown by the following entries:
Source 2513, fols. 87v-88v 2520, fols. 82v-83r
Locality L q L
Cairo 30;0 55;0 52;30 29;55 54;55
Damascus 33;27 60;0 39;50 33;30 60;0
Jerusalem 32;0 57;50 43;30 32;10 57;20
Mecca 21;20 67;0 - 21;30 67;0
Note that the distinctive value 29;55 for the latitude of Cairo was also used by al-Marrkush
(6.7), and that the distinctive value 33;27 for the latitude of Damascus was also used by al-
Mizz (9.2). Accurately computed for the coordinates displayed in MS ar. 2513, the three qibla
values are respectively 54;2, 28;51, and 39;42 using the exact formula and 54;4, 30;8,
and 40;43 using the standard approximate formula. Thus the stated qibla values are wretchedly
inaccurate, or based on other coordinates.20 al-Marrkush (6.7) states that the qibla value
52;30 for Cairo was widely accepted there, and it may be that it was conceived as a happy
compromise between Ibn Ynus first and second values, 52 and 53.
21 See King, Earliest Muslim Geodetic Measurements, pp. 226-227, on Ibn al-A{lams determination of
the latitude of Baghdad as 33;21.
308 part ii, chapter six
al-Marrkushs summa on spherical astronomy and instruments was the most important single
work on this subject to be compiled in medieval times (see already 2.7 and 6.0). It was also
the most influential, if only in Egypt, Syria and Turkey. I here describe only the tables presented
by al-Marrkush and discuss his remarks on the qibla at Cairo and on the twilight.
The treatise begins with a discussion of calendars and trigonometry, with approximate tables,
as well as a table displaying the solar longitude for the year 1275/76 and a star catalogue for
1282. al-Marrkush was of the opinion that the obliquity of the ecliptic oscillated between
23;53 and 23;33 and that there was an associated oscillatory motion of the equinoxes on the
ecliptic. For the present discussion it suffices to note that in all of his tables which are based
on a value of the obliquity he used = 23;35.
al-Marrkush tabulated () to two digits for each degree of (I.183). Nine of his ninety
entries are in error by -1 in the second digit, which suggests he rounded the entries in a table
of () to three digits. He used the approximate value Sin = Sin 23;35 24 instead of the
more accurate value 24;0,18 in his numerical example. He also tabulated 2(), the second
declination, in similar fashion (I.184). al-Marrkush then tabulated Tan () for = 23;35
to three digits for each degree of (I.185) see I-7.1.7. His entries are again considerably
less accurate than those in the corresponding table in the kim Zj. In particular he uses Tan
(90) = Tan 23;35 = 26;11;40, a value which can be obtained from interpolating between
correct values of Tan 23 and Tan 24. Ibn Ynus has the more accurate value 26;11,33,(23).
other early egyptian tables for timekeeping 309
al-Marrkush also tabulated () to two digits for each 0;15 of up to 23;35 (I.186/187).
Two rather unusual tables (see further I-7.1.7) presented by al-Marrkush (I.209-210)
display the functions:
f() = 1/12 Tan60 () and g() = 1/12 Tan60
to three and two digits respectively for each degree of arguments and . The purpose of these
tables is to determine the half excess of daylight using only another of Cotangents to base 12,
thus:
Sin d() = f() Cot12 and Sin d() = g() Cot12 .
The use of these functions reflects a strong sentimental attachment to the base 12, which was
characteristic of the muwaqqits of medieval Egypt.
al-Marrkushs next tables in the treatise display d(,) for each 6 of both arguments
(I.214-215). His entries for = 30 are slightly different from the corresponding entries in the
main Cairo corpus. His tables of () to three digits for each degree of (I.222-225) are less
accurate than those which can be derived from the corresponding values of () in the kim
Zj. He also tabulates () to degrees and minutes for each 6 of both arguments and (I.230-
235), whereas Ibn Ynus tabulated them for each 1 of both arguments. al-Marrkush does,
however, include a table of () for = 30 with entries for each degree of in degrees and
minutes (I.238-240), which again are less accurate than those in the kim Zj. Another table
of al-Marrkushs displays the length of the longest day in hours, minutes, and seconds for
each degree of up to 66 (I.246).
al-Marrkush presents a number of tables scattered throughout the remainder of his treatise
which relate the solar altitude and shadow to the time of day. All of the entries in these tables
are to two digits, and the functions of time have T = 1, 2, ... , 5 or 6sdh as arguments and
sometimes also the time of the afternoon prayer. The tables which are independent of terrestrial
latitude are of the functions:
h(T,H) H = 5, 10, ... , 90 (I.457)
z(12)(T,H) H = 5, 10, ... , 90 (I.253)
z(12)(T,x) x = Z(12) = 1, 2, ... , 12 (I.254)
z(12)(T,x), z(12)(T,x) x = Z(12) = 1, 2, ... , 36 (I.256-257)
ha(H), Sin ha(H), za(12)(H) H = 5, 10, ... , 90 (I.271)
The tables specifically for = 30 are of the functions:
h(T,) = 270, 300, ... , 90 (I.454)
z(12)(T,), z(12)(T,) = 270, 280, ... , 90 (I.428-429, 436-437)
h0(), () = 30, 60, 90 (II)
hq( ) = 270, 300, ... , 90 (II)
Finally there is a table displaying h(T,) for each equinoctial hour from 1 to 6 and each 30
of . Each of these small tables for timekeeping is of limited practical use. See further I-4.2.4.
The remaining tables in the treatise are specifically intended for use with sundials, and do not
concern the present study.
Muammad ibn al-Amr Fakhr al-Dn {Uthmn al-Kha} is an individual whose name is new
to the modern literature; his father was a Mamluk prince who flourished around 1450 (I-6.15.1).
His auxiliary tables, which contained about 825 entries, are partially extant in MS Vatican Borg.
ar. 217,2 (fols. 6r-7v), copied ca. 1500. They are less general in their application than other
later Egyptian and Syrian tables, being specifically intended for finding the hour-angle from
the solar altitude at the latitude of Cairo. The functions tabulated are related to those of al-
Khalls minor set of auxiliary tables, but the two sets may have been compiled inde-
22 See Kennedy & Kennedy, Islamic Geographical Coordinates, p. 111, and now King, Geography of
Astrolabes, p. 9 and n. 23.
other early egyptian tables for timekeeping 311
pendently. al-Kha} also tabulated ha(), and a(h) at the equinoxes. His table of the first
function is more accurate than the ones in the main corpus and the Muala Zj (4.4 and 6.6).
See further I-6.15.1.
MS Cairo MM 72, copied in 747 H [= 1346/47] by Ibn al-Kattn (5.5), contains a curious
set of tables preceded by an introduction of which unfortunately only the last page remains.
The function tabulated is not explained in the surviving portion of the introduction but in fact
it is the solar altitude for values of the time since sunrise and different values of the semi diurnal
arc, that is, h(D,T). Values of this function to two digits are given for the domains:
D = 104;36, 104, 103, ... , 76, 75;24 and T = 1, 2, ... , [2D] ,
where [2D] is the largest integer less than 2D. The underlying parameters are:
= 30;0 (Cairo) and = 23;35 ,
as in the main Cairo corpus. (Note that 104;36 and 75;24 are the extremal values of D for
these parameters.) For each double page of tables serving a value of D, a value of the function
ta(D) is given alongside the argument D at the head of the tables. The altitude of the sun at
each seasonal and equinoctial hour of daylight is also displayed for each value of D at the side
of the main tables. I suspect that these tables may be by Najm al-Dn al-Mir (6.5): more
information is in I-4.4.1 (illustrated).
MS Cairo MM 68, copied ca. 1500, contains an anonymous set of prayer-tables computed for
latitude 21, that is, for Mecca.23 The work bears no title and no date, but I do not doubt that
it was compiled in Egypt. Establishing the identity of the compiler of these tables would
probably solve some of the remaining problems associated with the main Cairo corpus. The
format of the tables in this Cairo manuscript is the same as that of the main corpus, and the
functions tabulated are:
H, D, ha, ta, Ta, 2N, (2n -r), r, s, s and .
The underlying value of is 23;35 and the tables are rather accurately computed.
The tables for latitude 21 in the corpus of prayer-tables for all latitudes preserved in MS
Princeton Yahuda 861,1 (8.1) are quite different from these, being very carelessly computed.
The table for in the Cairo manuscript is identical with that of Ibn Ynus for latitude 21
in MS Leiden Or. 143 of the kim Zj, containing the same slight errors. The tables for
twilight, that is, of the functions r and s, are based on parameters 19 and 17, and do not display
the error made by al-Mizz in his tables for Damascus (9.2). MS Cairo MM 68 also contains
a table of a function called al-al al-mulaq, the absolute base, defined by:
B() = Cos Cos / R .
Values are given to two digits for each 0;5 of argument up to = . (See further I-6.4.4.)
Fig. 6.10: The values above the table of the half arc of daylight represent the time taken by the sun to move
between the true and the visible horizons. [From MS Cairo MM 68, courtesy of the Egyptian National Library.]
At the top of the table of D() is a set of numbers for each of the signs (see Fig. 6.10) .
They are labelled daq}iq ikhtilf al-fq al-mar}iyya, the difference minutes of the visible
horizons, and it is clear that they are intended to represent the effect of refraction at the horizon
on the arc of daylight, that is 2(D), twice the difference minutes of the Cairo astronomers
(4.11). The values given for the twelve signs starting with Aries are as follows:
1;11 1;18 1;25 1;33 1;25 1;18 1;11 1;3 0;56 0;48 0;56 1;3
Thus the equinoctial value is intended to be 1;11 and the values for the summer and winter
solstices are respectively 0;22 more and 0;23 less. The equinoctial value corresponds to an
angle of depression below the horizon of approximately:
(1/2 1;11) cos 21 = 0;33 ,
but this is not what is intended. Consider the values of D for = 21 which are given by
al-Minf (8.2). These yield the following rounded values for 2(D), according to al-Minf
(A) and according to earlier scholars (B):
A: 1;6 1;13 1;20 1;27 1;20 1;13 1;6 0;59 0;52 0;45 0;52 0;59
B: 1;6 1;14 1;21 1;28 1;21 1;14 1;6 1;0 0;53 0;46 0;53 1;0
These values are close to those given in the Cairo manuscript. Actually, however, if we multiply
Ibn Ynus values by a factor of 2 22;40/30, rather than 2 21/30, which is the factor
underlying al-Minfs values, we obtain the set:
1;11 1;19 1;26 1;34 1;26 1;19 l;11 1;3 0;56 0;48 0;56 1;3 .
other early egyptian tables for timekeeping 313
Now these correspond very closely to those for Mecca in the Cairo manuscript. It remains only
to explain the number 22;40. Perhaps the values of 2(D) were originally intended for =
21;40,24 another widely-used medieval value for the latitude of Mecca, and a careless error
was made in the computation.
MS Cairo Sh 89,4 (fols. 29v-32v), copied in 1025 H [= 1616], is the only known copy of the
introduction to a set of prayer-tables for latitude 21, that is, Mecca, by Ibn al-Rashd (4.1).25
The tables are no longer contained in the manuscript but they are described in the text. The
latitude 21 is specifically mentioned, as well as the parameters 19 ands 17 for twilight (cf.
6.10). Ibn al-Rashd mentions ten tables of the following functions:
T(h) and t(h) for = 0, H, D, ha, ta, hb, Ta, s, r, .
The first table has its counterpart in the tables for Jerusalem in MS Princeton Yahuda 861,1
which I suspect were also computed by Ibn al-Rashd (8.1 and 9.5).
Added in proof: My notes to these tables indicate that MS Leiden UB Or. 2805, which I
have not consulted, may contain the same tables.
Two disordered copies of timekeeping tables for Cairo, MSS Cairo DM 45 and Cairo DM 153,
contain odd folios copied in the same hand ca. 1650 from a set of tables of the hour-angle
t(h,) computed for
= 32;0 (Jerusalem) and = 23;35 .
The tables have the same format as the Cairo corpus and are probably of Egyptian provenance.26
The entries are identical with the corresponding entries in al-Karaks tables of t(,h) (9.4).
Since al-Karak states in his introduction that the 14th-century Egyptian astronomer Ibn al-
Rashd had compiled a set of tables of t(h,) for an unspecified latitude I suspect these tables
for Jerusalem are by Ibn al-Rashd. See also I-2.1.5.
MS Cairo MM 33,2 (fols. 6v-7r), penned ca. 1600, is the only copy known to me of a table
for the afternoon prayer attributed to al-Bakhniq, the editor of the main Cairo corpus (4.1.5
and 5.6). The table is entitled jadwal ma{rifat fal d}ir al-{ar li-ru}s al-burj, table for
finding the hour-angle at the beginning of the afternoon prayer at the first points of the zodiacal
signs. The function tabulated is:
t a( , )
MS Cairo MM 43 is an Egyptian manuscript, perhaps dating from around 1600, which contains
al-Khalls universal auxiliary tables (10.5), complete with his introduction but without
mention of his name. The functions f and g are tabulated for values of up to 50 rather
than 55 and the function G is tabulated for values up to 55 rather than 59. These auxiliary
tables are preceded by tables of:
() ( = 23;31) and ha(H)
to two digits for each degree of argument, and are followed by a mixed bag of anonymous
spherical astronomical tables. The following functions are tabulated to two digits for each
integral unit of argument unless otherwise stated:
(1) B() ( = 30 and = 23;55), = 0;15
(2) max D() ( 23;30), to hours, minutes and seconds, = 1, 2, ... ,
66
(3) () ( = 23;35), = 6, 12, ... , 66, = 6
(4) arc Sin (x)
(5) ( )
(6) arc Tan12 (x)
(7) H(ha), ha = 1, 2, ... , 45 (for use on quadrants)
(8) () ( = 30), = 1, 2, ... , 60
(9) (), = 1, 2, ... , 23, 23;15, 23;30, 23;35
(10) arc Cot12 (x) (garbled)
(11) ha(H), H = 0;30
(12) Cot12 h, h = 0;30
(13) 1/12 Tan
(14) 1/12 Tan () ( = 23;35), 3 digits
(15) Vers , 3 digits
(16) Sin , 3 digits
(17) () ( = 23;35)
(18) () ( = 23;35, = 21, 24, 36)
(19) () ( = 23;35)
(20) n() ( = 23;35), = 1, 2, ... , 67, n = 1, 2, ... , 12
(21) G() ( = 30), = 1, 2, ... , 60
(22) d() ( = 30), = 1, 2, ... , 60 .
All of the tables are anonymous except no. 20 which is attributed to Ibn Ynus. There is also
a list of geographical coordinates and qiblas27 and tables for marking curves on astrolabes. On
tables nos. 1, 8, 13-14, and 21 see further I-6.4.3, 5.6.1, 7.1.7 and 6.8.1, respectively.
27 Overlooked in King, Mecca-Centred World-Maps, pp. 76-86.
other early egyptian tables for timekeeping 315
Three other Egyptian copies of al-Khalls universal auxiliary tables are in MSS Cairo MM
98,2, copied ca. 1700, Princeton Yahuda 861,2, and Cairo DM 758, copied ca. 1650. In the
first the tables are anonymous. In the second the auxiliary tables have been bound in the midst
of the tables of the Natja al-kubr (8.1): here they are incorrectly entitled Fal d}ir al-Khall,
al-Khalls hour-angle (tables). In the third the tables also bear the title Kitb Fal al-d}ir
and are attributed to al-Khall. Between the introduction and the tables themselves is an
anomalous table of hv() for Cairo attributed to Ibn al-Rashd (4.8), with a remark that for
other latitudes the azimuth of the bdahanj is also the rising point of Capricorn (!): see Fig.
VIIb-4.1.
Finally, MSS Cairo DM 644,1, copied ca. 1700, and Istanbul S. Esad Efendi Medresesi
119,2, date ?, are two copies of an anonymous Egyptian set of auxiliary tables entitled Fat
al-Karm al-Bq f ma{rifat al-d}ir wa-falihi fq, The Victory of God, the Noble and
Eternal, for Finding the Time Since Sunrise and Hour-Angle for all Latitudes. These tables
were lifted from al-Khalls auxiliary tables for timekeeping by the sun (10.3). See further
I-9.11.
6.15 The auxiliary tables of Ibn al-Mushrif, al-Mridn, al-Waf}, and al-f
At least five Egyptian astronomers compiled sets of auxiliary tables for solving the standard
problems of spherical astronomy.28 Tables of this kind containing a few hundred entries were
first devised by abash and Ab Nar in the 9th and 10th centuries, but the most outstanding
example is the set containing over 13,000 entries compiled by al-Khall in the 14th century
(10.7). abashs tables were known in Egypt: they are referred to by Ibn Ynus and were
incorporated into one recension of the 13th century Muala Zj (6.6). al-Khalls universal
auxiliary tables, as well as his minor set, were also known in Egypt (6.14). I have surveyed
the development of Islamic auxiliary tables, including the Egyptian examples noted below, in
I-9.
Jaml al-Dn {Abdallh ibn Khall al-Mridn (d. 1406) was one of the leading astronomers
in Cairo and/or Damascus in his time (I-9.6). His auxiliary tables, which contain 8,100 entries,
offer little practical advantage to the user because of the simple nature of the functions
tabulated. They are contained in MS Paris BNF ar. 2525,1 (fol. 1v-16v), copied ca. 1450, where
they precede an incomplete set of prayer-tables for all latitudes (8.1), and also MS Cairo K
4026 (late). See further I-9.6.
Ab Bakr ibn Ism{l known as Ibn al-Mushrif worked in Cairo in the first half of the 15th
century, and his name is new to literature.29 MS Cairo MM 241, penned ca. 1450, is a unique
copy of a set of auxiliary tables compiled by him containing over 15,000 entries. The tables
take advantage of the mathematical equivalence of the problems of determining the hour-angle
and the azimuth from solar or stellar altitude. See further I-9.8.
{Izz al-Dn {Abd al-{Azz ibn Muammad al-Waf} (d. ca. 1470) was one of the leading
astronomers in Egypt in the mid-15th century (I-9.9 and II-5.7). His auxiliary tables are extant
28 See already King, Universal Solutions from Mamluk Egypt and Syria, of which a new version is in VIb.
29 On Ibn al-Mushrif (I-9.8) see Cairo ENL Survey, no. C43. The individual with the same name mentioned
in King, Astronomy in Yemen, no. 12, is a Yemeni astronomer from the early 14th century.
316 part ii, chapter six
in MSS Vatican Borg. ar. 217,1 (fols. 1v-5v) and Istanbul Nuruosmaniye 2921,2 (fols. 22r-
26v), and contain about 1,500 entries. They are based on the function G() discussed at length
by Ibn Ynus, and, like those of Ibn al-Mushrif and al-Khall, exploit the mathematical
equivalence of the problems of determining the hour-angle and azimuth from solar or stellar
altitude. See further I-9.9.
Ibn Abi l-Fat al-f (d. ca. 1495) was another of the leading astronomers in Cairo in
the late 15th century.30 He prepared an extensive zj based on those of Ibn al-Shir (ca. 1350)
and Ulugh Beg (ca. 1430), adapting the tables for the geographical coordinates of Cairo.
Nevertheless in his own solar tables which precede his auxiliary tables in MS Oxford Seld.
Supp. 101 (Uri 1040) he prefered to use the parameters of Ibn Ynus derived five centuries
previously. (The title-folio of MS Leiden Or. 143 of the kim Zj bears biographical
information on Ibn Ynus in the hand of al-f: see Fig. 5.1) al-fs auxiliary tables contain
over 15,000 entries and are more suitable for finding the hour-angle from solar or stellar
altitudes for any latitude than for finding the azimuth. See further I-9.10.
MS Cairo DM 188,1 (fols. 1r-10r), copied ca. 1700, contains a set of prayer-tables entitled
Nuzhat al-abr f a{ml al-layl wa-l-nahr li-{ar kf-zy, Delights of the Eyes for
Astronomical Operations at Latitude 27, that is, for the city of Assiut. On the title page it
is stated that the tables were prepared by Muammad (ibn {Abd al-Qdir ibn Dalll) al-Waf}
(al-Asy),31 a student of Muammad ibn (Ab) al-Fat al-f (6.15).
The functions tabulated are the following, with identical format to the tables in the Cairo
corpus:
, H, d, D, , 2N, , h0, Z(12), za(12), ta, Ta, r, s, (2N-r), (r+D) and .
al-Asy was not a bad student: his tables are rather carefully computed and are more accurate
than the tables for latitude 27 in the Natja attributed to al-Waf} (8.1). However, al-Asy
preferred to use 23;35 for the obliquity rather than the more up-to-date value 23;30,(17) of
Ulugh Beg. His tables of r and s are based on parameters 19 and 17 respectively, and the
former is more accurately computed than the latter. His tables of ( = 27) and are less
accurate than those in the kim Zj (5.1). On the title folio of MS Cairo DM 188 there is
a note stating the approximation:
max d() 1/2 .
al-Asy would doubtless have disapproved of this: his own tables give the accurate value
12;51 for max d = d(90) ( = 27), and the value one can derive with this approximate formula,
namely 13;30, is in error by almost 2/3.
6.17 Ibn ghns table for determining the times of moonrise and moonset
MS Cairo MM 167,4 (fols. 158v-160r), copied in 989 H [= 1581], contains a short introduction
and table of two pages by Jaml al-Dn Ysuf ibn ghn al-Qiaj, otherwise known to us
as the author of a treatise on astrology.32 The table, displayed in Fig. 6.17, enables the user
to determine the time between sunset and moonset from the nocturnal arc, entered horizontally,
and the days elapsed since conjunction, entered vertically. Entries are given in degrees and
minutes for each 6 of nocturnal arc and each day from 1 up to 16. If more than 14 days of
the month have passed the instructions tell us to subtract 14 from the number of days passed
since conjunction and enter the remainder to find the time passed since sunset when the moon
will rise. The author acknowledges that the underlying procedure is approximate, and in fact
the entries f(2N,n) are determined by simple relation:
f = 2N/180 n ,
where is the mean daily relative motion of the sun and moon, 12;53.
Fig. 6.17: A unique table for finding the times of moonrise and moonset. Note that the first column on the left-
hand page should be the last column, as noted by the copyist. [From MS Cairo MM 167,4, fols. 159v-160r,
courtesy of the Egyptian National Library.]
32
On Ibn ghn see Cairo ENL Survey, no. C92; and ~hsanolu et al., Ottoman Astronomical Literature, II, pp.
228-229, no. 108.
318 part ii, chapter seven
CHAPTER 7
From a mathematical point of view it was difficult for the astronomers of medieval Egypt to
improve in any way upon the tables in the main Cairo corpus, these being in general very
carefully computed, and the changes in the obliquity of the ecliptic over the centuries had an
effect on the prayer-times that was negligible. Nevertheless, one Egyptian astronomer
endeavoured to modify at least the prayer-tables in the corpus for a more up-to-date value of
the obliquity (7.1). Certain muwaqqits with time on their hands succumbed to the temptation
to stretch some of the tables in the corpus to display values for each minute of solar longitude
instead of each degree (7.2). Others rearranged the format of the tables and added information
about timekeeping by the stars (7.5). Yet others were content to prepare some simple tables
for finding the solar longitude for a date in the Hijra calendar and supplement these with prayer-
tables lifted from the corpus (7.7-9).
From a practical point of view it was clearly more convenient to display the times of prayer
in almanacs with values arranged for each day of the year (7.4). For those who did not like
tables at all, the times of prayers for each day of the year were written out in words (7.6, 7.7
and 7.10). The Egyptian astronomers also modified the tables of the corpus so that the entries
were expressed in hours and minutes according to the Ottoman convention that sunset is 12
oclock (see 7.11 and 7.12-13; on the convention see 14.0).
Shams al-Dn Muammad ibn Nir al-Dn al-Minf was a muwaqqit at the Ghawriyya
madrasa in Cairo1 who lived in the mid 16th century and compiled a set of prayer-tables in
which he sought to modify the tables of the main Cairo corpus. Complete copies of these are
MSS Cairo DM 470 and 467, the former, in 50 fols., penned ca. 1570 by the authors son {Abd
al-Qdir (7.2 and 8.2), the second in 1144 H [= 1731/32]. The tables are entitled {Anwn al-
muhimmt f tarr al-awqt, The Epitome of Important Matters in Fixing Prayer-Times.
In the introduction the author claims that the tables are computed according to the new
observations and that he has reworked the tables of the Natja according to the new value
of the obliquity, meaning the value 23;30(,17) found by the astronomers of Ulugh Beg in
1 Muhammad al-Minuf is not mentioned in the modern bio-bibliographical sources before Cairo ENL Survey,
no. C120; and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 187-188, no. 88 (listing mainly Cairo
manuscripts). The suggestion in King, Astronomical Timekeeping in Medieval Cairo, p. 373, n. 68, that he
might be identical with Muhammad ibn Abi l-Fat al-f (see 6.15) is incorrect.
late modifications to the main cairo corpus 319
Samarqand. In fact, only a few of the simpler tables have been recomputed. It would be
interesting to know precisely what al-Minf meant by Natja: perhaps the complete set of
prayer-tables for Cairo appended to those of T, t and a in the edition of the main corpus by
al-Bakhniq (5.6)?
The other modification which al-Minf introduced was to incorporate corrections for
refraction at the horizon in all tables involving a period of time measured from sunrise or sunset.
In the introduction to his tables he mentions that because of the nature of light rays, an observer
can see more than half of the celestial sphere. According to Ibn al-Haytham, he says, a person
standing at the equator 31/2 cubits above the surface of the earth sees 0;4,26 more than half
of the celestial sphere. I have been unable to trace this statement in any treatise associated
with the illustrious 11th-century physicist (see, however, 9.3, also 14.9). The remark makes
little sense, especially in the light of what follows. al-Minf proceeds to state that Ibn Ynus
fixed the interval between the moment when the centre of the sun is on the true horizon and
the moment when its centre is on the apparent horizon as 0;47 at the equinoxes, increasing
0;5 for each sign to the summer solstice and decreasing similarly to the winter solstice. These
are the corrections D which are tabulated in certain late copies of the Cairo corpus (4.11),
including al-Minfs prayer-tables (see below). al-Minfs son {Abd al-Qdir wrote at length
about this function D and tabulated it for all latitudes (8.2).
al-Minf tabulates most of the functions found in the main corpus, modified as follows.
For those functions which involve horizon phenomena the function is tabulated for the sun
on the true horizon and then again for the sun on the apparent horizon, the second table having
the correction D taken into account. Another feature that is not found in the main corpus is
the instructions presented in marginal notes concerning the exact time of day for which to
determine the solar longitude to use as argument in the tables. Thus, for example, there is a
table of ta, the time from midday to the afternoon prayer: the instructions are to enter the true
longitude of the sun at the time of the afternoon prayer as argument. In the table of Ta, the
time from the afternoon prayer to sunset, the instructions indicate that the true longitude of
the sun at sunset should now be the argument and the value corrected with the difference
minutes and the minutes of the solar radius to find the time when the sun has disappeared
over the apparent horizon. This table is followed by another in which the difference minutes
are taken into account, and one need only correct for the solar radius to obtain the value
required. There is no table of the minutes of solar motion between midday and the other
prayer-times to facilitate the computation of the argument, such as is found in a note by al-
Waf} (5.7) and al-Ldhiqs tables (7.8).
al-Minfs work contains tables of the following functions (those adjusted for the difference
minutes are asterisked):
D, D*, ta, ha, Ta, Ta*, H, hv, d, , Tq, tq, Z(12), hq, , h0, 2Dh, 2D*h, , 2D*h,
s, r, (r+D), (D+s), 2N, 2N*, n, n*, , *, , D, s, , , and f .
The work concludes with a star table giving equatorial coordinates for the year 961 H [= 1551].
The functions d (and hence D, 2N, 2Dh, , etc.), (and hence H), , ha and ta, and and
( = 30) have been recomputed for = 23;30. The functions hq, Tq, tq, h0, r and s, and
hv, have been taken from the main corpus, being based on = 23;35. Functions such as n
(= 2N-(r+s)) are based on values computed with different parameters.
320 part ii, chapter seven
al-Minfs knowledge of spherical astronomy was such that he reproduced the garbled
entries:
(60) = 23;29 and h0(30) = 23;34
from earlier tables. Both these values should be the same as for latitude 30 (cf. 4.3), that
is, 23;30 for (60) and 23;35 for h0(30). Furthermore, he did not realize that Ibn Ynus
values for the difference minutes were mutually inconsistent. al-Minfs efforts to improve
the corpus were hardly successful, but he was the only Egyptian muwaqqit to attempt to
radically modify it.
Another work on timekeeping by al-Minf is a treatise entitled Nam al-yawqt f tarr
a{ml al-mawqt, Organizing the Gems concerning the Correction of the Operations of
Timekeeping, and is extant in MS Cairo MM 235,1 (fols. 1r-7r), copied ca. 1600 by his son
{Abd al-Qdir. In this al-Minf mentions the difference minutes and the way in which they
should be used, and cites the Tanq al-manir, the well-known work on optics by the late-
13th-/early-14th-century Baghdad scholar Kaml al-Dn al-Fris,2 rather than Ibn Ynus or Ibn
al-Haytham. See further 8.2.
7.2 The prayer-tables of {Abd al-Qdir al-Minf and Ibrhm ibn Qyitby
{Abd al-Qdir al-Minf3 was the son of Muammad al-Minf who modified the main Cairo
corpus (7.1). He wrote on the subject of refraction of the horizon (8.2) and also compiled some
prayer-tables in which the tables of various functions, mainly those tabulated by his father,
are stretched by linear interpolation to display values for smaller argument increments. In this
he was following an earlier tradition of al-Waf} and Ibn al-Mushrif (5.7).
MS Cairo DM 1101 was copied in the hand of al-Minf-Jr. in 974 H [= 1566/67] and
contains tables of the functions:
D + D, 2Dh, ta and s
with values to three digits for each 0;3 of . MS Cairo MM 66, copied in the previous year
by al-Minf-Jr., contains tables of () with values to three digits for each minute of argument,
and MS Cairo MM 45, copied ca. 1625 by Ibrhm ibn Qyitby (see below) contains similar
tables with entries to five digits (lifted from Ulugh Beg), both attributed to {Abd al-Qdir al-
Minf.
Ibrhm ibn Qyitby was a student of Shams al-Dn al-Minf who worked in Cairo ca.
1625.4 Several manuscripts preserved in Cairo contain prayer-tables computed and also copied
by him, similar to those of {Abd al-Qdir al-Minf. The functions tabulated in the various
manuscripts are as follows:
ta (DM 34 and DM 152), s (DM 153,1 and DM 740,2), D (DM 682),
Ta + D (DM 152) and D + D (DM 33)
See Fig. 7.2b for a sample of the first of these, and Figs. 7.2a and c for their title-folios. Some
2 On Kaml al-Dn see the article by Roshdi Rashed in DSB. See also n. 8:9.
3 On {Abd al-Qdir al-Minf see Suter, MAA, no. 479 (confused); Cairo ENL Survey, no. C121; and
~hsanolu et al., Ottoman Astronomical Literature, I, pp. 217-220, no. 97.
4 Ibrhm ibn Q}itby is not mentioned in the modern sources besides Cairo ENL Survey, no. C122, and
~hsanolu et al., Ottoman Astronomical Literature, I, pp. 264-265, no. 130 (lists only Cairo manuscripts).
late modifications to the main cairo corpus
Fig. 7.2a-b: The title-folio and an extract from the extended tables of the hour-angle at Fig. 7.2c: The title-folio of Ibrhm ibn Qyitbys
the beginning of the {ar. This manuscript belonged to the celebrated 17th-century extended tables of the difference minutes. This
astronomer and muwaqqit {Abd al-Ramn al-uln: see V-10. [From MS Cairo DM 34, manuscript also belonged to al-uln. [From MS
courtesy of the Egyptian National Library.] Cairo DM 682, courtesy of the Egyptian National
Library.]
321
322 part ii, chapter seven
other anonymous prayer-tables of the same kind are preserved in various other Cairo
manuscripts. For example, MSS Cairo DM 57,1 and 151,2, originally one codex of seven folios
copied ca. 1600, contain a set of tables of (al-d}ir li-l-salm) for each 0;6 of , based
on the assumption that the salm is 2 before daybreak, and MSS DM 545, 158 and 423, copied
about the same time, contain tables of (mali{ al-salm) for each 0;3 of based on the
assumption that the salm is 1 before daybreak.
MS Cairo DM 620,9 (fols. 58r-59r + 64r-72v), copied ca. 1450, contains a set of tables entitled
jadwal al-bq li-l-{ar, table of time remaining until the afternoon prayer, and attributed
to an individual called al-Qaymar, on whom I have no further information. The function
tabulated is simply:
a(h,) = ta() - t(h,)
and values are given for each integral degree of h such that ha() h H() for each degree
of . An extract is shown in Fig. 7.3. See 11.3 on two similar sets of tables for Damascus.
Fig. 7.3: An extract from al-Qaymars tables, serving Aries and Taurus / Virgo and Leo. [From MS Cairo DM
620,9, fols. 68v-69r, courtesy of the Egyptian National Library.]
late modifications to the main cairo corpus 323
These tables are followed (fols. 72r and 72v) by others of () and h0() (the entries, which
are given to two digits, are not corrupt as in other such tables in the Cairo corpus see 4.3)
and h(a,) for a = 30 (4.6).
MSS Cairo TR 114 (74 pp.) and ZK 154 (91 fols.), both from ca. 1700, are two copies of a
treatise entitled Itf al-abb bi-ma{rifat al-tawq{t wa-l-awqt wa-l-qibla bi-l-taqrb by
the 16th-century Egyptian astronomer Muammad ibn Abi l-Khayr al-usn.5 In this work al-
usn has included the main prayer-tables of the main corpus in the text of his treatise. For
each day of the Coptic year he gives the values of the functions:
D, ta, Ta, s, n and r ,
with values given to two digits and written in abjad notation. See 8.3 on some sophisticated
tables for twilight by al-usn.
MSS Paris BNF ar. 2545, copied ca. 1500 in an elegant hand, and Oxford Seld. Supp. 99 are
copies of an extensive set of prayer-tables entitled Natjat al-afkr f a{ml al-layl wa-l-nahr,
The Result of Thoughts about the Operations of Timekeeping by Night and Day, the same
title used by al-Ldhiq (7.8). These were compiled by a late-16th-century Egyptian astronomer
named Muammad ibn {Umar ibn iddq ibn {Umar al-Bakr al-Fawns.6 It is not difficult
to show that very little of consequence in this at first sight imposing work is original.
In the introduction to the tables the author notes certain approximate rules for determining
the standard functions of mqt, but makes no mention of the fact that similar rules had been
formulated by his predecessors Najm al-Dn al-Mir and al-Bakhniq see 2.5 and 2.9 above.
Each of al-Fawnss rules is intended for latitude 30, except those for functions ha and H,
of which he incorrectly states that they work well for latitudes between 0 and 40. His rules,
outlined in words in the text, are the following:
(a) d 1/2 ( + 1/4 )
(b) + 1/6
(c) hq 47 + + 1/4 + 1/6
(d) Tq 57 + + 1/2 + 1/6
(e) ta 52 + 1/2 1/6 ( > 0)
ta 52 - [ ||/4 + ||/6 ] ( < 0)
(f) ha 1/2 H + 1/10 [ max H - H ]
(g) hb 1/2 ha + 1/10 [ max H - ha ]
5 On Ibn Abi l-Khayr al-usn see Suter, MAA, no. 511; Brockelmann, GAL, SII, p. 485; Cairo ENL Survey,
no. C124; and ~hsanolu et al., Ottoman Astronomical Literature, pp. 255-262, no. 127.
6 On al-Fawns see Suter, MAA, no. 475; Brockelmann, GAL, II, p. 469, and SII, p. 485; Cairo ENL Survey,
no. D10; and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 237-238, no. 111.
324 part ii, chapter seven
late modifications to the main cairo corpus 325
Fig. 7.5a: An extract from al-Fawnss prayer-tables for day-time, displaying 19 different functions for each
day of the Coptic year, here @b 24 to Misr 6. [From MS Paris BNF ar. 2545, fols. 29v-30r, courtesy of the
Bibliothque Nationale de France.]
Fig. 7.5b: An extract from al-Fawnss prayer-tables for night-time, serving the end of Abb and the beginning
of Misr. The tables have a very distinctive format. [From MS Paris BNF ar. 2545, fols. 30v-31r, courtesy of
the Bibliothque Nationale de France.]
326 part ii, chapter seven
(6) 2N() - r() time from sunset to daybreak (al-m min ghurb al-shams il
ul{ al-fajr)
(7) () + 180 the amount of rotation at midnight, i.e., the oblique ascensions
of the ascendant at midnight (al-d}ir {ind nif al-layl)
(8) r( ) the amount of rotation at daybreak, i.e., the oblique ascensions
of the ascendant at daybreak (al-d}ir {inda l-fajr)
Approximate values of r() and s() are given at the top of each pair of facing pages. These
are expressed to the nearest 0;30 and are based on angles of depression of 18 and 16,
respectively. They are intended to be used for all the nights served by the given pair of pages.
The 2nd, 6th, and 8th functions are based on these approximate values of r and s. al-Fawns
did not copy the tables of r and s in the corpus because these are based on parameters l9
and 17. He gives no explanation why he adds 2 to the values of s(): perhaps he was not
really convinced that 16 was a better parameter for the evening twilight than 18. He also
names the stars which are culminating at nightfall, midnight, and daybreak, or gives what is
perhaps intended to be their distance from the meridian at these times, for each night of the
year. (See 5.5 for a different set of solar tables compiled by Ibn al-Kattn.)
The anonymous tables in MS Oxford Hyde 32 (Nic. 284), fols. 69v-81r, are related to those
of al-Fawns. The following functions are tabulated for each day of the Coptic year:
, Z(12), H, ha, , D, ta, Ta and 2N .
For the corresponding nights the altitudes of three prominent stars in the eastern and western
sky at daybreak are given, as well as the values of various functions relating to nightfall and
daybreak.
MS Paris BNF ar. 2571 contains an almanac for Cairo for the year 1015 H [= 1606/07], by
Amad ibn Muammad al-usn,7 entitled Taqwm al-sana al-qamariyya, literally, Ephemeris
for the Lunar Year. al-usn begins his almanac by paying tribute to his teacher, the Ottoman
astronomer Jaml al-Dn Ab D}d Sulaymn (ibn amza), otherwise known as Ibn
Bakhshsh,8 and also in passing mentions the Ottoman Sultn Amad Khn (reg. 1603-1617).
The almanac contains calendrical information and remarks on the nature of the lunar crescent
at first visibility for each month of the year. al-usn states that the lunar crescent whose
appearance marks the beginning of the year will be seen on a Sunday evening (laylat al-ithnayn
or night of Monday). The first day of the year would thus be a Monday, which would
correspond to April 30, 1606 (Julian calendar): modern tables give Muarram 1 as Sunday,
April 29 for that year. For the first and 15th day of the lunar month al-usn also gives the
values of:
D, ta, Ta, s, n and r
written out in words, and expressed in equatorial degrees, with parts of a degree expressed
as simple fractions. In Raman the time is also shown, based on the assumption that the
7 This al-usn is not mentioned in the modern sources.
8 On Ibn Bakhshsh see Azzawi, History of Astronomy in Iraq, pp. 181; Cairo ENL Survey, no. C117; and
~hsanolu et al., Ottoman Astronomical Literature, I, pp. 194-198, no. 95.
328 part ii, chapter seven
afy occurs 5 before daybreak. The solar longitude on Muarram 1 is taken as 48, which is
reasonable.
The entries in al-usns prayer-tables for a given day are generally mutually inconsistent.
The values of the functions ta, s and r are rounded from the corresponding entries in the main
Cairo corpus, but those for D, Ta and n have been haphazardly adjusted by what was probably
intended to be a correction for horizon phenomena. al-usns values of the nif qaws al-nahr
al-shar{, semi diurnal arc according to religious law, are in excess of the geometric semi
diurnal arc by approximately the following amounts:
VE: 0;3 SS: 0;55 AE: 0;51 WS: 0;43
He gives no explanation of these corrections.
MS Cairo M 87 contains a similar almanac for the year 1044 H [= 1634/35] attributed
to Muammad al-Kutub,9 which I have not investigated.
{Abd al-Qdir ibn Amad al-{Ajmw al-Azhar10 was a muwaqqit in the madrasa of Suln
asan in Cairo who compiled a set of prayer-tables entitled Nuzhat al-nir f ma{rifat m bayn
al-awqt min al-daw}ir, The Delights of Knowing the Intervals between the Prayer-times.
MS Paris BNF ar. 2578,2, dated 1024 H [= 1616], and also Cairo ZK 287, 203 fols., penned
1056 H [= 1646/47], are copies of this work, containing the prayer-times written out in words
for each day of the Coptic year. An extract is displayed in Fig. 7.7. Values are given for the
functions:
D, ta, Ta, s, n and r ,
Fig. 7.7: An extract from al-{Ajmws prayer-tables written as text. [From MS Paris BNF ar. 2578, fol. 46r,
courtesy of the Bibliothque Nationale de France.]
9 On Muammad al-Kutub see Cairo ENL Survey, no. D27; and ~hsanolu et al., Ottoman Astronomical
Literature, I, p. 276, no. 141.
10 On al-{Ajmw see Brockelmann, GAL, SII, p. 1018, and Cairo ENL Survey, D11.
late modifications to the main cairo corpus 329
and are those of the main Cairo corpus with the equatorial minutes rounded so that they can
be expressed as simple fractions.
his table for solar longitude by later muwaqqits. In MS Berlin Ahlwardt 5765 [= Wetzstein
1149, fols. 85v-88v] his introduction is followed by a set of prayer-tables for the latitude of
Nablus (11.7). I have no information on the life of al-Ldhiq, but his name indicates that his
family was at one time associated with Lattakia in Syria.
A similar work, with a slightly different introduction and different arrangement of the tables
is attributed to Muammad ibn Haykal al-Ikh, a muwaqqit at the Azhar Mosque in Cairo.12
al-Ikhs tables are preserved in MSS Cairo DM 60, 480, 739, and MM 81.
MS Alexandria 4441J is the only copy known to me of a set of prayer-tables for Cairo attributed
to an individual called Ibn Ab Rya, a muwaqqit at the Mu}ayyad Mosque in Cairo at the
beginning of the 18th century.13 The work is entitled Zuhrat al-afkr f a{ml al-layl wa-l-
nahr, The Beauty of Reflection on the Operations of Timekeeping by Night and Day, and
begins with an introduction of 13 chapters. A table of solar longitude for each day of the Coptic
year (with Virgo 16;12 for Tt 1) is is said to be computed for longitude 54;55 (7.8).
The other simple calendrical, astrological and solar tables in this work do not concern the
present study, save to mention one table displaying the lunar mansion corresponding to the
position of the sun, such as is found in MS Istanbul Badatli Vehbi Efendi 887 of the tables
of al-akm al-Ldhiq for Lattakia (11.9). This particular table is said to be computed for the
year 1117 H [= 1705/06] by Ibn Ab Rya.
The main set of prayer-tables is preceded by individual tables of (computed for = 23;30
but having 23;29 for = 60) and h0 (computed for = 23;35 but having 23;54 for =
30) (see 4.3 on this problem), as well as garbled tables of Cos () (called jayb tamm al-
mayl) and C() (called bu{d al-qur li-l-jayb (sic)) (see I-6.2 and 6.4). The main tables display
16 functions side by side for each degree of beginning with = 1. The first and third of
these, D and D, are called respectively nif qaws al-nahr al-sharq and al-gharb, semi diurnal
arc in the east and in the west. The former is simply the semi diurnal arc and the latter the
semi diurnal arc corrected for refraction at the horizon. The last function tabulated, which I
denote by [D() - min D], is called ziydat al-nahr, and measures the excess of the length
of daylight over its minimum at the winter solstice; I have not found this elsewhere. The
underlying value of is 23;35, and the tables are lifted from those in the main corpus rather
than those of al-Minf (7.1). The 16 functions tabulated are:
D, H, D, ha, ta, Ta, s, n, r, 2N, hq, D, Z(12), (2N-r), (D+2N-r) and [D - min D] .
The work concludes with a set of tables of:
, s, r, and * ,
as well as a star catalogue displaying the equatorial coordinates of about 285 stars for an
unspecified date.
12 On al-Ikh see Cairo ENL Survey, no. D3; and ~hsanolu et al., Ottoman Astronomical Literature, I,
pp. 348-349, no. 215.
13 Ibn Ab Rya is new to the modern literature.
late modifications to the main cairo corpus 331
The celebrated astronomer and author Riwn Efend al-Falak,14 who died in 1710, compiled
a lengthy volume on timekeeping called Dustr ul {ilm al-mqt wa-natjat al-nar f tarr
al-awqt, Statutes of the Principles of Timekeeping and the Results of Reconsidering the
Prayer-times. I have consulted MSS Cairo Azhar falak 8216, Berlin Ahlwardt 5710, Istanbul
S. Esad Efendi Medresesi 119,1, Istanbul Nuruosmaniye 2781, Cairo TR 116 and TR 276 of
this work. Apparently, the only original part is a set of tables for determining the times of the
suns entry into each of the signs: the treatment of the problems of spherical astronomy in the
introduction is standard and all the prayer-tables which follow have been lifted from earlier
sources.
These manuscripts contain all of the prayer-tables from the main corpus, and Riwn Efend
has made no effort to recompute even the simplest of them for the value of the obliquity used
in his Zj, namely, 23;30,17. This Zj al-Riwn, or al-Zj al-Mufd, of which I have examined
the copy preserved in MS Princeton Garret Hitti 1004, is based entirely on that of Ulugh Beg,
which in Cairo was available in Arabic in the recension of Ibn Abi l-Fat al-f (6.15). MS
Cairo DM 45 contains some tables computed by Riwn Efend giving the oblique ascensions,
a, at the beginning of the interval for the afternoon prayer, according to the new observation.
The new observation of Ulugh Beg was almost three centuries prior to Riwns activity
in Cairo.
MS Cairo DM 200 (7 fols., ca. 1750) contains a set of tables of the functions:
D, ta, Ta, s, n, r, and the solar longitude ,
tabulated side by side for each day of the Syrian calendar (equivalent dates in the Coptic
calendar are also given). Values are given to two digits. MS Cairo DM 531,1 contains sets
of sub-tables displaying values of the first six functions to one digit for each five days of the
Coptic calendar.
MS Cairo DM 46 (17 fols., ca. 1750) contains a motley collection of prayer-tables for Cairo,
some with entries to two digits taken from the main corpus, and others recomputed for Ulugh
Begs values of the obliquity, 23;30,17, with entries to three digits. Four of the tables display
the latitude of visible climate, , and its complement, , which is the altitude of the pole
of the ecliptic, as well as Cos60 and Tan60 , with values to three digits for each degree of
.15 The tables that were not lifted from the corpus display functions:
hb, Ta, Tb, (Ta+D), (Tb+D), a/b, 2N, n, (2N-r) and hq ,
and the tables that were lifted from the corpus display functions:
2N, n, s, r, tq, tq*, Tq, ha=30 and ha=60 .
No author is associated with any of them.
MS Istanbul Kandilli 424, copied in 1164 H (= 1750/51), contains (fols. 1v-4r) an anony-
mous set of tables displaying the longitude of the astrological houses for Cairo as a function
of the ascensions of the horoscopus, computed according to the Ptolemaic method,16 and
(fols. 5r-6v) an anonymous set of Ottoman-type prayer-tables for Cairo. I know of no other
copies of either set. The prayer-tables display seven functions side by side for each degree of
solar longitude beginning with Aries, and the times are given in equinoctial hours and minutes
according to the Ottoman convention (14.0). The tables are entitled jadwal s{t {aqrab al-
s{t, literally, tables of the hours of the clock-hand. The functions tabulated are the
following:
s {ish} the time of nightfall
j dhn Raman the time of the call to prayer in Raman
r fajr daybreak, the time of morning prayer
R shurq the time of sunrise
m uhr the time of the midday prayer
a {ar awwal the time of the first afternoon prayer
b {ar thn the time of the second afternoon prayer.
A note by the title states that the tables are computed for the setting of the upper limb of the
sun over the apparent horizon and the crossing of the entire solar disc across the meridian.
It is stated that these modifications take into consideration the correction for the difference
between the true and the apparent horizons and the radius of the sun, and that the tables are
based on the parameters of Ulugh Beg (i.e., = 23;30,17).
Comparison of the tables with a set of Ottoman-type prayer-tables computed for = 30;0
and = 23;30,17 and not modified for horizon phenomena reveals the following:
a) The tables of r and s are based on parameters 19 and 17, and the time j is always
20m before r.
b) The times m, a, and b are modified by a correction of approximately:
EQ: 4m SS: 6m WS: 3m ,
and the times r (and hence j) and R are modified by double these amounts. The radically
different amounts of this correction for the solstices suggest that the compiler used the
corrections for refraction associated with Ibn Ynus (4.11). Indeed if we add 0;15 (an
approximate value of the time for half the solar disc to set over the horizon) to Ibn Ynus
values we obtain
EQ: 0;47 + 0;15 = 1;2 4m
SS: 1;2 + 0;15 = 1;17 5m
WS: 0;32 + 0;15 = 0;47 3m .
(c) It is difficult to confirm whether or not the compiler of the tables introduced a correction
of about +0;15 +1m to the time m in order to ensure that the whole solar disc had
crossed the meridian at the time of the midday-prayer.
I have not investigated whether or not most of the entries in these tables can be derived from
the prayer-tables of al-Minf (7.1). No earlier tables of tb for Cairo are attested yet, but the
function hb was tabulated by al-Fawns (7.5).
16 On the houses see n. I-3:2.
late modifications to the main cairo corpus 333
Some manuscripts of al-Kutubs tables (7.14), such as MSS Cairo M 88 and M 142, contain
a set of prayer-tables for Cairo computed by the 18th-century Egyptian astronomer Raman
ibn li al-Khwnik.17 The following functions are tabulated in hours and minutes according
to the Ottoman convention, with values for each day of the Coptic year:
s, i, r, R, m, a and b .
The time of the imsk, i, is 20m before daybreak, r. The table for b occurs only in the first
source. I have not been able to check whether these tables are identical to those in MS Istanbul
Kandilli 424 (7.11).
MS Cairo TR 346, penned ca. 1875, is the only copy known to me of a set of prayer-tables
for Cairo attributed to asan al-Jabart (d. 1774), father of the celebrated historian {Abd al-
Ramn al-Jabart.18 The tables are entitled Mawqi{ {aqrab al-s{a murattaba {ala l-shuhr
al-Qibiyya, The Place where the Watch-hand Falls (at the Times of Prayer), arranged
according to the Coptic Months, and display the times:
s, i, r, R, m and ta
for each day of the Coptic year. The time of the imsk, i, is 20m before daybreak, r.
MSS Cairo DM 149, J 811,8, M 88 and M 142, DM 812, DM 1103 and 1104 contain
a set of tables for timekeeping prepared in the year 1150 H [= 1737/38] by {Abd al-Laf al-
Dimashq, known as al-Kutub.19 These display for the latitude of Cairo the times in hours and
minutes according to the Ottoman convention, before and after midday when the sun has a
given altitude and longitude, for each degree of both arguments: see further I-2.1.8 and the
extract illustrated there The entries are doubtless based on the entries in the tables of t(h,)
and T(h,) in the main corpus (4.5). al-Kutub states in his introduction that in his calculations
he has taken into consideration the apparent radius of the solar disc and the difference
minutes, as well as the increase in solar longitude from midday to the time in question.
17 On al-Khawnik see Brockelmann, GAL, II, pp. 471-472, and SII, p. 48; Azzawi, History of Astronomy
in Iraq, pp. 320-321; Cairo ENL Survey, no. D78; and ~hsanolu et al., Ottoman Astronomical Literature, I,
pp. 418-426, no. 279.
18 On Hasan al-Jabart see Brockelmann, GAL, II, p. 472, and SII, p. 487; Cairo ENL Survey, no. D91; and
~hsanolu et al., Ottoman Astronomical Literature, II, pp. 472-479, no. 314. On his son see the article al-
Djabart by David Ayalon in EI2.
19 On al-Kutub (I-2.1.8) see Cairo ENL Survey, no. D76; and ~hsanolu et al., Ottoman Astronomical
Literature, I, pp. 427-429, no. 281.
334 part ii, chapter eight
CHAPTER 8
In the sequel I discuss a set of late Egyptian prayer-tables for for each degree of latitude between
Mecca and Istanbul (8.1), as well as sets of such tables specifically for use in Alexandria,
Damietta, Rosetta and Crete (8.5-8). These are all clearly inspired by the tables in the main
Cairo corpus. I also discuss {Abd al-Qdir al-Minfs writings on refraction at the horizon
and his tables purporting to display the time taken by the sun to pass from the true horizon
to the visible horizon for all latitudes (8.2), and Ibn Abi l-Khayr al-usns tables for
computing the duration of twilight for all latitudes (8.3). al-Nabatts prayer-tables for the
pilgrim route between Cairo and Mecca (8.4) have unfortunately not survived. The much later
tables of Muammad if serve all localities of consequence in Egypt and the Hejaz (8.10).
These, and also the set for Cairo by {Abd al-Majd Efend printed in 1895 (8.9), bring our
discussion of one thousand years of astronomical timekeeping in Egypt to a close.
MS Princeton Yahuda 861,1, penned ca. 1600, contains a set of prayer-tables computed for
each integral degree of latitude between 21 (Mecca) and 41 (Istanbul). The tables are entitled
al-Natja al-kubr, which may be rendered Universal Prayer-tables. (The term natja means
calendar or prayer-tables in late medieval scientific Arabic.) On the title-folio the tables
are attributed to {Izz al-Dn al-Waf}, a muwaqqit at the Mu}ayyad Mosque in Cairo who died
about 1470.1 This attribution is highly doubtful, for reasons which will become clear below,
but at least some of the tables are due to him. The manuscript is carelessly copied in an untidy
hand, and is bound in some disorder. According to the title folio al-Waf}s tables also included
a set of auxiliary tables, but these are those of al-Khall see 10.7 and are described as
such in their sub-title. al-Waf} did compile some auxiliary tables of his own, which are extant
in another source see 6.15.
The manuscript begins with a short introduction on the use of the table displaying the solar
longitude for each day of the Coptic year. This is stated to have been computed for longitude
55 by al-Waf} (the entry for Tt 1 is Virgo 14;39). A star catalogue showing the right
ascension of 72 stars is likewise attributed to al-Waf}. Neither solar longitude table nor star
catalogue is dated. There follow two sets of prayer-tables, the first for Cairo, and the second
for all latitudes, including that of Cairo.
copied set of anonymous prayer-tables for various latitudes. This section of the manuscript is
unrelated to those which precede and follow and is clearly incomplete. It begins with a table
of for latitude 25, and continues with tables of:
D, H and
for latitudes 26 to 29 (except that for 27 ta is tabulated rather than H). There follow tables
of 24 different functions computed for the latitude of Cairo. The tables of D and H or ta for
latitudes 26 to 29 are virtually the same as those in the Princeton manuscript, being carelessly
computed. On the other hand, those of in the Paris manuscript are fairly accurately computed:
except for copyists errors (particularly for = 25), they are the same as the corresponding
tables in the kim Zj. All of the tables for latitude 30 are the same as those in the main
Cairo corpus.
Other functions tabulated in the Princeton manuscript are the following:
(a) meridian altitude for all latitudes: this is simply *() = 90+ () tabulated for two
digits for each degree of . Note that H(,) = [90+ ()] - .
(b) altitude for the afternoon prayer for all latitudes: this is simply ha(H) to two digits
for each degree of H from 1 to 90.
(c) rising times of the signs for different latitudes: the rising times for each pair of
zodiacal signs are given to two digits for each degree of from 1 to 60. The table
is based on = 23;35, and contains numerous errors. In MS Cairo MM 58 (I-5.5.1)
there is a similar table based on = 23;51,20 (Ptolemys value), which is attributed
to Ibn Ynus: the same table occurs in the Yemeni astronomical miscellany in the
manuscript in a private collection in Sanaa (12.4), without the false attribution.
(d) time since sunrise and hour-angle at the equinox for latitude 31: entries are given
for both functions for each degree of solar altitude from 1 to 58. The table is based
on latitude 32, not 31 as stated.
There is also an anonymous table of the latitudes of the stations on the Syrian pilgrim route
to Mecca. The entries are corrupt but were clearly lifted from the table in al-Khalls Damascus
corpus (10.9).
The celebrated Egyptian historian al-Jabart (d. ca. 1825)2 attributes two works entitled al-
Natja al-kubr and al-Natja al-ughr to the astronomer Riwn Efend (7.9), adding that
these works were well known and in common use in his time. These two titles could refer to
sets of prayer-tables (natja) for all latitudes (kubr, literally, larger) and for one latitude
(ughr, literally, smaller). No universal prayer-tables attributed to Riwn survive in the
manuscript sources. His prayer-tables for Cairo survive in numerous copies, several of which
are in his own hand (7.10). The possibility that Riwn was the compiler of the wretched tables
in MS Princeton Yahuda 861,1 cannot be excluded, but I consider it unlikely.
8.2 {Abd al-Qdir al-Minfs tables of corrections for refraction at the horizon at all latitudes
MS Istanbul Nuruosmaniye 2929,4 (fols. 22r-23v) contains two sets of tables compiled by {Abd
al-Qdir al-Minf (see 7.2 on his father) in 975 H [= 1567] displaying the correction for
2 Quoted in Dorn, Drei arabische Instrumente, pp. 32-33.
other late egyptian tables for timekeeping 337
refraction at the horizon for all latitudes (daq}iq ikhtilf m bayn al-ufq al-aqq wa-l-mar}
li-jam{ al-{ur).3 The first set (A) is according to the opinion of {Abd al-Qdir ibn
Muammad al-Minf, and the second set (B) is according to the opinion of earlier
astronomers but also computed by al-Minf. Both tables display the function D(,) for
arguments:
= 270, 300, 330, 0, 30, 60, 90 and = 1, 2, ... , 60.
Entries are given to four digits in A and three in B. No indication is given in the brief
introduction to the tables of precisely what al-Minf thought he was doing when he compiled
these tables. The two sets are based on the assumption that D varies linearly with , so that
the entries are merely proportional parts of the entries for = 60. The entries for = 30
in the two sets are:
= 270 = 0 = 90
A: 0;32,54,0 0;47,59,30 1;3,5,0
B: 0;32,0 0;47,0,0 1;2,0,0
Thus al-Minfs new tables (set A) represent an extension of his puerile attempt to improve
the values attributed to Ibn Ynus in other sources.
The notion that D is not without importance for the development of astronomical
timekeeping in Islam. We find in other sources sets of values of D for the latitudes of Mecca,
Crete, Damascus, and Istanbul, apparently based or intended to be based on this assumption.
See further 6.10, 8.8, 11.13 and 14.9.
MS Dublin CB 4067 contains the preface to a work entitled adaqat an-nir fi khtilf
al-manir, The Pupil of the Eye of the Observer of Parallax (!), by {Abd al-Qdir al-Minf.
In his preface the author mentions that the work contains an introduction, seven chapters, and
a conclusion, but the preface is followed by several pages of religious invocations and
repetitions of the name of the Prophet. al-Minf Jr. states that he wishes to discuss the problem
of the true and visible celestial circles using rigorous geometrical proofs. His father had told
him that the scholar Amad (sic) ibn Ynus had fixed the difference between the two horizons
(m bayna l-ufuqayn) as:
EQ: 47m SS: 62m WS: 32m
and that {Al} al-Dn Ibn al-Shir (9.3) fixed them as 2m more than Ibn Ynus, i.e.,
EQ: 49m SS: 64m WS: 34m .
al-Minf Jr. could not understand how Ibn al-Shir had arrived at these values. He states that
the only earlier astronomer who discussed this matter was Abu l-asan (sic) ibn al-Haytham,4
who stated that a person of height 31/2 cubits could see 0;4,26 more than half of the celestial
sphere. al-Minf Jr. could not understand what this meant, but he knew that Ibn al-Haytham
wrote two large volumes on parallax (!) which he had not seen and in which perhaps Ibn al-
3 On {Abd al-Qdir al-Minf see Suter, MAA, no. 479, and Brockelmann, GAL, SII, p. 486 (both confused);
and Cairo ENL Survey, no. C121.
4 On Ibn al-Haytham see the article in DSB by Abdelhamid I. Sabra. Prof. Sabra assures me that this subject
is not discussed in Ibn al-Haythams Optics. I have examined a manuscript in Oxford of a short treatise by Ibn
al-Haytham on the fact that the amount of the celestial sphere that is visible is more than half, but it does not
contain the numerical value given by al-Minf. However, several other manuscripts of two versions of this are
listed in Sezgin, GAS, VI, p. 260.
338 part ii, chapter eight
Fig. 8.2: Some mysterious tables in {Abd al-Qdir al-Minfs treatise Raf{ al-khilf. They purport to display
the correction to the arc of depression of the sun at daybreak and nightfall as a result of the effect of refraction
at the horizon, but most of the entries have been left blank, and it is not clear that our author knew what he
was doing. [From MS Cairo MM 123, fol. 7r, courtesy of the Egyptian National Library.]
other late egyptian tables for timekeeping 339
Haytham had explained the matter. However, he adds that Samaw}al Yay ibn {Abbs al-
Maghrib, in Part I of Book V of his work entitled Kitb al-Tahdhb f in{at al-nujm, had
made a table showing the amount of the celestial sphere visible for altitudes above the surface
up to 12,000 cubits, apparently using Ibn al-Haythams method.5 al-Minf Jr. mentions that
Abu l-asan (sic) ibn {Al ibn {Umar al-Marrkush (2.7) was asked in Rajab 694 H [= May/
June 1295] whether one could find the amount of the celestial sphere one could see standing
on the surface of the earth, and that he replied in the affirmative and had given geometrical
and numerical demonstrations which {Abd al-Qdir approved. Unfortunately, our authors
preface concludes without giving any more historical information.
Another treatise by {Abd al-Qdir al-Minf is preserved in MS Cairo MM 123, copied ca.
1600. This is entitled Raf{ al-khilf f {amal daq}iq al-ikhtilf, The Removal of the Difference
of Opinion on Operations with the Difference Minutes, and in it he mentions both Ibn Ynus
and Ibn al-Haytham again. It is not clear to me that al-Minf Jr. really knew what he was
talking about, but certainly the entire tradition of which he is at the tail end merits detailed
investigation. Some incomplete tables are found in this manuscript, whose purpose escapes me:
see Fig. 8.2.
8.3 Ibn Abi l-Khayr al-usns twilight tables for all latitudes
MS Cairo DM 1108,3 (fols. 5v-10v, copied in 1052 H [= 1642/43]) is a unique copy of a treatise
entitled Kashf al-karabt f taqq mas}il yatj ilayhi lib {ilm al-awqt, which means
something like The Exposure of Apprehensions concerning the Proper Handling of Methods
by the Student of Timekeeping, by Abu l-Khayr al-usn (see already 7.4). Part of the
introduction is devoted to a discussion of the difference minutes in which the values:
47 52 57 62 57 52 47 42 37 32 37 42
for each sign beginning with Aries are attributed to Ibn al-Shir (see 9.3). The remaining part
of the introduction is devoted to a discussion of the determination of the duration of twilight
by computation, without reference to any observations.
The work concludes with a set of tables displaying the duration of morning and evening
twilight r and s for each half-climate at the solstices and equinoxes. al-usn first gives values
of:
and Tan60
for each half-climate and then for the solstices gives values of:
Sin d, d, D, Dh .
5 On Samaw}al al-Maghribi (Suter, MAA, no. 302, and Cairo ENL Survey, no. G9), an extremely important
scholar of 12th-century Iran, see the article in DSB by Adel Anbouba, where his astronomical works are
unfortunately overlooked. On the introduction to his astronomical treatise see Rosenthal, Al-Asurlb and as-
Samaw}al on Scientific Progress. After this study was completed (the first time round), I obtained microfilms
of the Oxford and Leiden manuscripts of Samaw}als astronomical works and found that there is indeed a table
such as was described by al-Minf. al-Samaw}als astronomical works merit a separate publication, which will
contain much that is new to the history of Islamic science. See already Sezgin, GAS, VI, pp. 65-66. Parts of
the texts were investigated in my seminar on Arabic scientific manuscripts at Frankfurt during 2000-01. See
also the text to n. V-10:12.
340 part ii, chapter eight
other late egyptian tables for timekeeping 341
Then for each of the solstices and the equinoxes, as well as for the solar longitude for which
the sun passes through the zenith (i.e., such that () = ), he gives values of the quantities:
Sin H*, [Sin H* - 19], [Sin H* - 17], B, Vers tr, Vers ts, r and s ,
where tr and ts are the hour-angles at daybreak and nightfall, thus illustrating the steps by which
r and s were calculated. The tables are shown in Fig. 8.3. See further VIb-19 and also II-11.14
on the tables of {Abdallh al-alab for Aleppo.
MS Gotha A1412, which was lost from the Thringsche Landesbibliothek, Gotha, during World
War II, contained a set of four pages of prayer-tables compiled by {Abd al-Mun{im al-Nabatt
for the year 1041 H [= 1631/32].6 His prayer-tables, of which no other copy has come to my
attention, were intended to be used by pilgrims bound for Mecca, probably on the route from
Cairo. The Gotha catalogue records the introduction, which translates as follows:
I computed these useful tables, according to the latest observations, for finding the
times of prayer (ia al-awqt) by day and night at the stations on the pilgrim road
during the lunar year 1041.
Several other Arabic sources contain lists of geographical coordinates and qiblas for the stations
on the pilgrim roads from Cairo or Damascus to Mecca.7 As an example, I cite here the tables
in MS Paris BNF ar. 2560, fol. 164v, which apparently form part of the zj entitled al-Raw
al-{ir by the Damascus astronomer Ibn Zurayq (fl. ca. 1400).8 These display the latitudes
of some 32 stations on the pilgrim route from Damascus to Mecca, and the latitudes, longitudes,
and qiblas of 15 stations on the pilgrim route from Cairo to Mecca. For the few stations listed
in both tables the latitudes given are generally different. See also 10.9.
MS Cairo TR 354, copied ca. 1700, contains about 140 pages of tables, including a complete
set of tables of the functions:
t(h,), T(h,), and a(h,)
Fig. 8.3: The tables of Abu l-Khayr al-usn for determining the length of twilight for all latitudes. The tables
are of particular interest in that the compiler presents all of the auxiliary functions which he used. [From MS
Cairo DM 1108,3, fols. 9r-10v, courtesy of the Egyptian National Library.]
6 Gotha Catalogue, III, p. 63. On {Abd al-Mun{im al-Nabatt see ~hsanolu et al., Ottoman Astronomical
Literature, I, pp. 308-309, no. 167, citing the Gotha manuscript and a short treatise on timekeeping extant in
Berlin. I have seen a copy of an almanac by {Abd al-Mun{im al-Nabatt for the year 1073 H [= 1662/63],
consisting mainly of simple calendrical tables, but I forget where. He is not to be confused with {Abd al-Qdir
al-Nabatt (Cairo ENL Survey, no. D28, and ~hsanolu et al., op. cit., I, p. 279, no. 143), author of a treatise
on the agricultural and meteorological aspects (tawq{t) of the Syrian calendar, extant in MS Cairo MM 108,1,
or the latters son, {Al ibn {Abd al-Qdir al-Nabatt, author of a treatise on the algebra of inheritance and others
on simple timekeeping (Cairo ENL Survey, no. D28, and ~hsanolu et al., op. cit., I, pp. 286-287, no. 151).
7 Some examples are mentioned in King, Mecca-Centred World-Maps, pp. 84-86.
8 On Ibn Zurayq see n. 11:4.
342 part ii, chapter eight
computed to two digits for each degree of both arguments. The underlying parameters are found
by inspection to be:
= 31 (Alexandria) and = 23;35 .
The maximum solar altitude argument is thus 82 rather than 83 as in the main Cairo corpus.
The three functions are tabulated side by side as in al-Bakhniqs edition of the Cairo corpus
and in MS Cairo MM 71 of al-Khalls and Shihb al-Dn al-alabs tables for Damascus
(5.6, 10.5 and 11.2). Following these extensive tables is a single table of h0 based on the
parameters = 31 and = 23;30.
On the title-page it is stated that the tables are due to Ibn Ynus, but the person who wrote
this knew so little about the illustrious Egyptian astronomer that he called him Ibn Ynus al-
Mawil (i.e. from Mosul), confusing him with an {Irq scientist also named Kaml al-Dn
ibn Ynus.9 It is easily shown that the tables are not due to Ibn Ynus the Egyptian.
Firstly, the tables are generally less accurately computed than those in the Cairo corpus.
Secondly, the values of t and T in MS Cairo TR 354 are related by those in the table of D()
for = 31 found in MS Princeton Yahuda 861,1 of the prayer-tables for all latitudes, which
are ultimately due to Najm al-Dn al-Mir (8.1 and 6.5). Again, the tables of T(h,) for altitudes
19 and 17 in MS Cairo TR 354 show suspicious resemblance to the tables of r() and s()
for = 31 in MS Princeton Yahuda 861,1.
MS Cairo M 216,3 (fols. 14r-18v, 23r-87r, copied ca. 1800), contains a different set of
tables of:
T(h,) and t(h,)
for parameters:
= 31 and = ?.
The tables are incomplete and bound in disorder. Tables of the two functions are copied on
facing pages.
MS Cairo DM 1207,1 (fols. 1r-57v) is a copy of a set of prayer-tables for latitude 31
prepared in 1168 H [= 1754/55] by usayn, a muwaqqit in Rosetta, from a copy in the hand
of the author Riwn, dated 1090 H [= 1679/80]. The tables are not related to those in MS
Cairo TR 354 discussed above. The work begins with a brief introduction with calendrical,
solar, and astrological tables, followed by a set of geographical tables (fol. 18r) in which we
find the following entries for various cities in the Delta:
L
Cairo 54;55 30; 0
Damietta 53;50 31;25
Maalla 54;30 31; 0
Alexandria 51;54 30;58
Rosetta 52;50 31;20
(Mecca 67; 0 21; 0)
The prayer-tables (fols. 21v-35r) display the following functions:
, d, , h0, hq, Z(12), H, D, D, ha, ta, Ta, s, r, 2N, n, B, C, hq, , , s and r .
9 On Kaml al-Dn ibn Ynus see n. 7:2 above. Similar confusion also occurred in the 20th century: see
King, Ibn Ynus and the Pendulum, pp. 43-47.
other late egyptian tables for timekeeping 343
Values appear to be based on = 31 ( uncertain), although the table of Z(12) and the second
table of hq are lifted from the Cairo corpus and are based on = 30 (with q = 37). The tables
for twilight are based on parameters 19 and 17 but the tables are not the same as the
corresponding ones in MSS Cairo TR 354 and Cairo M 216. The first table of hq is based
on parameters = 31, and I was not able to investigate the underlying value of q. The function
D represents the time between midday and when the sun has set over the visible horizon and
appears to be computed using:
D = D + D + 0;16 ,
where D are the difference minutes for Cairo (4.11). The work concludes (fols. 37v-57r)
with an undated star catalogue.
MS Cairo ZK 324, copied ca. 1800, is another set of prayer-tables for latitude 31 containing
the same introduction, calendrical tables, and most of the prayer-tables that are in MS Cairo
DM 1207. MS Cairo DM 1211, copied in 1101 H [= 1689/90] in the hand of Riwn Efend,
contains some of these prayer-tables, namely:
Z(12), H, ha, ta, Ta, s , r, 2N, n , , and s ,
with a note by the first that it is for = 30 not 31.
MS Cairo M 216, fols. 19r-22v, copied ca. 1700, contains another set of prayer-tables for
= 31 displaying functions:
D, ta, Ta, ha, s, r and n .
Clearly Riwn Efend had some input in certain of these tables, but I am not inclined to
attribute the main tables of (T,t,a) for latitude 31 to him without new evidence.
MS Cairo M 241,1 (fols. 1r-33r), copied 1245 H [= 1829/30], and several other inferior copies
preserved in the Egyptian National Library, contains a set of prayer-tables for the latitude of
Damietta compiled by the Egyptian astronomer Qub al-Dn al-Maall in 1088 H [= 1677/
78].10 The following functions are displayed to two digits for each degree of :
H, , d, D, ha, ta, Ta, s, r, n, , r and ,
and based on the parameters:
= 31;25 and = 23;35 ,
with 19 and 17 for twilight. Two different conventions are used for the argument . In the
tables of Ta, , r and , the arguments are as in the main Cairo corpus. In those of H, ,
d, D, ta, s, r and n, the vertical arguments are still 1 to 30 downwards and 0 to 29 upwards,
but Aries 1 corresponds to = 0. Thus the first entry in the table is for the vernal equinox
and no entries are given for the solstices. The upward argument is absurd, and implies that
Virgo 29 corresponds to the autumnal equinox. The hapless soul who arranged these tables
in this way so that the first entry was for = 0 rather than = 1 inadvertently rearranged
the table of ha twice, since the first entry is for = 359! The entries for and d (and also
10 On Qub al-Dn al-Maall see Cairo ENL Survey, no. D37; and ~hsanolu et al., Ottoman Astronomical
Literature, I, pp. 312-313, no. 173 (lists other new sources).
344 part ii, chapter eight
H and D) are fairly accurately computed, but the table of is carelessly computed. The tables
of ta and Ta, which have two different formats, show corresponding error patterns, but some
of the entries in each are garbled.
In an Egyptian manuscript which in 1939 was in the Institut fr Geschichte der Medizin
und der Naturwissenschaften in Berlin, according to a description by Willy Hartner,11 there
were some tables giving the solar azimuth as a function of the altitude for an unspecified latitude
(I-5.1.4), as well as a table of h0 for = 31;25 (not found in MS Cairo DM 106 see I-4.8.7)
and another relating to the azimuth of Mecca, probably of the function hq. Perhaps the azimuth
tables were computed for latitude 31;25, which was the standard value for Damietta. Elsewhere
in the former Berlin manuscript there were some other spherical astronomical tables for this
same latitude, as well as a table of the corrections for refraction at the horizon Hartners
Minuten des Abstandes, this last table being attributed to an individual identified as al-Hind.12
MS Cairo DM 1207,2 (fols. 57v-58v, copied ca. 1850) see Fig. 8.7 contains a set of tables
of the functions:
2N, D and d ,
computed for latitude = 31;20 (Rosetta). The tables are copied in an atrocious hand and
are very carelessly computed. Consider, for example, the following values of d compared with
the correctly computed values d (for = 23;35):
15 30 45 60 75 90
d 3;51 7;43 10;17 12;51 14; 8 15;25
d 3;38 7; 8 10;21 13; 0 14;47 15;25
Note that the value given for argument 30 is about one-half that given for 90, and that the
values for 15, 45, and 75 were probably derived from those for 30, 60, and 90 by linear
interpolation. See also 11.9 for another example of simplistic interpolation of this kind.13
al-Jaml Ysuf ibn Ysuf, Kilrj (major domo) of asan @gh, was a student of Riwn
Efend who became rznmej in Cairo and later worked in Mecca and the Yemen; he died
in 1153 H [= 1741/42].14 One of his works was a set of prayer-tables for Crete, extant in the
unique copy MS Cairo DM 834, 70 fols., copied in 1120 H [= 1708/09], perhaps by the author.
In his introduction Ysuf Kilrj states that he prepared these tables because there were none
Fig. 8.7: Some messy prayer-tables for Rosetta, displaying the duration of night and half the true length of
daylight. [From MS Cairo DM 1207,2, courtesy of the Egyptian National Library.]
available for Crete. He mentions the traditions relating to refraction at the horizon attributed
to Ibn al-Haytham and Ibn Ynus and refers to a universal table displaying D(), although
without mentioning al-Minf (8.2). His introduction ends with the poem on the prayer-times
which he attributes to al-Shfi{ (2.1). After some tables for calendar conversion and solar
longitude and velocity Ysuf Kilrj presents tables with the format of the main Cairo corpus
of the following functions, mostly with values to three digits:
, d, , h0, B, C, hq, H, D, D, (D+ D), ha, ta, Ta, Z(12), Z(7), za(7), zb(7),
(r+D), s, r, 2N, n, Dh, 2Dh, , 2Nh, tah, Tah, sh, rh and (2N-r)h .
These are followed by tables with entries for each five days for the functions:
sh, (2n-r)h, rh, 2Nh, (r+D)h, tah, Tah and 2Dh .
The work concludes with tables of the functions:
, (*), s and r ,
and a star catalogue for 1120 H [= 1708/09] as well as an unusual kind of table displaying
the solar longitude for which various stars become visible in the west. The parameters
underlying the prayer-tables are:
= 35;30 (Crete) and = 23;30,17 .
346 part ii, chapter eight
The maxima and minima for D are 1;13,12,0 and 0;37,52,0. The corresponding values of D
for = 35;30 according to the theory of {Abd al-Qdir al-Minf (8.2) are 1;13,22 and 0;37,52.
The parameters used for twilight are 19 and 17. The azimuth of the qibla underlying the table
of hq has not been investigated. See also 14.10 on some Ottoman-type prayer-tables for the
latitude of Khania in Crete.
8.9 {Abd al-Majd Efends prayer-tables for Cairo and al-Rashds table for finding the qibla
MS Cairo Azhar falak 18082 consists of a set of prayer-tables for Cairo printed there in 1312
H [= 1894/95], and compiled by {Abd al-Majd Efend.15 These display the times of daybreak,
sunrise, midday, the afternoon prayer, sunset and nightfall in ifranj, that is, European time.
At the end of this work there is a list of geographical coordinates and qiblas of several cities,
giving the value 134;32,58 (from the north) as the qibla for Cairo. It is stated that these values
are taken from a book by Ism{l Psh al-Falak.16 This qibla is about 11/2 north of the actual
(i.e., modern) qibla for Cairo, but is about 71/2 south of the qibla favoured by the medieval
Egyptian astronomers (q = 53) (4.7).
Following this list of coordinates is a discussion of the use of a compass for finding the
qibla. Then {Abd al-Majd Efend presents a table for determining the qibla which he states
was compiled by Muammad al-Rashd al-Falak.17 This purports to give the time of each
day of the Coptic year when the sun indicates the azimuth of Mecca. The times are given
according to the Ottoman convention that sunset is 12 oclock (14.0), and the underlying
azimuth of Mecca which gives the closest fit to the entries in the table (maximum: 4;19h, mean:
3;40h, and minimum: 3;0h) is about 54, measured from the meridian. Thus it appears that al-
Rashd simply converted the table of tq in the Cairo corpus to hours and minutes, Turkish time.
MSS Cairo K 3792-3802, copied ca. 1900, consist of eleven small booklets of prayer-tables
compiled by Muammad Efend if, an official of the Egyptian Customs in {Abbsiyya,
Cairo: see Fig. 8.10a-b.18 The tables display the times:
r (fajr), R (shurq), ({d), z (zawl), m (uhr),
a ({ar awwal), b ({ar thn), S (maghrib ifranj), s ({ish})
Generally, separate tables are given for Arab ({arab), i.e., Turkish, and local mean time
(ifranj). Where the time for the {d prayer is given it is usually between 20 and 30m after sunrise.
The localities served by these tables are Cairo ( = 30); Assiut (27); Qusayr (26;7); Sohag
15 On {Abd al-Majd al-Falak see Azzawi, History of Astronomy in Iraq, p. 325. He does not seem to be
listed in ~hsanolu et al., Ottoman Astronomical Literature.
16 On Ism{l Psh al-Falak see Azzawi, op. cit., pp. 328-329. He does not seem to be listed in ~hsanolu
et al., Ottoman Astronomical Literature.
17 This al-Rashd is apparently not mentioned in the modern sources.
18 See Cairo ENL Survey, no. D194.
other late egyptian tables for timekeeping 347
Fig. 8.10a: An extract from Muammad Efend ifs prayer-tables for latitude 25 and Turkish time ({arab),
serving August and September. The entries are in minutes, with the appropriate hour indicated above a diagonal
slash in the first entry of each column. [From MSS Cairo K 3792-3802, courtesy of the Egyptian National
Library.]
Fig. 8.10b: These tables display the shadow at the uhr for every two days of the year, also for latitude 25.
The entries are in feet and decimal fractions thereof. [Same source.]
and Akhmim (26;30); Qus (26); Edfu and Esna (25); Medina (both 24 and 24;10); Markaz
al-Durr (?) (23); Korosko (Markaz al-sidr) (22;30); Wadi Halfa (22); Mecca (both 21;28
and 21;30); and Jabal al-r (in the Sinai) (28;5). Some other funtions are tabulated for certain
latitudes.
348 part ii, chapter nine
CHAPTER 9
The first Muslim astronomer known to have worked in Damascus (as well as Baghdad and
Samarra) was abash al-sib (ca. 850).2 He wrote at least three zjes, one of which, called
al-Zj al-Dimashq, the Damascus Zj, survives and has been studied by Ted Kennedy and
Marie-Thrse Debarnot.3 abash did work of considerable interest, and is to be regarded as
the most original astronomer of the early Islamic period, as well as one of the most prolific.
His early work in mathematical astronomy shows considerable Indian influence, but this is
supplanted by Greek influence and then he lived for 100 years we find him solving higly
complicated problems which went beyond the interests of earlier traditions. His work was in
a sense experimental for the Islamic world was not yet ready for some of his achievements,
some of which had far less influence in later Islamic astronomy than they shoud have had.
Thus, for example, he compiled a set of tables for constructing horizontal sundials for a range
of latitudes (I-4.1.1); he compiled the first set of universal auxiliary trigonometric tables for
solving problems of spherical astronomy for all latitudes (I-9.1); he was the first to tabulate
the standard trigonometric functions that we know today; he conceived an astrolabe based on
a projection preserving azimuth and distance to the celestial pole;4 and he was the first to solve
the qibla problem accurately by the so-called method of the zjes, the mathematics of which
underlies the cartographic grids for world-maps centred on Mecca preserving direction and
distance to the centre, known from three Safavid examples.5
Shortly after the time of abash, we find al-Battn (ca. 910) making serious observations
at Raqqa and compiling a zj which, as a competent epitome of Ptolemaic mathematical
astronomy with a distinct Islamic flavour, was to exert far more influence in medieval and
Renaissance Europe than it actually did in the Islamic world.6 This work marks the successful
appropriation and synthesis of Ptolemaic planetary astronomy and Greek and Indian spherical
astronomy and trigonometry, a process to which abash had been the main contributor.
The only astronomers know to have prepared zjes in Syria between the time of abash and
al-Battn and the 14th century, when Damascus became the leading centre of astronomy in
the Islamic (and indeed in the entire) world are:
1 I have attempted an overview of astronomy in Syria in Paris IMA 1993-94 Catalogue, pp. 386-395. See
also my Astronomy of the Mamluks.
2 On abash see n. I-9:1.
3 Kennedy, Zj Survey, pp. 151-154, and Debarnot, Zj of abash.
4 See Kennedy & Lorch & Kunitzsch, Melon Astrolabe.
5 See King, Mecca-Centred World-Maps, esp. pp. 61-64 and 241, and also VIIc.
6 On al-Battn see n. I-4:11.
early syrian tables for timekeeping 349
(1) Ibn al-Dahhn (ca. 1170) who worked in the service of the Ayyubid ruler al al-Dn
(Saladin);7
(2) Mu}ayyad al-Dn al-{Ur (ca. 1250), who later assisted in the observations at Maragha
in Iran under the direction of Nar al-Dn al-s;8
(3) Muyi l-Dn al-Maghrib (ca. 1250), who also went on to work in Maragha;9 and
(4) Ibn al-Lubd (ca. 1250), who worked in both Damascus and Cairo.10
The zjes of Ibn al-Dahhn, al-{Ur and Ibn al-Lubd are no longer extant, and the zj which
al-Maghrib compiled in Damascus, called Tj al-azyj, Crown of the Zjes, survives but has
not been studied in detail. A fragment from the Zj of Ibn al-Dahhn preserved in a Yemeni
manuscript is discussed in 9.1. The recovery of al-{Urs Zj would contribute to our knowledge
of the development of Islamic planetary theory, since al-{Urs work influenced Ibn al-Shir
(see below). Likewise a detailed study of the Zj of al-Maghrib and his later zjes prepared
at Maragha, would also contribute towards clarifying our knowledge of astronomy in Syria
and its influence on the Maragha school. Important for our understanding of the influence of
Andalus and Maghrib influence in Syria is the recension for Damascus of the Zj of the 13th-
century Tunisian astronomer Ibn Isq (13.1).
In the 14th century a number of astronomers flourished in Syria, mostly associated with the
Umayyad Mosque in Damascus.11 Ibn al-Sarraj, who worked in Aleppo apparently without
institutional backing, specialized in devising new astronomical instruments.12 Foremost
amongst Damascus scholars were Ibn al-Shir, who made substantial contributions to planetary
theory,13 and al-Mizz14 and al-Khall,15 whose interests were in astronomical timekeeping.
Each of these 14th-century astromomers devised different kinds of quadrants and other devices
for solving problems in spherical astronomy.16 There was especial interest in instruments which
would provide solutions for all latitudes (VIb). Both Ibn al-Shir and al-Mizz had studied
astronomy in Alexandria before taking up their positions at the Umayyad Mosque, and the
Egyptian influence is apparent in their interests and achievements.
al-Khalls extensive tables for timekeeping (Ch. 10), which were used in one form or
another from the 14th to the 19th century, replaced an earlier set compiled by al-Mizz (9.2).
Ibn al-Shir also prepared some prayer-tables (9.3). Another student of al-Mizz was al-Karak,
who worked in Jerusalem and adapted the tables of Ibn al-Rashd to form the main Jerusalem
corpus (9.4).
Damascus was razed to the ground by the Mongols under Timurlang in 1401. There was
little creative astronomical activity of any consequence in Syria thereafter: the centre of Islamic
planetary astronomy moved to Samarqand, and the centre of astronomical timekeeping moved
back to Cairo for a century or two before it finally moved to Istanbul.
Fakhr al-Dn Ab Shuj{ Muammad ibn {Al known as Ibn al-Dahhn, who worked in
Damascus ca. 565 H [ ca. 1170], was an astronomer associated with the celebrated ruler al
al-Dn (Saladin).17 He compiled a zj, which is unfortunately no longer extant. However, a
fragment survives in the Sanaa miscellany of al-Afal (12.4), and consists of a set of tables
giving the values of the following functions at the summer solstice:
d, dh, H, Z(12) and za(12)
for some 120 cities.18: see Fig. 9.1. Values are given to two digits. The underlying value of
is 23;35. The relation between the last two functions is simply:
za = Cot ha = Cot H + 12 = Z + 12.
Fig. 9.1: An extract from the beginning of Ibn al-Dahhns tables. Notice that the place-names are not visible
on this page. A reconstruction of the entire table and the geographical coordinates underlying it would be possible
and even worthwhile; but where are the volunteers? [From Varisco & Smith, eds., al-Afals Anthology, p. 151,
courtesy of Professor Daniel Varisco.]
17 On Ibn al-Dahhn see Suter, MAA, no. 310; Kennedy, Zj Survey, no. 89; and King, Astronomy in Yemen,
p. 37. Add to II n. 10:17: This Ibn al-Dahhn is not to be confused with {Imd al-Dn ibn al-Dahhn, chief
of the astronomers in Mosul ca. 1285, for a reference to whom see Cook, Muslim Material on Comets and
Meteors, p. 147, and Rada, Comets in Arabic Literature, p. 45.
18 See now the facsimile in Varisco & Smith, eds., al-Afals Anthology, pp. 151-154.
early syrian tables for timekeeping 351
The azimuth of the qibla measured from the meridian is also given,19 as well as the vowelling
of the Arabic names of the cities. Not all of the city names are visible on the microfilm which
I have used, but there seems to be no entry for Mecca (as one can see from the qibla-values,
if not the names). Sample significant entries for other cities (values of are calculated from
those of H) are:
Locality d max H [ ] q
Cairo 14;3 84;20 [29;15] 54;39
Jerusalem 16;5 80;35 [33;0] 41;21
Damascus 16;40 80;5 [33;30] 28;29
Ibn al-Dahhn is known to have visited Cairo, but his value for the latitude of that city,
confirmed by the value of d (so that it is not a scribal error for 29;55), is too small by 3/4.20
Likewise his value for the latitude of Jerusalem is too large by about 1. (However, a later
qibla table compiled for Jerusalem is based on the parameter 41;30, rather than on values which
can be derived using a more accurate latitude for Jerusalem (see further 9.5).) Of the Islamic
geographical tables investigated by the Kennedys (1.3), only the table in al-Brns Qnn
has the same latitudes for these three cities as Ibn al-Dahhn. However, Ibn al-Dahhns qibla
values do not correspond to those that can be computed from al-Birunis coordinates using
either the exact or the standard approximate formulae.
The astronomer Shams al-Dn Muammad ibn Amad ibn {Abd al-Ram al-Mizz (1291-
1343), whose family originated from a suburb of Damascus, studied under Ibn al-Akfn in
Alexandria, and then moved back to Damascus where he eventually worked in the Umayyad
Mosque.21 He was a fine instrument-maker, and several instruments of his construction and
treatises on their use have survived.22 He also compiled some tables for timekeeping. The main
set appears to be no longer extant, but, according to his student al-Karak (9.4), it consisted
of tables of the hour-angle t(h,) computed for latitude 33;27, that is, Damascus: see I-2.1.3.
al-Mizzs minor set of tables survives in MS Cairo MM 62, copied ca. 1400. The tables are
not accompanied by any introduction, and the functions tabulated are:
, H, Z(7), d, D, , 2Dh, 2N, h0, , , , ha, ta, Ta, r and s .
The entries are arranged with the format of the main Cairo corpus (see Fig. 9.2), and are based
on al-Mizzs distinctive parameters:
= 33;27 and = 23;33 ,
on which see below. They are very accurately computed, although al-Mizz used an
interpolation system in the fourth columns of his tables of r and s which failed to take account
of the actual behaviour of these functions for solar longitudes in Libra and Pisces. This means
Fig. 9.2: An extract from al-Mizzs prayer-tables showing the sub-tables for the time from the beginning of
the {ar to sunset and the duration of evening twilight (based on a solar depression of 17). [From MS Cairo
MM 62, fols. 8v-9r, courtesy of the Egyptian National Library.]
that the corresponding entries in his hour-angle tables would also be in error. The 15th-century
astronomer Jaml al-Dn al-Mridn23 states that al-Mizz used parameters 20 and 16 for
morning and evening twilight, but the tables of r and s in this Cairo manuscript are based on
19 and 17.
MS Florence Laurenziana 24 is the only known copy of a recension of the lkhn Zj for
Mosul, prepared by D}d al-Mawil.24 On fols. 1r-1v there is an incomplete set of tables that
are unrelated to the zj of al-Mawil. On fol. 1r there is a simple table for finding Easter and
another of () with entries to two digits for each degree of and based on = 23;33. On
fol. 1v there is the first page of a set of prayer-tables for an unspecific latitude. Entries are
given to two digits for each degree of from 1 from 30. The functions tabulated are:
H, D, , h0, ha, ta, 2Dh, , s and r ,
and the values are those of al-Mizz for Damascus.
The latitude 33;27 for Damascus is attested in MS Paris BNF ar. 2513, fols. 87v-88v, of
the Egyptian 13th-century Muala Zj, and from MS Paris BNF ar. 2520, fol. 72r, of another
rescension of the same work we learn that both parameters were derived by observations made
by an individual referred to as al-Sharf, to be identified with the 10th-century Baghdad
astronomer Ibn al-A{lam (6.6). Ibn al-Shir also used these parameters in his early work: they
are found, for example, in Chs. 18 and 114 of his treatise on the complete quadrant (MS
Damascus hiriyya 3098). In the year 765 H [= 1363/64] Ibn al-Shir derived a new pair
of parameters, for which al-Khall computed a new corpus of tables for Damascus (Ch. 10)
to replace those of al-Mizz.
23 I forget where I read this, but surely in an instrument treatise by al-Mridn (6.15).
24 On the lkhn Zj see Kennedy, Zj Survey, no. 6. D}d al-Mawil, not mentioned in the modern
literature, states that he was asked to prepare the recension by al-abb (the medic) {s and that it took him
twenty days to translate the introduction into Arabic.
early syrian tables for timekeeping 353
The Damascus astronomer Ibn al-Shir (1306-1375/80), author of two zjes and an extensive
treatise on theoretical astronomy in which he explained his new models for the sun, moon,
and planets,25 also compiled a set of prayer-tables for latitude 34, possibly intended for the
new Mamluk city of Tripoli (accurate latitude 3427). The evidence for this is contained in
a short treatise attributed to him, preserved in MS Leiden Or. 1001, fols. 108r-113r, which
was originally intended as an introduction to the tables. The tables themselves are extant in
a different manuscript (see below).
Ibn al-Shir begins his introduction by stating that one of his friends asked him to prepare
a set of prayer-tables for latitude 34. The functions:
H, Z(7), D, 2N, ha, ta, Ta, r, s, and tq
are specifically mentioned (fol. 110v:16-19). Ibn al-Shirs description of the method of
finding the value of a given function for a particular solar longitude indicates that the functions
were tabulated side by side. Thus they would resemble the tables of al-Khall in appearance,
rather than those of Ibn Ynus and al-Mizz. He also notes how to compute:
, d, , 2Dh and za(7)
from the entries in the tables, indicating that he did not actually tabulate these functions.
MS Cairo DM 1170,2 (fols. 11r-22v), copied about 1700, is the only known copy of an
anonymous set of prayer-tables for latitude 34 which answer the description of Ibn al-Shirs
introduction in the Leiden manuscript: see Fig. 9.3. The tables on fols. 14v-19r are preceded
by an introduction, of which the beginning is missing, and by calendrical and solar longitude
tables (fols. 11r-14r), and they are followed by a star catalogue for the end of the year 1033
H [= 1624] and an almanac (fols. 19v-22v). According to a note in the margin of the star
catalogue (fol. 21r) the entries in this, and probably also the other introductory tables, were
adopted by Muammad ibn al-ulwn26 for 1033 H [= 1623/24] from another set compiled
by {Al} al-Dn ibn Nir al-Din al-arbulus, a muwaqqit at the Umayyad Mosque in
Damascus, for the end of 984 H [= 1577]. Values of the following functions:
H, Z(7), D, 2Dh, ha, ta, Ta, 2N, s, r, n and tq ,
computed to two digits, are displayed side by side for each degree of beginning with Capricorn
1. The underlying parameters are:
= 34;0 and = 23;31.
I have not been able to investigate the parameters used for twilight or the qibla. Additional
tables display the functions:
d, and , s and r
for the same and . It seems to me most likely that this collection of tables are the prayer-
tables of Ibn al-Shir.
Added in June, 2002: Whilst finalizing the illustrations for this study I realized that in the 1970s
I had not included mention of the tables for latitude 34, specifically intended for Tripoli,
25 On Ibn al-Shir see the references cited in n. I-2:15.
26 On Muammad ibn Ibrhm al-ulwn al-usn al-Shfi{ see Cairo ENL Survey, no. D22; and ~hsanolu
et al., Ottoman Astronomical Literature, I, pp. 280-281, no. 146.
354 part ii, chapter nine
Fig. 9.3a: An extract serving solar longitudes in Taurus and Cancer from a set of prayer-tables for latitude 34
probably due to Ibn al-Shir. [From MS Cairo DM 1170,2, courtesy of the Egyptian National Library.]
Fig. 9.3b: An extract for solar longitudes in Pisces from an an anonymous set of prayer-tables for Tripoli (latitude
34;0). [From MS Cairo M 228,3, fols. 16v-17r, courtesy of the Egyptian National Library.]
early syrian tables for timekeeping 355
contained in MS Cairo M 228,3, fols. 9r-9v + 16r-21r, copied ca. 1500.27 I cannot compare
these now with those in the Cairo DM manuscript. Suffice it here to note that the functions
tabulated are:
H, D, 2Dh, ha, ta, Ta, tb, 2N, s, r, n, and tq ,
and to present an extract in Fig. 9.3b. The functions tb and n are referred to as m bayn al-
uhr wa-khar al-{ar and m bayn al-dhnayn, respectively. Clearly, the two sets, and
possibly also those in 11.6, merit a new investigation.
The treatise in the Leiden manuscript also contains a description of how to compute the
solar longitude from the date, the time of day from shadow lengths, the solar altitude at the
beginning of the afternoon prayer, and the times of rising and setting of the moon: all these
methods are approximate though this is not stated. For example, to find the time elapsed since
sunrise when the shadow is z (measured to base 7), the author suggests:
T { 42 / [ (z + 7) - Z(7) ] } sdh .
Equivalent versions of this Indian approximate formula were used by various Islamic
astronomers from the 8th century onwards,28 but one is suprised to see an astronomer of the
calibre of Ibn al-Shir advocating it. Likewise, the method outlined by him for finding the
altitude at the beginning of the afternoon prayer involves the approximation:
ha 1/2 [ H + 1/2 1/6 ] ,
and the reader is warned against using:
ha 1/2 H + 1/2 1/6 .
Approximate methods of this kind are also given in Ibn al-Shirs treatise on the complete
quadrant (MS Damascus hiriyya 3098, e.g. Chs. 114 and 173). For similar approximations
in other sources, see the references in 1.4.
In his treatise on planetary theory entitled Nihyat al-su}l (MS Oxford Marsh 139, fol. 7r),
Ibn al-Shir mentions that the person who stated that the visible portion of the heaven is 0;4,26
more than the invisible portion was in error. (Shams al-Dn al-Minf (7.1) associates that value
with the illustrious Ibn al-Haytham: see also 14.9.) He adds that he had observed the half arcs
of visibility of the sun and stars and found that they were greater than the arcs found by
calculation by an amount exceeding 2/3. In what purports to be a set of notes due to the Egyptian
astronomer {Izz al-Dn al-Waf} the values of the difference minutes for Cairo (4.11) are
attributed to Ibn al-Shir (5.7, also 8.3), but I feel that this attribution is somewhat suspect.
In a study of Ibn al-Shirs contributions in astronomical timekeeping it would be amiss
not to mention the magnificent sundial which he constructed for the main minaret of the
Umayyad Mosque. The remains of this instrument are today preserved in the garden of the
Archeological Museum in Damascus, but in its original position there is a faithful replica made
by the late-19th-century muwaqqit al-anw (11.13). The sundial serves to regulate time with
respect to each of the five times of prayer.29
MS Leipzig UB 808, fols. 3r-93r, penned in 805 H [= 1402], is an apparently unique copy
of a set of tables for timekeeping compiled by Zayn al-Dn Ab Bakr al-Karak30 for the latitude
of Jerusalem. He appears to have been a student of al-Mizz (9.2) and can thus be dated to
the mid 14th century. The introduction to his tables on fol. 2v is of considerable relevance to
the present study and I present a translation:
In the Name of God, the Merciful and Compassionate. The shaykh and imm Zayn
al-Dn Ab Bakr ibn Muammad ibn Ayyb al-Tamm known as al-Karak, muwaqqit
in Sacred Jerusalem may God have mercy on him said the following in the first
section (of his book), which concerned the compilation of tables of time since sunrise
and the hour-angle for latitude 32 north. When I saw that the virtuous shaykh, scholar,
and calculator Jaml al-Din Abu l-{Abbs Amad ibn {Umar ibn Ism{l ibn
Muammad ibn Ab Bakr al-f al-Maqs may God have mercy on him had
compiled tables of the time since sunrise for latitude 30, which require further
calculation to find the hour-angle, and that our teacher the virtuous shaykh and
meticulous scholar Shams al-Dn Ab {Abdallh Muammad ibn Amad ibn {Abd al-
Rahm al-Mizz may God Almighty have mercy on him had compiled tables of
the hour-angle for latitude 33;27, which also require further calculation to find the
time since sunrise, and that the shaykh Shams al-Dn Ibn al-Rashd may God have
mercy on him had put the altitude at the head of each table and the hour-angle
opposite the solar longitude, I wanted to participate with them in such compilations.
So I put the significant (functions) together to facilitate the work of the observer, and
tabulated the time since sunrise and the hour-angle opposite the altitude with the solar
longitude at the head of the tables, beginning at the first point of the ascending zodiacal
signs (i.e., the winter solstice). The same tables can be used for the descending signs
in the opposite direction (since the last table is for the summer solstice). I did this,
asking help from God and placing my trust in Him ... .
al-Karaks tables do indeed display the functions:
T(,h) and t(,h)
for each degree of both arguments. Note that values are given for both solstices and the
equinoxes. For each value of , D() is also tabulated. The underlying parameters are:
= 32;0 and =23;35 .
The entries, which are given to two digits, are rather accurately computed. Note that al-Karak
does not actually say that he computed both of the functions himself. Indeed, it may be that
he only computed the time since sunrise from someone elses hour-angle tables. On fols. 94v-
123v of the Leipzig manuscript, in a different hand, there is another set of hour-angle tables
arranged in the same way as al-Karaks tables, and also showing D(). The title and colophon
indicate that they are for latitude 31;40. In fact they are for 32;0 and the entries are the same
as those in al-Karaks set. A note in the incomplete introduction to this second set on fol. 94r
30 The Leipzig manuscript of al-Karaks tables is mentioned in Brockelmann, GAL, SII, p. 156, after Leipzig
Catalogue, p. 261.
early syrian tables for timekeeping 357
mentions the completion of the minaret on a mosque in Ramla in 797 H [= 1395]. The value
31;40 is used for Ramla in several Islamic geographical tables.31
al-Karak was clearly impressed with al-Maqss tables of T(h,) for Cairo (5.4), and al-
Mizzs tables of t(h,) for Damascus (9.2). It is a pity that he does not state the latitude for
which Ibn al-Rashd compiled his hour-angle tables. His name is not otherwise associated with
any of the various hour-angle tables that I have found in the manuscript sources. It seems to
me likely that Ibn al-Rashd compiled a set of tables of t(h,) for Jerusalem, which al-Karak
then used to compile his tables with a different format. Now fragments of tables of t(h,) for
latitude 32 exist in MSS Cairo DM 45 and DM 153 (6.12) with identical corresponding entries
to al-Karaks tables. If my supposition is correct, al-Karaks tables for Jerusalem are merely
an adaption of Ibn al-Rashds tables, which were probably compiled in Cairo. See also the
next section.
The tables for latitude 32 in MS Princeton Yahuda 861,1 of the Natja al-kubr attributed
to al-Waf} (8.1) differ from those for other latitudes in such a way that it is immediately
obvious that they were lifted from various independent sources. The following functions are
tabulated for this latitude:
D, 2Dh, 2N, H, ha, ta, hq, Tq, r, s, n, , and .
The entries in certain of the tables are reasonably accurately computed and in others are very
carelessly computed. My investigation of the tables reveals the following.
(1) The table of D, which is carelessly computed, is the same as the corresponding table
in MS Oxford Marsh 676 (Uri 944 = 995) due to Najm al-Dn al-Mir (6.5).
(2) The table of 2Dh, the length of daylight in equinoctial hours, is a particularly wretched
specimen. Some of the entries, which are badly garbled, are in hours and minutes and
others are in hours and degrees.
(3) The table of 2N, which is rather accurately computed, is based on values of D other
than those in (1). In fact it is based on the values of D used by al-Karak to compile
his tables of T(,h) from (Ibn al-Rashds ?) tables of T(h,) (9.4).
(4-6) The tables of H, ha and ta are likewise rather accurately computed.
(7-8) The tables of hq and Tq are based on a value of q for some particular locality, although
neither parameter nor locality is specified for either function. The latitude 32 is the
only latitude other than 30 for which hq and Tq are tabulated in the Princeton
manuscript. However, in the title of the first function tabulated for latitude 32, namely
D(), the cities of Gaza, Ramla and Jerusalem are specifically mentioned, and by
inspection I find the underlying value of q to be 41;30 (measured from the meridian).
The two tables are reasonable accurately computed. Now in the geographical tables in
MS Princeton Yahuda 861,1 the following entries for localities in Palestine are given:
L q q (acc.) q (approx.)
Gaza 56;30 32;0 52;40 42;59 43;41
Ramla 57;20 32;20 45;0 39;37 40;31
Jerusalem 58;35 32;10 45;51 35;58 37;5
Nablus 58;15 32;20 37;0 36;41 37;45
(Mecca 67;0 21;0 - - -
In each of several earlier Egyptian and Syrian sources I have consulted,32 the qibla
values have been carelessly computed, and in some cases miscopied to boot. However,
in none of these sources do we find the qibla value 41;30. It was probably derived
for Jerusalem. Indeed it was perhaps based on the value 33;0 for the latitude of
Jerusalem, as found, for example, in the tables of Ibn al-Dahhn (9.1). Notice that if
we take
L = 56;30 and =33;0 with LM = 67;0 and M = 21;30 ,
the accurate and approximate values of q are 41;37 and 42;26, and if we take:
L = 57;0 and = 33;0 with LM = 67;0 and M = 21;40 ,
then the accurate and approximate values of q are 40;31 and 41;28. Ibn al-Dahhn
himself gives 41;21 for the qibla at Jerusalem.
(9-11) The tables for twilight are based on parameters 20 and 16 for morning and evening
and are reasonably accurately computed. The entries are the same as the corresponding
ones in al-Karaks tables of T(,h) for h = 20 and 16. All of the other twilight tables
in the Princeton manuscript are based on parameters 19 and 17.
(12) The table of is carelessly computed, but is based on a different set of values of d
than was used to compile the table of D (see (1) above).
(13) The table of s is based on a more accurate set of values of than (12), using:
s() = ( + 180) + s() .
(14) The table of (), displaying the time from sunset to the salm, is ostensibly based on:
() = 2N() - r() ,
so that the salm was to be performed precisely at daybreak rather than a few minutes
before as was the case in Egypt (4.10). However, the last four columns of entries are
badly garbled.
(15) Finally, amongst the tables for latitude 31 in the Princeton manuscript there is one
displaying the functions t(h) and T(h) for each degree of h at the equinoxes. The
underlying latitude is 32 not 31 as stated, and the tables originally formed part of
the more extensive set of tables compiled by al-Karak. The entries are badly garbled,
From these investigations it is clear that there existed a set of carefully-computed prayer-tables
for Jerusalem displaying at least the functions:
D, 2N, H, ha, ta, hq, Tq, r, s, n and ,
thus complementing al-Karaks extensive set of tables for timekeeping.
32
Namely, the geographical tables in MSS Paris BNF ar. 2513, fol. 89r, and Paris BNF ar. 2520, fol. 82v,
of the Muala Zj; MS Oxford Seld. A30, fols. 155r-157v, of the Zj of Ibn al-Shir; MSS Paris BNF ar. 5968,
fols. 162v-163r, of the anonymous Dastr al-munajjimn; and Gotha A1403 of the derivative zj of Ibn Zurayq.
See further King, Mecca-Centred World-Maps, pp. 76-84, on the unhappy state of such tables.
the damascus corpus of al-khall 359
CHAPTER 10
An impressive corpus of astronomical tables was used in Damascus from the 14th to the 19th
century. The tables in their entirety contain over 40,000 entries. Their compiler was a
Damascene astronomer named Shams al-Dn al-Khall, a man recognized only in the 1970s
for his remarkable competence in computation.1 al-Khalls corpus is more clearly-defined than
the Cairo corpus, by which it was obviously inspired. We know who compiled the tables, and
we are even provileged to have an earlier work of his which shows how his thinking had
progressed. And the question why we had to wait to the late 20th century to have a clear picture
of his achievements is simple to answer: his tables were rehashed and modified over the
centuries and numerous of the manuscripts bear the names of later astronomers who simply
appropriated them.2
Damascus was home to a vigorous tradition of astronomy in the 14th century, centred on
the Umayyad Mosque in the heart of the city.3 This tradition was but one of many colourful
aspects of life under the Mamluks, who ruled Egypt, Syria, Palestine and the Hejaz from the
mid-13th to the early 16th century. A team of astronomers called muwaqqits (that is, professional
timekeepers), which included al-Khall, was employed at the Mosque. Their main duty was
to determine the times of prayer so that the muezzins could perform the call to worship from
the minarets at the appropriate time. In addition, they specialized in the regulation of the lunar
calendar and the determination of the qibla. The interests of some of the muwaqqits extended
far beyond these somewhat mundane tasks (V).
The chief muwaqqit at the Umayyad Mosque during the 1360s and 1370s was {Al} al-Dn
{Al ibn Ibrhm called Ibn al-Shir, known today for his innovative and successful research
on geometric models for the sun, moon and planets (9.3). An earlier Damascus astronomer
was al-Mizz, who started the business of compiling prayer-tables for the city (9.2). As we
shall see, both al-Mizz and Ibn al-Shir influenced the work of al-Khall. Both had themselves
1 Before the 1970s al-Khall was known only from the bio-bibliographical references in Suter, MAA, no.
418, Brockelmann, GAL, II, pp. 156-157, SII, p. 157, and Sarton, IHS, III:2, pp. 1526-27. My first studies on
his tables were al-Khalls Auxiliary Tables (1973) and al-Khalls Qibla Table (1975). A summary of
his achievements and a critical list of the available manuscripts of his works appeared in DSB, XV (Supp. vol.
I), pp. 259-261. See also Cairo ENL Survey, no. C37.
2 Appropriated is a nicer word than plagiarized: see n. 4:12 above. The Damascus of the astronomers,
like Cairo, was a small world and it was not possible for astronomers to pass off as their own tables that
were computed by their illustrious predecessor.
3 This activity is surveyed in King, Astronomy of the Mamluks, and idem, Astronomie en Syrie. On
timekeeping in particular see my Astronomical Timekeeping in Medieval Syria.
360 part ii, chapter ten
been influenced by developments in Egypt in the 13th century (6.0). al-Khall recomputed al-
Mizzs tables for the new parameters (local latitude and obliquity of the ecliptic) derived by
Ibn al-Shir in 765 H [= 1363/64], namely:
= 33;30 and = 23;31.
His corpus of tables for solar timekeeping and the regulation of the times of prayer for Damas-
cus continued to be used in one form or another until the 19th century. In addition, al-Khall
prepared a set of auxiliary tables for solving the standard problems of spherical astronomy for
all latitudes, as well as a universal qibla table.
The work of the early-14th-century Aleppo astronomer Shihb al-Dn Amad ibn Ab Bakr
called Ibn al-Sarrj, which represents the high point of Islamic instrumentation, was also known
to al-Khall.4 Indeed, the tables of al-Khall and the instruments of Ibn al-Sarrj are singularly
important developments in the particularly Islamic tradition of providing solutions to the
problems of spherical astronomy for all latitudes, which I have documented elsewhere (VIa-
b). However, al-Khall makes no reference to any of his predecessors except Ab {Al al-
Marrkush (10.8).
In later centuries al-Khalls universal auxiliary tables (10.7) were also used in Egypt,
Tunisia and Turkey. As far as we know, al-Khall had no influence on the development of
astronomy or trigonometry in Europe. There in the 15th and 16th centuries auxiliary trigono-
metric tables were developed by a series of astronomers of whom the most important were
Regiomontanus and Magini (I-10.1-2). Yet despite the sophistication of these European tables
they are not as useful for practical purposes as those developed by al-Khall. His tables would
have been of great interest to European astronomers of the from the 15th century right through
to the early 20th century, had they been known to them.
7 On Sharaf al-Dn al-Khall see Suter, no. 427, and Cairo ENL Survey, no. C38.
362 part ii, chapter ten
Location of the various tables of the main corpus in the available manuscripts
I al-Khalls prayer tables
II al-Khalls hour-angle tables
III al-alabs azimuth tables
IV al-Khalls universal auxiliary tables
V al-Khalls qibla table
MS I II III IV V
Berlin Ahlwardt 5754-6 * - - * *
Berlin Ahlwardt 5758 * - - - -
Berlin Ahlwardt 5759/5771 * - - - -
Berlin Ahlwardt 5760/5772 * - - - -
Berlin Ahlwardt 5739 - - - * -
Beirut PC - - - * -
Cairo MM 71 - * * - -
Cairo MM 43 - - - * -
Cairo DM 184 * - - - -
Cairo MM 98 - - - * -
Cairo M 255,7 * - - - -
Cairo DM 758 - - - * -
Cairo K 8525 * * * - -
Cairo M 228,5 * * - * -
Cairo M 173 * * - - -
Cairo TR 129 * * - - -
Cairo DM 1007 * * - - -
Damascus 3116 - * - - -
Damascus 10387 (?) * - - - -
Damascus 9227 - - * - -
Damascus 9233 * * - - -
Damascus 4893 * - - - -
Escorial r. 931,8 - - - * -
Gotha A1412 * - - - -
Istanbul Ayasofya 2590 - - - * -
Istanbul Esat Efendi 1990 * - - - -
Istanbul Hamidiye 1453,3 - - - * -
Istanbul Serez 1914 - - - * -
Leipzig 814 * - - - -
London BL Add. 9599,31 - - - * -
Mosul al-Muammadiyya 129 - * - - -
Oxford Seld. Supp. 100 * * - - -
Oxford Marsh 39 (Uri 1042) - * - - -
Oxford Marsh 95 (Uri 961) * - - - -
Paris BNF ar. 2558 * * - * *
Paris BNF ar. 2560,11 - - - - *
Princeton Yahuda 861,2 * - - * -
the damascus corpus of al-khall 363
8 These were prepared in 1993 in the process of the compilation of my detailed study mentioned in n. 1:23
above. I prefer them to the descriptions presented in 4.2.
9 The manuscript (MS 4091) is part of a collection of treatises in the same hand (MSS 4090-92). The Arabic
foliation runs from 159 to 181 skipping a folio between 165 and 166. The modern foliation is also incorrect.
The notes here follow the correct foliation.
10 I am most grateful to my colleague Dr. Sonja Brentjes for obtaining a microfilm of this manuscript.
364 part ii, chapter ten
No title folio.
1v-4r: al-Khalls introduction to his universal auxiliary tables.
4r-5v: Sharaf al-Dn al-Khalls notes on the use of his uncles tables.
6r: Title for al-Khalls universal auxiliary tables (jadwal fal al-d}ir al-fq).
6v-46v: The universal auxiliary tables (including a sub-table for = 21;30). For each
latitude various localities are named. The tables run to latitude 55 with a page
ruled for 56 without entries.
47r: al-Khalls introduction to his qibla table.
47v-55r: The qibla table.
55v: Blank.
56r: Scribbles.
56v-58v: al-Khalls introduction to his universal auxiliary tables (incomplete), in hand
B.
59r: Cotangent table to base 10 (argument 1-90).
59v: Sine table.
60r-61r: Notes on Cotangents.
61v: Table of solar declination and solar altitude at the {ar (attributed to al-
Khall).
62r: Blank.
62v-94r: Special tables for the {ar (hand C, hand A resumes on fol. 64r).
94v: Blank, no colophon.
MS Escorial r. 931,8, fols. 171r-211v 13
171r: Title folio (Kitb al-D}ir wa-fal al-d}ir wa-l-samt) attributed to al-Khall.
171v-209r: al-Khalls universal auxiliary tables (no sub-table for = 21;30, latitude
argument only up to 50).
209v: Table of solar declination ( = 23;35 !).
No colophon. Dated 933 H [= 1526].
MS Cairo MM 43:
Notes: The manuscript contains some 50 unnumbered folios and the tables are unattributed.
It is of Egyptian provenance and the copyist is recognizable as {Al ibn Muammad al-Dalm
and thus datable to ca. 1450. The tables listed here are followed by a variety of other tables
for trigonometry, spherical astronomy and instrument construction (some universal, some
specifically for the latitude of Cairo). There is also a table of geographical coordinates.
1r: Miscellaneous scribblings. No title.
1v-5r: (al-Khalls) introduction to his universal auxiliary tables. Definitions of the
functions f and g (fol. 5r).
5v: Tables of solar declination ( = 23;31) and universal {ar.
6r-24r: (al-Khalls) tables of the first and second functions ( = 1-50 including
21;30 and 33;30) and the third function.
MS Cairo M 228,5:
Notes: All of the tables are anonymous. The manuscript is undated, but is copied in a legible
naskh script which is estimated ca. 1500. The folios have been bound in disorder, but the
universal auxiliary tables and the hour-angle tables are complete.
1r-3r: No title. (al-Khalls) introduction to the universal auxiliary tables.
3r-3v: Notes on the determination of the qibla and the use of the sine quadrant by
an unidentified astronomer Zayn al-Dn {Umar al-Zuhr (Cairo ENL Survey,
no. C79).
4r: Tables of solar declination ( = 23;31) and universal {ar.
4v-7v: (al-Khalls) tables of the first and second functions ( = 1-7 only).
8r: (al-Khalls) hour-angle tables (h = 79-80 only).
8v: Notes on the determination of the qibla.
9r-9v: Prayer-tables for Tripoli (latitude 34) (incomplete, only 6 functions for
Capricorn and 6 for Aquarius): see fols. 16r-21r and 9.3.
10r-13r: (al-Tzns) tables for converting from the Hijra to the Syrian calendar
(incomplete).
13v-14r: (al-Tzns) table of solar longitude.
14v-15r: Table of solar mean motion.
15v: Blank.
16r-21r: Continuation of prayer-tables for Tripoli (12 functions for the remainder of
Aquarius up to Gemini): see contents of fols. 9r-9v above.
21v-25v: Tables of ascensions (for Damascus).
26v: Cotangent tables (bases 12 and 7).
27r-28r: Table of solar altitude at the {ar (for each 0;15 of H).
28r-60v: (al-Khalls) hour-angle tables (h = 1-78 only).
61r-85r: Remainder of (al-Khalls) first two functions ( = 8-54 including =
21;30 and 33;30).
85v-86r: Geographical tables as follows: (a) list of localities, mainly in Egypt, with
their coordinates and qiblas; (b) list of localities giving latitudes only; and
(c) list of localities on the Egyptian pilgrim route with coordinates and qiblas.
86v-98r: Tables of (al-Khalls) third function (complete).
98v-101r: Miscellaneous notes on calendars, scribblings, notice of possession by
Muammad al al-Dn (1)117 H [= 1705/06] (fol. 98v).
The set of tables contained in the unique source MS Dublin CB 4091 was compiled before
the main Damascus corpus. It consists of three parts, which are considered in turn (see already
I-9.4 and I-5.3.1 on the tables, here considered in more detail). al-Khall refers to these tables
in the introduction to his universal auxiliary tables (10.7).
the damascus corpus of al-khall 367
nisbat al-taful). Next find the difference between the Sines of the meridian altitude
and the instantaneous altitude by subtracting the Sine of the altitude from the Sine
of the meridian altitude. Enter with the base (as horizontal argument) in the universal
hour-angle table until you come opposite the difference between the two Sines (as
vertical argument), using linear interpolation if (the difference) is not equal to the
argument. The result will be the hour-angle if the base is integral. If it is not integral,
round the fraction to unity and determine the hour-angle by interpolating linearly for
the difference between the two Sines. Then find the first difference from the
appropriate box and recall the fraction which you rounded: if it was one-half or one-
third or more or less, add one-half or one-third or more or less of the first difference
to the hour-angle.
Example. The base is 24;15. We round (the fraction) to unity and obtain 25. The
difference between the two Sines is 29;30. With this we find the hour-angle, which
is 65;48. We find that the first difference is 1;32, and the amount we added to the
fraction to make it whole is 0;45. We add one-half and one-quarter of the difference,
namely 1;9, to the hour-angle and the result is 66;57, which is the required hour-
angle.
If we enter with the base (as horizontal argument) until we come opposite [the Sine
of] the meridian altitude or the Sine of the meridian altitude of the opposite degree
of the ecliptic, the result will be the half arc of daylight or of night (, respectively).
(To determine) the time from midday to the {ar, find the altitude at the {ar from
the special table and then find the corresponding hour-angle. The time remaining until
sunset is the excess of half the arc of visibility over the time from midday to the {ar.
(To determine) the durations of morning and evening twilight (al-iatn), enter with
the base (as horizontal argument) until we come opposite the excess of the Sine of
the meridian altitude of the opposite point of the ecliptic over the Sine of 17 for
evening twilight and 19 for morning twilight. The resulting hour-angles are the times
between evening and morning twilight and midnight: subtract either of them from the
half nocturnal arc and the remainder will be the duration of twilight. God knows best.
The calculation was completed in the Holy Month of Muarram in the year 763 H
[= November, 1361].
The main tables are preceded by two small tables, the first of which is for the function:
*() = 90 + () ( = 23;31 )
labelled al-ghyt fqiyya ba{d isq al-{ar minh, meridian altitudes for all latitudes when
the latitude is subtracted. The second is for the Sine function, with entries to two digits for
each degree of argument. To find t(h,,) al-Khall proposes first using the table of *() to
find H(,), thus:
H(,) = *() -
and then using the Sine table to find Sin H and Sin h to determine their difference:
H(h,,) = Sin H(,) - Sin h.
Using the two main sets of auxiliary tables one should then be able to find B(,) and finally:
t(h,,) = V { B(,), H(h,,) } .
the damascus corpus of al-khall 369
al-Khall does not mention that his tables can also be used to determine the solar azimuth,
but there is an obscure reference in the introduction to his major tables which indicates that
he was aware of this possibility. The procedure would be to form the quantity:
H(h,,) = Sin ( + h ) - Sin (),
with which the azimuth (measured from the north point) is given by:
a(h,,) = V { B(,), H(h,,) } .
Although only one copy of these original auxiliary tables has come to light, as opposed to the
several extant copies of the universal table, it is clear that their circulation was not limited.
They were known even in Tunis and Cairo (see 10.11).
Table 10.3b
List of formulae for spherical astronomy
(MS Dublin CB 4091, fol. 17v)
Notes: a text Sine (!); b text Cosine (!); c text Cosine (!)
the damascus corpus of al-khall 371
The fourth relation allows a determination of the hour-angle from the difference between the
Sines of the meridian and instantaneous altitudes:
base Sine of altitude = unity auxiliary Sine ,
whence:
Vers t = [Sin H - sin h] = R2 [Sin H - Sin h] / [Cos Cos ]
= [Sin H - Sin h] Vers D / Sin H = Vers D - Sin h Vers D / Sin H ,
or, in the simplest possible terms:
Vers t = Vers D - j .
Rules such as these, and the accompanying technical terminology, were standard knowledge
amongst contemporaneous Muslim astronomers.
(argument) 24 and you subtract one third of the difference from it, namely 0;34:
the remainder is 41;19, which is the azimuth. If (the difference) had been marked
positive we would have added them. Know this!
The simplest formula for compiling such a table of a(H,h) is:
a(H,h) = arc Sin { [ Sin h - Sin h0(H) ] Tan / Cos h } ,
where:
Sin h0(H) = R Sin ( H - ) / Sin ,
h0(H) being the altitude of the sun in the prime vertical on a day when it has meridian altitude
H.
Although several other solar azimuth tables are known from the Islamic sources (I-5), mostly
with arguments h and ; al-Khalls is the only one with arguments H and h. Tables of the
time since sunrise were often computed for arguments H and h, the advantage being that only
about 3,000 entries have to be computed, as compared with about 10,000 entries in tables with
arguments h and . al-Khalls azimuth table was soon replaced by Shihb al-Dn al-alabs
solar azimuth table (11.2) which had the same format as al-Khalls hour-angle tables and
which was occasionally included in copies of the main Damascus corpus.
10.3* The universal auxiliary tables for finding the solar azimuth
The following tables are mentioned by al-Khall in the introduction to his main universal
auxiliary tables (10.4) but the unique surviving copy come to my attention only in January,
2001.15 I have mentioned them briefly in I-8.5.1a and I-9.4a, and show here how they fit into
al-Khalls development.
MS Bursa Haraiolu 1177,4 (fols. 72r-90r), copied ca. 1450, is a unique copy of a set
of tables by al-Khall, not identified as the compiler, and entitled simply Jadwal al-samt li-
kulli rtif{ f {ur al-aqlm al-sab{a. The set contains sub-tables for the following functions:
f () = R Sin () / Cos and g(h) = Sin h Tan / R
for the domains:
and h = 1, 2, ... , 90 ,
= 1, 2, ... , 48, as well as 33;30 (Damascus) ,
and:
K(x,h) = arc Sin { R x / Cos h }
for the domains:
x = 1, 2, ... , 59 and h = 0, 1, ... , n(x) ,
where n(x) is the largest integer such that x Cos n(x). The functions f and g are labelled
jayb sa{at al-mashriq and iat al-samt by al-Khall, and it is clear that they represent Sin
() and k(h) see 5.0 and 8.5. The value of underlying the first table is about 23;30, possibly
23;31.
The instructions (see Fig. 10.3*) describe how to determine the ta{dl al-samt using the first
two tables, and then how to find the azimuth from the third table. These universal auxiliary
tables represent the ultimate solution to the problem of determining the solar azimuth a(h,,).
15 The manuscript is mentioned in ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 805.
the damascus corpus of al-khall 373
Fig. 10.3*: The introduction to al-Khalls tables and the first two sub-tables of f and g for latitudes 1 and
2. See also Figs. I-9.4*a-b. [From MS Bursa Haraiolu 1177,4, fols. 72v-73r, courtesy of the Genel
Ktphanesi.]
It is clear that all al-Khall had to do to compile his splendid tables described in 10.7 was
to replace the table of f () with a new one of f() and rewrite the instructions!
There is a note on fol. 90v of the Bursa manuscript in the hand of (the Damascus astronomer)
Muammad ibn Muammad al-Tzn stating that he had looked at this table in the year 867
H [= 1462/63].16 al-Tzn himself compiled a set of tables for sundial construction for different
latitudes which is extant in MS Vatican ar. 105,3.17
10.4 The main Damascus corpus: the introduction and solar tables
The full title of the Damascus corpus of tables for timekeeping is given on MS Paris BNF
ar. 2558, fol. 1r. It translates: Tables of the hour-angle and the operations of night and day
for latitude 33;30 computed by al-Shaykh Shams al-Dn Muammad ibn Muammad ibn
Muammad al-Khall, muezzin at the Umayyad Mosque, followed by a table for finding the
azimuth of the qibla and a universal table which can be used in all localities.
Fol. 1v of MS Paris BNF ar. 2558 contains virtually all that remains of an introduction to
16 On al-Tzn see Suter, MAA, no. 450; Brockelmann, GAS, II, p. 160, and SII, p. 484; Cairo ENL Survey,
no. C95; and also 10.4.
17 These are not listed in any reference work from Suter, MAA, to Rosenfeld & ~hsanolu, MAIS (no. 903).
374 part ii, chapter ten
some tables for calendar conversion and for finding the solar longitude.18 The bulk of the
introduction and the tables themselves have apparently been removed and replaced by a later
set (fols. 2v-9r) in a different hand. At this time fol. 1v was apparently glued to an empty fol.
2r. More recently this introductory fragment was separated from fol. 2r with the result that
a mirror image of part of the text may be read. Although the introduction and the missing tables
are anonymous they are most probably the work of al-Khall himself; they were already
outdated in 1408, when the manuscript was copied. It is clear from this fragment that al-Khall
had prepared a table for converting Hijra lunar dates into dates in the Syrian solar calendar
and another table displaying the solar longitude for each day of an artificial Syrian year which
needed to be corrected from an auxiliary table. This correction provided the necessary
adjustment for leap years in the Syrian calendar and was tabulated as a function of the excess
Hijra years over al-Khalls epoch (not stated) reckoned modulo 60.
The new introduction and tables in MS Paris BNF ar. 2558, fols. 2v-9r, are the work of
al-Tzn (10.3*). They are similar in conception to those of al-Khall, and tabulate the Syrian
dates corresponding to the beginnings of each lunar month from 851 to 1000 Hijra. al-Tzn
then presents a table of solar longitude of the shabaka variety for an artificial Syrian year to
which a correction must be applied. The correction is to be taken from a sub-table copied around
the main table; values are given for each year after 850 Hijra from 1 to 150 (corresponding,
of course, to years 851 to 1000 Hijra). al-Tzn then goes on to discuss various timekeeping
operations which need not concern us here.
Solar tables arranged according to the date in a solar calendar are attested in other Islamic
sources both earlier and later, but none of these has been investigated for structure or accuracy.19
It is likely that al-Khalls solar longitude table would have been computed using the solar
tables in the Zj of Ibn al-Shir, and that the entries would have been accurate to within a
few minutes.
MS Paris BNF ar. 2558, fol. 50v, contains a simple table for correcting the solar longitude
at midday to yield the solar longitude for any other time of day. Corrections are given as
a double-argument function of the daily solar motion (Arabic, buht) and the time t before
or after midday. Values to one digit (minutes) for each 3 of t from 3 to 180, and each unit
of the daily solar motion from 57 min/day to 62 min/day are given. The underlying formula20
is simply:
= t / 360.
MS Paris BNF ar. 2558, fol. 64r, contains a table of solar declination () with values to degrees
and minutes for each degree of solar longitude. The underlying value of the obliquity is 23;31.
Finally, in MS Paris BNF ar. 2558, fols. 18v-20r, there is a set of equatorial coordinates
(,) for 81 stars. This table is not the work of al-Khall: the entries are, as stated in the
heading, those of Ibn al-Shirs star catalogue, with 0;15 added to the normed right ascen-
sions.21 Of course the correction for precession should be added to the stellar longitudes, so
we have here a very crude and rather unsuccessful attempt to update the tables. A precessional
18 On calendars in Islamic astronomy see n. I-1:3.
19 See nn. 1:43.
20 See Kennedy, Zj Survey, p. 162b.
21 On Ibn al-Shirs star catalogue see Kennedy, Zj Survey, p. 164a.
the damascus corpus of al-khall 375
correction of 0;15 corresponds to about 15-20 years using the standard medieval values (1
in 66 or 70 solar years) for this motion. Ibn al-Shirs catalogue is for 765 H [=1363/64].
For our present purposes it is important to realize that al-Khall assumed that the user of
his tables for timekeeping knew either the longitude or declination of the sun or the stellar
declination.
al-Khalls corpus contains a complete set of tables of the hour-angle t(h,) for Damascus based
upon the parameters:
= 33;30 and = 23;31.
There is no mention of these tables, or the prayer-tables (10.6), in the introduction. Values
of the hour-angle are given for arguments:
h = 1, 2, ... , 80 and = 271, 272, ... , 359, 0, 1, 2, ... , 90.
Note that because of the format of the tables no entry is given for the winter solstice, that is,
= 270. There are about 3,000 entries. Extracts are shown in Figs. I-2.1.4a-b.
The standard of al-Khalls computational accuracy is remarkably high. He may have used
his first set of auxiliary tables (10.3) to compile these hour-angle tables. The procedure would
be equivalent to:
t(h,) = arc Vers { [Sin h() - Sin h] / B() } .
Alternatively he may have tabulated the functions B() and C() for this latitude and then used
the simple relationship:
t(h,) = arc Cos { R [ Sin h - C() ] / B() } .
The hour-angle tables are contained in MS Paris BNF ar. 2558, fols. 20v-50r, and at least seven
later copies (see 10.2). In MS Cairo MM 71,1, copied ca. 1600, triplets of entries (t,T,a) for
Damascus are given for each pair of arguments (h,). The values of both t and T are attributed
to al-Khall, and the values of a are attributed to Shihb al-Dn al-alab (11.2). Some tables
for the time remaining until moonset based on al-Khalls tables of T(h,) for the sun are
described below (10.11).
MSS Paris BNF ar. 2558, fol. 33r, and Oxford Seld. Supp. 100, fol. 16v, of the corpus contain
two tables by al-Khall without instructions for their use. The two tabulated functions are
labelled qaws al-al bi-l-mur and qaws al-al bi-ghayr mur, arc of the base with / without
a movable marker. These functions are denoted by b1 and b2, and inspection reveals that for
al-Khalls parameters:
b1() = arc Sin { B() } and b2() = arc Tan { B() }.
where:
B() = Cos () Cos / R .
These tables, discussed in more detail in I-6.5.1, were designed to facilitate the marking of
B() on an instrument called the sine quadrant (Arabic: al-rub{ al-mujayyab), a computational
device which in different forms was popular among Muslim astronomers from the 9th to the
19th century. Complicated trigonometric computations can be made with the sine quadrant
without any calculations whatsoever.
376 part ii, chapter ten
The main set of prayer-tables compiled by al-Khall, for which there is likewise no introduction
(10.5), consists of tables displaying some twelve functions for each degree of solar longitude,
beginning with Capricorn 1. The tables are entitled jadwal a{ml al-layl wa-l-nahr, that is,
tables for the operations of night and day, and are based on the same parameters: an extract
is shown in Fig. 10.6. The functions tabulated are:
Fig. 10.6: The sub-table for solar longitudes in Aquarius in al-Khalls corpus of prayer-tables. [From MS Paris
BNF ar. 2558, courtesy of the Bibliothque Nationale de France.]
the damascus corpus of al-khall 377
Wetzstein 1138 (Ahlwardt 5754-6), fol. 59r). The use of these bases is remarkable.
No explanation is provided in the manuscript and none is forthcoming from this
author.22
(c) Tables of the function ha(H), which is universal as it is independent of terrestrial latitude
(al-{ar al-fq) (e.g., fol. 64r of the Paris copy). This is computed to two digits for
each degree of H from 1 to 90, and is based on the standard formula. al-Khalls
values of ha() (see above) may have been computed using this table, although they
would have been more accurate had they been computed directly.
(d) Tables of two functions which we label b1() and b2() for use in timekeeping
calculations with quadrants (only in the Paris copy of the corpus, fol. 33r): see already
10.5.
(e) Most of the sources (e.g., fols. 104v-105r, 17v-18r, 15v-16r, 16v-17r, of the Paris copy)
contain four tables of ascensions, with values in degrees and minutes for each degree
of ecliptic longitude. The functions tabulated are:
(), (), s() and r().
Each table is arranged in twelve columns of thirty entries, with the name of the zodiacal
sign at the head of each column and a vertical argument running from 1 to 30. The
actual arrangement of the columns differs for each table, beginning with the columns
for the quadrant of the ecliptic at the beginning of which the function tabulated is
between 0 and 90. Thus the first columns in the table are for Capricorn, Aries, Libra,
and Cancer, respectively.
The tables of r and s are based on the relations:
r() = ( H) = () - r() and
s() = ( H) = 360 - [ (180 - ) - s() ] .
In the first case H is the ascendant at daybreak, and in the second it is the ascendant
at nightfall.
(f) The Paris copy (fols. 18v-20r) contains a star catalogue attributed to Ibn al-Shir listing
the equatorial coordinates of 81 stars.
Each of the sub-tables in this set is rather carefully computed, with the exception of the tables
for the {ar (noted above).
al-Khalls most original contribution was his second set of auxiliary tables for solving all of
the standard problems of spherical astronomy for all latitudes:22a see I-9.5 and VIa-9.1 and
the extracts illustrated there. Over a dozen copies are now known, of which the best is on fols.
61v-104r of the superior Paris manuscript. al-Khall tabulated three auxiliary functions, each
to two sexagesimal digits. For the first two, the horizontal argument is terrestrial latitude, and
the functions are denoted by ff and gf, where f is the local latitude. The vertical argument is
22 The same holds for the use of 10 by Ibn al-@dam in the 10th century: see I-4.2.3. On the use of base
20 see n. I-1:35, and the text to n. II-10:39a.
22a The first study of these was in King, al-Khalls Universal Tables (1973).
the damascus corpus of al-khall 379
defined in the instructions, and the functions are called al-maf al-awwal and al-maf al-
thn, first and second functions, respectively. The third function, which is denoted by K,
has a quantity called jayb al-tartb, perhaps best rendered as the auxiliary Sine, as horizontal
argument. The vertical argument is again defined in the instructions. The jayb al-tartb used
here is not the same as the one defined in the minor auxiliary tables above (10.3).
The functions are not explained in the instructions, but are in fact:
f() = R Sin / Cos , g() = Sin Tan / R , and
K(x,y) = arc Cos { R x / Cos y },
for appropriate domains (see below), and the trigonometric functions are to base R = 60. The
functions are defined in a marginal note in MS Cairo MM 43, fol. 5r:
The first function is the quotient of the Sine of an arc divided by the Cosine of the
latitude. The second (function) is the product (musaa, lit., rectangle) of the Sine
of the arc and the Tangent of the latitude, (both expressed) to base sixty (sittniyyan)
The function tabulated in the table labelled jayb al-tartb (il jayb al-tartb) is the
complement of the arc (i.e. inverse Cosine) of the quotient [multiplied by a factor of
sixty] of (the jayb al-tartb) and the Cosine of the arc.
The first two functions are tabulated for the domains = 1, 2, ..., 90, and = 1, 2, ...,
55, as well as 21;30 (Mecca) and 33;30 (Damascus). In some sources (e.g. MS Paris ar.
2558) there is no table for Mecca, and in others (e.g. MSS Escorial r. 931 and London BL
Add. 9599,31) the argument runs only to 50. The arguments for the third functions are x
= 1, 2, 3, ... , 59 and y = 0, 1, ... , n(x), where n(x) is the largest integer such that x Cos
n(x). The values for each function are arranged in columns each of thirty entries. On a given
double opening in the first set of tables one finds values of f and g for two consecutive values
of . In the second set one generally finds values of K for four consecutive values of x. However,
when the values of K are, in modern terms, non-real, no entry is given in the table. Columns
which would be empty are omitted; thus, for example, the values of K for x = 52, 53, ... , 59
fit on a double opening in MS Paris ar. 2558, fols. 103v-104r. If the values of K were displayed
in a single table it would be trapezoidal in shape.
The first two tables each contain 5,040 entries and the third about 3,500, all computed with
remarkable precision. Most of the entries are accurate, and the error in the remainder is usually
only 1 in the second digit, and occasionally 2. Errors in the K table are slightly more frequent
than in the f and g tables, and larger isolated errors do occur. Some entries in the table of
K(x,y) for small values of x and large values of y are rather inaccurately computed. This author
is unable to explain how these errors might have arisen.
However, Glen Van Brummelen, armed with an impressive combination of mathematical
and statistical skills and not lacking in historical sensitivity, has investigated these error patterns
in an attempt to establish precisely how al-Khall computed his tables; this was a topic beyond
the scope of the present study, and I cannot claim to understand all that Van Brummelen has
published on this subject.23 He would be the first to admit that al-Khalls tables still actually
defy explanation. He has, however, applied various statistical procedures in his investigation
23 Van Brummelen, al-Khalls Auxiliary Tables, and idem & Butler, Interdependence of Astronomical
Tables, pp. 46-48.
380 part ii, chapter ten
of the structure of the auxiliary tables and to our joint disappointment, for I do not think
it was too much to hope for was also unable to come up with precisely the procedures that
were used by al-Khall. The interested reader should consult Van Brummelens paper, of which
I here cite the conclusion (using the mathematical notation of this study):24
We have found in al-Khalls auxiliary tables an interesting mix of ingenious insights
and rudimentary errors. The g() table was computed from two-place sine and tangent
values on an interpolation grid of 5 of , and some of the remaining entries were
filled in using a well chosen variant of linear interpolation. Many of the entries, how-
ever, seem to be unexplainable other than by simple guesswork. The f() table was
computed from the g() table by a clever application of prosthaphairesis. The K(x,y)
table was calculated by using only two places for the value of Cos y, followed by an
extremely careful and precise calculation of the quotient Rx / Cos y and the arc
Cosine.
Van Brummelen has shown that the errors in the entries for f and g in a large part of the
tables, if not throughout, are linked by the following prosthaphaeretical relation:
f( n ) = Cos n g(,n) ,
and he suggests that al-Khall, having computed the table of g() and noted that f() = R
= 60, used this relation to generate a major part of the table of f(). When he published this
in 1991, I found the implications surprising. However, we now have independent confirmation
of the fact that al-Khall did tabulate g() first, because he simply took over the table from
his universal auxiliary tables for finding the solar azimuth.
We are on safer ground regarding what al-Khall intended with his tables. The rules for
their use which he outlines in words will now be compared with the medieval formulae listed
in I-1.2. Most of the rules can be explained by transforming standard medieval expressions
into a form which is equivalent to al-Khalls rule, given the definitions of the functions f,
g, and K. The arguments to be used in the tables are considered positive, and although al-
Khalls introduction (MS Paris ar. 2558, fols. 61v-64r) gives detailed instructions to cover
all possible cases, modern notation renders these superfluous in the commentary below.
In the name of God, the Merciful and Compassionate: Praise be to God, by whose
grace good things are achieved, and to whose might the earth and the heavens are
obedient, and by whose wisdom the months and the times (of prayer) are known, and
to whose order(s) the sun, moon and stars are subservient. May God bless our Lord
Muammad, master of brilliant miracles, and his family and his companions, (all of
whom) possess great virtue and noble qualities, with a blessing such that he who
pronounces it be saved from misfortunes in life and in death.
The aspect of astronomy most deserving of investigation is that by which the times
of the five prayers and the qibla can be found and which deals with other information
useful for guiding oneself on land and sea. The instruments which achieve this are
very numerous, and the best of them are those which are easy to use and which are
not specific to one latitude, but there are only a few people who can use these properly.
Finding (the times of prayer and the qibla) by means of calculation is more precise
24 Van Brummelen, al-Khalls Auxiliary Tables, p. 689.
the damascus corpus of al-khall 381
but it involves more work, and only someone skilled in this discipline can do it
properly. Thus if a table is prepared by calculation it is better to use such a table than
to use instruments, although (??) the table is easier to comprehend than (the
instruments).
Previously I had compiled two universal tables, from one of which one could
determine the times, and from the other, the azimuth. Now God has enabled me to
compile a table for finding all that could be found from those two tables and more
besides. In addition, it is simpler than them. Praise and thanks be to God.
Section on the half diurnal or nocturnal arc for the sun or a star. Find the longitude
of the sun from an ephemeris or some other tables and use it to find the declination
from the declination table, or find the declination of the star corrected for the epoch
in which you are working from the table of stellar coordinates. Then look for the page
on which is written the local latitude and find the value of the second function
corresponding to the solar or stellar declination. If the declination is in the first, second
or third column, take the value of the function from the first, second or third column,
respectively. Enter the result in the table of the third function (as the horizontal
argument) and the declination as the (vertical) argument. What you find in the first,
second or third box will be the half diurnal arc if the declination is southerly and the
half nocturnal arc if it is northerly. If one of them is subtracted from 180, the result
will be the other one.
Section on the hour-angle. Find the value of the first function on the page
corresponding to the terrestrial latitude with the altitude as (vertical) argument and
the value of the second function as before in the section on the half arc. If the
declination is southerly add the two values, otherwise take the difference. The result
is the auxiliary sine. Enter it in the table of the third function with the solar or stellar
declination as (vertical) argument, and you will find the hour-angle. If the declination
is northerly and the first function is less than the second, subtract the result from 180
and the remainder will be the hour-angle. If the two functions are equal then the hour-
angle will be 90.
Section on the maximum altitude, the local latitude, the altitude of (the sun) and
the hour-angle at the {ar, the time remaining from the {ar until sunset, and the
duration of evening and morning twilight. Add the northern declination to 90 and
subtract the southern declination: the result will be the sum of the maximum altitude
and latitude. Then find the local latitude by observation or otherwise and subtract it
from this sum: the result will be the maximum altitude. If it is greater than 90 subtract
it from 180: the remainder will be the maximum altitude to the north of the zenith
(i.e. measured from the north point). If the maximum altitude has been found by
observation and it is subtracted from the sum, the remainder will be the local latitude.
If the latitude is added to the maximum altitude and the difference between the sum
and 90 is taken, the result will be the solar or stellar declination. If the maximum
altitude is northerly with respect to the zenith, subtract it from 180 and subtract the
remainder from the sum: the result will be the local latitude, or add the latitude to
the remainder, subtract 90 from the sum, and the remainder will be the declination.
382 part ii, chapter ten
Next enter with the maximum altitude of the sun in the universal {ar table and you
will find the altitude of the {ar. Find the hour-angle for this altitude: it will be the
time from midday to the {ar. Subtract this from the half diurnal arc: the remainder
will be the time from the {ar to sunset. Next you find the value of the first function
on the page for the latitude corresponding to (vertical argument) 17 for nightfall or
19 for daybreak. Then add to it the previous value of the second function for the half
diurnal arc if the declination is northerly, otherwise subtract it from the arc: the result
will be the auxiliary sine. Enter this in the table of the third function and the declination
as (vertical) argument. Subtract what you find from the half nocturnal arc: the result
will be the duration of twilight.
Section on the rising amplitude, the altitude in the prime vertical, and the azimuth
corresponding to a particular altitude. Find the value of the second function for latitude
45 with (vertical argument equal to) the declination and enter the result as argument
in the table of the third function with the local latitude or its complement as vertical
argument, and you will find the complement of the rising amplitude or [complement
of] the altitude in the prime vertical. Another method: Find the value of the first
function for the solar or stellar declination on the page corresponding to the latitude
and enter whatever it is as (vertical) argument in the table of the second function for
latitude 45 or in the page corresponding to the latitude. The corresponding argument
will be the rising amplitude or the altitude with no azimuth.
Then find the value of the first function for the solar or stellar declination on the
page corresponding to the latitude, and find the value of the second function for the
altitude. Add the two values (if the declination is) southerly and take the difference
between them (if it is) northerly. The result is the auxiliary sine: enter it in the table
of the third function with the altitude as the (vertical) argument and you will find the
azimuth measured from the meridian. If the declination is southerly the azimuth is
southerly. If it is northerly and the value of the first function is greater than that of
the second, the azimuth will be northerly; if it is less, then (the azimuth) will be
southerly. If the two values are equal then there is no azimuth.
Section on finding the declination by observation given that the latitude is known
and finding both the latitude and the declination. Determine the meridian or the prime
vertical either by means of an azimuth found from an altitude observation (samt al-
irtif{) or by using the Indian circle or a qibla-box (uqq al-qibla) or by some other
well-known means.25 When one of these two lines is known use it to find the azimuth
of a particular altitude by measuring the altitude at that time to obtain (a pair of) altitude
and azimuth (values). Then look in the table of the third function for a box such that
if the azimuth is entered in it (as one argument) the altitude would be the corresponding
(vertical) argument: the auxiliary sine is (the argument) above the box. Then find the
value of the second function with the altitude (as argument) on the page for the latitude.
Subtract the value of the second function from the auxiliary sine for southern
25 On the Indian circle see n. 2:23. On qibla-boxes in Egypt ca. 1300 see Schmidl, Earliest Arabic Sources
on the Compass, and King, Mecca-Centred World-Maps, pp. 113-114.
the damascus corpus of al-khall 383
declinations, or add the two for northern declinations if the azimuth is northerly,
otherwise take the difference between them. The result is a value of the first function,
so find the corresponding argument: it will be the declination.
If you do not know either the latitude or the declination, make an observation of
the azimuth and altitude and find the auxiliary sine as before, then observe another
azimuth and altitude and find the auxiliary sine (for this pair) also. If the azimuths
are in the same direction, take the difference between the two, otherwise add them:
the result will be the ia. Next find the values of the second function for latitude
45 and for each of the altitudes and take the difference between them. Enter with
the result in the (table of the) second function for latitude 45 and find the
corresponding argument: the result will be the ta{dl. Then look for a page of (the tables
of) the second function so that if you enter with the ta{dl as argument, the value is
the ia: (you do this) to obtain the latitude at the top of the page. Then find the value
of the second function for the larger altitude on that page and add to it the auxiliary
sine for the larger altitude if the azimuth is northerly, otherwise take the difference
between them: enter the result in (the table of) the first function on the (same) page,
and find the corresponding argument to obtain the declination.
Conclusion on operations with stars. Enter the longitude of the sun in the table of
(normed) right ascensions and you will find the (normed) right ascensions for that day.
Subtract from these the half diurnal arc if possible, otherwise add a complete revolution
(of 360 and then subtract): the result will be the oblique ascensions. If you add the
half diurnal arc to the (normed) right ascensions, the result will be the (oblique)
ascensions of the opposite point of the ecliptic. If the sum is greater than one revolution
then the excess will be the ascensions of the opposite point. Subtract the half diurnal
arc of a star or its eastern hour-angle from its ascensions if this is not possible add
a revolution to the ascensions: the result will be the ascensions of midheaven when
the star is rising or when its altitude was measured. If the half diurnal arc or the hour-
angle is added to the ascensions (of the star) the result will be the ascensions of
midheaven when (the star) is setting or when its altitude was measured.
If the ascensions of midheaven are found by these operations or by means of a star
which is culminating, subtract from them the ascensions of the opposite point of the
ecliptic and the remainder will be how much of the night has passed. If the ascensions
of the opposite point are added to the ascensions of midheaven, the culmination will
be before sunset by the amount of the sum (lit. increase). If you subtract the ascensions
of midheaven from the oblique ascensions, the remainder will be the time remaining
of the night. If you add the ascensions of midheaven to the oblique ascensions, the
culmination is after sunrise by the amount of the sum (lit. increase).
Section on the declination of a star. Look in the table of the third function for a
box such that if you enter in it with the longitude of the star measured from the
beginning of Capricorn or Cancer the corresponding (vertical) argument would be the
latitude (of the star): you obtain the auxiliary sine above that box. Then enter in (the
table of) the second function for latitude 49 the altitude of the star, and double the
result and add it to the auxiliary sine if the longitude of the star and its latitude are
384 part ii, chapter ten
in the same direction, otherwise take the difference. Enter one half of the result as
a value of the first function for latitude 37 and find the corresponding argument: it
will be the declination of that star. The direction of the declination is the same as the
direction of the longitude and the latitude (when these are) in the same direction. If
they are in different directions and if the auxiliary sine is less than the value of the
second function, I mean the doubled value, then the direction of the declination is the
same as that of the longitude. If they are equal then there is no declination and no
direction.
Section on the ascensions of a star. Enter with the latitude of the star in (the table
of) the first function for latitude 37, double what you find and keep it in mind. Next
enter the declination of the star in (the table of) the second function, double what you
find and keep it in mind. Then add the two quantities which you kept in mind if the
latitude of the star and its declination are in different directions and take the difference
between them if they are in the same direction: the result will be the auxiliary sine.
Enter this in the table of the third function with the declination of the star as the
(vertical) argument, and see what you find in the table. If the latitude and the
declination are both northerly and the value of the first function is more than that of
the second, and the longitude is between Capricorn and Cancer, then the result is the
ascensions of the star. If the longitude is between 3 (signs) and 9 subtract the result
from 360: the remainder will be the ascensions. Likewise for (both latitude and
declination) southerly. If the value of the first function is less than that of the second
and the longitude is between 3 and 9, then subtract the result from 180, but add it
to 180 if the longitude is between 3 and 9 to obtain the ascensions. Likewise for both
(latitude and declination) in the south if the value of the first quantity is greater than
that of the second. If the latitude is northerly and the declination is southerly and the
longitude is greater than Capricorn, then the result is the ascensions of the star, and
if (the longitude) is less then subtract the result from 360 to obtain the ascensions.
If the latitude is southerly and the declination is northerly and the longitude is less
than Cancer, subtract the result from 180, and if (the longitude) is more, add the result
to 180 to obtain the ascensions of the star. God knows best.
Commentary:
(i) To find the half arc of daylight, D, or night, N, of a celestial body with declination , given
the terrestrial latitude , use:
D or N = K{ g(), }
for \% 0, respectively. From F7 in I-1.2 for all :
D = 180 - N = 90 + arc Sin { Tan Tan / R }
so that:
D or N = arc Cos { Sin Tan / Cos } for < or > 0, respectively.
(ii) To find the hour-angle, t, given the solar altitude, h, for any declination and terrestrial
latitude, use:
t = K { [ f(h) - g() ], }.
To illustrate the correctness of this procedure, we transform F12 in I-1.2 into:
t = arc Cos { R [ R Sin h / Cos - Sin Tan / R ] / Cos }.
the damascus corpus of al-khall 385
The declination may then be found from f() = g(h) - x, as in (vii) above. These procedures
are easily verified by projecting the celestial sphere onto the meridian plane.
(ix) To find the declination of a star, , from its ecliptic coordinates (,), first find , the
longitude difference between the star and the nearest solstice, and then solve for x in:
= K(x,) .
The declination can then be found by solving:
f37() = 1/2 { 2g49() + x } .
To explain this procedure, let denote the obliquity of the ecliptic. The modern formula26 can
be transformed thus:
sin / sin = sin cos / sin + cos sin .
In order to adapt this formula to serve al-Khalls tables, write:
R Sin / Cos 1 = 1/2 { 2 Sin Tan 2 / R + Cos Sin / R }.
It follows that the values 1 and 2 are defined by:
1/ cos = sin and 2 tan = cot .
2 1 2
For al-Khalls parameter = 23;31, the accurate values of 1 and 2 are 37;4 and 48;58:
he used the approximations 37 and 49.
(x) To find the normed right ascension of a star, use:
= K { [ 2f37() - 2g49() ], }
where is the normed right ascension measured from = 0 or = 180. To justify this,
we write the modern formula for the right ascension as:
= arc Sin { R [ Sin Cot / R - R Sin / Sin ] / Cos },
so that, with 1 and 2 as defined in (ix) above:
= arc Cos { - R [ 2 Sin Tan 2 / R - 2 R Sin / Cos 1 ] / Cos }
= arc Cos { R [ 2 R Sin / Cos 1 - 2 Sin Tan 2 / R ] / Cos },
al-Khalls handling of these problems can, by any criteria, only be described as brilliant!
Three copies of al-Khalls corpus contain a qibla table, namely, MSS Paris ar. 2558, fols.
53v-61r, copied 811 H [= 1411]; Berlin Ahlwardt 5754 (Wetzstein 1138, fols. 47v-55r), ca.
1700; and Paris BNF ar. 2560,11, fols. 148v-159v, ca. 1750.27 The tables in the first two of
these sources are preceded by a short introduction but in the third source there is no introduction
and indeed no indication of the nature of the table.
The function tabulated by al-Khall is here denoted by q, although in fact it defines the
acute angle which the arc joining a given locality to Mecca makes with the local meridian.
(See Fig. VIa-3.1.) Values of the function q(,L) are given in two sexagesimal digits for the
domains:
= 10, 11, ... , 56 and 33;30, and L = 1, 2, ..., 60.
26 Smart, Spherical Astronomy, p. 40.
27 My detailed study listed as King, al-Khalls Qibla Table (1975) contains a critical edition of the table
based on the first Paris manuscript and the Berlin one (pp. 87-94). The second Paris manuscript came to my
attention too late to be used for the edition (see ibid., p. 84, n. 10).
the damascus corpus of al-khall 387
Fig. 10.8: An extract from al-Khalls universal qibla table serving latitudes 28-33. The entries in light (actually
red) ink indicate that the qibla is northerly. [From MS Paris BNF ar. 2558, courtesy of the Bibliothque Nationale
de France.]
For each value of there is a sub-table in which L is entered vertically see Fig. 10.8 for
an extract. The double vertical argument runs from 7 to 66 (longitudes west of Mecca, for
which the qibla is easterly) and from 127 to 68 (longitudes east of Mecca, for which the
qibla is westerly). These two sets of longitudes correspond to a change in L from 60 to 1,
since the longitude of Mecca is taken as 67;0, and the underlying latitude of Mecca is found
by inspection (see below and also 10.9).
The entries for each latitude are arranged in two columns of thirty entries, and two facing
pages of the manuscripts serve six consecutive latitudes. There are 2,880 entries in the entire
table.
In the introduction (see below) he mentions that the qibla is southerly if the entry in the
table is in red ink and northerly if it is in black ink. In the first sub-tables all entries are northerly,
and in the remaining five sub-tables all entries above the broken line are also northerly; all
other entries are southerly.
al-Khalls qibla values are remarkably accurate: the errors in most of the table are less
than 0;2, although larger errors do occur in certain sections. It is reasonable to assume that
al-Khall would have computed a matrix of values for, say, each degree of and each 5 of
L, and that he would have then used an interpolation scheme to compute intermediate values.
388 part ii, chapter ten
The error patterns in the table confirm this, and the interpolation scheme used, which was more
sophisticated than a second order scheme, generally works rather well. Note, however, the
relatively large errors in the entries for small values of L and values of close to M. For
these arguments q is very sensitive to small changes in L. Another minor point to be noted
is that for L = 1 (L = 66, 68) the entries for values of from 21 down to 17 are identical
to the entries for from 22 up to 26, and the qibla differs only in direction for these values
close to M. It is clear from the errors in these values that the entries for from 21 to 17
were copied from those for from 22 through 26, which were computed.
It is of interest that al-Khall has tabulated q(,L) for the latitude of Damascus, 33;30. The
latitude of Baghdad is also close to this, but the only reason for computing such a table would
appear to be that al-Khall had already tabulated his auxiliary functions for this latitude and
he used them to compile his qibla table. On the other hand, he also tabulated q(,L) for latitude
56, but as far as we know, he computed the auxiliary functions only up to latitude 55.
The introduction to the qibla table (fols. 52v-53r of the first Paris copy) reads:
In the Name of God the Merciful and Compassionate. Praise be to God who made
us pray towards His Sacred House (i.e., the Ka{ba). May God bless our Lord
Muhammad, who was sent (as a Prophet to institute) the religion of Islam, His Family,
His Companions, and His Wives noble persons (all) with a blessing that will heal
whosoever utters it of sickness and make him want to be in the Abode of Peace (i.e.
Paradise).
I have not seen a better method for (determining) the azimuth of the qibla than the
one mentioned by Ab {Al al-Marrkush in the sixty-seventh chapter of his Treatise
on the Sine (Quadrant) (Rislat al-Jayb). If the moon is at some instant (directly)
above the Ka{ba beginning to be eclipsed then people looking at it from (different)
parts of the earth will see it in different parts of the sky relative (to the meridian of)
the (particular) location. One person will see before it culminates, another will see
it culminating in the south or the north, another will see it after it culminates. Anyone
who faces the moon at that time will be facing the qibla. If the declination of the moon
at that time is known, and it is equal to the latitude of Mecca, and the hour-angle
corresponding to its position (relative to the meridian) in any locality is known, and
it is equal the difference in longitude (between the locality and Mecca), the altitude
of the moon will be known, and similarly its azimuth, that is, the azimuth of the qibla.
If the azimuth of the qibla is subtracted from 90, the remainder will be the azimuth
of the qibla measured from the local meridian (al-inirf).
Almighty God has enabled me to compile a table for finding the meridian azimuth
of the qibla from the longitude and latitude of the locality on the basis that the longitude
of Mecca may God honour it is 67 and the latitude is 21;30. The way to use
this table is to look for the page on which is written the latitude of the locality and
to enter in the first or second argument with its longitude. Then move vertically and
horizontally until the column and line meet and you will find the azimuth of the qibla
for that locality. If the longitude is in the first longitude argument column then take
the first entry for the azimuth, and similarly for longitudes in the second column. The
the damascus corpus of al-khall 389
28 On al-Mizz see n. 9:21 above. On his quadrants see n. 9:22 and X-6.
29 See King, Earliest Qibla Methods, p. 90, and idem, Mecca-Centred World-Maps, pp. 57-59 and 335-
336, on these procedures.
390 part ii, chapter ten
Unfortunately al-Khall does not describe the method by which he compiled his table.
However, he does remark that he knows of no better method for finding the qibla than the
one advocated by Ab {Al al-Marrkush (6.7).30 This involves first finding the altitude h of
the zenith of Mecca above the local horizon using:
Sin h(,L) = Sin ( + M ) - Vers L Cos M Cos / R2 ,
and then finding the corresponding azimuth q using:
q(,L) = arc Cos { R [ Sin h Tan / R - R Sin M / Cos ] / Cos h }.
al-Khalls universal auxiliary tables can also readily generate qibla values for a particular
value of . First calculate h(,L), an operation simplified by the fact that:
Sin h = c1 + c2 Cos L,
where c1 and c2 are constants. Then compute the quantity:
x(,L) = g(h) - f(M)
using the first two auxiliary tables and enter this as horizontal argument in the third table since:
q(,L) = K { x(,L) , h }.
Although it cannot be established that this procedure was actually followed by al-Khall, by
using his auxiliary tables it is possible to generate qibla values which are less accurate than,
the same as, or more accurate than his own.31 Van Brummelen has noted that the errors in
al-Khalls table of f() for = M are not reflected in his qibla table, and he agrees that al-
Khall did not use his auxiliary tables to compute the qibla table.32
Some copies of the Damascus corpus, for example, MS Paris ar. 2558, fol. 51v, contain a table
of geographical coordinates copied from the handwriting of ... al-Mizz with associated qibla
values computed by Shams al-Dn al-Khall: see Fig. 10.9.33 In this copy the first four entries
(Mecca, Medina, Cairo and Alexandria) are doubtless later additions, firstly because the title
in the Egyptian copy MS Cairo MM 167, fol. 204r (dated 989 H [= 1581/82]), indicates that
the localities belong to the climate of Syria to the (further) edge of al-{Irq, but secondly
because the qiblas are carelessly computed. The entries in the table as found in the Paris and
Cairo copies are presented below. Clearly the final qibla values offer some measure of control
over copyists mistakes.
The final column in the edited table displays values computed using linear interpolation in
al-Khalls qibla table. It is clear that these are slightly more accurate than those in the list
of qiblas, which indicates that al-Khall did not use his qibla table to compile them.
30 For details of al-Marrkushs procedure and an attempt to recompute al-Khalls qibla values see King,
al-Khalls Qibla Table, pp. 99-108.
31 See the examples listed ibid., pp. 106-108.
32 Van Brummelen, al-Khalls Auxiliary Tables, pp. 688-689, and idem & Butler, Interdependence of
Astronomical Tables, pp. 46-48.
33 Kennedy & Kennedy, Islamic Geographical Coordinates (pp. xxii sub KHL and xxv sub MIZ) have
determined that the list is related to an early set of geographical tables recorded by Abu l-Fid} (fl. Hama, ca.
1310 see n. 2:33). This table is now published in King, Mecca-Centred World-Maps, App. F7.
the damascus corpus of al-khall 391
Fig. 10.9: The geographical tables in al-Khalls Damascus corpus. [From MS Paris BNF ar. 2558, fols. 51v-
52r, courtesy of the Bibliothque Nationale de France.]
392 part ii, chapter ten
Notice that the qibla value 29;4 for Damascus (in error by only +1) is at variance with the
value 28;57 which underlies the table of solar altitude in the azimuth of the qibla in al-Khalls
prayer-tables (10.6). Different approximate formulae as well as the exact formula have been
applied to a number of sets of medieval geographical coordinates for Damascus and Mecca
to no avail. Two pairs of values, however, which yield precisely 28;57 using the exact formula
are:
L = 60; 0 and = 33;33 LM = 67; 0 and M = 21;30
L = 60; 0 and = 33;24 LM = 67; 0 and M = 21;20
The value 33;33 is not attested in any known geographical tables, but the value 33;24 is found
in an early anonymous zj which unfortunately has no coordinates for Mecca,34 and is also used
by Ibn al-Mushrif (I-7.3.1 and 9.8 and II-9.10). The value 21;20 was used by Ibn al-Shir
for Mecca. Perhaps al-Khall used the second set of parameters to derive 28;57 for the qibla
at Damascus; perhaps he did not.
Opposite this geographical table in the first Paris manuscript (and also in MS Paris BNF
ar. 2560,13, fol. 164v, ca. 1750) is a list of stations on the pilgrim road from Damascus to
Mecca (manzil al-ijz al-Sharf {ala l-darb al-Shm) with their respective latitudes: see
also Fig. 10.9.35 It is unlikely that al-Khall made these measurements himself, and the list
should be compared with similar ones in other sources (e.g. MS Cairo MM 167, fol. 203v).
In some (e.g. Cairo MM 167, fol. 203r), we find also a list of pilgrim stations on the road
from Cairo to Mecca. See also 8.4.
There can be no doubt that for his calculations al-Khall used a sexagesimal multiplication
table displaying products:
m n for m, n = 1, 2, 3, ... , 59 .
Such tables had been in use amongst Muslim astronomers since at least the 10th century. He
may also have made use of a table of sexagesimal quotients m/n for the same arguments. The
34 Kennedy & Kennedy, op. cit., p. 473 (sub UTT). See also n. I-7:28.
35 On the pilgrim routes see EI2, III, p. 34 (in the article adjdj by Bernard Lewis), and Brice et al., eds.,
Atlas, p. 22.
394 part ii, chapter ten
more extensive Islamic tables for arithmetical operations with sexagesimals were probably
compiled after his time. See I-1.2.
The contemporary Zj of Ibn al-Shir contains tables of the Sine function (base 60) to four
sexagesimal places for each degree of argument, and the Tangent function (base 60) to three
places for each half degree. Both tables also display first differences. In addition Ibn al-Shir
tabulated the Cotangent to bases 7 and 12 to three places for each degree of argument and
the inverse Cotangent function.36
Tables of the Sine and Cotangent function with values for each minute of arc computed
to four or five sexagesimal places had been available since the 10th century. The muwaqqits
of Cairo also used tables of the Sine and Cotangent for each minute of argument to three
places.37 Such tables would have been available to al-Khall. Yet the only table of the Sine
function in the available manuscripts gives values to only two sexagesimal digits for each
degree of argument (MS Dublin CB 4091, fol. 1v), and, rather surprisingly, is not without
mistakes. This may be compared with the values of al-Khalls function g() (10.7) for =
45, which is identical to the Sine function. Van Brummelen has determined that al-Khall
used Sine and Cotangent tables with values to two sexagesimal places in the computation of
his auxiliary functions.38
It is also possible that al-Khall used tables of the Secant function and the inverse Sine /
Cosine function since these functions are fundamental to his tables. But while such tables are
attested in the Islamic sources,39 they do not occur in any known copies of the Damascus corpus.
Only two other trigonometric tables are found in the various manuscripts of al-Khalls
tables, both relating to the Cotangent function. In the first (MS Paris ar. 2558, fol. 51r) the
base used is 20. Values are given to two digits for each 1 of altitude from 1 to 80 (this
being the maximum integral solar altitude at Damascus). It is likely that al-Khall had a gnomon
divided into 20 units for which such a table could have been useful. I know of only one other
attestation of the use of base 20 in the medieval sources.39a
In the second trigonometric table (MS Berlin Wetzstein 1138 (Ahlwardt 5754-6), fol. 59r)
the base used is 10. Values are given to two digits for each 1 of altitude from 1 to 90. No
reason is given for the choice of 10 as base, not even in the instructions (fols. 60r-61r) on the
use of the shadow tables. The base 10 was also used in the trigonometric tables of the 16th-
century Istanbul astronomer Taqi l-Dn ibn Ma{rf.40 In the tables in the Damascus corpus
the entries are expressed sexagesimally, whereas in Taqi l-Dns tables they are expressed
decimally.
Possibly al-Khall used some kind of interpolation scheme to compile his tables, which
contain some 40,000 entries. Several descriptions of second-order interpolation schemes are
known from the Islamic sources,41 including one advocated by the 10th-century Cairo
astronomer Ibn Ynus and ideally suited for generating tables of trigonometric functions.42
Nevertheless Glen Van Brummelens detailed analysis of the error patterns in al-Khalls tables
has not cast any light on the procedures that he used.43
Since al-Khalls auxiliary tables for solar timekeeping survive in their original form only in
a single manuscript (MS Dublin CB 4091), it is probable that they were not widely used in
Syria. However it is known that the tables travelled as far as Tunis, Tlemcen, and Cairo, where
they were either extended or modified. In each case the attribution to al-Khall was suppressed
(or at least is not found in the surviving copies).
Tunisian redaction
MS Cairo DM 689 contains an extensive set of auxiliary tables copied in an elegant Maghribi
hand ca. 1600.44 They conclude with a star catalogue dated 801 H [= 1398], and appear to
have been compiled in Tunis. The title folio, instructions, and first few tables are missing from
the manuscript, which begins with the last page of a set of tables displaying the solar longitude
for each day of a period of four Syrian years. The tables for solar timekeeping follow.
The main functions tabulated are al-Khalls B(,) and V(x,y). In fact these Tunisian
tables are merely an extension of al-Khalls tables, and the corresponding entries in both sets
are the same. The Tunisian tables of B are computed for each degree of from 1 to 48 and
also 21;40 (Mecca) and various other non-integral latitudes between 30 and 40 intended to
serve localities in Ifdrqiyya and the Maghrib and perhaps also Sicily. al-Khall had a separate
table for latitude 33;30 (Damascus), and the Tunisian set has separate tables for latitudes:
30;30, 31;30, 32;30, 33;30, 33;40, 34;30, 35;30,
36;30, 36;40, 37;10, 37;30, 38;30 and 39;30.
The Tunisian tables of V(x,y) are simply those of al-Khall rearranged so that the horizontal
argument is increasing.
The tables of B are preceded by tables of the function Sin H() (called jayb al-ghya)
computed to two digits for each degree of and the latitudes between 30 and 38 for which
B is tabulated. These were probably used by the anonymous Tunisian astronomer to compile
his tables of B(), since:
B(,) = 1/2 [ sin H(,) + sin H(*,) ],
where * = 180 + . The underlying value of is 23;35, whereas al-Khall used 23;31.
However, the change in hardly affects the values of B() given to two digits. The Tunisian
tables also contain a table of Sin H() for = 0, which is simply Cos (). See further 13.4.
41 See n. I-1:21. MS Cairo Zakiyya 917,4 (pp. 50-54, copied 1163 H) contains some notes on interpolation
attributed to al-Khall but I have been unable to obtain a microfilm of it.
42 King, Astronomical Timekeeping in Medieval Cairo, pp. 354-357, and Hamadanizadeh, Islamic
Interpolation Schemes, p. 146.
43 See the text to nn. 10:23-24.
44 Listed in Cairo ENL Survey, no. F30.
396 part ii, chapter ten
Tlemcen extract
MS London BL Or. 411,2 contains the only known complete copy of a commentary on the
astronomical poem of the late-14th-century Maghribi writer {Abd al-Ramn ibn Muammad
al-Jdar.45 This anonymous commentary was written in Tlemcen and contains several spherical
astronomical tables computed for parameters:
= 35;0 (Tlemcen) and = 23;35 .
The same manuscript also contains interesting historical accounts of trepidation, twilight, and
the obliquity of the ecliptic. Among the spherical astronomical tables two display the functions:
Sin H() and 1/2 B(),
and another displays the function:
t { 1/2 B, (Sin H - Sin h) }
for the same latitude. Unfortunately it has not been possible to compare these tables with the
Tunisian corpus of auxiliary tables (see above), from which they may have been lifted. See
further 13.6.
Egyptian version
Two Egyptian manuscripts (Istanbul S. Esad Efendi Medresesi 119,2 and Cairo DM 644,1)
contain an anonymous set of auxiliary tables entitled Fat al-Karm al-Bq f ma{rifat al-dir
wa-falihi fq (The Victory of God the Noble and Eternal for Finding the Time since Sunrise
and Hour-Angle for all Latitudes).46 The main functions tabulated are:
B(,) = 1/2 B(,) = 1/2 Cos () Cos / R
for the domains:
= 1, 2, ... , 90; = 3, 6, ... , 27, 28, 30, 32, 33;30, 36, 38, ... , 48
and:
V(x,y) = arc Vers { R y / 2x }
for the domains:
x = 19, 20, ... , 30 and y = 0;5, 0;10, ... , 60;0.
The first function is labelled al-al, (half) the base and the arguments in the table of V are
called al-al al-mu{addal, the modified base, and fal al-jaybayn, the difference between
the two Sines. The entries in both tables were simply coped from al-Khall, in whose set B
is tabulated for some 50 values of and the increment in the argument y for the tables of V
is 0;10 rather than 0;5. Also, this later Egyptian set gives no horizontal differences in the tables
of V, but the simplified format makes them slightly easier to use. See further I-9.11.
each to three digits for each degree of argument. The second and third functions are computed
for parameters:
= 30;0 (Cairo) and = 23;35 .
The main set, incomplete in the Vatican manuscript, displays the function:
V(x,y) = arc Vers { R y / x } ,
with values to two digits for the domains:
x = 47;37, 47;57, ... , 51;57 and y = 1, 2, ... , Y(x),
where Y(x) is a certain maximum defined below (see an extract in Fig. I-9.9b). The instructions
for finding t(h,) indicate that one should first find B() and z(h,) (i.e., Sin H() - Sin h) and
enter these arguments in the main table. This is valid since:
t(h,) = V { B(), z(h,) } .
The argument x runs in intervals of 0;20 between the limits of B() for Cairo, and the argument
y runs in unit intervals up to Y(x), the greatest integer less than the value of Sin H(x). The
table originally contained about 825 entries, but in the Vatican manuscript only the page for
y > 31 remains. The entries are fairly accurately computed.
Being devised for latitude 30 al-Kha}s tables are not as useful as al-Khalls minor
auxiliary tables for solar timekeeping. It is probable that al-Kha} knew of al-Khalls tables,
but the formers tables were redundant anyway because by the 13th century there existed a
complete set of tables of t(h,) for Cairo. See further 1-6.15.1 and II-6.8.
Prayer-tables
With the passage of time only al-Khalls calendrical tables and the solar longitude table needed
to be updated. The decreasing obliquity (23;30 observed in ca. 1440 by Ulugh Beg47 and
23;29 used by later Ottoman astronomers48) was not considered sufficient cause to recompute
the corpus, or even the prayer-tables.
After the time of al-Tzn, the Damascus muwaqqit Ibn al-Kayyl provided a new set of
calendrical and solar tables to precede al-Khalls tables for timekeeping.49 These are extant
in numerous copies including MSS Berlin Ahlwardt 5758 (Sprenger 1858), 5759/5771
(Wetzstein 1146), and 5760/5772 (Wetzstein 1148), none of which credit al-Khall as the
author of the main tables. The same is true of the 19th-century Damascus muwaqqit Muammad
ibn Muaf al-anw,50 whose recension is contained in several copies including MSS
Damascus hiriyya 9233, Cairo M 173, Cairo TR 129, and Cairo DM 1007. al-anw,
however, modified the tables so that the entries displayed actual times in hours and minutes
according to the Ottoman convention that sunset is 12 oclock. Thus al-Khalls tables survive
in three versions: the original, and the two recensions of Ibn al-Kayyl and al-anw. See
further 11.4 and 11.13.
Besides these modifications to al-Khalls corpus, various new sets of prayer-tables were
compiled for a variety of localities after the 14th century (see further 11). Mention may be made
of the anonymous prayer-tables for latitude 34;20 (Tripoli ?) in MS Cairo M 228,3; for
latitudes 41;15 (Istanbul) and 36 (Aleppo) in MS Cairo M 255,6; for latitude 33;45 (Tripoli)
in MS Damascus hiriyya 4893; and for latitude 41;0 (Istanbul) and 21;30 (Mecca) in MS
Cairo DM 184. All of these are based on obliquity 23;31, which betrays the influence of Ibn
al-Shir and al-Khall. On the other hand, {Abd al-Fatt al-Dss tables for latitude 34;30
(Lattakia) in MSS Aleppo Awqf 911 and Leiden Or. 2808,2 are based on obliquity 23;35,
which betrays Egyptian influence, and have the format of the Cairo corpus. The later tables
in MS Aleppo Awqf 943 for latitude 35;50 (Aleppo) are based on obliquity 23;30, which
is the parameter of Ulugh Beg (rounded to minutes).
In MSS Oxford Seld. Supp. 100 and Paris BNF ar. 2521 there are separate tables of various
supplementary functions in the same format as the Cairo corpus. These functions are:
(r+D), (D+2N), (2N-s), (2N-s+D), (D+s), and (D+2n-r) .
Although not attributed to al-Khall, the entries are based on those in his main set of prayer-
tables.
Similarly, MS Leipzig 814, fols. 31r-31v, is one remaining folio of a set of prayer-tables
for Damascus which was drawn up using al-Khalls tables. The functions displayed are:
Z(12), H, D, za(7), , , d, ha, ta, Ta, hb, tb, 2N, N, , s, r, n, tq, and 2Nh .
No other examples of these prayer-tables are known. The function tb is labelled khir waqt
al-ikhtiyr, the last time for choosing (to begin the afternoon prayer) and hb is called irtif{
al-{ar al-thn, the altitude at the second afternoon prayer. In al-Khalls main set tb is called
m bayn al-uhr wa-khir waqt al-{ar, the interval between midday and the last time for the
afternoon prayer.
Hour-angle tables
al-Khalls tables of t(h,) are supplemented in various copies by a set of tables for T(h,).
Although it is not certain that these are the work of al-Khall himself (they are not contained
in MS Paris ar. 2558), evidence that they might be is contained in MS Cairo MM 71, where
both the tables of t(h,) and T(h,) are attributed to al-Khall while the table of a(h,) is
attributed to Shihb al-Dn al-alab. al-anw, besides converting the prayer-tables as
mentioned above, also converted al-Khalls hour-angle tables so that the entries (given for
solar altitudes in both the east and west) were expressed in hours and minutes according to
the Ottoman convention, as, for example, in MSS Cairo M 173 and DM 1007. See further
11.13.
Finally, in MS Damascus hiriyya 7387, fols. 57v-60r, of an abridgement of the Zj of
Ibn al-Shir by {Abd al-Ram al-Qazwn of Damascus (fl. ca. 1610),51 there is a set of tables
entitled jadwal al-dir li-rtif{ al-shams wa-yu{raf minhu qaws al-ru}ya min al-makth idh
dakhalta bi-nar al-burj ..., displaying the time remaining till moonset for lunar altitudes:
h = 7, 8, ... , 16,
and each degree of ecliptic longitude (symmetrically arranged). The entries are from al-Kha-
lls tables of T(h,), and, as stated in the title, are to facilitate calculations relating to lunar
crescent visibility. The tables can be used to find the arc of visibility (qaws al-ruya), that is,
the altitude of the moon above the horizon at sunset, from the difference in setting times of
the sun and the moon (qaws al-makth). The longitude of the point of the ecliptic which sets
with the moon, , is computed using the longitude and latitude of the moon and the latitude
of the locality, and then the difference in setting times can be determined using tables of oblique
ascensions.52 The tables can then be used to find the lunar altitude h corresponding to this time
s, with the modified longitude entered as argument.
52 For the procedures see King, Tables for Lunar Crescent Visibility, pp. 186-189.
53 On al-alab see n. I-2:42.
54 On this see 5.6 (illustrated).
400 part ii, chapter ten
(Note the different format of these tables in MSS Damascus hiriyya 9227 and Berlin
Wetzstein 1138 (Ahlwardt 5754-6).) See further 11.3.
Egyptian copies
MSS Cairo MM 43, MM 98, DM 758, and Princeton Yahuda 861,2 are Egyptian copies of
these tables. Evidence exists that the leading Egyptian astronomers in the 15th century, men
such as {Izz al-Dn {Abd al-{Azz ibn Muammad al-Waf}55 and Shams al-Dn Muammad
called Ibn Abi l-Fat al-f,56 had seen the tables. From a passing remark by al-Waf} in
his treatise Khulat al-durar (MS Manchester Rylands 361, fol. 45r) it can be inferred that
the expressions first and second functions (al-maf al-awwal and al-thn) were well
known to contemporary Egyptian astronomers. al-Waf} advocates the use of these functions
without even defining them. See further 6.14.
The Egyptian astronomer Muammad ibn Amad al-Bilbays wrote a commentary on the
auxiliary tables,57 stating that he found al-Khalls instructions to be long-winded. This
commentary, extant in the unique MS Cairo DM 442, copied ca. 1700, includes Sharaf al-Dn
al-Khalls instructions about the use of his uncles tables to compute functions for predicting
lunar crescent visibility (also found in MS Berlin Wetzstein 1138 (Ahlwardt 5754-6).
MS Cairo MM 167,8, copied in 989 H [= 1581/82], contains a fragment of an anonymous
Egyptian treatise on the sine quadrant with many numerical examples, mostly for latitude 30
(i.e. Cairo) and obliquity 23;35. The author deals at length with the functions f and g and
the determination of the qibla (M = 21). He mentions various earlier Mamluk astronomers
such as {Al} al-Dn bugh al-Baklamsh, (Taqi l-Dn) Ab hir, and Jaml al-Dn
{Abdallh ibn Khall al-Mridn,58 but not al-Khall. MSS Cairo MM 177,2, and Cairo DM
624,2, both copied ca. 1700-1750, are two other treatises dealing with al-Khalls functions
f and g, but neither of them mentions his name.59
Maghribi copy
MS London BL Add. 9599,31 is a late Maghribi copy of al-Khalls universal auxiliary tables.
This could be taken as evidence that it also travelled to the Islamic West, but the possibility
of a Maghribi copyist working in, say, Cairo, should not be excluded. See further 13.4.
Turkish copies
MS Istanbul Hamidiye 1453/3 is a copy of the tables prepared about 1465 in Edirne, while
MS Istanbul Ayasofya 2590 contains another copy of the tables compiled in 1491 by the
55 On al-Waf} (Suter, no. 437, and Cairo ENL Survey, no. C61) see n. I-9:29.
56 On al-f (Suter, nos. 447/460, and Cairo ENL Survey, no. C98) see n. I-9:30.
57 On al-Bilbays see Cairo ENL Survey, no. C80.
58 On these see Cairo ENL Survey, nos. C53, C56, and C47 (also Suter, MAA, no. 421).
59 Ibid., nos. C145 and D232.
the damascus corpus of al-khall 401
Istanbul muwaqqit Muammad ibn Ktib Sinn,60 preceded by a Turkish translation of al-
Khalls instructions. An unnumbered manuscript formerly (ca. 1970) in the private collection
of the late Professor Buhairi, American University of Beirut contains tables of f and g for
= 41 (Istanbul) and also of K for x = 41. The hapless individual who put these three tables
together was unaware that the argument x is in no way related to the latitude. See further 14.6.
CHAPTER 11
After the destruction of Damascus in 1401 there was little original in Syrian astronomy, but
the activity certainly continued. In the early 15th century Shihb al-Dn al-alab prepared a
redaction for Damascus of the lkhn Zj (Maragha, mid 13th century);1 at the end of that
century {Abd al-Ramn al-li prepared a redaction for Damascus of the Zj of Ulugh Beg
(Samarqand, ca. 1430);2 and in the mid 16th century Ibn al-Kayyl prepared a redaction for
Damascus of al-Kshs Khqn Zj (Samarqand, ca. 1438).3 Recensions of the Zj of Ibn al-
Shir (Damascus, ca. 1350) were prepared in the mid 16thcentury by Ibn Zurayq,4 around 1600
by al-Qazwn, and even in the 18th century by al-Mukhalllt.5 There was also some activity
in astronomical timekeeping. Shihb al-Dn al-alab contributed some solar azimuth tables
to the main Damascus corpus (11.2), and Ibn al-Kayyl compiled some calendrical and solar
longitude tables which were used with al-Khalls prayer-tables until the 19th century (11.4).
Various sets of prayer-tables prepared in order to serve other cities in Syria such as Aleppo,
Tripoli, Nablus and Lattakia (11.5-9). In the 19th century al-anw made an attempt to revive
traditional astronomy in Damascus, and brought up to date the calendrical and solar tables,
which were used alongside al-Khalls prayer-tables (11.13).
In MS Paris BNF ar. 2521, fols. 69v-75r, penned in the 16th century, there is an anonymous
set of prayer-tables which the title states are for latitude 33;6. Elsewhere in the same
manuscript (fols. 14v-16r), there are prayer-tables based on those of al-Khall (discussed in
10.11), for parameters = 33;30 and = 23;31. The anonymous set gives values of the
following functions for each degree of , a page of tables serving each zodiacal sign:
2Dh, , H, D, ta, ha, Ta, n, 2N, s and r .
1 On al-alab (5.2) see n. 11:6. His Zj is preserved in MS Cairo M 226,1 (see Cairo ENL Survey, no. C69/
2.1.12).
2 On al-li see Suter, MAA, no. 454; Cairo ENL Survey, no. C87; and ~hsanolu et al., Ottoman
Astronomical Literature, I, pp. 204, 205, and 225.
3 On Ibn al-Kayyl (11.4) and n. 11:8. This work of his is extant in MS Dublin CB 4677.
4 On Ibn Zurayq see Suter, MAA, no. 426; Mayer, Islamic Astrolabists, suppl., p. 296; Cairo ENL Survey,
no. C116; and ~hsanolu et al., op. cit., I, pp. 155-158, no. 79. On various recensions of Ibn al-Shirs Zj see
my article Ibn al-Shir in DSB.
5 On al-Qazwin see Cairo ENL Survey, no. D38; and ~hsanolu et al., op. cit., I, pp. 301-302, no. 160. On
al-Mukhallalt see Cairo ENL Survey, no. D90, and ~hsanolu et al., op. cit., II, pp. 544-545.
later syrian tables for timekeeping 403
It is immediately apparent that the tables are very corrupt, and analyzing them is made difficult
by the fact that the individual who compiled them did not understand the argument system
used in the tables from which he was copying. Thus the entries in his tables do not display
the symmetry which the tabulated functions naturally enjoy.
The table of is based on = 23;35, but the values of H are not consistent with the
corresponding values of or any particular value of . It seems, however, that the compiler
wished his tables to work for = 33;30, because most of his values are lifted from al-Khalls
prayer-tables. It is also curious that the second digit in these entries for D assumes values only
between 0 and 5, which suggests that the units used are sixths of a degree.
An analysis of the entries for a particular longitude will suffice to illustrate their deficiency.
We consider = 45, and first note that the corresponding entries for = 135, which should
be the same, differ in several instances. The entries 13;23 ( = 45) and 13;25 ( = 135)
for 2Dh differ from al-Khalls value 13;30. The entry 16;26 for is accurate for = 23;35,
and the entry 72;56 for H corresponds to = 33;30. For D our table has 100;2 ( = 45),
which seems to mean 100;20, and 100;3 ( = 135), that is 100,30: the accurate value is
about 101;15. The entries 55;38 and 45;48 for ta and Ta are al-Khalls values for = 46,
but the entry 37;25 for ha is his for = 45. The entries 108;59 and 157;34 for n and 2N
are likewise al-Khalls for = 45. The 23;4 ( = 45) and 22;56 ( = 135) for s differ
from al-Khalls 22;50 ( = 45) and 22;54 ( = 46), but the 25;50 ( = 45 and 135)
for r is al-Khalls for = 46.
The tables are thus full of inconsistencies, and it was doubtless from tables of this kind
originally computed for Cairo that those described in 7.6 were copied out in words. Two sets
of these corrupt prayer-tables now preserved in Syria, MSS Aleppo Amadiyya 1310 and
Damascus 9227, appear to be based on tables computed for Cairo.
The Damascus astronomer Shihb al-Dn Amad ibn Ibrhm al-alab (d. 1455)6 computed
a set of tables of the function a(h,), measured from the prime vertical, for each integral degree
of h and each degree of . The underlying parameters, that is, the values used for and ,
are the same as those used by al-Khall (see 10.5-6). These tables are preserved in MSS Cairo
K 8525, copied in al-alabs own hand and thus datable ca. 1425, Cairo MM 71,1, copied
ca. 1600, and Damascus 9227. The first manuscript contains al-Khalls prayer-tables, and al-
Khalls values of t(h,) are tabulated side by side with al-alabs value of a(h,). In the
second the function a(h,) is tabulated separately, and in the third, triplets of values (t,T,a)
are given, the first pair of functions being attributed to al-Khall.
al-alabs azimuth tables, which are fairly accurately computed, are similar in conception
to those of Ibn Ynus for Cairo, computed over four centuries previously (4.4 and 5.2), and
to the later azimuth tables computed for Alexandria (8.5), but they have al-Khalls format.
al-Khall had also compiled azimuth tables for the latitude of Damascus, but having arguments
H and h rather than h and (10.3).
6 On al-alab see Suter, MAA, no. 434; Brockelmann, GAL, II, p. 159; and Cairo ENL Survey, no. C69.
404 part ii, chapter eleven
MS Damascus hiriyya 9227 of Shihb al-Dn al-alabs solar azimuth tables (11.2) also
contains a set of tables attributed to Muammad ibn Barakt ibn al-Buraw, on whom I have
no further information.7 These display a(,h), the time to the beginning of the interval for the
afternoon prayer as a function of solar longitude and solar altitude in the eastern sky over
Damascus. The entries are based on the simple relation:
a(,h) = ta() - t(h,) ,
and for each degree of the arguments for h are the integral values such that ha() h H().
A marginal note by al-anw (11.13) explains the method of computing the entries in the
table from those of ta() and t(h,). See also 7.3 on some similar tables for Cairo.
In MS Berlin Ahlwardt 5754/5/6, fols. 62v-94r, there are some anonymous tables appended
to others by al-Khall (10.5) of a function which is described as al-maktb bi-l-sawd al-
bq li-l-ghurb wa-bi-l-amra li-l-{ar, time remaining until sunset if written in black ink
or time remaining until the beginning of the afternoon prayer if written in red ink. The tables
have the same format and the same argument domains as al-Khalls hour-angle tables, and
it is understood that the altitudes are eastern. The tabulated function is simply:
f(h,) = T(h,) for h < ha() and a(h,) for h ha() .
Note that the tables of a in the Damascus and Berlin manuscipts have different formats.
MSS Damascus hiriyya 4893, Istanbul Esat Efendi 1990, Berlin Ahlwardt 5759/5771
(Wetzstein 1146), ca. 1700, and 5760/5772 (Wetzstein 1148), ca. 1800, are copies of al-
Khalls prayer-tables (10.6) in which the main tables are preceded by an introduction and
some calendrical tables by the mid-16th-century Damascus astronomer Ibn al-Kayyl.8 There
is no mention of al-Khall.
7Ibn Barakt, whose full name is given in MS Damascus hiriyya 9227, is not listed in the modern sources.
8On Ibn al-Kayyl see Suter, MAA, no. 474; Brockelmann, GAL, II, p. 469; Cairo ENL Survey, no. D60;
and ~hsanolu et al., Ottoman Astronomical Literature, I, p. 127, no. 59.
later syrian tables for timekeeping 405
MS Cairo M 255,6, copied 1060 H [= ca. 1650], contains al-Khalls introduction to his
prayer-tables followed by some anonymous prayer-tables for latitudes:
= 36 (Aleppo) and 41;15 (Istanbul).9
Mention of these was omitted from the first version of this text, and all that is available to
me now is a photo of an extract from the tables for Aleppo (Fig. 11.5). This shows tables of
the same kind as those of al-Khall but for = 23;35 rather than his more up-to-date value
23;31. The functions tabulated are:
H, Z, D, ha, ta, Ta, N, n, r, s, , d and tq .
The quantities ta and Ta are labelled iat al-{ar and iat al-ghurb, expressions which I
do not recall having seen elsewhere (though see below). I have not investigated the parameters
used for r, s and q. See also 14.8.
Fig. 11.5: An extract from an anonymous set of prayer-tables for the latitudes of Aleppo and Istanbul. [From
MS Cairo M 255,6, fols. 78v-79r, courtesy of the Egyptian National Library.
MS Aleppo {Uthmniyya 14, copied ca. 1750, contains another anonymous set of prayer-
tables for Aleppo. Values of the following functions are displayed in equatorial degrees and
minutes for each degree of from Capricorn 1:
D, ta, Ta, s, n and r
and I have not investigated the underlying parameters. The entries are written in Arabic num-
erical notation. The five functions are labelled iat al-zuhr, iat al-{ar, iat al-maghrib,
iat al-{ish}, and iat al-fajr.
MS Damascus hiriyya 4893 contains calendrical and solar tables due to Ibn al-Kayyl (11.4),
followed by some anonymous prayer-tables for the latitude of Tripoli, stated to be 34. (See
also 9.3 on two earlier sets of prayer-tables for the latitude of Tripoli, whioch may ber related
to each other and to some of these.) The functions tabulated are:
D, ta, Ta, s, n, r and za(7) .
The entries are arranged in six columns of thirty, beginning with the column for Capricorn/
Aquarius, although the entries for D() begin with the column for Aries/Virgo. Analysis reveals
firstly that the table of D() is based on the parameters:
= 33;45 and = 23;31.
Some of the entries are garbled in such a way as to give the impression that they were copied
from another set with a different format: the original table appears to have been rather
accurately computed. Secondly, certain of the other tables, including those of ta, r and s, are
the same as those for latitude 34 in MS Princeton Yahuda 861,1 of the Natja attributed to
al-Waf} (8.1). Thus they are supposedly based on a value of 23;35 for , but the entries were
so carelessly computed that the value of is of little significance. Finally, the table of n()
is based on the above-mentioned table of D() (computed for = 33;45 and = 23;31) and
the anonymous values for r() and s() in the Princeton manuscript (computed for = 34;0
and = 23;35)!
MS Berlin Ahlwardt 5765 [= Wetzstein 1149, fols. 85v-88v], copied ca. 1850, consists of
anonymous prayer-tables compiled for the latitude of Nablus, taken as 32.10 The tables are
preceded by a short introduction in which the title Natjat al-afkr f a{ml mawqt al-layl
wa-l-nahr (compare the title of al-Ldhiqs tables for Cairo in 7.8) is mentioned. The
functions tabulated for each degree of solar longitude are:
D, ta, Ta, s, n and r ,
expressed in equinoctial hours and minutes. Values are displayed side by side in columns, and
are written in modern Arabic numerals in an untidy hand see Fig. 11.7 for an extract. They
are generally inconsistent for a given solar longitude. Consider for example, the entries for
Gemini 1 ( = 61), which are:
D: 7;4 h ta: 3;56 h Ta: 3;3 h s: 1;49 h n: 8;40 h r: 1;33 h .
10The only medieval source which gives 32;0 for Nablus (32;10 is a common value) is labelled qiys
(measurement) by Abu l-Fid}: see Kennedy & Kennedy, Islamic Geographical Coordinates, pp. xxx and
235. Given the crude nature of the tables, this may not be significant: possibly 32 was simply rounded from
32;10.
later syrian tables for timekeeping 407
Fig. 11.7: An extract from some messy prayer-tables for Nablus. [From MS Berlin Ahlwardt 5765 [< Wetzstein
1149], fol. 87v, courtesy of the Deutsche Staatsbibliothek (Preuischer Kulturbesitz).]
Notice that the sum of these values, which should correspond to a complete day, is over 26
hours.
MSS Aleppo Awqf 911 and Leiden Or. 2808(2) are copies of a work entitled Asna l-ghyt
f {ilm al-mqt, The Best Ends in Timekeeping, compiled by {Abd al-Fatt ibn Ibrhm
al-Mlik al-Azhar al-Ds,11 an Egyptian astronomer who was a student of Riwn Efend
(7.10). After a brief introduction in 3 bbs al-Ds presents a set of prayer-tables for parameters:
= 34;30 (Lattakia) and = 23;35.
The format of the tables is the same as that of the main Cairo corpus. The functions tabulated
are:
, d, H, D, ha, ta, Ta, s, r, D, B, C, h0, 2N, and .
The values of D are those for Cairo found, for example, in MS Cairo DM 108 of the main
Cairo corpus (4.11). The work concludes with a table of the solar motion between the prayer-
times as in the tables of Shams al-Dn al-Ldhiq (7.8), and a star catalogue for 1160 H
[= 1747].
11 On al-Ds (I-4.8.7, etc.) see Brockelmann, GAL, SII, pp. 1017, where the manuscript of his prayer-tables
which is now in Leiden is listed; Cairo ENL Survey, no. D62 (there incorrectly as al-Dayst without a dot); and
~hsanolu et al., Ottoman Astronomical Literature, II, p. 514, no. 352 (now as al-Daysa).
408 part ii, chapter eleven
Cairo M 92, copied ca. 1750, contains a small zj by an individual named Muammad ibn
{Abd al-Mamd al-Ldhiq compiled in Lattakia in 1170 H [= 1756/57].12 This contains tables
of three spherical astronomical functions, namely, , , and , computed to two digits for
each degree of and based on the parameters:
= 34;30 (Lattakia) and = 23;35.
MS Istanbul Badatli Vehbi Efendi 887 contains a selection of anonymous planetary tables
as well as others for and attributed to al-akm al-Ldhiq.
The tables are of interest in that the values in the table of () have been derived by linear
interpolation between the entries for = 0, 30, 60, and 90, which are rather accurately
computed.13 I know of no other Islamic ascension tables as crude as this, although see 8.6 on
some late Egyptian tables also based on a simplistic interpolation scheme. The table of d(),
on the other hand, is not based on such a scheme, and so al-Ldhiq had tables of () and
d() at his disposal with which he could have compiled a proper table of (), but he preferred
his own curious technique.
MS Cairo DM 184, penned in 1171 H [= 1757/58], is the only copy known to me of a treatise
on astronomy compiled in the early 17th century by Muammad ibn Mamd al-Manshr,
otherwise known for his work in the Qur}nic sciences.14 al-Manshr presents various
calendrical and planetary tables and also four sets of prayer-tables, for Damascus, Cairo, Mecca,
and Istanbul. The tables for Damascus are those of al-Khall (10.6), based on parameters:
= 33;30 and = 23;31 ,
and are attributed to him. The functions tabulated include various ones not generally displayed
in the main Damascus corpus (10.6):
Sin H, Sin , and Sin .
al-Manshr tabulates some twenty six functions side by side for each degree of starting
at Capricorn 1, namely:
H, Sin H, Z(12), Z(7), D, 2Dh, , d, , Sin , ha, ta, Ta, hb,
2N, N, 2Nh, , s, r, , Sin , h0, n and tq .
In the column for Sin ( > 0) al-Manshr tabulates al-Khalls values. However in the
column for Sin ( < 0) a completely different function appears, which I am unable to identify.
Sample values from this garbled table are:
1 89;3 31 76;31 61 165;25
15 89;1 45 69;43 75 60;29
30 88;56 60 62;31 91 57;36
The prayer-tables for Cairo are taken from the main corpus, although this is not stated, and
are thus based on the parameters:
= 30;0 and = 23;35 .
The columns are ruled for functions:
H, D, ha, ta, Ta, s, r, 2N and n ,
but only the values for D, ta, Ta, s and r have been filled in.
The prayer-tables for Mecca and Istanbul are not attested in any other source known to me.
The functions tabulated are:
H, D, ha, ta, Ta, s, r, and 2N .
It is stated that the underlying latitudes are 21;30 and 41;15 but in fact the tables for Istanbul
are based on latitude 41;0. The obliquity is 23;31 in both cases, and the parameters for
morning and evening twilight are 19 and 17. It seems unlikely that al-Manshr compiled
the prayer-tables for Mecca and Istanbul himself. However, in view of the fact that both sets
are based on obliquity 23;31, a value used in mqt tables only as far as we know by
al-Khall and Ibn al-Shir, it seems probable that these tables were compiled in Damascus.
al-Manshr states in his introduction that his calendrical tables are based on the techniques
of {Abd al-Laf ibn Ibrhm al-Kayyl (11.4), which were in turn taken over from al-Suy
al-Jall ibn al-Kaml (?).15 In his discussion of the determination of the prayer-times he refers
to the statement of Sharaf al-Dn al-Khall concerning the observations made on Jabal Qsiyn
with his uncle Shams al-Dn al-Khall.
15 This can hardly be a reference to the famous 16th-century polymath Jall al-Dn al-Suy (Suter, MAA,
no. 449; Cairo ENL Survey, no. 103). His treatise on the references to astronomy in the Qur}n and adth has
been published in Heinen, Islamic Cosmology. However, al-Suyu is not known to have written anything of
consequence on the calendar. On his poem on the midday shadows for each month of the Coptic year see III-
9.7a.
16 See the article Shafa_ on twilight in EI .
2
410 part ii, chapter eleven
(b) the interval between the time to extinguish the candles and the time to start fasting in
Raman (bayn al-afy wa-l-imsk) is 0 25 minutes; and
(c) the interval between the time to start fasting and daybreak (bayn al-imsk wa-l-fajr)
is 0 45 minutes.
I am not sure what these intervals mean, as the time between the appearance of the zodiacal
light and the true dawn can hardly be 0;45 + 0;25 + 0;45 = 1;55 hours. See 4.10 on these
institutions as cultivated in Cairo.
MS Cairo M 81,1 (fols. 1r-84v), copied ca. 1900, contains a set of tables of the function T(,h)
for altitudes in the east and west, and with entries in hours and minutes for each degree of
both arguments. The underlying latitude is found by inspection to be = 32 (Jerusalem), and
values of D() are given on each sub-table of entries: see Fig. I-2.2.6. These tables are followed
by a set of tables of the function t(h,) with values in degrees and minutes, for parameters:
= 32;0 (Jerusalem) and = 23;35 .
These tables are copied without the solar longitude arguments, and without the degrees of the
entries, except at the head of each column. These are probably the tables computed by Ibn
al-Rashd (I-2.1.5 and II-6.12). The manuscript concludes with tables of the three functions:
N, d and h0
for latitude = 32;10 (Nablus).
I suspect that all of these tables in the Cairo manuscript were copied from an original in
the Khlidiyya Library in Jerusalem.
Muammad ibn Muaf ibn Ysuf al-antw al-anda}t} al-Azhar al-Dimashq was a
muwaqqit in the Umayyad Mosque at Damascus who died in 1889, and thus represents the
last period of medieval Islamic astronomy in Syria.17 Numerous of the astronomical
manuscripts in the hiriyya Library in Damascus are written in his hand: in particular, two
manuscripts are actually attributed to him. The first, MS 9353, is a collection of tables for
constructing sundials, in which he used on the parameters ( and ) of Ibn al-Shir. It was
al-anw who by accident broke the sundial of his illustrious predecessor which embellished
the main minaret (al-{ars) of the Umayyad Mosque in Damascus: whilst trying to realign it:
he constructed a new one with identical markings which is still in position on a balcony of
the minaret.18 The second, MS 9233, contains a brief introduction explaining the calendrical
and solar tables which form a small part of the manuscript (pp. 1-21). These are followed by
a complete set of al-Khalls hour-angle tables, prayer-tables and tables of ascensions (pp. 22-
156), but the only reference to al-Khall is a statement that it was he who computed the tables
of Cotn H() and Cotn ha() (n = 7 and 12). These are not contained in MS Paris ar. 2558 of
al-Khalls prayer-tables (10.2). MS Damascus 9233 concludes with some tables (pp. 157-161)
for the lunar mansions, computed for the years 1292 and 1294 H [= 1875/77]. Finally, MS
7387 contains a set of tables attributed to al-anw displaying the function T(,h) expressed
in equinoctial hours and minutes for each integral degree of both arguments. I have little doubt
that these were computed from the corresponding values of T(h,) in equatorial degrees and
minutes in the tables of al-Khall (cf. 10.5) by a simple division by 15. Note that the format
of the two sets of tables is different.
MS Cairo M 173, copied ca. 1900, is the only one of the manuscripts associated with al-
anw that I have located which contains a set of prayer-tables for Damascus with times
expressed according to the Ottoman convention (14.0). Here:
s, i, r, R, m and a
are displayed side by side for each degree of starting with Capricorn 1. The time of the
imsk, i, is always 15m before daybreak, r.
Several of the manuscripts containing tables associated with al-anw, such as MSS
Damascus hiriyya 3116, fols. 1v-2v, 9233, p. 37, 10076, fol. 16r, and 7388, fol. 16v, contain
a table computed by him entitled daq}iq al-ikhtilf wa-nif qur al-shams li-{ar lm-jm lm,
the difference minutes and the radius of the sun for latitude 33;30, which purports to
represent the interval in equatorial minutes between the time when the centre of the sun is on
the true astronomical horizon, and the time when the solar disc has disappeared below the
visible horizon at Damascus. Values are given for each degree of to one or two digits and
are derived by linear interpolation from three significant entries, namely:
EQ: 67 SS: 84 WS: 51 .
In two of the sources, MSS Damascus hiriyya 9233, p. 9, and 7388, fol. 16v, it is stated
that the value used for the radius of the sun is 0;15, and if this is subtracted from the entries
in the table the resulting values are the time taken for the centre of the sun to move between
the two horizons, that is, the D() of earlier Egyptian sources (4.11, 5.7, 6.10, 7.1, 8.2, and
8.8). Indeed, if we subtract 15 minutes from the above values we obtain:
EQ: 52 SS: 69 WS: 36 .
If we multiply the corresponding values for Cairo, namely, 47, 62, and 32 minutes, by the ratio
of the latitude of Damascus to that of Cairo, namely, 33;30/30, according to the method of
{Abd al-Qdir al-Minf (8.2), we obtain:
EQ: 52;29 SS: 69: 14 WS: 35;44 .
which, when rounded, are the same as those underlying al-anws tables. (If we assume that
the distance between the true horizon and the circle 0;15 below the visible horizon is 0;40
+ 0;15 = 0;55, then the time in equatorial minutes taken by the sun to move between these
two positions is actually 66 at the equinoxes and 74 at both solstices.)
In MS Damascus 9233, p. 9, there are some notes in the hand of al-anw which explain
the use of this table to modify the various functions of timekeeping which relate to horizon
phenomena. The procedures he outlines are along the same lines as those of Muammad al-
Minf (7.1). al-anw concludes with the statement that all the muwaqqits use a fixed amount
to adjust for horizon phenomena. They call this quantity, which does not vary throughout the
412 part ii, chapter eleven
year, tamkn. al-anw complains that the muwaqqits do not know why they do this and they
fall into error. What surprises him more, he says, is that they do not know the truth and will
not accept it from one who does. al-anw was very much in control of his subject, but his
hapless colleagues represent the ultimate decline of Syrian astronomy, and the end of the
brilliant tradition associated with al-Mizz, Ibn al-Shir, and al-Khall.
In 1971 the Director of the Umayyad Mosque in Damascus told me that he had never heard
of al-anw, the last traditional muwaqqit at the Mosque not dead a century, let alone al-
Khall. The manuscripts in the nearby hiriyya Library had just been catalogued for the first
time, but the name of al-Khall does not appear there: the tables attributed to Ibn al-Kayyl
and al-anw were listed but had not inspired any interest locally.
MS Aleppo Awqf 943, copied ca. 1800, contains an anonymous set of anonymous auxiliary
tables for computing the hour-angle (I-6.14.4), and a set of prayer-tables for parameters:
= 35;50 (Aleppo) and = 23;30
attributed to {Abdallh al-alab, muwaqqit at the Umayyad Mosque in Aleppo ca. 1750.19
The same prayer-tables occur in MS Aleppo Awqf 962. The functions tabulated are the
following:
N nif qaws al-layl al-aqq
D nif qaws al-nahr al-mar}
m waqt al-uhr al-shar{
R ul{ al-shams bi-l-ufuq al-mar}
H ghyat irtif{ al-shams
Sin H jayb al-ghya
Z = Cot H ill al-ghya
ha irtif{ al-{ar al-awwal
Sin ha jayb al-{ar al-awwal
Sin H - Sin ha al-fal bayn al-jaybayn
Vers ta sahm d}ir al-{ar
ta fal d}ir al-{ar
a waqt al-{ar al-awwal
hb, ... , b ditto (6 functions) for al-{ar al-thn
Vers (D + r) sahm fal d}ir al-fajr
D+r fal d}ir al-fajr
r waqt al-fajr
Values are given for each degree of beginning with Cancer 0. The times are given according
to the Ottoman convention, and are expressed in hours, minutes and seconds, and values of
other functions are given to two digits. Separate tables display values of:
19 On {Abdallh al-alab (I-6.2.7) see ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 560-561,
no. 398.
later syrian tables for timekeeping 413
Figs. 11.14a-b: An extract from {Abdallh al-alabs prayer-tables, showing the first half of the entries (a),
and his tables of the absolute base (90 values) and the difference between the Sines of the solar meridian altitude
for 180 plus the solar longitude and the Sine of 19 (b). [From MS Aleppo Awqf 943, courtesy of the
Amadiyya Library.]
The late Syrian treatise on the 15 times of day with religious significance preserved in MS
Aleppo Awqf 970 (see already 2.11) concludes with a set of auxiliary tables for timekeeping.
The functions are labelled mayl, sahm al-mayl (= sahm al-al al-mulaq), and nisba sahmiyya,
and I have not investigated their character. Values appear to be given to six decimal places
(!), and the underlying parameters are stated to be:
= 35;45 (Aleppo) and = 23;28.
Some numerical examples follow, as well as a computation of the qibla at Jerusalem as 45;27.
The anonymous author states that the scholars of astronomy and geometry have established
that the visible horizon is half a degree below the true horizon. At the solstices the legal day
is 2 longer than the day derived mathematically, the difference being twice the sum of the
solar radius and the time taken by the sun to move between the two horizons, both 1/2. For
the summer and winter solstices he advocates respectively differences of 3 and 4. This means
that he is assuming the following corrections for the difference minutes:
EQ: 1/2 SS: 1 WS: 11/2
For = 36 and h = 1/2, the accurate values are:
EQ: 0;37,5 SS: 0;42,41 WS: 0;42,31 .
Thus the underlying theory is the (inappropriate) one used by earlier Egyptian astronomers (see,
for example, 4.11 and 7.1) and some earlier, but still late, Syrian astronomers (11.8 and 11.13).
yemeni and hejazi tables for timekeeping 415
CHAPTER 12
In the Yemen there was considerable activity in mathematical astronomy between 900 and
1700.1 More than 15 zjes were compiled and several sets of tables for timekeeping were
prepared during this period. This colourful Yemeni tradition is of interest to the history of
science not least because it was based mainly on Egyptian, {Irq, and Tunisian sources, some
of which are no longer extant.in their original form. The earliest Yemeni tables which survive
are those on the Muaffar Zj of Muammad ibn Ab Bakr al-Fris, based on Iranian sources.2
About 700 H [ 1300] the Yemeni astronomer Abu l-{Uql compiled an enormous corpus of
tables for timekeeping for the latitude of Taiz (12.1); he was obviously more influenced by
the Mamluk Eyptian tradition. There are tables for instrument construction and timekeeping
in the two treatises on instruments and mathematical astrology by the late-13th-century Rasulid
Sultan al-Ashraf (12.2). The late-14th-century Rasulid Sultan al-Afal has left us a disordered
pile of tables in his astronomical miscellany (12.4). Some simpler tables are found in a
remarkable Yemeni ephemeris from 1405 (12.5). An early-14th-century Yemeni astronomer
apparently named Ibn al-Mushrif, who cannot be identical with the early-15th-century Egyptian
astronomer of the same name (I-9.8, etc.), may have compiled a complete set of tables for the
latitude of Zabid (12.1), incorporating some of Abu l-{Uqls tables for Taiz. Another Yemeni
astronomer named {Abdallh ibn al D{ir (12.6) compiled a set of tables for timekeeping
for Sanaa, of which only a fragment survives. The prayer-tables of al-Thbit (12.7), compiled
for the Yemen (latitude of Sanaa), are for timekeeping by lunar mansions. Those of al-Wsi{
ca. 1940 (12.12) give no hint of any awareness of the achievements of his predecessors in
Rasulid times.
As far as I know, no astronomical works of scientific consequence were compiled in Arabia
outside the Yemen. (See, however, the highly-significant treatise on folk astronomy by Ibn
Raq (2.3).) Some prayer-tables for Mecca (6.10 and 11.10) were probably compiled in Cairo
and Damascus, respectively. Some late prayer-tables for Mecca and Medina have been located
(12.8-10).
1 A survey of Yemeni astronomers and their surviving works, based on about 100 manuscripts, is in King,
Astronomy in Yemen. New information is contained in my review of the facsimile edition of al-Afals miscellany
cited in n. 12:9.
2 On al-Fris (I-7.1.7 and II-2.2) see King, Astronomy in the Yemen, no. 9.
416 part ii, chapter twelve
The extensive corpus of tables for timekeeping entitled Mir}t al-zamn, The Mirror of Time,
was compiled for the latitude of Taiz by the Yemeni astronomer Muammad ibn Amad al-
abar, known as Abu l-{Uql, ca. 1300.3 Most but not all of the tables in this corpus are
extant in MSS Berlin Ahlwardt 5720 (Mq. 733,3), copied in Mukhkha in 1209 H [= 1795].
MS Milan Ambrosiana C84 is a late anonymous abridgement of the tables, and a manuscript
in a private collection in Sanaa of an astronomical miscellany by the Yemeni Suln al-Afal
compiled ca. 1375 (12.4) contains a few isolated tables from the corpus. Abu l-{Uql also
compiled two zjes. The first, called the Mukhtr Zj, is extant in MS London BL Or. 3624,
copied in 1008 H [= 1599], and is based mainly on an early zj of Ibn Ynus. The second is
no longer extant, but a few tables from the Zj of Abu l-{Uql which are not from the Mukhtr
Zj are found in the Sanaa manuscript.
The title folio and beginning of the introduction are missing from the Berlin manuscript,
and the tables in it are in fact anonymous. The attribution to Abu l-{Uql is secured from
references in the Sanaa manuscript. The introduction (fols. 16r-23v) is in two maqlas of nine
and seven bbs dealing with the various operations of timekeeping by day and night, respec-
tively. All the tables are based on the parameters:
= 13;40 (Taiz) and = 23;35 .
The tables in the Berlin manuscript are bound in disorder, but we can distinguish the following
main sets:
(a) Tables of t(h,) and T(h,), the hour-angle and time since sunrise as a function of solar
altitude and solar longitude (fols. 128v-166r). See further I-2.1.2, with extract.
(b) Tables of H(h,), the longitude of the horoscopus as a function of solar altitude (fols.
26r-31v, 33r-10v, 122r-123v). See further I-3.1.1and the extract illustrated there.
(c) Tables of coordinates of fixed stars and lunar mansions, and tables of h(H), the altitudes
of various fixed stars at daybreak as a function of the longitude of the horoscopus (fols.
32r-32v, 24r-25v, 108v-120v). See further below and also I-4.5.1 (illustrated).
(d) Some spurious tables for Sanaa, computed by Ibn D{ir (12.6).
Various minor tables by Abu l-{Uql relating to timekeeping are contained in the various
manuscripts mentioned above. Firstly, the Sanaa manuscript contains an anonymous set of
spherical astronomical tables displaying the functions:
, 2D, 2Dh, H, 2N and ,
side by side for each degree of . Values are given to two digits and the underlying parameters
are:
= 13;40 (Taiz) and = 23;35 .
Details of the rising, culmination and setting of certain fixed stars, as well as agricultural
information, are also given for each 1 of , roughly corresponding to each day of the year.
It is stated that this information is based on the positions of the stars in 631 H [= 1233/34].
The tables are in fact based on those of the Mukhtr Zj (MS London BL Or. 3624, fols. 176v-
184r), in which each of the functions:
Fig. 12.1: A table displaying the oblique ascension of the horoscopus at daybreak, most probably by Abu l-
{Uql. [From a manuscript in a private collection in Sanaa; taken from Varisco & Smith, eds., al-Afals
Anthology, p. 317, with permission of Professor Daniel Varisco.]
, 2Dh, H, d and
is tabulated to two digits for each degree of for latitudes 13;0 (Aden), 13;40 (Taiz), 14;0
(Zabid), and 14;30 (Sanaa). Tables of the functions:
H, Dh and d
computed for = 13;40 and = 23;35 also occur on fol. 107r of the Berlin copy of the Mir}t
al-zamn. I conclude that the tables in the Sanaa manuscript were also compiled by Abu l-
{Uql. The additional information on the stars and on cultivation of crops formed the basis
of an almanac, extant in several copies, of which I have examined MS Milan Ambrosiana C46
(fols. 52-58).4
Elsewhere in the Sanaa volume there is a table of the functions:
Z(7) and ha(H)
with entries to two digits for each degree of H from 1 to 90. It is stated that this table is
taken from the Mir}t al-zamn by Abu l-{Uql, and the same table occurs on fol. 108r of
the Berlin copy.
A table for twilight which is probably by Abu l-{Uql is extant only in the Sanaa manuscript
(see Fig. 12.1), where it is not actually attributed to him. It is, however, rather accurately
computed for his parameters and is based on other tables compiled by him. The function
tabulated is simply the oblique ascension of the horoscopus at daybreak, r(), as in the Cairo
and Damascus corpuses (4.10 and 10.6), and the underlying parameter for morning twilight
4 See now Varisco, Yemeni Almanac.
418 part ii, chapter twelve
is 20. The function is called al-d}ir min al-falak {inda ul{ al-fajr, literally, the amount
the celestial sphere has revolved at daybreak, and is defined by:
r() = () - r() = () - T(20,*) .
Values of () for the above parameters are given in the second twilight tables described
below, and a separate table of () with the same entries occurs in MS Paris BNF ar. 2523,
fol. 95v, of an anonymous late-14th-century Yemeni Zj (12.3). The table of () for Taiz in
MS London BL Or. 3624 of Abu l-{Uqls Mukhtr Zj is based on = 13;40. Abu l-{Uql
tabulated t(h,) and T(h,) for each degree of both arguments (see above), and his table of
T(h,) for h = 20 is found on fol. 138r of the Berlin manuscript. The values of r() in the
Sanaa manuscript can be derived from the appropriate values of () in the same Berlin
manuscript and Paris BNF ar. 2523, and of T(h,) for h = 20 again in the Berlin manuscript.
The other twilight tables of Abu l-{Uql are found on fols. 24v-25v, 32r-32v, etc., of the
Berlin copy. They show the altitudes at daybreak of about eight prominent stars in the eastern
and western sky as a function of the longitude of the horoscopus. For each degree of H, values
to two digits are given for the stellar altitudes, and also for (H) and the duration of morning
twilight r(H), and for the corresponding solar longitude. I have not investigated the parameters
used for the various stars, but there is a table of right ascensions of thirty stars on fol. 24r
of the Berlin manuscript which Abu l-{Uql states he used for the twilight tables. On fol. 108v
there is another table giving the ecliptic coordinates and declination, and, for the latitude of
Taiz, the meridian altitude, half arc of visibility, and the co-ascendants and co-descendants of
the stars. No doubt these are the coordinates which underlie all of Abu l-{Uqls tables for
timekeeping by the stars. The star catalogue was originally dated, but the year has been
deliberately erased. However, the stellar positions correspond to about 1300. In the Sanaa
volume there is a similar star catalogue compiled for a later date, with an additional set of
entries displaying the meridian altitudes at Zabid: this table is stated to have been taken from
a Zj by Ibn al-Mushrif, otherwise unknown to the literature for he cannot be identical with
the early-15th-century Egyptian astronomer with this name (I-9.8). In the Sanaa manuscript
there are also various astrological tables by Abu l-{Uql, some of which originally belonged
to the Mir}t al-zamn, such as tables of the longitudes of the astrological houses.5
In all, the Taiz corpus contains over 100,000 entries, and is the largest single set of tables
for a specific locality known to have been compiled by a medieval astronomer. Abu l-{Uql
was well named as al-sib, the calculator, although this actually meant simply the
astronomer in some circles.
The major astronomical works of the Rasulid Sultan al-Ashraf (reg. 1295-96),6 compiled whilst
he was still a prince, were his treatise on the construction of astrolabe and sundials, as well
as the qibla-compass,7 extant in the autograph MSS Cairo TR 105 (149 fols.) and a late copy
5 On the houses see n. I-3:2.
6 On al-Ashraf (I-3.3.1, etc.) see Suter, MAA, no. 394, and King, Astronomy in Yemen, no. 8.
7 See now Schmidl, Early Arabic Sources on the Compass.
yemeni and hejazi tables for timekeeping 419
thereof in a library in Tehran (unidentified), and his astrological work extant in MS Oxford
Bodleian Hunt. 233, copied in the 14th century (?). See I-3.3.1, 4.1.4, 4.2.5, and 5.4.6, on various
rather unusual tables contained in these.
In the former treatise the qibla for Taiz, Zabid, and central Yemen is said to be 27-29
W. of N. (p. 145 of the Cairo copy). The modern value for Taiz is 26;47! This good news
is somewhat vitiated by the fact that the diagram of the qibla-compass in the same manuscript
shows the qibla of Taiz at 20 West of North and the qibla at Aden at 20 East of North: see
Fig. X-9.2.2.
The anonymous Yemeni Zj preserved in MS Paris BNF ar. 2523 (late 14th century), which
was compiled in Taiz in 775 H [= 1374], contains a table of T(H,h) (fols. 97v-100v). Entries
are computed to two digits for the domains:
H = 46, 47, ... , 90, and h = 1, 2, ... , H .
In the same work there is a table (fol. 86r) of the function h(T,H) computed for the domains
T = 1, 2, ... , 11sdh and H = 51, 52, ... , 90. Entries are given to two digits. Note that both
tables are adequate for use in the Yemen, where the minimum value of the solar altitude is
about 50. On these tables see also I-2.5.3 and I-4.3.3.
MS Cairo TR 274 (152 pp.) is an anonymous Yemeni almanac for the year 808 H [= 1405/
8 On al-Afal and his astronomical miscellany (I-2.1.2) see King, Astronomy in Yemen, no. 18.
9 See Varisco & Smith, eds., al-Afals Anthology, and also my review in Yemeni Update (www.aiys.org/
webdate/kngrev.html), which includes of a survey of research over the past 20 years on Yemeni astronomy.
10 This has been studied in Varisco, Yemeni Almanac II. The entire text of the almanac is reproduced in
Varisco & Smith, eds., al-Afals Anthology, pp. 97-109.
420 part ii, chapter twelve
Fig. 12.4: An extract for the beginning of the sign of Libra of the almanac in al-Afals collection of treatises.
[From a manuscript in a private collection in Sanaa, photo courtesy of Professor Daniel Varisco.]
Figs. 12.5a-b: An extract from the ephemeris (a) and the associated astrological consequences (b) for the first
few days of Raman, 808 H. [From MS Cairo TR 274, courtesy of the Egyptian National Library.]
MS Berlin Ahlwardt 5720 of the corpus of tables for Taiz (12.1) contains some additional tables
of a less sophisticated type for the latitude of Sanaa (fols. 121v, 124r-127r). These are all for
timekeeping by the stars, and I suspect that they were originally part of a larger corpus of such
tables for Sanaa. The functions tabulated are: (a) the longitude of the horoscopus as a function
of the altitude of 16 stars; (b) the longitude of the horoscopus as a function of each hour of
visibility of 18 stars; and (c) the altitudes of 18 stars at each hour of visibility. The hours of
visibility represent twelfths of the arc of visibility for each star and hence differ for each star.
In the heading of the first table it is stated that the stellar coordinates were based on the
observations of Ibn Ynus, but the name of the compiler and the date for which he computed
his stellar coordinates have been deliberately erased. However, it is just possible to read the
name {Abdallh ibn al D{ir,14 which also appears on fol. 121r in reference to some
instructions on the use of the third table. On these tables and their author see further I-3.2.3,
and also I-3.4.1 and I-4.6.1.
MSS Berlin Ahlwardt 5769 (Mq. 733,2), fols. 8v-15v, copied in Mukhkha in 1209 [= 1795],
Vatican ar. 962, fols. 13r-19r, and Sanaa GML majm{ 27,1, contain a set of prayer-tables
compiled for the latitude of the Yemen by Muammad ibn {Abd al-Laf al-Thbit. MS Algiers
Fagnan 1485,3, which I have not examined, contains the same tables.15 al-Thbit was a Syrian
who lived in Zabid, and he compiled these tables in 1047 H [= 1637/38].16
For each day of the Syrian year al-Thbit tabulates the solar longitude and which 13th part
of the appropriate lunar mansions is:
(a) culminating at sunset,
(b) rising at daybreak,
(c) culminating at midnight, and
(d) culminating at daybreak.
Entries are given as M1, M2, ... , M13, where M is the name of a given mansion. A rearrangement
of al-Thbits column of entries for the lunar mansions rising at daybreak is found in an
almanac called isb al-Shibm, which was still in use in the aramawt in the mid 20th
century.17
14 On Ibn D{ir (I-3.2.3, etc.) see King, Astronomy in Yemen, no. 31, and now Ihsanolu et al., Ottoman
Geographical Literature, I, pp. 76-78, no. 39.
15 Algiers BN Catalogue, p. 409.
16 On al-Thbit see King, Astronomy in Yemen, no. 33.
17 This table is published on p. 435 of Serjeant, S.W. Arabian Almanac, where it is implied that it displays
the lunar mansions corresponding to the position of the sun. See King, Astronomy in Yemen, no. 44.
yemeni and hejazi tables for timekeeping 423
al-Thbit further tabulates the hours of day and night, giving entries to two digits, where
the second represents thirtieths of an hour, as well as the shadows to base 61/2 corresponding
to the solar altitude at midday and at the beginning of the afternoon prayer, giving entries to
two digits, where the second now represents twelfths of a shadow digit. The midday shadows
are computed according to a linear zigzag scheme,18 with a maximum of 55/12 at the winter
solstice and a minimum of 22/12 (south) at the summer solstice. Note that these extremal values
are equivalent to 10 and 4 digits for a gnomon of 12 digits, which suggests that al-Thbit
converted these from base 12 to base 61/2. In fact, for = 15 the extremal values are closer
to 10 and 2 digits, al-Thbits afternoon shadow lengths are always 22/12 feet more than (the
absolute value of) his midday shadow lengths. See further III-4.7.
MS Cairo Sh 89,11-15 (fols. 157r-322v), copied 1070 H [= 1659/60], is the only known copy
of some scientific treatises by usayn ibn Shm al-Hattr, otherwise known only for his
writings on religious topics.19 Some of the former, which deal with the standard topics of zjes,
the astrolabe, and trigonometric and astrolabic quadrants, as well as arithmetic, are here copied
in al-Hattrs own hand. Scattered throughout all of his treatises on astronomy are some
spherical astronomical tables computed for = 25;0 (Medina).
Some of the tables are based on = 23;35 and others are based on = 23;30(,17?), and
I have not systematically investigated their accuracy. The functions tabulated include:
, 2, , , , , H, d, D, N, ta, Ta, s, r, n and ,
and al-Hattr also presents some simple trigonometric tables, and some tables for trepidation
(fols. 203r and 218v-220r).20 An extract from the prayer-tables is shown in Fig. 12.8. MS
Fig. 12.8: This extract from al-Hattrs prayer-tables for Medina shows the sub-tables for the functions N()
and n(), that is, the length of night from sunset to sunrise and the length of darkness of night from nightfall
to daybreak. [From MS Cairo Sh 89, fols. 295v-296r, courtesy of the Egyptian National Library.]
18 Such a crude procedure is virtually unknown in medieval Islamic astronomy. For one example see the study
by Kennedy and Ukashah cited in n. 1:9.
19 On al-Hattr see Brockelmann, GAL, II, p. 523, and SII, pp. 543 and 1039; Cairo ENL Survey, no. D42;
and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 388-389, no. 252.
20 See n. I-6:22.
424 part ii, chapter twelve
Leiden Or. 2538 is another copy of one of al-Hattrs treatises entitled al-Sirj al-wahhj
which also contains some of the tables listed above.
MS Cairo K 4002, 40 fols., penned in 1249 H [= 1833/34], is the only known copy of a set
of tables for timekeeping compiled for the latitude of Mecca, taken as 21;45, by the Ottoman
astronomer usayn usn ibn Amad ab.21 The main set of tables displays the time of day
according to the Ottoman convention as a function of solar longitude, with entries for solar
altitudes in the east and west given in hours, minutes and seconds. See further I-2.2.4
(illustrated). These are followed (fols. 35v-38r) by a set of prayer-tables for the same latitude.
For each degree of solar longitude the following functions are displayed with entries to minutes:
2D, 2N, m, a, b, s and r
I have not investigated the parameters used for twilight.
for each degree of solar longitude. The first three of these five tables are computed to two
digits, but the remaining two are computed to four digits ( = 23;35) and were clearly lifted
from some earlier source.
I have not investigated parameters used for morning and evening twilight. The entries for
z (uhr) are in fact simply (12-Dh). The entries for m (nif al-nahr, midday) have been made
earlier than z by a few minutes (about 8m at the equinoxes), and the entries for R, sunrise,
have been made earlier by double this correction (about 18m at the equinoxes). These
corrections represent a generous allowance for horizon phenomena (see also 14.15). The entries
for all functions but R and m correspond quite closely to recomputed values.
The work Kanz al-thiqt f {ilm al-mqt, The Treasure of Reliable Scholars for the Science
of Timekeeping, prepared by Shaykh {Abd al-Wsi{ ibn Yay al-Wsi{23 for the Imm Yay
was published in Cairo (Maba{at ijz) in 1358 H [= 1939/40]. It contains two sets of tables,
the first for finding the feria of a date in the Muslim calendar (from 1358 H to 1596 H) as
well as the corresponding day number of the solar year, and the second for reckoning the prayer
times, apparently based on = 15 (Sanaa) although the actual latitude of Sanaa is 15;23.
In this second set of tables the following information is given for each day of the solar year:
manzilat al-shams, the 13th division of the lunar mansion of the sun
li{ awwal al-layl, the ascendant at the beginning of night (in signs and degrees)
ma{lim al-zir{a, information about the rising or setting of a prominent star or
constellation
daraj al-shams fi l-burj, the solar longitude (to the nearest degree)
al-m min al-shuhr al-rmiyya, the date in the Syrian calendar
aqdm al-uhr / al-{ar wa-bannuhu wa-s{atuhu wa-daq}iquhu, the shadow length in
feet and fingers at the midday / afternoon prayer and the corresponding times in
hours and minutes
ul{ al-fajr, daybreak
shurq al-shams, sunrise.
The times of daybreak, sunrise, midday and the afternoon prayer are given in hours and minutes
according to the Ottoman convention. Daybreak is made to precede sunrise by 1h10m throughout
the year, which corresponds to a parameter of about 17 at the equinox. The shadow lengths
are to base 7 and the fractional twelfths are called bann, fingers, as in al-Thbits prayer-
tables (12.7). The shadow lengths are computed according to a very primitive scheme. al-
Wsi{s shadow has the constant value 4/12 feet from Taurus 8 to Cancer 18. Since the midday
shadow in the Yemen actually becomes southerly in the summer, al-Wsi{s values represent
a gross distortion. The equinoctial and solstitial values are given as:
VE: 2 SS: 4/12 AE: 14/12 WS: 5 ,
23 On al-Wsi{ see King, Astronomy in Yemen, no. 47. A more detailed investigation of his shadow-schemes
is in III-4.10.
426 part ii, chapter twelve
CHAPTER 13
I know of no tables relating to astronomical timekeeping from al-Andalus, other than the simple
shadow-schemes of the kind found in the 10th-century Calendar of Cordova, which I have
discussed elsewhere (III-5.1). We can, however, be sure that some tables for mqt were
compiled there, not least because the earliest European tables of this kind, some from medieval
Chrisitan Spain, show some Islamic influence (I-10.1). It is therefore to the Maghrib that we
turn next, a region whose distinctive astronomical tradition was influenced successively over
the centuries by Baghdad, then by al-Andalus and then by Egypt and Syria, whence came also
the Zj of Ulugh Beg.
Only recently has the history of astronomy in the Maghrib been surveyed.1 According to
the late 14th-century Tunisian historian Ibn Khaldn,2 the zj that was commonly used in his
time (in Tunis or in Morocco or both?) was that of Ibn Isq, who worked in Tunis in the
early 13th century. A manuscript of this zj was discovered only in 1978 and has been subjected
to detailed study in Barcelona.3 An abridgement of it was made by Ibn al-Bann} of Marrakesh
ca. 1300, and this survives in a number of copies.4 Four 14th-century Maghribi zjes have
survived, namely three zjes by Ibn al-Raqqm, compiled in Tunis,5 and a small zj by al-
Qusann, compiled in Fez.6 This last is of particular interest because it contains planetary
equation tables based on Indian rather than Ptolemaic theory. The later Tunisian zjes of usayn
Qu{a and Muammad Sanjaqdr are based on the 15th-century Zj of Ulugh Beg of Samarqand.7
Also, a recension of the 14th-century zj of Ibn al-Shir was prepared for Algiers.8 The three
Tunisian sets of tables for timekeeping described here (13.2-4) were probably also compiled
in the 14th century, and reflect the influence of the Syrian tradition in astronomical timekeeping.
Certain later Maghribi sources are full of surprises (13.5-8).
In the past hundred or so years Getan Delphin, Henri-Paul-Joseph Renaud, and Georges
Colin conducted a series of studies on astronomy in medieval times in what is now Morocco.9
1 See King, Astronomy in the Maghrib, already to some extent superseded by various works of the Barcelona
school, such as Sams, Maghribi zjes, and idem, Astronomical Observations in the Maghrib. See also King
& Sams, Islamic Astronomical Handbooks and Tables, pp. 60-64.
2 Franz Rosenthal, Ibn Khaldn: The Muqaddimah An Introduction to History, 2nd edn., 3 vols., Princeton,
N.J.: Princeton University Press, 1967, III, pp. 133-137.
3 On Ibn Isq see n. I-6.17, and especially the new studies of ngel Mestres there cited.
4 On Ibn al-Bann} see n. I-6:18.
5 On Ibn Raqqm see n. I-6:19.
6 On al-Qusann (Suter, MAA, no. 371) and his tables see Kennedy & King, Indian Astronomy in
Fourteenth-Century Fez.
7 See n. I-4:20.
8 Cairo ENL Survey, nos C30/2.1.19 and F66.
9 Cf. Delphin, Lastronomie au Maroc, and Colin & Renaud, Ab Miqra{.
428 part ii, chapter thirteen
Their investigations suggested that the most celebrated muwaqqits there such as the 13th-century
scholar Ab Miqra{ (2.0) were innocent of the developments in astronomical timekeeping
further East (which the French scholars did not know about anyway). However, the numerous
sets of prayer-tables mentioned below, mostly unknown to modern scholarship before the
1970s, show that there was more activity in mqt in Morocco than was previously thought
and that some of it was rather sophisticated.
The mqt specialists of the Maghrib were involved with three times of day, two of which
were not generally used by their colleagues further East. These three times are graphically
displayed on an elegant 14th-century sundial from Tunis.10 The first of these was the u
prayer, beginning around mid-morning, or, more precisely, at a time before midday equal to
the time after midday of the beginning of the time for the {ar.11 The u was also considered
important in Ottoman Turkey (Ch. 14). The second was the ta}hb, a preparation for the
communal prayers on Friday, which took place one equinoctial hour or 15 equinoctial degrees
before midday.12 The third time is the beginning of the uhr, which was taken not as
immediately after midday as it was in the East but rather as the moment when the shadow
of any object z has increased by one-quarter of its length over its midday minimum Z. This
definition of the uhr is attested in some early Eastern Islamic sources (see, e.g., 2.4), but
remained standard in Maghribi and Andalus practice.13 The corresponding solar altitude hz
is defined by
hz = arc Cotn { Z(n) + 1/4 n } .
In this chapter I use the subscripts , t and z to refer to the u, ta}hb and uhr.
13.1 The spherical astronomical tables for Tunis in the Zj of Ibn Isq
MS Hyderabad @afiyya 298, copied ca. 1400, is a unique copy of a voluminous (ca. 400 pp.)
Syrian recension of the Zj of Ibn Isq.14 During my brief encounter with this manuscript in
Hyderabad in 1978, I was able to record the existence of some auxiliary tables for calculations
in timekeeping and computing ascensions also found in earlier works of Ibn al-Zarqlluh and
the later works of Ibn Isq, Ibn al-Bann} and Ibn al-Raqqm: see further I-6.9.1* and 7.1.5*.
The only other tables specifically relating to timekeeping in this manuscript are tables of the
functions:
2Dh, , H and *,
with entries in degrees and minutes for each degree of based on parameters:
= 36;40 (Tunis) and = 23;33.
Two tables whose titles indicate that they were to display r() and s() in hours and minutes
have no entries (see further XI-5.3). Another table, unique of its kind, displays the half-excess
10 See King, Tunisian Sundial, for an analysis of this instrument and (pp. 193-196) a first interpretation
of the institutions represented by the markings (including the u, ta}hb, uhr and {ar), accomplished with
the help of contemporaneous Tunis tables for timekeeping. More information is in IV-4-6.
11 On the u see ibid., pp. 189 and 193-194, and IV-5.
12 On the ta}hb see ibid., p. 190, and IV-6.2.
13 Cf. Wiedemann & Frank, Gebetszeiten. On the origin of this definition see IV-4.4.
14 On Ibn Isq see n. I-6:17.
maghribi tables for timekeeping 429
of the arc of visibility of the moon for each of the seven climates and for each degree of its
declination from 1 to 30 (see further Fig. VIa-7.2).
Whilst this manuscript may be somewhat disappointing for the history of astronomical
timekeeping it provides exciting new material for the history of most other aspects of
mathematical astronomy and astrology.
MS Berlin Ahlwardt 5724 (Wetzstein 1150), fols. 1r-41r, copied ca. 1700, contains an
introduction and a set of tables for timekeeping computed for the latitude of Tunis. The work
is anonymous but is entitled {Umdat al-nur f mawqt al-layl wa-l-nahr, roughly The
Support of those who look closely of the Times of Night and Day. The author dedicated his
work to the afid ruler Ab Fris (al-Mutawakkil) {Abd al-{Azz ibn Abi l-{Abbs Amad
(reg. 1394-1434).
The first tables (fols. 11r-13r) are for calendar conversion and for finding the solar mean
longitudes (from 600 H to 900 H) and equations. These are followed (fols. 13v-14r) by Sine
and Cotangent (base 12) tables, and (fols. 14v-15v) by tables of:
, H and D
for parameters:
= 37 (Tunis) and = 23;35 ,
which also underlie the remaining tables. There follow (fols. 16r-28v) tables displaying the
hour-angle t(H,h) for each degree of both arguments (h H) (I-2.3.5, illustrated). A full set
of prayer-tables follows (fols. 29r-41r), in which the following functions are displayed:
h0, , Sin , , Dh, 2Dh, ht, hz, ha, (D+s), (D+2N-r), and ,
as well as k(h) and sec h for azimuth calculations (I-8.1.3 and 8.4.1). There is also a separate
table (fols. 3v-38r see Fig. 13.2) with entries to one digit for the functions:
D, (D+s), (D+2N-r) and (D-r) .
Those tables with entries for 0 begin with Capricorn 1. The tables for the solar altitude
at the time of the uhr and ta}hb are based on the definitions noted in 13.0. The tables for
twilight are based on parameters 20 and 18 for morning and evening, respectively. The entries
in all of the tables are rather accurately computed to two digits. The remainder of the tables
in the Berlin manuscript are based on latitude 36;40 for Tunis and are the same as those in
the Zj of usayn Qu{a (see the next section).
A second set of prayer-tables for the latitude of Tunis, based on parameter 36;40, is attested
in a number of sources. The prayer-tables for latitude 37;0 in MS Berlin Ahlwardt 5724 (13.2)
are followed (fols. 42r-53r) by this second set, as well as a star catalogue for the year 773
H [= 1371/72] attributed to Ibn al-Shar (fols. 53v-54v), and a table for reckoning time from
stellar altitudes (fols. 55r-56r), on which see I-2.4.1. The prayer-tables for latitude 36;40 also
occur in MS Princeton Yahuda 147c, pp. 113-132, of the later Zj of usayn Qu{a, based
430 part ii, chapter thirteen
Fig. 13.2: An extract from the prayer-tables in the main Fig. 13.3: An extract from anonymous prayer-tables for Tunis
Tunis corpus. [From MS Berlin Ahlwardt 5724, fol. 36r, serving Aquarius / Leo. [From MS Berlin Ahlwardt 5724, fol.
courtesy of the Deutsche Staatsbibliothek (Preuischer 56v. [Courtesy of the Deutsche Staatsbibliothek (Preuischer
Kulturbesitz).] Kulturbesitz).]
mainly on the 15th-century Zj of Ulugh Beg, and the prayer-tables are likewise not original.
The tables for latitude 36;40 include some for the ta}hb and uhr like those for latitude
37;0. The functions tz and za denote the time from midday to the uhr and from the uhr
to the {ar. Altogether these prayer-tables display the following functions, reasonably accu-
rately computed to two digits for each degree of solar longitude:
(Berlin and Princeton): ht, H, Sin H, Z(12), zz(12), za(12), hz, tz, za,
ha, ta, Ta, s, (n-45), n, (2N-r), r
(Princeton only): D, , Dh, (D+2n-r), and (D+s) .
The entries begin with those for either Capricorn 1 or Cancer 1. An extract from part of the
tables in the Berlin manuscript is shown in Fig. 13.3. The function (2N-r), representing the
time from sunset to daybreak, is referred to as qaws al-layl al-shar{, length of night as defined
by religious law, an expression also used in the tables for latitude 37;0. The function
(n-45) is referred to as m bayn al-shafaq wa-btid} al-dhn, the time between nightfall
and the beginning of the prayer-call. This means that the muezzin would perform a prayer-
call three hours before daybreak. Probably this particular table was intended for use only in
the festive nights of Raman, to announce the time for the sur. A similar table for Damascus,
allowing more time for sleep, is contained in MS Berlin Ahlwardt 5758, fol. 49v, of the
Damascus corpus (10.2).
A manuscript in the Topkap Library, whose number is not available to me, is another
anonymous copy of some of these tables without introduction or title. The following functions
are represented:
ha, ta, Ta, Z(12), zz(12), za(12), , Sin , s, (n-45), n, (2N-r) and r .
It is noted that the tables of and Sin are based on = 37 rather than 36;40.
maghribi tables for timekeeping 431
MS Cairo DM 689 contains an extensive set of auxiliary tables copied in an elegant Maghribi
hand ca. 1600.15 The tables conclude with a star catalogue dated 801 H [= 1398], and they
appear to have been compiled in Tunis. The title folio and instructions are missing from the
manuscript, which begins with the last page of a set of tables displaying the solar longitude
for each day of a period of four Syrian years (Fig. I-1.2d). The auxiliary tables are simply
those of al-Khall for timekeeping by the sun (10.3), with a few additional tables of for
the latitude of Mecca, taken as 21;40, and various unspecified localities in the Maghrib, al-
Andalus and Sicily (?) with latitudes:
30;30, 31;30, 32;30, 33;40, 34;30, 36;30, 36;40, 37;10, 37;30, 38;30 and 39;30
and also tables of Sin H() for these latitudes, from which those of B() are easily derived.
See further I-6.2.5, I-6.3.3, and I-9.7 (illustrated).
MS Cairo TR 338,2 (pp. 204-287), penned ca. 1850, is the only known copy of a set of prayer-
tables for various latitudes in what is now Morocco.16 The tables are preceded by a short
introduction in which the author is named as Ab {Abdallh Muammad ibn Muammad ...
al-Jannd al-Anr al-Andalus, and the work is dedicated to the ruler (called imm and suln)
Nir al-Dn Ab {Abdallh Muammad ibn {Abdallh ibn Ism{l, surely the {Alaw sharf
Muammad III ibn {Abdallh ibn Ism{l who ruled Morocco during 1171-1204 H = 1757-
1790].17
al-Jannds tables begin with a table of solar longitude for each day of the Christian year.
According to this table the sun enters Aries 0 on March 8-9, and in the title of this table and
of all those which follow it is stated that the operations are all li-arakat y, which shows that
al-Jannd accepted that the equinox had moved 18 from the sidereally-fixed Aries 0, as a
result of the (false) notion known as the trepidation of the equinoxes.18 A table of geographical
coordinates promised in the introduction is no longer contained in the manuscript. The
remaining tables consist of three sets for latitudes:
34 (Meknes, Zarhun, Fez, and Sakka), 31 (Marrakesh, Kairouan (!), Alexandria, etc.),
and 30 (Sijilmasa, Cairo) ,
with underlying obliquity = 23;30. An extract is shown in Fig. 13.5. The functions displayed
are in each case:
, H, ha, Sin ha, C, B, d, tah, 2Dh, 2Nh, Dh, (D+s)h, (N-r)h and Nh .
Values of the first seven functions are given to two sexagesimal digits, and all times are given
in hours, degrees and sexagesimal minutes thereof, except the lengths of day and night, which
are tabulated to one extra digit.
Fig. 13.5: An extract from al-Jannds prayer-tables for the Maghrib. [From MS Cairo TR 338,2, pp. 258-259,
courtesy of the Egyptian National Library.]
One of the numerous commentaries on al-Jdars poem on timekeeping Rawat al-azhr (2.0,
also I-6.15.2)19 is an anonymous work entitled Nat}ij al-afkr. This commentary is of interest
for the historical information it contains about the medieval Islamic tradition of obliquity
measurements, twilight parameters, and theories of trepidation.20 It is extant in two copies, MSS
London BL Or. 411,2, and Cairo K 4311, dated 1183 H [= 1769/70]. The first of these contains
a set of tables for timekeeping scattered in the text of the commentary; the second contains
no tables. All of the tables based on local latitude are computed for:
= 35;0 (Tlemcen) .
The tables include some for the solar mean motion and the solar equation, simple trigonometric
functions, and a star catalogue. The tables relating specifically to timekeeping display the
functions:
H, , , D, Dh, hz, ha ta and Ta ,
with values given in degrees and minutes for each degree of solar longitude. There are also
tables of various auxiliary functions for use in determinations of the solar azimuth and hour-
angle see I-6.15.2 and I-8.1.4.
MS Algiers Fagnan 1472 contains six pages of prayer-tables for the latitude of Algiers. I have
not been able to inspect the manuscript, but according to the catalogue,21 the prayer-times are
arranged in eight columns, values are given for each day of the year, and the manuscript itself
dates from the end of the 18th century.
MS Cairo FT 9,1, penned ca. 1900, is the only copy known to me of an extensive
anonymous almanac and set of prayer-tables for the latitude of Algiers, stated to be 36;40.22
The introduction to the work is in Turkish, and the author pays his respects to Shaykh Vef
(14.3) and {Al Efend (the latter presumably made some alterations to Shaykh Vefs
almanac),23 and then proceeds to mention Ulugh Beg, {Al al-Maghrib (perhaps Ab {Al al-
Marrkush of 6.7 is intended), Ab Miqr{ (2.0) and Ibn al-Bann} (13.0). His almanac
contains extensive calendrical tables, and a table for the solar declination in a four-year cycle,
as well as three sets of tables for timekeeping. The first of these displays for each day of each
month in the Syrian calendar the duration of day (al-nahr) and night (al-layl). The second
displays likewise see Fig. 13.7 the following ten functions:
2D, 2N, (2D+r), (2N-r), r, D, tz, ta, Ta, s .
The third and fourth functions are labelled al-nahr al-shar{ and al-layl al-shar{, as opposed
to the first and second, labelled ... al-{urf. In both of these sets values are given in hours and
minutes. The third set of tables see Fig. IV-5.5 displays the time in hours and minutes,
according to the Turkish convention, of the three institutions, the anjq, the {ar, and the imsk.
These tables are not based directly on those in the second set. The term anjq, which here
refers to a time shortly before midday, is new to me. These tables merit more detailed
investigation: see IV-5.4.
Fig. 13.7: An extract from the anonymous prayer-tables for Algiers. See also Fig. IV-5.5. [From MS Cairo FT
9,1, fols. 34v-35r, courtesy of the Egyptian National Library.]
with entries to four digits for each degree of , and based on the parameters:
= 34;48 (Sfax) and = 23;30,17 .
As this book was going to press I came across some photocopies of photos from a Maghribi
manuscript labelled Ms. 7, v. 3. Zijes. Manuscript, with the pages numbered 446 to 474.
The original document dates from about 1800. I have no recollection of the source of these
documents. There are some interesting tables for timekeeping, which I can only briefly list
here: see Figs. 13.8*a-d.
These tables may have been taken from a larger set; certainly, there is no indication of a
compiler or a date. The tables are stated to serve latitude 34;10, which is not an attested value
for any location in the Maghrib, and may well be the result of a new set of observations for
Fez, whose latitude is accurately 3405, but was traditionally taken as 33 or 33;40, so that
was certainly need for a new measurement.25 The underlying obliquity is 23;29, which may
also have been derived by observation, although it is an attested Ottoman (and European) value.
See I-6.4.20 and 6.10.17 for related tables.
The first group of tables displays the following functions side-by-side for each degree of
' (1 to 90):
, Sin , , Sin , d, Sin d,
arc Sin C, C, arc Sin B, and B,
Figs. 13.8*a-d: Four extracts from the remarkable anonymous prayer-tables for Fez that survive in an apparently
unique copy, whose location is a mystery to me. Possibly the compiler is al-Warjni see I-6.4.20. [From a
set of photocopies of a manuscript labelled Ms. 7, v. 3, Zijes, Manuscript, provenance unknown.]
436 part ii, chapter thirteen
with values to two digits. The functions arc Sin C and arc Sin B are called qaws bu{d al-qur
and qaws al-al, respectively; on their purpose see I-6.5.
The second group displays in a similar fashion, but now for from 270 to 90, the functions:
H, Cot12 H, Sin H, ha, Cot12 ha, and Sin ha .
The third set displays two sets of functions labelled ia and s{t (sometimes sawyi{),
also for from 270 to 90. The ia are for daybreak (al-fajr) and sunrise (al-shurq), both
measured from midnight, and for the afternoon prayer (al-{ar), sunset (al-ghurb), and
nightfall (al-{ish), all three measured from midnight, followed by the hour-angle at nightfall
(fal d}ir al-{ish). The s{t are for the same five times, now expressed in hours and minutes,
followed by the time between nightfall and midnight (m bayna l-{ish wa-nif al-layl). The
s{t are written in Maghribi Hindu-Arabic notation, whereas the ia are in abjad notation.
Apart from a single table showing the solar longitude to the nearest degree from the date
in the Western calendar, the remaining tables are of the same format, and are of a kind not
known from any other source. A series of functions:
ha (irtif{ al-ar), tah (iat al-{ar), (D+s)h (iat al-{ish),
(N+r)h (iat al-fajr), and dh (nif al-fala)
are tabulated for each half degree of solar declination (entered vertically) for each half degree
of latitude (entered horizontally) from 30 to 35;30, then 35;45. The parameters underlying
the entries for evening and morning twilight are stated to be 18 and 19, respectively. Again
the entries are in Maghribi Hindu-Arabic numerals. Since only the first of three pages for
the last function is present in the photocopy at my disposal it may be that the set originally
served yet other functions.
In the light of the existence of this late set of sophisticated tables for timekeeping from the
Maghrib, and no less the set described in 13.5, it is perhaps all the more remarkable that H.-
J.-P. Renaud, who wrote the first serious historical essays on Maghribi astronomy (see 13.0),
was not aware of the existence of any tables of this kind.
A motley collection of corrupt approximate rules for the prayer-times is contained in a late
Maghribi source that I have consulted, but forgot where. The rules outlined are as follows:
ta = 50 + 1/4 d or (45 + 1/4 - 1) + 1/4 d (where = 24!)
or 51 + 1/4 d or 4/7 [90 - 1/2 d]
r = s + 3 .
The anonymous author also explains how to find the time of the u using the fact that
t = t a.
turkish tables for timekeeping 437
CHAPTER 14
The only astronomical activity amongst the Muslim community in Anatolia before the rise of
the Ottomans is restricted to some legendary accounts of observations in specific madrasas
of Kirshehir and Ktayha documented by Aydn Sayl,1 and the sojourn of the well-known
astronomer Qub al-Dn al-Shrz, who compiled some of his astronomical works in Konya,
Sivas and Malatya in the middle of the second half of the 13th century.2 If my assumption that
the two sets of tables discussed below (14.1-2) are of Seljuk Anatolian provenance is correct,
then clearly further research in this field would be worthwhile.
The later Ottoman tradition in astronomy has until recently been generally neglected by
historians, despite the vast amount of source material. The popular Ruznme of Shaykh Vef
(14.3) has never been properly studied, in contrast with the less widely-used Rznme (almanac)
of Darendel (14.4), on which three descriptive studies were published in Europe between 1676
and l804.3 Some brief notes on Ottoman astronomical works were published by J.-B. Toderini
in 1789,4 and an eloquent study by Adnan Advar, La science chez les Turcs Ottomans, alas
devoid of any serious scientific content, was published in Paris in 1939. Otherwise, apart from
a series of articles on specific topics that were published over many decades by Aydn Sayl
and his former student, Sevim Tekeli,5 the notes which follow represented in the 1970s the
first serious attempt to investigate the activities of the astronomers of Ottoman Turkey.6 Now,
however, we have a monumental survey of the available literature prepared by Ekmeleddin
~hsanolu and his colleagues in Istanbul, which will serve as a starting-point for all future
research.7
In the sequel (14.3-14), I describe several Ottoman sets of tables for timekeeping in general
and prayer-tables in particular. These are of two kinds: in the first, time intervals are expressed
in equatorial degrees and/or equinoctial hours, and in the second the times of day are expressed
in equinoctial hours according to the convention that sunset is 12 oclock. This latter
convention, which in the 1970s was still used in the Levant, and more especially in Arabia,
1 Sayl, The Observatory in Islam, pp. 253-255.
2 Ibid., pp. 215-217, etc., on his association with the Observatory at Maragha. See also the article by Seyyed
Hossein Nasr in DSB. MS Paris BNF ar. 2516 of his Tufa shhiyya on mathematical cosmology is in his own
hand and dated Sivas, 684 H [= 1285/86]: see Cairo ENL Survey, no. G27.
3 See nn. 14:20-21.
4 Jean-Baptiste Toderini, De la littrature des Turcs, traduit de lItalien en Franais par Cournant, 3 vols.,
Paris, 1789, I, pp. 147-148 and 406.
5 A list of Sayls publications is in Sayl Memorial Volumes, I, pp. 22-29. Many of Tekelis publications
are listed in ~hsanolu et al., Ottoman Astronomical Literature, II, pp. 956-957.
6 See King, Astronomical Timekeeping in Ottoman Turkey.
7 ~hsanolu et al., Ottoman Astronomical Literature, and also Ottoman Mathematical Literature.
438 part ii, chapter fourteen
including the Yemen, seems to be of Ottoman Turkish origin, although, of course, the Islamic
day has always been considered to begin at sunset.8 It is the reckoning of time from sunset
in equinoctial hours rather than seasonal hours that characterizes Ottoman practice. An obvious
disadvantage of this convention in modern usage is that clocks must (in theory) be reset every
evening at sunset.9 In the sequel I use prime notation to indicate those times which are expressed
in this convention (see also 1.3 above).
The most popular prayer-tables prepared for Istanbul were the rznmes of Shaykh Vef
and Darendel, already mentioned above. I show (14.3 and 14.4) that they are based on different
values for the latitude of Istanbul, and that the entries are expressed in different units. Both
of these almanacs also contain calendrical tables, information on religious festivals, the seasons,
and astrology, none of which con-cerns the present study. More sophisticated prayer-tables
were prepared for Istanbul and Edirne, probably in the 16th and 17th centuries (14.7, 14.8, and
14.11). Particularly impressive are the contributions to our subject by Muammad ibn Ktib
Sinn and Taqi l-Dn (14.5 and 14.9). The extensive corpus of tables for timekeeping by the
sun compiled by li Efend (14.12) in the late 18th century, by which time Ottoman
astronomers had come into contact with European astrology and had translated the Zjes of
Cassini and Lalande into Turkish and Arabic and adapted the tables to their own meridians.
Various Ottoman-type prayer-tables for other localities in the Ottoman Empire, such as Cairo,
Aleppo, Sanaa, Algiers, Mecca, Crete and Yarqand, have been located in the manuscript sources
(see 7.11, 11.11, 12.11, 13.7, 12.10, 14.14 and 3.17, respectively). Doubtless other Ottoman
sets of tables of this type for yet other localities are available amongst the several thousand
mainly-uncatalogued astronomical manuscripts preserved in Turkey.
8 See Charles Pellats EI2 article Layl and Nahr, esp. p. 708b.
9 See Kurz, European Clocks, pp. 83-84.
turkish tables for timekeeping 439
on an approximate formula different from the standard approximate one (see 1.4 and, for
example, 3.3), and they correspond most closely to values recomputed with parameter 20. I
have not been able to determine the underlying formula. Whereas the table has values:
26;30 25;39 32;56
for arguments = 271, 1, and 90, recomputation with the accurate formula and parameter
20 gives:
27;46 25;57 35;11
and recomputation with the standard approximate formula and parameter 20 gives:
25;24 25;59 36;30 .
This anomalous set of entries of course affects the values of n (=2n-r-s). Otherwise the tables
are rather accurately computed.
The Rznme-yi Vef} exists in dozens of manuscript sources, of which I have examined MSS
Istanbul Hamidiye 842, Istanbul Nuruosmaniye 2914,1, Paris BNF turque 186, 187, 188, 194,
turque supp. 537, Vienna 1426 and 1427, and several copies in Cairo, including Cairo KhMT
3: see Fig. 14.3a-b. See The identity of the compiler of this almanac has not previously been
established with any certainty: the work has previously been ascribed to Shaykh Vef, a
celebrated saint who lived in the time of Muammad II and Byezd II and died in the year
1491,12 and to Shaykh Vef} Muammad, who wrote a history of Murd III ca. 1585. A new
source, MS Cairo K 4037, copied ca. 1700,13 contains an anonymous commentary on the
Rznme and confirms the former attribution, providing the following biographical information
(fols. 1v-2r):
Actually his name is Muaf ibn Amad ibn al-rawi (?) al-{saw, known as Vef.
This is as found in his handwriting on some of his copies (of the Ruznme), according
to the statement of the author of the (book entitled) al-Sh[aq]}iq.14 [Note: the rest
of these notes are probably taken from that work.] Vef was the name of his father,
but (the son) was also known by (this name). He studied mysticism under Shaykh {Abd
al-Laf al-Quds and with him completed the stages of the order.15 (al-Quds) author-
ized him (to give) instruction (himself). (Shaykh Vef) may God Almighty have
mercy on him was a anaf, versed in all of the phenomenological and esoteric
sciences and also in the science of timekeeping. He had noble behaviour because of
his blessing. He also had a complete knowledge of music and rhetoric in poetry and
prose. He was contemporary with the Suln Muammad al-Fti [reg. 1444-1446]
and he died in the time of the Suln Byezd (II) [reg. 1481-1512] in Constantinople.
12 See Cairo ENL Survey, no. H2; and ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 51-54, no.
19 (with references to many more manuscripts).
13 For the Arabic text see Cairo ENL Catalogue, II, pp. 234-235.
14 This is the biographical encyclopedia al-Shaq}iq al-nu{mniyya f {ulam} al-dawlat al-{Uthmniyya by
ashkprzde (d. 1561 see the EI2 article by F. Babinger updated by Christine Woodhead).
15 See the article Taawwuf in EI .
2
turkish tables for timekeeping
Fig. 14.3a-b: More extracts from another copy of the Rznme of Shaykh Vef: calendrical tables and mystical interpretations of the Ka{ba,
mentioning the rijl al-ghayb, intermediaries between God and Man (b), and the first two of twelve pages of tables for the times of prayer
(c). [From MS Cairo KhMT 3, fols. 1v-2r and 2v-3r, courtesy of the Egyptian National Library.]
441
442 part ii, chapter fourteen
Fig. 14.4b-c: The Oxford copy of Darendels prayer-tables, apparently dated 1795. [From MS Oxford MHS
84-33, courtesy of the Museum of the History of Science.]
(This was) after he had journeyed on the pilgrimage by sea. The Christians took him
prisoner and incarcerated him in the Citadel of Rhodes.16 He compiled these tables
in prison. The amr Ibrhm Bey ibn Qarmn17 bought him from (the Christians),
then he went home to Constantinople and died there. He has a prayer room and mosque
which are known, and his tomb is in front of the mosque. It is famous and is visited
(by the people) to obtain blessing from him.18 The date of his death is the year eight
hundred and ninety six of the Hijra of the Prophet [= 1490/91] ... .
16 On the history of Rhodes under the Ottomans see the EI2 article Rodos by S. Soucek.
17 Perhaps the grandson of the better known Ibrhm Beg (d. 1464): see the EI2 article ^armn-Oghullari
(^armnids) by F. Smer, esp. pp. 619 and 622.
18 ~hsanolu, Ottoman Astronomical Literature, I, p. 51.
turkish tables for timekeeping 443
All copies of the Rznme contain calendrical tables. However, MSS Istanbul Hamidiye 842
and Nuruosmaniye 2914,1, Paris BNF turque supp. 537, and Vienna 1426 contain in addition
a set of prayer-tables. In MS Vienna 1426 there are twelve sub-tables, one for each month of
the Syrian year. In this particular Paris manuscript the tables for each month have been copied
on both sides of a single folio. The solar longitude corresponding to each day of the year is
given in the argument column in MS Vienna 1426, and in a separate table in the Paris
manuscript, which is described below. The functions tabulated are the following, and the values
are expressed in equinoctial hours (s{t) with fractions thereof in equatorial degrees (daraja):
2D nahr, length of daylight
2N layl, length of night
D uhr, time from sunrise to midday or midday to sunset
ta {ar, time from midday to the afternoon prayer
Ta maghrib, time from the afternoon prayer to sunset
s {ish}, duration of evening twilight.
The entries are written in Eastern Arabic numerical characters. Pretty designs (MS Vienna
1426) or the Arabic word khl meaning empty (MS Paris BNF turque supp. 537) are used
for units which are zero. Half degrees are indicated by the symbol . The parameters underlying
these prayer-tables are:
= 41;30 and 23;29 ,
with 19 and 17 for morning and evening twilight. In other Ottoman tables the latitude 41;30
is used for Edirne (14.7), but it seems certain that the Rznme-ye Vef} was intended for use
in Istanbul, where the latitude is in fact 41.19
In MS Paris BNF turque supp. 537 there is a separate table of two functions for regulating
the ifr and imsk in Raman, that is, the time of the evening meal after the days fast and
the beginning of fasting the next day. The functions are labelled iat-i ghurb and iat-
i ur (sic for sur) and values are given to one digit for each day of the Syrian year. The
solar longitude is also given for each day. In fact the main functions are simply Ta and r,
expressed in equatorial degrees.
MS Istanbul Esat Efendi 1973, fol. 148v, contains a table entitled jadwal mqt imskiyya,
which means table for the time of fasting in Raman. It is stated that the tamkn, the time
before daybreak when one should start fasting, is 12m. The entries in the table are given in
hours and minutes according to the Ottoman convention, for each degree of from Capricorn
1 to Gemini 30. The entries are symmetrically arranged, but both vertical arguments run from
1 to 30, whereas the upward arguments should run from 0 to 29. The entries are 10m (not
12m) less than the times for daybreak given in the Rznme-yi Vef}. The remaining tables
in this manuscript consist of planetary tables for Istanbul based on the parameters of Ulugh
Beg.
19 See n. 14:25.
444 part ii, chapter fourteen
The Rznme of Darendel has been published three times, by Velschius (1676), dOhsson
(1787), and Navoni (1804).20 From an examination of the material published by dOhsson and
Navoni, as well as of MSS Istanbul Air Efendi 470, Istanbul Kandilli 440, Oxford MHS 84-
33, and an unnumbered manuscript in the Yale Medical School Historical Library, New Haven,
Conn., it is clear that at least these copies are recensions of the same work. The relatively small
number of surviving manuscripts suggests that this rznme was less popular than that of
Shaykh Vefa (14.3). Navoni and dOhsson attributed the almanac to an individual named
Darendel (Arabic, al-Darandaw), as did Toderini, who stated in 1789 that the almanac had
been in use in Turkey for a century.21 Muammad ibn {Umar ibn {Uthmn Darendel22 is also
known as the author of a treatise on the astrolabic quadrant,23 and he died in 1739: his name
indicates that he or his family was associated with the town of Darende in central Anatolia.24
Judging by the number of extant copies his almanac was less popular than that of Shaykh Vef
(14.3).
The lithograph copy of Darendels almanac published by dOhsson and also the Yale and
Oxford manuscripts are in the form of scrolls, about 10 cm wide and 1 m long: see Fig. 14.4a-
b. The non-numerical material is written around the tables in Osmanli Turkish, and the entries
in the tables are written in Eastern Arabic numerals as in the Rznme-ye Vef}. An entry such
as 2512 (in Arabic numerals) indicates 25 minutes past 12 oclock. Zero is represented by a
dot in the number 10, but by three dots when it is one of the units of an entry. One-half
is represented by a tick or by /. Time is reckoned in equinoctial hours and minutes, according
to the Ottoman convention.
The prayer-tables of Darendel consist of six sub-tables for each pair of zodiacal signs: see
Fig. IV-5.4. The vertical argument in each generally runs from 1 to 30, but extra entries enable
the user to find the prayer times for each day of the solar year using a second argument column.
Nine functions are tabulated side by side in each sub-table. These are:
2D nahr, the length of daylight
2N layl, the length of night
M uhr, the time of midday
a {ar-i awwal, the earliest time for the first afternoon prayer
b {ar-i thn, the earliest time for the second afternoon prayer
s {ish}, the time of nightfall
r imsk, the time of daybreak
tq qibla, the time when the sun is in the azimuth of Mecca
awa, mid-morning
14.5 Muammad ibn Ktib Sinns tables for timekeeping by the stars
Muammad ibn Ktib Sinn was a muwaqqit in Istanbul in the late 15th century.28 One of his
works is an enormous table for timekeeping by the stars, preserved in MSS Istanbul Ayasofya
2710 and Istanbul Topkap T 3046 (Ahmet III 3515). The table occupies some 500 pages and
contains about 240,000 entries, and is arranged so that one can feed in both the normed right
ascension of a star which is culminating and the solar longitude, and read off:
(1) the time since sunset,
(2) the time remaining until sunrise,
(3) the time remaining until daybreak, and
(4) the time remaining until midday .
The tables are computed for latitude = 41;0 (Istanbul): see further I-2.7.2. On a much smaller
Turkish table for timekeeping by the stars, or rather, one particular star, see I-2.8.1.
25 On this see my notes in Byzantine Astronomy, pp. 117-118; and King, Geography of Astrolabes, p.
12.
26
See n. 14:40.
27
Navoni, Rouz-name des Turcs (cited in n. 14:20), p. 58.
28
On Muammad ibn Ktib Sinn (I-2.7.2) see Suter, MAA, no. 455; Cairo ENL Survey, no. H8; and ~hsanolu
et al., Ottoman Astronomical Literature, II, pp. 84-90, no. 46.
446 part ii, chapter fourteen
MS Istanbul Hamidiye 1453,3 (fols. 232v-266v) contains al-Khalls universal auxiliary tables
(10.7) and was copied in Edirne in 869 H [= 1464/65] by {Umar ibn {Uthmn al-usayn. The
tables follow a number of treatises on timekeeping and quadrants (see, for example, 2.5 and
2.6) and are preceded by al-Khalls instructions written in Arabic (see below). The functions
f and g are tabulated only for arguments:
= 20, 21, ... , 49, as well as 21;30 (Mecca) and 33;30 (Damascus) .
The main tables are preceded by a small table (fols. 233v-234r) displaying the qibla as a
function of local latitude and longitude, q(,L), which, as I have shown elsewhere, is
ultimately of Abbasid origin;29 a table of () to two digits, which, although for = 90 it
has entry 23;32, a value of nowhere else attested in the known Islamic sources, is apparently
based on = 23;31; and tables for finding the solar longitude from the date in the Muslim
and Syrian calendars (the solar longitude for @dhr 1 is given as Pisces 20;23). The main
tables are followed by some simple trigonometric tables of the functions:
arc Sin (x), arc Vers (x) , Cot12 h and Tan60 h ,
as well as a table of max d() based on = 23;35.
MS Istanbul Ayasofya 2590 is a second Ottoman copy of al-Khalls universal auxiliary
tables in a recension prepared by the muwaqqit Muammad ibn Ktib Sinn (14.5) for the
Suln Byazd, and dated 897 H [= 1491]. Ibn Ktib Sinn translated al-Khalls introduction
into Turkish and then copied the tables in their entirety. He also copied al-Khalls parameters
19 and 17 for twilight in the introduction, although elsewhere (as in his treatise on the
astrolabe preserved in MS Istanbul Ayasofya 2708) he advocated 20 and 16. The main set
is preceded by others of () and 2() computed to three digits for = 23;35 and followed
by tables of ha(H) and hb(H).
In a late Ottoman miscellany of astronomical treatises preserved in an unnumbered
manuscript formerly (ca. 1970) in the private collection of the late Professor Buhairi of the
American University of Beirut, there are tables of f() and g() for = 41 (Istanbul) and
a single table of K(x,y) for x = 41. There are no accompanying instructions, and the hapless
individual who thought fit to copy these three tables together was unaware that the argument
x is unrelated to the local latitude.
MS Oxford arab. e. 93, fols. 3v-25r, contains an undated set of anonymous tables of the
functions t(,h) and T(,h) expressed in equatorial degrees: see I-2.2.2 (illustrated). Values are
given for each degree of starting at the winter solstice and each degree of h up to [H()],
and the city of Edirne is specifically mentioned at the head of some of the tables. The underlying
parameters are:
= 41;30 and 23;30 .
There are no prayer-tables accompanying this set, but notice that the Rznme-ye Vef} for
Istanbul (14.3) is based on the same parameters. See also 14.11 on some other tables for Edirne.
MS Cairo M 255,6, copied 1060 H [= ca. 1650], contains al-Khalls introduction to his
prayer-tables and some anonymous Syrian prayer-tables for latitudes:
= 36 (Aleppo) and 41;15 (Istanbul).30
Mention of these was omitted from the first version of this text, and all that is available to
me now is a photo of an extract from the tables for Aleppo (Fig. 11.5). This shows tables of
the same kind as those of al-Khall but for = 23;35 rather than his more up-to-date value
23;31. The functions tabulated for Aleppo are:
H, Z, D, ha, ta, Ta, N, n, r, s, , d and tq ,
MS Istanbul UL T1824,1 (fols. 3r-9r) contains another set of prayer-tables for Istanbul, based
on the parameters:
= 41;15 and = 23;30 .
No compiler is mentioned. Notice that the value of the latitude is the same as that used by
Taqi l-Dn in his early work (14.9), and that the value of the obliquity is the one usually
associated with astronomers who relied on Ulugh Beg (see, for example, 7.1). The set begins
with a table of () computed to two digits for each degree of . The main prayer-tables display
some 22 functions tabulated side by side for each degree of from Capricorn 1 to Gemini
30, as in the tables of al-Khall (10.6). Values are given to two digits. The functions tabulated
are the following:
Dh, C, H, D, 2Dh, ha, hb, ta, tb, Ta, Tb, 2Nh,
s, (2N-r), r, (r+D), (D+ta), hq, tq, , B and b1.
MSS Paris BNF ar. 2544,14 and Cairo DM 36 contain an extract from these tables displaying
only the functions:
D, 2Dh, ta, Ta, s, (2N-r), r and .
No compiler is mentioned in either source.
These tables are rather accurately computed. The parameters used for twilight are 19 and
17, and I have not investigated the value of the qibla underlying the tables of hq and tq. Note
that in the geographical tables of Taqi l-Dn (14.9), we find the entries:
Istanbul : 41;15 L: 60;0
Mecca 21;30 77;0
so that the qibla at Istanbul, computed according to the accurate formula, would be 41;10 E
of S. On the tables of B, b1 and C see I-6.4.9, I-6.5.2 and I-6.10.7. The fact that a table of
b1 was included in the set I take to be further evidence that its compiler was inspired by the
tables of al-Khall.
MS Istanbul Hamidiye 842 contains another set of less extensive prayer-tables based on the
same parameters, but with slightly different entries. Values of the following functions are given
to two digits in modern Arabic numerical notation for each degree of from Capricorn 0 to
Gemini 29:
H, Z(12), D, ha, ta, Ta, s, r, and 2Nh .
MS Cairo MM 100 (13 fols., ca. 1750) contains some anonymous prayer-tables for Istanbul
arranged according to the Byzantine months, which I have not investigated.
Taqi l-Dn (ca. 1525-1585)31 was the principal astronomer associated with the founding of
the short-lived observatory in Istanbul.32 He wrote a number of works of considerable interest,
which have not yet received the attention which they deserve. MS Istanbul Kandilli 208 is a
copy of some of Taqi l-Dns works in his own hand, and contains a aylasn table displaying
the hour-angle and time since sunrise as functions of the meridian altitude and the instantaneous
altitude for parameters:
= 41 (Istanbul) and = 23;30 .
I see no reason yet to doubt that this was computed by Taqi l-Dn himself (see further I-2.3.6).
It may be that Taqi l-Dn also compiled the prayer-tables discussed in 14.8 since he used the
latitude 41;15 for Istanbul in his earlier works, such as his treatise on sundial construction.
In his later work, however, he used latitude 40;58 for Istanbul and obliquity 23;28,54: these
are the results of his new Murd Khn observations (see, for example, I-6.4.8).
In his Zj entitled Jardat al-durar wa-khardat al-fikar, of which I have examined MS
Istanbul Esat Efendi 1976, Taqi l-Dn advocates the parameters 19 and 17 for twilight and
then mentions certain corrections (ta{dl) which should be applied to the length of twilight or
the diurnal arc. The following is a translation of the relevant section in Ch. 98 of the Zj:
For a day of fasting it is necessary to modify the duration of twilight and the length
of daylight by the quantities which I determined by observation and calculation. This
is explained in detail in my treatise called al-Qawl al-mar} fi l-nahr al-shar{. In
brief, I say that if Raman occurs when the sun is at the beginning of Aries then
the amount which must be added to the semi diurnal arc in the west is 1;3, likewise,
Taurus 1;20, Gemini 1;43, Cancer 1;39, Leo 1;23, Virgo 0;59, Libra 0;54,
Scorpio 0;54, Sagittarius 0;39, Capricorn 0;29, Aquarius 0;59, and Pisces 1;3.
To the semi diurnal arc in the east should be likewise added, in the order of signs:
0;56, 0;58, 1,8, 1;15, 1;16, 1;8, 1;6, 1;7, 1;6, 1;5, 0;59, and 0;58. In the
middle of the signs one should use linear interpolation. Modifying the length of
daylight obviates the need to modify the duration of morning twilight, because the
latter is defined in terms of the apparent horizon and the light of the upper edge of
the disk of the sun (jib al-shams). The size (of the sun) never exceeds the sum of
the two corrections, that is, a maximum 2;51 in Gemini, which is about 3, and a
minimum 1;34 in Capricorn, which is approximately 11/2, an amount that is not at
all small (i.e., negligible).
31 On Taqi l-Dn see n. I-2:35. See also the next note.
32 See Sayl, The Observatory in Islam, pp. 289-305, and also nver, Istanbul Observatory (in Turkish).
turkish tables for timekeeping 449
Unfortunately I do not know of any copies of Taqi l-Dns treatise al-Qawl al-mar} ... . The
title means something like A Statement as Clear and Visible as Light of Day concerning the
Length of Daylight as Defined by the Religious Law.33 It would be interesting to know how
he derived these two sets of values. Suffice it to say for the time being that the mean of the
two sets is suspiciously close to the values given by {Abd al-Qdir al-Minf (8.2) for = 41.
These mean values are:
1;0 1;19 1;26 1;27 1;20 1;4 1;0 1;1 0;53 0;47 0;59 1;1 ,
and al-Minf has:
1;4 1;11 1;18 1;25 1;18 1;11 1;4 0;57 0;51 0;44 0;51 0;57 .
Note that al-Minf Jr.s tables for refraction were compiled in Cairo in 975 H [= 1567] and
Taqi l-Dns Zj was compiled in Istanbul about 990 H [= 1582].
In a short treatise preserved in MSS Istanbul Kandilli 208 (fols. 89v-90r) and Istanbul
Kandilli 176, of which the former is in his own hand, Taqi l-Dn claims to discuss the difference
of opinion between the Cairo muwaqqits concerning the precise length of daylight, nightime,
and twilight. He makes no mention of the nature of the difference of opinion, but does state
that the visible horizon is 0;2,13 below the true horizon. This number is related to the numerical
tradition attributed to Ibn al-Haytham (see 7.1, also 9.3). Otherwise the treatise contains no
information of historical or scientific consequence.
MS Istanbul Kandilli 441 of the shorter version of li Efends tables for timekeeping
(14.13), copied in 1266 H [= 1849/50], contains a simple table for evening twilight (p. 62)
and a set of geographical coordinates (p. 73), both of which are attributed to Taqi l-Dn. The
following values are given for the duration of twilight in equatorial degrees for each zodiacal
sign:
23 25 29 32 29 25 23 23 24 25 24 23
If we round the values given in MS Istanbul UL T1824,1 (14.8), which are based on Taqi l-
Dns first value 41;15 for the latitude of Istanbul and parameter 17, we obtain:
23 25 29 31 29 25 23 23 24 25 24 23 .
All but the fourth value correspond precisely to Taqi l-Dns values.
Since there is a dearth of Ottoman geographical tables that have been published,34 and since
the Kennedys did not use any in their survey of Islamic geographical tables, I present the table
attributed to Taqi l-Dn in full:
Locality L
1 Istanbul 41;15 60; 0
2 Gallipoli 41;15 58; 0
3 Mitilene 39; 0 55; 0
4 Salonica 39;30 50; 0
5 Kavalla 40;30 46;30
6 Cos 38; 0 55; 0
7 Izmir 38; 0 56;20
33 No copies of this work are mentioned in ~hsanolu et al., Ottoman Astronomical Literature, I, pp. 199-
217, but Taqi l-Dns treatises listed as nos. 3, 12 and 16 (pp. 204, 209-210 and 211) relate to the same subject
and would repay investigation.
34 This table is now featured in Sezgin, GAS, X, pp. 182-184. See also King, Mecca-Centred World-Maps,
p. 86-87, on another Otoman table giving only the qiblas of 90 localities.
450 part ii, chapter fourteen
MS Cairo KhMT 2, copied about 1750, consists mainly of a set of planetary tables based on
the Zj of Ulugh Beg and a few spherical astronomical tables specifically for = 41, that is,
Istanbul. Included amongst these (fols. 55v-58r) is a set of prayer-tables for latitude 38, locality
unspecified. The functions tabulated are:
H, 2D, D, ta, ta/b, tb, Tb, s, 2N, n and r ,
and the values are given in Eastern Arabic numerical notation to the nearest half degree, one-
half being indicated by the symbol . I have not investigated the parameters underlying the
values of s and r. The function ta/b is refered to as bayn al-{arayn, and displays the hour-angle
at a certain time about halfway between the two {ar prayers. The tables display some
discrepancies: for example, the values of tb and Tb do not total D, and the values of s and n
and r do not total 2N.
MS Istanbul Hsn Paa 1286, copied around 1700, contains an anonymous set of tables
(fols. 66r-74r) of the functions:
2N, N, B, C, H, , d, , h0, Dh, D, 2D, and ,
computed for parameters:
= 42 (Edirne) and = 23;31 (?).
On other tables for Edirne see 14.7.
turkish tables for timekeeping 451
MS Cairo KhMT 2,1, copied about 1750 (14.10), contains a flyleaf (fol. 1r) on which part of
a aylasn table has been copied. The table is entitled jadwal fal al-d}ir and displays the
hour-angle t(H,h) for the argument domains:
H = 71;29, 71,70, ... , 48 and h = 0, 1, ... , H .
Values are given to two sexagesimal digits. Since the entry for H = 48 and h = 0 is 90;0,
the underlying parameter is:
= 42;0 (with = 23;29) .
Not all of the entries have been filled in, and the page of tables for < 0 is not contained
in this manuscript. There is no mention of the locality for which the table was prepared; this
may have been Edirne. See also I-2.3.7.
MS Istanbul Kandilli 196 is the only copy known to me of a set of hour-angle tables for Istanbul,
computed in 1095 H [= 1684] by Amad Efend, known as Mir al-Islmbl, the Egyptian
in Istanbul.35 The function t(h,) is tabulated for the domains:
h = 1, 2, ... , 72 and = 1, 2, ... , 90 ( \ 0)
and the underlying parameters are:
= 41;0 (Istanbul) and = 23;30 .
The tables have the same format as those in the main Cairo corpus (4.0) and are quite accurately
computed: see already I-2.1.7. Amad Efend prefered to use obliquity 23;30, as in the late
Egyptian tradition (7.1), rather than Taqi l-Dns value 23;28,54 (14.9). A century later li
Efend recomputed t(,h) for latitude 41 using Taqi l-Dns parameter (14.13).
In February, 2003, as I was preparing the final illustrations for this book, I came across a
negative of some prayer-tables in a Cairo manuscript that were not yet described in the text.
Checking my 1986 Survey of the Cairo manuscripts, I found that I had promised a description
in SATMI. The following must suffice.
MS Cairo DMT 2,1 (fols. 1r-26v, 1202 H [= 1787/88]) is a copy of a set of prayer-tables
by Ism{l Fahm, son of the better-known Ibrhm aqq who was the author of a well-known
encyclopaedia entitle Ma{rifat-nme.36 Five other manuscripts are preserved in Turkey. The
tables are preceded by an introduction in Turkish and the work is dated 1193 H [= 1779] and
dedicated to the father. The underlying latitude is stated to be 38;30, and apparently no locality
is specified. However, in the colophon of some of the copies mention is made of the village
35 I did not find mention of Amad Efend or these tables in ~hsanolu et al., Ottoman Astronomical Literature.
36 On the author see Cairo ENL Survey, no. H39; and ~hsanolu et al., Ottoman Astronomical Literature,
II, 627-628 (no. 462), where the other manuscripts of these tables are listed. On his father see Cairo ENL Survey,
no. H38.
452 part ii, chapter fourteen
Fig. 14.12*: An extract from the prayer-tables of Ism{l Fahm. [From MS Cairo DMT 2,1, courtesy of the
Egyptian National Library.
of Tillo, where the father, originally from Erzerum, was living at the time. Ekmeleddin
~hsanolu kindly informs me that Tillo is in the district of Aydnlar in S. E. Turkey, directing
me to www.tillo.net. The tables display for each degree of each sign the functions:
2Dh, 2Nh, z, a, s and i .
I have not investigated the underlying twilight parameters.
li Efend Mi{mr the architect (fl. ca. 1700) compiled an extensive corpus of tables for
timekeeping by the sun, extant in numerous copies.37 MSS Princeton Yahuda 353, Istanbul
Air Efendi 224, Istanbul Kandilli 219, Vienna 2379 (Mixt. 989), Cairo M 151 and 215, and
37 On li Efend see n. I-2:23.
turkish tables for timekeeping 453
Cairo K 18199, are copies of this corpus; all but the last are complete. The title is simply al-
jadwal al-kabr, the large table, which is appropriate since the corpus contains over 80,000
entries. li Efends tables, and simpler versions thereof, were widely used in Istanbul in
the 19th century.
The main tables are for timekeeping by the sun; the remainder is for regulating the times
of prayer. All are based on the parameters:
= 41;0 (Istanbul) and = 23;28,54 .
This distinctive value of the obliquity is the value derived by Taqi l-Dn (14.9), who worked
in Istanbul two centuries prior to li Efend.
In the main tables of the corpus li Efend tabulated the hour-angle and time since sunrise
T(,h) and t(,h) for each degree of l starting at the winter solstice and for each degree of
h up to [H()]. A given sub-table displays entries for one particular solar longitude, with the
times for solar altitudes in the east and west expressed in equatorial degrees (al-d}ir, fal al-
d}ir), equinoctial hours (al-mustawiya), seasonal hours (al-zamniyya), and in equinoctial
hours according to the Ottoman convention (al-muwfiqa). Entries are given to three
sexagesimal digits in each case and are very carefully computed. The way in which the tables
were compiled is clear from the nature of certain auxiliary functions also tabulated by li
Efend. See further I-2.2.3. Various functions are displayed around each of these sub-tables,
often in different coloured inks. These are the following:
H, 2D, 2N, a, b, r and s .
The parameters used for twilight are 19 and 17, and the entries, which but for the first function
H are in hours, are expressed to three sexagesimal digits.
At the end of the main set of tables the functions:
2D al-nahr length of daylight
r al-fajr time of daybreak
a = Ta al-{ar al-awwal the first afternoon prayer
b = Tb al-{ar al-thn the second afternoon prayer
2N al-layl length of daylight
s al-shafaq time of nightfall
are tabulated separately (the functions T being labelled al-bq ila l-ghurb, time remaining
until sunset, and the times al-muwfiqa, meaning time according to the Ottoman convention),
with entries expressed in equinoctial hours to three digits. A given page of tables serves a pair
of zodiacal signs symmetrical with respect to the solstices.
MS Istanbul Husrev Paa 232 contains a set of tables attributed to Muammad diq Jihn-
gr, muwaqqit in the Mosque called Wlidat al-Suln (Valide Sultan Camii) in Istanbul.38 This
displays the two times when the sun has a particular altitude on the assumption that midday
is 12 oclock. Here we see European influence at work. Values are given in equinoctial hours
to three digits and are derived from those of li Efend. Anonymous copies of these same
tables are contained in MSS Istanbul Badatli Vehbi Efendi 990, Istanbul Lala Ismail 287,
Istanbul Kandilli 220, 440 and 441, Cairo M 120, and anonymous copies of similar tables
38 Ibid., II, pp. 561-562, no. 399, and also p. 456 and 684. See the illustration in King, Astronomical
Timekeeping in Ottoman Turkey, pl. 9.
454 part ii, chapter fourteen
with entries to two digits rather than three are contained in MSS Istanbul Esat Efendi 1979,
Istanbul UL T1963 and T1964.
MS Baghdad Awqf 325/12248 contains some 50-odd pages of tables attributed to Jihngr.
I have not been able to inspect this manuscript, but it is catalogued under the title ikhtilf m
bayna l-ufuq al-aqq wa-l-mar}, which refers to corrections for refraction at the horizon.
This title probably relates to something written on the first page of the manuscript rather than
the main tables.
On the flyleaf of MS Istanbul Lala Ismail 287, copied in 1195 H (= 1781), is a set of numbers
for each of the zodiacal signs labelled daq}iq iat al-qur li-{ar mm-alif, minutes of the
correction for the solar disc at latitude 41. The numbers are as follows, in the order of signs:
3 2 2 2 2 2 3 4 5 5 4 3.
MS Istanbul Kandilli 522 contains a set of prayer-tables for the year 1290 H [= 1873/74]
computed for the latitude of Khania in Crete (see already 8.8 on some other tables for Crete).39
The following functions are tabulated for each day of the year, expressed in equinoctial hours
and minutes:
R, m, a, s and r .
The underlying parameters are found by inspection to be:
= 35;20 and 23;29 ,
and the parameters used for morning and evening twilight are 19 and 17. The entries for
m, a and s correspond very closely to recomputation without taking horizon phenomena into
consideration. However both the times r and R, daybreak and sunset, are earlier than the
recomputed times by an amount which varies between 16m at the equinoxes and 20m at the
solstices. This appears to be a generous correction for refraction and the size of the solar disc
(see also 12.11), half of which the anonymous calculator should also have applied to the times
m and a.
At the end of the treatise on the trigonometric quadrant by the 19th-century Ottoman astronomer
al-Geds,40 printed in Istanbul in 1311 H [= 1893/94], there is a set of imsk tables for different
latitudes. The time of the imsk in hours and minutes, Turkish time, is displayed for each 2
of solar longitude and for latitudes:
21 (Mecca), 24 (Medina) and 35, 36, ... , 45.
I have not investigated the underlying value of the solar depression and the definition used
39 This source is mentioned in ~hsanolu et al., Ottoman Astronomical Literature, II, p. 807.
40 On Sulaymn Murd ibn {Umar ibn Amad Sa{d al-Geds see ~hsanolu et al., Ottoman Astronomical
Literature, II, pp. 601-602, no. 436. A translation of this treatise and a commentary were published in
Wrschmidt, Gedosi ber den Quadranten. Wrschmidt mentions the tables (p. 154) and apologizes for not
reproducing them. Copies of the tables were kindly prepared for me by the late M. Alain Brieux of Paris.
turkish tables for timekeeping 455
for the imsk. In Gedss treatise the parameters used for morning and evening twilight are
19 and 17, and the imsk is taken as being 16m before daybreak.41
14.16 Sa{d Beg Zdes auxiliary tables for computing the prayer-times for all latitudes
It is appropriate that we should conclude the present study with a few remarks on a set of tables
compiled in Istanbul in the 19th century for computing the prayer-times for all latitudes. The
compiler of these tables, which are extant in MS Istanbul Kandilli 226, was one Sa{d (or
possibly Sayyid) Beg Zde Mr Ibrhm, on whom I have no further information.42 The tables
are preceded by an introduction in Turkish and followed by some worked examples. The
notation used by Sa{d Beg Zde is rather unfortunate: most of his functions are called ta{dl,
equation or correction. Also he uses logarithms to base sixty expressed sexagesimally.
Nevertheless his tables are of interest because they are really still part of the medieval tradition,
and indeed bear some resemblance to the auxiliary tables of al-Waf} compiled several
centuries previously (6.15). One of the two known copies of al-Waf}s tables (MS
Nuruosmaniye 2921,2) is preserved in Istanbul, and I consider it likely that Sa{d Beg Zde
had seen such a copy. He advocates the use of parameters 20 and 17. He also gives tables
for finding the time of prayers at the two major religious holidays in Islam (alt-i {dayn),
based on the assumption that the prayer should take place when the solar altitude is 6 in the
east. I know of no other tables for this prayer-time, or reference to this particular definition
in the Islamic sources. I have outlined the principle underlying these tables in I-9.13.
The first set of tables displays three functions of , with values for each degree of argument
expressed to two sexagesimal digits in the case of the first and to three digits in the case of
the second and third. These functions are:
() ( = 23;28) mayl awwal first declination
L(H) = log{Sin H} ta{dl-i uhr equation of midday
?? ta{dl-i mayl equation of the declination
The first main table displays values of the following functions for each degree of argument
({adad) H, or H* in the case of the table for twilight, from 1 to 90. Values are given to three
sexagesimal digits, except for the last function whose values are given to two digits.
La(H) = log{Sin H - Sin ha(H)} ta{dl-i {ar-i awwal equation of the first afternoon
prayer
Lb(H) = log{Sin H - Sin hb(H)} ta{dl-i {ar-i thn equation of the second afternoon
prayer
Lr(H) = log{Sin H* - Sin 16} ta{dl-i shafaq equation of nightfall
Ls(H) = log{Sin H* - Sin 20} ta{dl-i fajr equation of daybreak
L(H) = log{Sin H - Sin 6} ta{dl-i alt al-{dayn equation of the prayers at the
two religious festivals
S(h) = Sin h ta{dl-i irtif{ equation of the altitude
(No values are given for Lr(H) for H* < 16, etc.)
The second main table displays the following functions for each 0;15 of argument x (ta{dl-
i awwal) up to 120 in the case of the second and third functions, and up to 60 in the case of
the remainder. Values are given to two sexagesimal digits except for the second function whose
values are given to three:
S-1(s) = arc Sin (s) ta{dl-i thn second equation
L(x) = log x ta{dl-i mu{addal modified equation
V(x) = 1/15 arc Vers (x) s{t-i d}ir hours of time
C(x) = /15 arc Cos (x)
1 uhr-i shaml midday for northern latitudes
12h - C(x) uhr-i janb midday for southern latitudes
A major disadvantage of these tables, as we shall see, is that one must find V from L in a
table where both functions are tabulated with x as argument. I introduce the function
V*(x) = 1/15 arc Vers (alog x) ,
which Sa{d Beg Zde would have done well to tabulate. The third main table displays the
function:
G*(,) = log G(,) ,
called ta{dl-i {ar, the equation of the latitude. Values are given to three significant
sexagesimal digits for each 5 of and each 1 of up to 90.
The prayer-times expressed according to the Ottoman convention can be found using these
functions in the following way:
uhr: m = V* { L(H) + G*(,) }
{ar 1 and 2: ta/b = m + V* { La/b(H) + G*(,) }
shafaq: s = m - V* { Ls(H*) + G*(, ) }
fajr: r = m + V* { Lr(H*) + G*(,) }
alt-i {dayn: = m + 12 h - V* { L(H) + G*(,) }
To find the hour-angle corresponding to a given solar altitude h, form:
L { H(H,h) } = L { S(H) - S(h) }
and then:
t(h,H,,) = V* { L { H(H,h) } + G(,) } .
Sa{d Beg Zde gives numerical examples for = 41 and = 60. The prayer-times agree
with the corresponding entries in the prayer-tables of li Efend (14.13) to within a minute
of time. He also demonstrates how to compute the qibla, and using coordinates:
Istanbul L: 56;33 : 41;0
Mecca 68;30 21;15
he somehow derives the value 66;31 (measured from the prime vertical). Accurately computed
for these coordinates the qibla is 59;16. As noted above (14.4) the qibla at Istanbul is in fact
closer to 611/2 S of E.
Postscript: When I first started writing Part II of this book some 30 years ago, tables were
available from the local authorities in each region of the Islamic world displaying the prayer
times. Some examples are presented in Fig. V-13.1 and in the article M_t ii. Astronomical
Aspects in EI2. Nowadays one can download tables for any locality from the Internet: see,
for example, http://prayer/al-islam.com (2003).
introduction 457
Part III
A survey of arithmetical
shadow-schemes
for time-reckoning
458 part iii, chapter one
In numerous medieval Islamic treatises on traditional folk astronomy and legal texts various
simple arithmetical shadow-schemes are found for regulating the times of the uhr and {ar
prayers. Less frequently, there occur schemes for the shadows at each of the seasonal hours
of daylight. This material represents a development virtually independent of the highly
sophisticated astronomical timekeeping ({ilm al-mqt) practiced by the astronomers of Islamic
world, and it is here investigated for the first time.
As we shall see, some of this material derives from Hellenistic folk astronomy, notably,
from the shadow schemes which have been studied by Otto Neugebauer. Also an approximate
Indian rule for reckoning time of day from shadow lengths had considerable influence in Islamic
folk astronomy, to the extent that it even underlies the definitions for the uhr and the {ar
which became standard in later practice. The Yemeni sources provide the most original
materials, but all of the material discussed here has a distinct Islamic flavour. In passing I
mention the shadow-schemes in medieval European folk astronomy, which were also studied
by Neugebauer.
Otto Neugebauer (1899-1990) was the leading historian of the exact sciences in the 20th
century. His monumental contributions to ancient Babylonian and Greek astronomy and
mathematics are well known, his contributions to Islamic, Byzantine and medieval European
astronomy perhaps less so. The interested reader should consult Noel Swerdlow, Otto
E. Neugebauer, Proceedings of the American Philosophical Society 137:1 (March 1993), pp.
137-165, or a summary available on the Internet under www.nap.edu/html/biomems/
oneugebauer.html, as well as the bibliography compiled by Janet Sachs and Gerald Toomer
in Centaurus 22 (1979), pp. 257-280.
But Neugebauer should also be remembered for his contributions to folk astronomy,
Hellenistic, Ethiopic and medieval European. Most historians of science have no conception
of folk science, but the fact that it flourished alongside the exact sciences wherever these were
cultivated, not least in two of the most significant civilisations in human history, that is, the
Hellenistic and the Islamic, suggests that perhaps we historians of science should take it
seriously. For an introduction, the interested reader may consult any of Neugebauers
publications listed below:
Astronomische Papyri aus Wiener Sammlungen: II. ber griechische Wetterzeichen und
Schattentafeln, Sitzungsberichte der philosophisch-historischen Klasse der sterreich-
ischen Akademie der Wissenschaften 240:2 (1962), pp. 29-44.
On Some Aspects of Early Greek Astronomy, Journal of the American Philosophical
Society 116 (1972), pp. 243-251, repr. in idem, Essays, pp. 361-369.
Ethiopic Astronomy and Computus, Vienna: sterreichische Akademie der Wissenschaften,
1979.
Abu Shakers Chronography A Treatise of the 13th Century on Chronological,
Calendrical, and Astronomical Matters, written by a Christian Arab, preserved in Ethiopic,
Vienna: sterreichische Akademie der Wissenschaften, 1988.
462 part iii, acknowledgments
Astronomical and Calendrical Data in the Trs Riches Heures, (originally published 1974),
repr. in idem, Astronomy and History Selected Essays, New York, etc.: Springer, 1983,
pp. 507-520.
(For writings by other scholars on Islamic folk astronomy see the notes to 1.1 below.)
Most people could only stand in awe of Neugebauers academic achievements, but those
who knew the man will remember him most for his warmth and his inimitable humour. He
for one knew that eagles fly along the Bosphorus, could identify those who liked a good Spiel
und Tanz, and recognized that Kindi played a Doppelhorn. I was happy that he could see
the first published version of this study, presented to him on the occasion of his 90th birthday.
This appeared as A Survey of Medieval Islamic Shadow Schemes for Simple Timereckoning,
Oriens 32 (1990), pp. 191-249.
Most of the sources for this study were gathered during my survey of the scientific
manuscripts in the Egyptian National Library conducted during the years 1974-76. My activities
at the American Research Center in Egypt during 1972-79 were made possible mainly by the
Smithsonian Institution and (U.S.) National Science Foundation, and my research on Islamic
folk astronomy during 1983-85 by the (U.S.) National Endowment for the Humanities. This
support is gratefully acknowledged. It is also a pleasure to record my appreciation to the
Egyptian National Library for the privilege of working with their vast manuscript holdings,
and my thanks are also due to the other libraries where the sources are preserved.
To my friend Daniel Varisco I owe much of my enthusiasm for folk astronomy and most
of the references to modern practices relating to irrigation. Dr. Abdallah Nasif (University of
Riyadh) kindly provided me with copies of relevant sections of his thesis. Dr. Tom Zuidema
(University of Illinois) drew my attention to the basic documentation on the Gresik sundial.
For remarks on the penultimate draft of this study and for much more I am indebted to
Ted Kennedy.
Since the original study was published, two new studies have appeared that deal with
timekeeping by night using the lunar mansions. These are Miquel Forcada, Mqt en los
calendarios andaluses, al-Qanara 11 (1990), pp. 59-69, and Petra Schmidl, Volksastrono-
mische Abhandlungen aus dem mittelalterlichen arabisch-islamischen Kulturraum, doctoral
dissertation, Institute for History of Science, Johann Wolfgang Goethe University, Frankfurt
am Main, 2003.
In this new version I have simplified the footnotes by using the bibliographical abbreviations
listed at the beginning of this book and inserted cross-references to I-II and IV.
table of contents 463
TABLE OF CONTENTS
CHAPTER 1
INTRODUCTION
In the medieval Islamic world there were two separate traditions of astronomical knowledge.
The first was a simple, non-technical, essentially practical tradition, devoid of theories. The
second was a mathematical tradition in which theories, tables and computation featured
prominently. Since the distinction between the traditions is not generally recognized we shall
begin by considering each of the traditions separately.
First, the folk tradition.1 The Arabs of the Peninsula in the time before Islam possessed
an intimate knowledge of the apparent motions of the sun, moon, and stars across the heavens,
the months and seasons, the changing night sky throughout the year, and associated meteo-
rological phenomena. A distinctive feature of this tradition was the division of the year into
13-day periods (anw}) defined by the (acronychal) settings and (heliacal) risings of the 28
lunar mansions (manzil).2 Another was the custom of describing the time of day with reference
to the length of a persons shadow.3 The Arabs also had names for various divisions of the
day and night, including names for the seasonal hours, that is, the twelve hours of daylight
and the twelve hours of night.4
With the advent of Islam in the early 7th century, the Muslims expanded out of the Peninsula
to establish a commonwealth from Spain to Central Asia. At various stages they incorporated
some of the simple and non-technical procedures that they found in use in these regions. The
resulting Islamic tradition of folk astronomy was widely practiced until the 19th century, and
vestiges of it can be detected to this day in rural areas from Mauritania to Indonesia.
Besides the obvious fact that any man in the street could master the fundamentals of folk
astronomy, there were two main reasons for its appeal. First, the sun, moon and stars were
mentioned in the Qur}n, and a knowledge of the basics of their apparent courses constituted
1
All Arabic sources dealing with folk astronomy per se and compiled prior to ca. 1050, as well as all previous
studies thereon, are surveyed in Sezgin, GAS, VII, pp. 336-370. Some of the sources were seriously investigated
for the first time in Nallino, Scritti, V, pp. 152-197. A contribution of a more philological kind on the subject
is the article Layl and nahr by Charles Pellat in EI2. A new overview is Varisco, Islamic Folk Astronomy.
Other recent contributions, particularly by Paul Kunitzsch and Miquel Forcada, are mentioned below. On certain
distinctive features see also nn. 1:2-4 and on various practical applications see n. 1:9.
2
On the lunar mansions in Islamic astronomy see the introductory articles Anw} by Charles Pellat and
Manzil by Paul Kunitzsch in EI2, repr. in idem, Studies, XX, and the overview in Varisco, Islamic Folk
Astronomy. On the Andalus tradition in particular see now Forcada, Anw} Books in al-Andalus.
3
Arent J. Wensinck et al., Concordance et indices de la tradition musulmane, 8 vols., Leiden: E. J. Brill,
1936-1988, (repr. in 4 vols., Leiden: E. J. Brill, 1992), IV, pp. 78-79, and V, pp. 211-213, lists references to
the terms ill and fay} for shadow in the adth literature. See also 2.1 below.
4
On the seasonal hours in general see, for example, Neugebauer, HAMA, III, p. 1069; idem, Ethiopic
Astronomy and Computus, pp. 167-170; Kennedy, Zj Survey, p. 141; and IV-2.1. On these names (which
are not mentioned in Pellats article cited in n. 1:1) see IV-8.
466 part iii, chapter one
part of a standard religious training. Even a distinctively Islamic cosmology developed over
the centuries.5 Second, the Qur}n and Islamic tradition advocate three duties incumbent on
every Muslim which are connected with astronomy, at least with the moon, the sun, and the
earth, namely:
Observance of a sacred month of fasting and of other religious festivals regulated by a strictly
lunar calendar in which the beginnings of the months are determined by the first visibility
of the lunar crescent;6
Performance of five ritual prayers at times determined by the suns position relative to
the local horizon, defined in terms of shadows during the day, by sunset and sunrise,
and by twilight phenomena;7 and
Performance of various ritual obligations, including prayer, but also recitation of the Qur}n,
the call to prayer, slaughter of animals, and burial of the dead, in a sacred direction towards
the Ka{ba in Mecca (called in Arabic, qibla).8
The basic stipulations of each of these duties correct timing and correct orientation could
be achieved by applications of the techniques of folk astronomy, with virtually no computation
beyond simple arithmetic.9 These procedures were advocated by the scholars of the sacred law
and are to be found recorded not only in legal texts, but also in works on folk astronomy and
popular works such as lexicons and encyclopaedias. They involved respectively:
Actual sighting of the lunar crescent and, in the case of bad weather, counting months of
roughly 291/2 days each in pairs;10
Timekeeping by day using rough estimates of shadow lengths (as described in the present
study) and by night using the risings and settings of the lunar mansions; direct observation
of sunset, nightfall, daybreak, and sunrise; and
Adopting as indicators of the local sacred direction the north or south points or the risings
and settings of various prominent stars or of the sun at the equinoxes or solstices.11
The early Muslims must have been particularly happy to come across arithmetical rules for
determining the shadows at midday and at the seasonal hours: these, as we shall see, they used
not only to formulate precise definitions for the times of the daylight prayers prescribed in
Islam, but also in certain circles for timekeeping in general.
Second, mathematical astronomy. In the 8th and 9th centuries the Muslims enjoyed an
encounter with a very different approach to astronomy.12 This they found documented mainly
5
On this Islamic tradition see Heinen, Islamic Cosmology. On the corresponding scientific tradition see
Jachimowicz, Islamic Cosmology.
6
See, for example, the articles Raman by M. Plessner and awm by C. C. Berg in EI1, updated by
the editors in EI2.
7
See the article M_t, i: ritual and legal aspects by A. J. Wensinck in EI2.
8
See the article ^ibla, i: ritual and legal aspects by A. J. Wensinck in EI2.
9
See King, Islamic Folk Astronomy, A-B, for overviews of the folk astronomical tradition and the ways
in which it was applied to regulating the calendar, timekeeping, and determining the sacred direction towards
Mecca.
10
The best source of information on this subject is still Renaud, Sur les lunes du Ramadan (mainly on
the practice in the Maghrib).
11
See my forthcoming monograph Sacred Geography of Islam, summarized in the article Makka. As Centre
of the World in EI2, repr. in King, Studies, C-X, for a survey of these procedures, and VIIa on the consequences
for the orientation of Islamic religious architecture.
12
For an overview of the mathematical tradition in Islamic astronomy see King, Islamic Astronomy. On
introduction 467
in texts of Indian and Greek provenance. Here the emphasis was on careful observation,
preferably with instruments; geometrical models for the sun, moon and planets; and on
computation. Greek and Indian astronomers, like their Babylonian predecessors, had a passion
for tables and no fear of extensive calculations. Muslim astronomers adopted this earlier Greek
and Indian material and modified it for their own use. They conducted new observations,
compiled new tables, devised new instruments, and made progress in all aspects of mathemati-
cal astronomy within its ancient and medieval framework. Only since the middle of this century
has it become clear to us to what extent they too had a passion for tables and instruments,
and the basic documentation of their achievements is still in progress.
Muslim astronomers working in the mathematical tradition also treated the three problems
mentioned above. Their solutions included:
Developing limiting conditions on such quantities as the apparent elongation of the moon
from the sun, the difference in setting times of the moon and sun, or the height of the moon
above the local horizon at sunset, and preparing tables based on such conditions for
predicting lunar crescent visibility in specific latitudes;13
Applying trigonometric formulae, either complicated accurate ones or handy approxima-
tions, for reckoning time from solar and stellar altitudes and compiling tables for
timekeeping by the sun and stars for different latitudes, sometimes containing tens of
thousands of entries, based upon such formulae (I-II); also developing instruments such as
the astrolabe, quadrant and sundial for the same purpose (X).
Developing cartographic, geometric and trigonometric solutions to the problem of determin-
ing the direction of one point on a sphere to another; tabulating the direction of Mecca as
a function of terrestrial longitude and latitude; listing the qiblas of cities throughout the
Islamic world; and devising cartographic techniques for reading the qibla directly from
maps.14
These approaches are a far cry from the methods advocated by the legal scholars. Only when
we realize how different they were do we find ourselves in a position to assess Islamic science
on its own terms.
After the formative period of Islamic astronomy in the 8th to 10th centuries there developed
regional schools with their own specific interests and authorities.15 It was in the Yemen in
particular that the two traditions of astronomy flourished side by side. Likewise in the Maghrib
there were several widely-respected authorities in folk astronomy, this in spite of the
the beginnings of this tradition see Pingree, Indian Influence and Greek Influence, and more recently Saliba,
Arabic Science and the Greek Legacy. Pingrees article {Ilm al-hay}a in EI2 provides an overview of the
activities of the astronomers in all fields except those related to Islamic ritual (on which see the overview in
King, Islamic Folk Astronomy, A-B). For the manuscript sources from the early period see Sezgin, GAS, VI.
13
See King, Tables for Lunar Crescent Visibility, on the earliest tables, and the article Ru}yat al-hill
in EI2 for an overview of Muslim activity in this field.
14
An overview is in the article ^ibla (astronomical aspects) in EI2, repr. in King, Studies, C-IX. More
information is in idem, Mecca-Centred World-Maps.
15
On the schools in the Islamic East see Kennedy, Exact Sciences in Iran, A-C; on those in Egypt and
Syria, the Yemen, and the Maghrib see King, Astronomy of the Mamluks; idem, Astronomy in Yemen, and
idem, Astronomy in the Maghrib, respectively. On astronomy in al-Andalus see Sams, Ciencias en al-Andalus,
and idem, Studies, and many other publications of the Barcelona school.
468 part iii, chapter one
availability there of sophisticated tables relating to mathematical astronomy. In Egypt and Syria
there were apparently no such authorities.
Now what concerns us here is the second obligation to pray at the right times, particularly
in the case of the two prayers performed during the day. It seems that in the first centuries
of Islam the times of prayer were regulated by the muezzins who performed the call to prayer
five times a day. They were chosen for the excellence of their voices and their good character;
they then had to learn the basics of folk astronomy, which was all that they needed in addition
to perform their task adequately. However, the 13th century witnessed the rise of the institution
of the muwaqqit, that is, the astronomer associated with a mosque for the purposes of regulating
the times of prayer, as well as the prediction of lunar crescent visibility and the determination
of the qibla (V). Some impressive scientific achievements were made by certain muwaqqits,
particularly in Egypt and Syria. Tables displaying the times of prayer for each day of the year
for specific localities became commonplace. Instruments mainly sundials, astrolabes and
quadrants also bore markings for the times of prayer. We do not know how the muwaqqits
functioned in such proximity to legal scholars who were advocating simple techniques of folk
astronomy; our sources are silent on this point. Certainly I am not aware of any legal text in
which it is suggested that one should consult an astronomer on the prayer-times or use any
of the astronomical tables or instruments that were available for this purpose. It would be nave
to suppose that there was any reason why a legal scholar should have consulted an astronomer.
In this study I present a survey of all known Islamic shadow-schemes for regulating the
time of day in general and the times of the midday and mid-afternoon prayers in particular.
This material belongs almost exclusively to the folk tradition in Islamic science, and most of
it has never been studied previously. It has been found mainly in books dealing with the
regulation of the prayer-times by non-technical means (kutub al-mawqt), in popular almanacs,
and in books on the sacred law of Islam. Not a few of the schemes are recorded as curiosities
in works belonging to the mathematical tradition, or simply as marginalia or notes on flyleaves.
I make no claim to have exhausted the available material (especially legal texts), but I have
included all the relevant material currently known to me. I have not come across any Turkish
or Persian sources, so the Ottoman heartland and the Islamic East are not well represented in
my survey.
1.2 The seasonal hours and the times of the uhr and {ar prayers
A minority of our sources displays shadow lengths for each seasonal hour of the day during
each month of the year, but most of the Islamic shadow-schemes that I shall discuss were
intended to be used primarily for regulating the times of the midday (uhr) and afternoon ({ar)
prayers. It is not clear how these were regulated during the first few decades of Islam: the
Qur}n and the adth do not between them provide precise definitions (IV-1-3). But from the
8th century onwards, the times of these two prayers were defined in terms of shadows, or
more specifically shadow increases: see Fig. 1.2a.16 The uhr began shortly after midday
On the definitions of the prayers in the astronomical sources see also Wiedemann & Frank, Gebetszeiten;
16
Fig. 1.2a: The standard definitions of the times of the beginning of the uhr and the beginning and end of the
{ar.
when the shadow was observed to increase, and the {ar began when the shadow had increased
over its midday minimum by the length of the gnomon. Variant definitions for the uhr in terms
of shadow increases are also attested. Likewise, the end of the interval permitted for the {ar
is often defined as being when the shadow has increased by twice the length of the gnomon.
As I have shown in detail elsewhere (IV), these definitions represent practical means of linking
the prayers to the seasonal hours, the link being provided by an approximate Indian formula
for timekeeping first attested in the Islamic sources of the 8th century see below and also
2.1, 2.3 and 3.2.
Underlying most of the schemes is the implicit notion that one should measure ones own
shadow in terms of the length of ones feet. The most common Islamic base for the gnomon
length was 7 (qadams or feet) although sometimes 6, 61/2 and 62/3 feet were used. These values
between 6 and 7 were thought to correspond to the ratio of the height of a man (qma) to the
length of his feet.17
A gnomon the height of a man is still in situ in the courtyard of the 7th-century Mosque
of Janad in the Yemen see Fig. 1.2b. It is known locally as the stick ({a) of Mu{dh ibn
Jabal, who was appointed by the Prophet Muammad as the first governor of the Yemen.18
17
Kennedy, al-Brns Shadows, I, pp. 68-79, and II, pp. 25-31, ad al-Brn, Shadows, Ch. 7. In the
mathematical tradition two other bases were used, namely 12 (iba{s or digits) and less frequently 60: see
Schoy, Schattentafeln, and Kennedy, Zj Survey, p. 140; and also the last paragraphs of I-1.1 and II.1.3.
18
See also King, Astronomy in Yemen, p. 8.
470 part iii, chapter one
Fig. 1.2b: The gnomon in the courtyard of the Mosque of Janad. [Photo courtesy Professor Daniel M. Varisco.]
The 13th-century legal scholar al-Aba (see 4.1) states explicitly that he measured the midday
shadows in Janad for every 13 days of the Syrian year, that is, for each of the anw}. I do
not know the age of the present gnomon, but there has clearly been a gnomon there for centuries.
Usually the schemes are presented in words or mnemonics in standard medieval Arabic
alphanumerical (abjad) notation.19 The schemes for the hours were sometimes represented in
tabular form, but underlying them we also find trivial arithmetical schemes. Indeed, virtually
all of the material is arithmetical, devoid of any influence of trigonometry. I have included
a few shadow-lists that were observed or derived by calculation; these stem from the murky
and unclearly-defined border between the two traditions.
Only rarely do competent astronomers mention these primitive schemes. The celebrated
scientist al-Brn (see 2.4) devoted a whole book to the topic of shadows,20 and in this he
discussed two Indian schemes, but only because he found them interesting, not because he
19
On this convention see Irani, Arabic Numeral Notation, and also Berggren, Islamic Mathematics, pp.
41-42.
20
An uncritical edition of the Shadows is listed under al-Brn, Shadows. For translation and commentary
see Kennedy, al-Brns Shadows.
introduction 471
would have approved of their use. On the other hand, two masters of spherical astronomy, Najm
al-Dn al-Mir and Ibn al-Shir (see 9.2 and 8.2), do advocate using the simple formula of
Indian provenance mentioned above as a handy approximation, and they can be forgiven for
doing this because in spite of its simplicity the formula is remarkably accurate for most practical
purposes (see 1.3). Before we consider the Islamic material in detail we should briefly mention
the earlier traditions (1.4).
The approximate method for timekeeping advocated by al-Fazr (see 2.3) and the equivalent
shadow-schemes for the hours in several later sources investigated below (see 2.1, 3.2, etc.)
are actually of Indian origin21 although this is not explicitly stated (see, however, 2.4). It is
equivalent to the formula:
T 6 n / (z + n) .
Note that, by definition, for T = 0, z , and for T = 6, z = 0. The formula is the simplest
means of assuring that these two boundary conditions are satisfied. For the general situation
0 < T < 6 the formula provides a reasonable approximation.22 This was the formula mentioned
in 1.1 and 1.2 that was used to derive the standard definitions of the times of the daylight prayers
in Islam.
Simple arithmetical schemes representing the shadow lengths at midday for each month of the
year, and tables displaying the shadows at each seasonal hour for each month of the year or
each zodiacal sign also based on a simple arithmetical scheme, are attested already in early
Greek astronomy, and their use continued in the Byzantine, Coptic, and Ethiopic traditions,
as well as in the medieval Islamic world and Europe. It is to Otto Neugebauer that we owe
the first detailed investigation of this material.23
21
On this formula in the already known Islamic sources, see Pingree, al-Fazr, pp. 121-122 (al-Fazr),
Kennedy, al-Brns Shadows, II, pp. 116-121 (al-Brn), and IV-2.4. In this study I refer to this formula as
the standard Indian formula although the Muslim astronomers also used an accurate formula for timekeeping
also of Indian provenance see Davidian, al-Brn on the Length of Day, for details. There is a substantial
secondary literature on timekeeping by shadows in Babylonian and Indian astronomy, much of which seems
to confuse shadows with shadow increases.
In XI, the standard approximate formula is the trigonometric formula, also apparently of Indian origin,
referred to in n. 4:5 below.
22
An investigation of its accuracy is in Davidian, op. cit., p. 334. See also XI-A2.
23
See Neugebauer, Astronomische Papyri; idem, Early Greek Astronomy, pp. 243-246; and esp. idem,
HAMA, II, pp. 737-746.
For the Ethiopic tradition, based on Hellenistic schemes, with numerous absurdities (e.g., 70 instead of 0
for the twelfth hour, resulting from the misreading of the Greek zero-sign meaning no shadow as an omicron,
that is, 70), see idem, Ethiopic Astronomy and Computus, pp. 209-215.
See also Kennedy, Overview of the History of Trigonometry, esp. pp. 5-6, for the first comparison of some
of the approximate rules discussed in the present study.
The absurdity of attempting to analyze these purely arithmetical schemes from a modern astronomical point
of view is well illustrated in Bremner, The Shadow Table in Mul.Apin. In this 7th-century B.C.E. source,
472 part iii, chapter one
The midday shadows are usually symmetrically arranged so that seven values suffice to
define them for the whole year, thus for the months of the Julian calendar (type M):
XII XI/I X/II IX/III VIII/IV VII/V VI
or for the zodiacal signs between the winter and summer solstices (type Z):
/ / / /
/
The midday shadow-scheme:
2 3 4 5 6 7 8
th th
appears to be of early (5 or 4 century B.C.) Greek provenance. Later modifications include
inter alia:
2 3 4 6 8 10 13 and 1 2 3 4 5 7 8,
as well as:
0 11/2 3 41/2 6 71/2 9 .
Also attested are pairings of all the months or signs to yield six rather than seven significant
values. The sources usually do not specify the length of the gnomon casting the shadows or
the locality for which the values are intended. For the hours it is assumed that the shadows
increase at the same rate throughout the year, thus:
hours of daylight: 5/7 4/8 3/9 2/10 1/11
increase each hour: 1 2 3 4 10 (=1+2+3+4)
cumulative increase: 1 3 6 10 20
Each of the classical and medieval traditions was remarkably prone to errors by those who
conceived of the tables and yet more by those who copied them, but this feature by no means
lessens the interest of the material for the history of science. In his monumental History of
Ancient Mathematical Astronomy Neugebauer was able to cite just one example of a shadow-
scheme from a published Islamic source,24 and, as might be expected, the unpublished sources
offer many more.
a time function T in degrees (or units of 4 minutes) is tabulated for each cubit of the shadow z from 1 to 10.
The values of T(z) are simply T1/z where T1 are the times for z=1 at the equinoxes (75) and the solstices (60
in summer and 90 in winter).
24
Neugebauer, HAMA, II, p. 743 (on the scheme listed in 6.3 below).
25
The various calendars in use in the medieval Islamic world are discussed, for example, in Adolf G.
Grohmann, Arabische Chronologie, Leiden: E. J. Brill, 1966; the article Zamn by Willy Hartner in EI1;
Kennedy, Zj Survey, p. 139; and the article Tarikh iv by Benno van Dalen in EI2.
introduction 473
b) midday shadow-schemes for the months based on Greek schemes, also attested in Coptic
sources, and modifications thereof (7.2, etc.);
c) shadow tables for the hours and for each month based on Greek models (6.2, 6.7, 9.7a,
etc. see also 4.9a on a possible Himyaritic scheme);
d) midday shadow-schemes for each 13-day period of the year (anw}) (4.1, etc.);
e) simple lists of midday shadows for a given locality, derived by observation (3.2, 4.1, etc.);
f) simple lists of midday shadows derived by calculation (for given latitude and obliquity)
(5.1, etc.); and
g) simple shadow tables for the hours and for each zodiacal sign or for each few days of the
year based on calculation with an approximate trigonometric formula (only in 4.2).26
26
See I-2.5 and 4.3, for examples of more extensive tables based on this formula, and also XI-3 and 4.
474 part iii, chapter one
Fig. 1.4a: Solstitial and equinoctial shadows for a tropical locality such as the Yemen (here, specifically Taiz).
equinoxes is = 90 - . Thus the shadow lengths at the winter solstice, equinoxes, and summer
solstice are respectively:
n cot ( - ) , n cot , and n cot ( + ) .
Fig. 1.4a shows the solstitial shadows for a tropical locality ( < ) such as the Yemen or Mecca:
these vary between almost a full gnomon length north at midwinter to a small fraction of the
gnomon length south at midsummer. For the convenience of the reader, I present in Table 1.5
the approximate shadow lengths at the solstices and equinoxes for the standard gnomon lengths
7 and 12 for latitudes corresponding to the Yemen, Mecca, Cairo, Aleppo, Cordova and Istanbul
(assuming = 231/2). Values are rounded to the nearest digit, as in most of our sources. It
is not always sensible to try to establish the latitude for which a given scheme might have been
computed.27
Table 1.5
Latitude Midday altitudes Midday shadows
n=7 n = 12
WS EQ SS WS EQ SS WS EQ SS
15 511/2 75 981/2 6 2 1(S) 10 3 2(S)
211/2 45 681/2 92 7 3 0(S) 12 5 0(S)
30 361/2 60 831/2 9 4 1 16 7 1
36 301/2 54 771/2 12 5 2 20 9 3
38 281/2 52 751/2 13 5 2 22 9 3
41 251/2 49 721/2 15 6 2 25 10 4
27
See Neugebauer, HAMA, II, p. 739, on the problems.
introduction 475
In our sources fractions are invariably given as combinations of unit fractions of the form
n (n < 9) so that, for example (see Tables 4.1 and 4.3), /12 is expressed as nif suds, that
1/ 1
is, /2 of /6, and /6 is expressed as nif wa-thulth, that is, 1/2 plus 1/3. The one exception to
1 1 5
the unit fractions is 2/3 (Arabic has a dual form). This is a feature of simple arithmetic in
medieval Islam, reminiscent of the practice in ancient Egypt, and it persisted not least because
of the fact that Arabic has no names for unit fractions with denominator greater than ten.28
Occasionally numbers with fractions are written sexagesimally in the standard manner of the
astronomers; a number rendered in the form a b in the texts is rendered a;b in the
commentary, and this means a + b/60.
See Neugebauer, ESA, pp. 74-77, and the article {Ilm al-isb by Abdelhamid I. Sabra in EI2, esp. III,
28
pp. 1140-1141.
476 part iii, chapter one
CHAPTER 2
EARLY SOURCES
There are various references to shadow lengths in the statements attributed to the Prophet
Muammad (known as adth). I shall not attempt to give a survey of these; rather, I shall
restrict my attention to just two examples.1
First, the Prophet is reported to have said that the angel Gabriel led him in prayer on two
consecutive days. On the first day they prayed the uhr when the shadow was equal to the
width of a thong of a sandal and the {ar when the shadow of every object was the same as
its length. These situations correspond to z = ( << n) and z = n. On the second day they
prayed the uhr when the shadow of every object was the same as its length and the {ar when
the shadow of every object was twice its length. These correspond to z = n and z = 2n.
As pointed out by al-Brn, these definitions in terms of shadow lengths could not be applied
universally.2 Only in tropical latitudes can the midday shadow be relatively small, and in non-
tropical latitudes the solar altitude does not always attain the 45 necessary to ensure that
sometime in the afternoon the condition z = n will be satisfied. The definitions for the uhr
in terms of shadow increases that were adopted in later practice, namely:
z = 1/12 n , 1/5 n or 1/4 n ,
ensure that the shadow has increased by an observable amount. The definition for the {ar in
terms of shadow increases which was adopted by three of the four major legal schools, namely,
z = n, and that adopted by the fourth (the anafs), namely, z = 2n, can be applied in any
habitable latitude. The adth in a sense justifies both of these amounts n and 2n for the
{ar. Yet other evidence indicates that the reason for adopting these definitions was to regulate
the prayers by the seasonal hours. The formula:
T = 6 n / (z + n)
yields the following shadow increases at the seasonal hours:
T 0 1 2 3 4 5 6
z 5n 2n n 1/ n 1/ n 0
2 5
so that the definitions z = n for the {ar (and occasionally for a prayer at mid-morning called
the u) represent practical means of linking the {ar to the beginning of the tenth hour (and
the u to the beginning of the fourth). Likewise, the definitions:
z = 1/12 n , 1/5 n or 1/4 n
are practical means of relating the uhr to one-half or one hour after midday.
1
See IV-1 and 4.1-2, for more information on these two adth. For references to shadows in the adth
literature see n. 1:3.
2
al-Brn, Shadows, p. 168, and Kennedy, al-Brns Shadows, I, p. 218, and II, pp. 137-138.
early sources 477
The second report about the Prophet states that he prayed the uhr when 5 < z < 7 in winter
and when 3 < z < 5 in summer. Since the winter midday shadow in Mecca and Medina is about
71/2 (n = 7), this cannot be a description of a daily routine.
al-Aba (see 4.1) relates a tradition (MS Cairo DM 948, fol. 41r) on the authority of Yay
ibn Sa{d3 about the fourth Caliph {Al ibn Ab lib.4 The latter to whom numerous
achievements of a mildly scientific nature are attributed in Islamic legend is reported to have
said that the day on which the midday shadow at Basra was 61/2 feet (the shadow of any object
being then equal to its length) marked the beginning of winter. I do not know of any other
sources in which winter is thus defined5 (see also 4.1 on a specific date for the winter solstice
as opposed to the middle of winter attributed to {Al). The report continues to list the shadow
lengths at each eighth of the year; these are shown in the following table:
Table 2.2
Season Days elapsed Midday shadow
Beginning of winter 0 61 / 2
midwinter 45 9 - 1/ 8
end of winter 46 61 / 2
middle of spring 45 31 / 2
beginning of summer 46 11/ 2+ *
midsummer 45 1
/2+
end of summer 47 (leap year: 48) 11 / 2 +
midautumn 45 31 / 2 +
end of autumn 46 61 / 2
Since for Basra, 30, the shadow (base 61/2) actually varies between
83/4 (WS) , 33/4 (EQ) , and 3
/4 (SS),
and so the values given in the text are reasonable. Note that at the latitude of Mecca the midday
shadow at the winter solstice rather than the middle of winter is equal to the length of
the gnomon (see Table 1.5 and Section 3.3).
Ibn Raq reports that, according to {Al, (the end of) the ninth hour of daylight marks the
beginning of the {ar prayer see 3.2.
3
On possible candidates with this name, both having a connection with al-{Irq, see Sezgin, GAS, I, p. 407
(on Ab Sa}d Y. b. S. al-Anr, a tbi{ from Medina, d. 760 in Basra (!)), and p. 293 (on Ab Ayyb Y.
b. S. al-Umaw, b. ca. 732 in Kufa, d. 809 in Baghdad, author of a work on the campaigns of the Prophet).
4
On {Al ibn Ab lib see the article by L. Vecchia Vaglieri in EI2. On various achievements in the sciences
attributed to him see Sezgin, GAS, VII, pp. 8, 10, 11, 26, and 184; Cairo ENL Survey, no. B1; King, A Medieval
Arabic Report on Algebra before al-Khwrizm and n. IV-2:8.
5
See, however, Lane, Lexicon, V, p. 1916a, where the word ill, shadow, is associated with the beginning
of winter.
478 part iii, chapter two
2.3 al-Fazr
The extant fragments of the writings of the mid-8th-century astronomer al-Fazr, now gathered
and interpreted by David Pingree, display the eclectic nature of early Islamic astronomy.6 For
timekeeping al-Fazr advocated:
T 6 n / (z + n) (n = 12),
a formula which was to be extremely influential in later Islamic folk astronomy and in the
formulation of the practical definitions for the times of the daylight prayers which are still in
use to this day (see 2.1, 3.2, 4.1, 6.4, 6.7, 8.2, 9.2, etc.).
2.4 al-Brn
Two numerical shadow-schemes are presented by al-Brn in his treatise On Shadows.7 The
account of the first is garbled, and it was interpreted by E. S. Kennedy thus:
T 1 2 3 4 5 6
[z-12] 60 24 12 2;50 0 ?
[z 72 36 24 14;50 12 0]
However, it seems more likely that the text should be restored to yield a scheme as follows:
T 1 2 3 4 5 6
[z] 60 24 12 [6] 2
[2 /5] 0
This emendation requires only inserting a value 6 for the fourth hour and reading wa-khumsayn
(= 2/5) rather than wa-khamsn (= 50) after the 2 for the fifth hour. Now the scheme corresponds
precisely to the Indian formula for n = 12.
The second shadow-scheme mentioned by al-Brn is also Indian. It relates the shadow
increase to the passage of time in muhurtas, one-fifteenth divisions of the time between sunrise
and sunset. The values given are as follows:
T 1/14 2/13 3/12 4/11 5/10 6/9 7/8 71 / 2
z 90 60 12 6 5 3 2 0
The same scheme is described in al-Brns India, with 96 for the first entry.8
6
On al-Fazr see the article by David Pingree in DSB. For his use of this formula see Pingree, al-Fazr,
pp. 121-122.
7
On al-Brn see the article by E. S. Kennedy in DSB. On this passage see al-Brn, Shadows, pp. 158-
159, and E. S. Kennedys translation and commentary, I, pp. 195-196, and II, pp. 121-122, as well as idem,
Overview of the History of Trigonometry, esp. pp. 5-6.
8
Sachau, Alberunis India, pp. 338-339.
hejazi sources 479
CHAPTER 3
HEJAZI SOURCES
Ab {Abdallh Muammad ibn Raq ibn {Abd al-Karm wrote a treatise on the calendar, lunar
mansions, and simple time-reckoning, which is extant in the unique MS Berlin Ahlwardt 5664
(Landberg 108), 71 fols., copied ca. 1350. The work appears to have been compiled in Mecca
in the 11th century.1 Of particular relevance to the present study are, first, a set of midday
shadows at Mecca for each of the Coptic months and, second, a scheme for reckoning the
seasonal hours by shadow increases.
Ibn Raq quotes Ab {Al {Araqa, a muezzin at the Mosque of {Amr in Fustat, as having
stated that Amad al-Rz, a muezzin at the Sacred Mosque in Mecca, had observed the sun
over a period of seven years and had determined the shadow lengths for the Coptic months
starting with B}na (X), expressed in terms of the height of a man (qma). His results (fols.
12r-12v) are as follows:
Table 3.2a
Month Shadow
X 1
/3 1/6 (S)
XI 0
XII 1
/6
I 1
/ 3 + ( 1 / 2 1/ 6) a
II 1
/2
III 2
/3 + 1/ 4
IV 1 + (1 /6 1 /8 )
V 1 - (1 /2 1 /8 )
VI 1
/3 + 1/ 4
VII 1
/4 + 1/ 8
VIII 1
/6
IX 0
a
text has 1/3 + 1/2 + 1/6
Since the latitude of Mecca is about 211/2, the meridian altitudes at the winter and summer
1
On Ibn Raq see King, Astronomy in Yemen, no. 2, also II-2.3 and IV-1.5. His treatise, as well as those
of al-Aba and al-Fris (nn. 4:1 and 4:4), are currently being investigated in Schmidl, Islamische
volksastronomische Abhandlungen.
480 part iii, chapter three
solstices, measured from the south point, are 45 and 92, and one would expect maximum
shadows in the north of about 1 qma and in the south of about 1/30 qma. I have no information
on the date of al-Rz.
Elsewhere in his treatise (fols. 28v-29r see Fig. IV-2.1) Ibn Raq presents the scheme:
T 1 2 3 4 5
z 39 19 /2
1
6 /2
1
3 /4
1
1+ ,
and he uses gnomon length 61/2. (The value for T = 5 is given as qadam rji.) If we use the
standard Indian formula for T = 5, 4, and 3 seasonal hours, we obtain for this value of n:
z = 13/10 , 31/4 , and 61/2 ,
as in the text. When T = 2 and 1, z is large compared with n, and if we apply the very crude
approximation:
T 6n / z ,
we obtain for these values of T and n:
z = 191/2 and 39 .
Each of Ibn Raqs values is thus explained. The above approximation is nowhere attested,
explicitly or implicitly, in any of the other Islamic sources known to me.
Ibn Raq adds a remark that (the end of) the ninth hour marks the beginning of the {ar
prayer according to {Al ibn Ab lib (see 2.1). It was this remark that first led me to investigate
the standard definitions of the times of the daylight prayers in Islam and find that they were
practical means of regulating the prayer-times in terms of the seasonal hours.
The last part of Ibn Raqs treatise (fols. 40r-71v) consists of an almanac arranged according
to the lunar mansions rising at daybreak. For each 13-day period, various kinds of information
are given, including the midday shadow lengths at Mecca. Part of the almanac is missing (some
20 fols. between fols. 43 and 44), and the shadow lengths that I have been able to extract are
shown in Table 3.2b.
3.3 al-Shill
MS Cairo DM 130,2 (fols. 14v-15v, copied ca. 1800 see Fig. 3.3) contains a short treatise
by Muammad ibn Ab Bakr ibn Amad al-Shill B {Alaw, a scholar from the Hadramawt
who worked in Mecca in the 17th century.2 al-Shill discusses the disagreements between
scholars about the midday shadow lengths throughout the year at Mecca and Medina and then
presents his own opinion on the shadows at Mecca, whose latitude he takes as 21.
In brief, al-Shill states that the sun is overhead at Mecca on the fifth day of the sign of
Gemini and the twenty-sixth day of the sign of Cancer.3 The sun is overhead (in tropical
localities, i.e. where < ) when its declination equals the local latitude. For solar longitude
65 and = 23;35 (the standard value, which I would have expected al-Shill to have used)
I compute the declination as 21;16; for longitude 64 the corresponding value is 21;4, which
2
On al-Shill see Brockelmann, GAL, II, p. 502, and SII, p. 516; Cairo ENL Survey, no. D47; and King,
King, Astronomy in Yemen, p. 59.
3
Cf. the discussions in Lee, Astronomical Tables of Al Farsi (1822), p. 265 (Sanaa); and King, al-Bazdaw
on the Qibla in Transoxania, p. 19 (Samarqand).
hejazi sources 481
Fig. 3.3: The full text of al-Shill dealing with shadow-lengths. [From MS Cairo DM 130,2, fols. 14v-15v,
courtesy of the Egyptian National Library.]
482 part iii, chapter three
is close to his value for the latitude of Mecca. However if we compute with longitude 64 and
obliquity 23;30 (rounded from the value of Ulugh Beg) the declination is exactly 21, which
is very nice. (Had al-Shill taken the trouble to measure the latitude he would have no doubt
been surprised to find that it is closer to 21;30.) He goes on to say that the small increase
(in midday shadow) (al-ziyda al-ughr) occurs between these two days and reaches a
maximum of 25 minutes (i.e. sixtieths of a foot, where the gnomon is 7 feet long). He states
that the daily increase in the shadow is 1 minute. The large increase (al-ziyda al-kubr)
occurs at the winter solstice, and no value is given. al-Shill states that the daily shadow changes
in minutes in the signs are as follows:
/// : 2 /// : 3 /
: 4 ,
and from this one could infer that the shadow at the winter solstice was about:
4 1 + 30 2 + 30 3 + 30 4 + 30 3 + 30 2 = 424 minutes,
that is, about 7 feet (see 2.2). This is reasonable because the solar meridian altitude is close
to 45 at this time. The author promises a table displaying the midday shadow lengths at Mecca
for (solar longitudes in) each degree of each zodiacal sign, but there is no table appended to
the treatise in the Cairo manuscript.
3.4 Tj al-Dn
MS Damascus hiriyya 5588 (16 fols., copied ca. 1900) contains some prayer-tables for
Medina by Tj al-Dn, muwaqqit at the Sacred Mosque there. This individual is otherwise
known only by a commentary on the almanac of Shaykh Vef for Istanbul (ca. 1475), of which
the unique Cairo copy is dateable to ca. 1700.4
One of Tj al-Dns tables (fol. 14v) displays midday shadows in feet (aqdm) and tenths
thereof (a{shr) for some days of the solar year for the latitude of Medina .... The entries
beginning with Knn II (= January) are shown in the following table.
Table 3.4
Month I II III IV V VI
Day
1 7 0 5 3 3 5 1 9 0 8 0 2
10 6 6 4 6 3 0 1 5 0 5 0 1
20 6 0 4 1 2 5 1 2 0 3 0 2
For = 25 the maximum shadows for bases 7 and 61/2 are 7.9 and 7.3. For = 24, a more
common medieval value for Medina, the shadows are 7.6 and 7.1. The use of decimal fractions
is not attested elsewhere in Islamic tables except in the trigonometric tables of the 16th-century
Istanbul astronomer Taqi l-Dn.5
4
Cairo ENL Survey, nos. D239 and H2.
5
King, Islamic Multiplication Tables, B, pp. 413-415.
hejazi sources 483
In the unique copy of the prayer-tables for Medina by Tj al-Dn (see 3.4) there is an
introduction by an individual named al-Amrn or al-Imrn, who is new to me. In this (fol.
8r, also 2r) we find the following shadow-scheme for the uhr:
9 7 5 3 2 1 1 2 3 5 8 10
advocated for the (Syrian ?) months, with 7 to be added for the {ar. Elsewhere (fol. 4r), the
following scheme is given for the hours:
1 2 3 4 5 6 7 8 9 10 11
28 18 9 6 3 0 3 6 9 18 28
as in the treatise of Ibn Tmart (see 6.1). The units are aqdm, but no indication is given whether
these are shadow lengths or shadow increases.
484 part iii, chapter four
CHAPTER 4
YEMENI SOURCES
4.1 al-Aba
Ibrhm ibn {Al al-Janad al-Aba worked in Janad in the Yemen in the mid 13th century
and compiled an extensive treatise on folk astronomy.1 This is extant in several copies, notably
MS Baghdad Awqf 2982/6276, apparently executed in Taiz in 680 H [= 1281/82]. I have used
MS Cairo DM 948,1, copied in 1320 H [= 1902/03] but nevertheless reliable.
al-Abas treatise is a valuable source of information on Islamic folk astronomy. Of
particular interest to the present study is the fact that he advocates (fols. 38r-38v) the shadow-
scheme:
T 1 2 3 4 5
z (in multiples of n) 5 2 1 1
/2 1
/5
which corresponds precisely to the Indian rule. See also 2.1, 2.2, 3.2, and 6.4.
al-Aba concludes his treatise with the statement (fol. 51v) that it is the custom of those
astronomers who author works (on folk astronomy) to conclude their treatises with a statement
on the (midday) shadow at the location where they work. He therefore adds to his treatise a
list of shadow lengths observed by him at 13-day intervals throughout the Syrian year and he
specifically states that he made the observations in the Mosque of Janad in the year 654 H
[= 1256]. The values are displayed in Table 4.1. al-Aba neglects to mention the length of
the gnomon, but for the latitude of Janad (about 13;40) his solstitial values correspond best
to 61/2. His values are expressed in qadams and occasionally also a smaller unit qr, not usually
used as a unit of length (see also Table 4.5).2 Some of the values (indicated by an asterisk
in Table 4.1) are labelled sh, a term I have not encountered elsewhere but which seems
to mean just less than.3 The first value is for XII 15, which al-Aba says is the date of
the winter solstice according to {Al ibn Ab lib (see 2.2).
1
On al-Aba see King, Astronomy in Yemen, no. 5, and IV-1.5, etc. His treatise is currently being
investigated by Petra Schmidl: see n. 3:1.
2
On the use of qr as a unit of weight see Hinz, Islamische Masse und Gewichte, p. 27, and the same authors
article ^r in EI2. For another attestation of the term as a unit of length, this time in a 16th-century treatise
on timekeeping, see Wiedemann, Schriften, II, p. 1039. Wiedemann, on what authority it is not clear, states that
it is equal to a fingers breadth.
3
Thus, for example, the value for XII 15 is khamsat aqdm ill thumn sh, and the value for X 11 is
thalthat aqdm sha. The verb shaa means, amongst other things, to become short, to run out, to decrease,
to dwindle. The term sh thus has the opposite meaning of the more common term rji, just more than
(see, for example, III-3.2 and IV-7.5). Since sometimes the adjective sh is applied only to a negative
fractional part of the entry I have preferred not to use the notation n- in Table 4.1.
yemeni sources 485
Table 4.1
Date Mansion Midday Shadow
XII 15 (WS) 21 5 - 1 /8 *
XII 28 22 5 - 1 /3 *
I 10 23 41 / 3
I 23 24 3 + / 2 + 1/ 4 + q
1
II 5 25 31 / 4
II 18 26 2 + 1 /2 q
III 3 27 2 - ( 1 / 2 1 / 6) +
III 16 (VE) 28 11/ 3 +
III 29 1 1 - 1 /6 +
IV 11 2 0
IV 24 [3] deest
V7 4 1 - 1/6 + q [S]
V 20 5 1 - (1/2 1/6)* [S]
VI 2 6 11/2* [S]
VI 15 (SS) 7 1 + (1/2 1/6) + [S]
VI 30 8 1 + (1/2 1/6)* [S]
VII 12 9 1 - 1/6 + [S]
VII 25 10 1
/3 + (1/2 1/6)* [S]
VIII 7 11 0
VIII 20 12 1
/2
IX 2 13 1 + (1 /2 1 /6 )
IX 15 (AE) 14 1 + 1 /3 - q
IX 28 15 2 + 1/ 6
X 11 16 3*
X 24 17 31 / 2
XI 6 18 4 + (1 /2 1 /6 )
XI 19 19 4 + 2 /3 - (1 /2 1 /8 )
XII 2 20 5 - 1 /8 +
4.2 al-Fris
The Yemeni astronomer Muammad ibn Ab Bakr al-Fris,4 author of a zj compiled for the
Rasulid Sultan al-Muaffar ca. 1295, also wrote a treatise on folk astronomy entitled Tufat
al-rghib. This is extant in the unique complete copy MS Milan Ambrosiana Griffini 37
(unfoliated, copied ca. 1500).
In this work, which was compiled in Aden, al-Fris presents the following shadow-scheme
for the hours, written out in words (see further below):
Table 4.2a
T 1 2 3 4 5 6
Sign
/ 26 12 7 4 21 / 2 1
/ 28 13 8 5 3 21 / 2
/ 31 15 9 6 5 4
32 151/2 10 7 51 / 2 5
4
On al-Fris (Suter, MAA, no. 349 and 349N and Kennedy, Zj Survey, no. 54) and his works see King,
Astronomy in Yemen, no. 6. Another anonymous, incomplete copy of the Tufat al-rghib is MS Berlin Ahlwardt
5731. This treatise is currently being investigated by Petra Schmidl: see n. 3:1.
486 part iii, chapter four
He then states that an approximate scheme for the shadows at the hours throughout the year
is:
28 17 15 7 3 0.
As can be seen from comparing these values with those in Table 4.2a, this scheme is
approximate indeed.
Elsewhere his work al-Fris presents the following values of the midday shadows for
various solar longitudes, also written out in words; he specifically mentions the town of Aden,
for which he uses latitude 13:
Table 4.2b
Sign Shadow Direction
/ 1 + 1 /4 + 1 /5 + 1/ 6 N
/ 1
/6 + 1/6 1/10 N
10 / 25 0 -
/ 4
/5 + 1/2 1/ 6 S
1 + 1 / 4 + 1 / 3 1/ 5 S
/ 3 + 1/6 + 1/6 1/10 N
/
4 + 1/2 + 1/10 N
0 5 + 1/ 5 N
In a section on the prayer times al-Fris uses z = 7 and 14 to define the beginning and
end of the {ar, so that he is using n = 7. Now if we take al-Friss parameters:
= 13 and = 23;35,
the meridian altitudes at the summer solstice, equinox, and winter solstice measured from the
south point are approximately
1001/2 77 531/2
and the corresponding shadows to base n = 7 are:
1;18 S 1;37 N 5;11 N .
These correspond quite closely with al-Friss values in Table 4.2b, these being equivalent
to:
1;19 S 1;37 N 5;12 N .
To compute the shadows at the seasonal hours si al-Fris has used a simple approximation
to find the corresponding altitudes hi equivalent to:
sin hi = sin H sin ( i 15 ) ,
a procedure derived originally from Indian sources and used by certain Muslim astronomers.5
For the solstices, using = 13 and n = 7, I compute the values:
26.6 12.4 7.2 4.3 2.3 1.3
32.9 15.9 10.1 7.2 5.7 5.2
so that al-Fris has in general truncated the results of his calculations.
5
See n. 1:21. The history of this formula, as well as various tables and instruments based on it, in both the
Near Eastern and the European traditions, is traced in XI. See already Goldstein, Ibn al-Muthann on al-
Khwrizm, pp. 82-83 and 207-209 (on al-Khwrizms use of the formula), and King, al-Khwrizm, pp. 7,
11.
yemeni sources 487
The Rasulid Sultan al-Ashraf (reg. 1296-97) compiled two major works on astronomy. The
first was a lengthy treatise on the construction of astrolabes and sundials and also on the use
of the magnetic compass. The second was an equally lengthy volume on astrology, which also
contains tables for timekeeping for Sanaa and an agricultural almanac.6
This second work, entitled Kitb al-Tabira f {ilm al-nujm, survives in the unique MS
Oxford Hunt. 233, copied ca. 1400, and in the almanac on fols. 97r-108v, shadow lengths for
the uhr and the beginning and end of the {ar are given for each half-month of the Syrian
year. The values are written out in words at the side of the almanac; they are reproduced in
the following:
Table 4.3
Notes: All fractions are additive unless otherwise stated.
An asterisc denotes that an approximation has been made in a particular set of entries.
A double asterisc denotes that there is an error in the calculation of either of the entries for the {ar.
Month uhr {ar 1 {ar 2
I 5 11 2/3 18 2/3
4 1/ 2 11 1/6 18 1/6
II 3 1 /2 1/ 4 10 1/4 1/6 17 1/4 1/6
3 9 2/ 3 16 2/3
III 2 1/ 4 8 2 / 3 1/ 4 a 16 2/3 1/4
1 1/ 2 8 15 1/5 **
IV 1
/ 2 1/ 4 7 1/3 1/10 14 1/3 1/10 **
1
/4 6 2 /3 1/ 4 13 2/3 1/4
V 1
/3 7 14
1
/2 7 1 /4 1/ 6 14 1/4 1/6 **
VI 1 1/2 1/10 7 2/ 3 14 2/3 *
1 1/10 7 1 /2 1/ 4 14 1/2 1/4 *
VII 1 7 2/ 3 14 2/3
1
/ 2 1/ 4 7 1/ 3 14 1/3 **
VIII 1
/5 6 1 /2 1/ 3 13 1/2 1/3 *
1 7 14 **
IX 1 1/2 1/10 7 14 2/3 b **
1 1 /2 1/ 4 8 2/ 5 15 2/5 **
X 2 1 /3 1/ 4 9 1 /4 c 16 1/4
3 1/ 3 10 17
XI 4 1/ 6 10 1/2 17 1/2 1/3 **
4 1 /2 1/ 4 11 1/4 1/6 18 1/4 1/6
XII 5 1/ 6 11 1/2 1/3 18 1/2 1/3
5 1/ 4 11 2/3 1/4 18 2/3 1/4
a
text: 3 /3 /4; text: fraction partially illegible (2 for numerator is clear); text: 6 /4
2 1 b c 1
We notice immediately that, apart from a few calculating errors, the entries for the end of the
{ar are larger than those for the beginning of the {ar by 7, and that the latter are larger than
those for the uhr by 62/3. This is curious indeed, and I have no explanation to offer. Elsewhere
in his treatise al-Ashraf presents a table of cotangents to base 62/3, and in his timekeeping tables
he uses obliquity 24 (14;30 for the latitude of Sanaa). For these parameters the maximum
On al-Ashraf (Suter, MAA, no. 394) and his astronomical works see King, Astronomy in Yemen, no. 8, and
6
King, Yemeni Astrolabe. His shadow-scheme was brought to my attention by Daniel Varisco, whose detailed
study of al-Ashrafs almanac is now published as Varisco, Yemeni Almanac. See also n. 4:12.
488 part iii, chapter four
midday shadow is 5;18, which corresponds quite well to the maximum value in the table, viz.
51/4. Note that the fact that the midday shadows around the summer solstice are southerly is
overlooked in the computation of the shadows at the {ar.
4.5 B Makhrama
A late Ottoman navigational manual (see 10.2) contains a table of shadow lengths, said to be
{al manzil al-Shibm al-shaykh {Abdallh ibn {Umar B Makhrama, that is, in accordance
with the lunar mansions as described by ... B Makhrama. This individual was a Yemeni
scholar active ca. 1500, and is otherwise known by a treatise on folk astronomy entitled al-
Shmil, which survives in a unique manuscript.8 I have not noticed the shadow table in this
treatise. The epithet al-Shibm indicates his familys provenance from Shibm in the
aramawt.9
In the table the 28 mansions are divided in four columns between the seasons (al-Han{a
= M5 = first mansion of spring), and the shadow lengths in qadams at midday and the {ar
are written out in words. The difference is invariably 7, which is clearly the value used for
n. The entries are displayed in the following table:
Table 4.5
M Z M Z
6 5 + 1/ 8 20 1 - 1/20
7 5 - 1/ 6 21 1 - 1/ 5
8 4 + 1/ 3 22 1
/2
9 4 - 1/ 5 23 1
/8 e
10 3 + 1/ 8 24 1
/3
11 2 + 1/2 - 1/30 a
25 1 - 1/ 6
12 2 - 1/ 6 26 1 + 2/ 5
7
On the astronomical activity of al-Afal see King, Astronomy in Yemen, no. 18, and II-12.4.
8
King, Astronomy in Yemen, no. 26.
9
Another Shibm of renown in Yemeni folk astronomy is listed ibid., no. 43.
yemeni sources 489
13 1 + 1/ 5 27 2 + 1/30
14 2
/3 28 2 + 2/ 3
15 1
/8 b
1 3 + 1/ 3
16 1
/ 4 + q - 1/ 4q c 2 4
17 1
/3 - 1/20 d 3 4 + 1/2 + 1/30
18 1 - 1/ 8 4 5 - 1/20
19 1 - /301
5 5f
a
text: qadamayn wa-nif ill thulth {ushr; b text adds remark: on the sixth day of this mansion the midday
shadow is zero; c here q = qr (text: rub{ qadam wa-qr ill rub{) see 4.1; d for the {ar: 8 - 1/3 of 1/2
of 1/10; e for the {ar: 7 + 1/3; f for the {ar: 11 + 1/6
Notice that the extremal values are 5 + 1/8 and 1 - 1/30 (S), which are mutually inconsistent.
They correspond most closely to latitude 13, commonly used for Aden: for this latitude the
approximate values are 5 + 1/6 and 1 + 3/20 (S).
In some notes appended to MS Cairo DM 948 (fols. 101v-102v) there are some values of
the midday shadow (base not stated) at Aden, given for each ten days of the Syrian year. These
are specifically attributed to B Makhrama. They are followed by an incomplete, anonymous
set for Taiz (with n = 61/2), also for each ten days of the year (fols. 102v-103r).
4.6 al-awsh
A Yemeni treatise on folk astronomy entitled Mift al-asrr f {ilm al-falak al-dawwr is
attributed in both of the extant manuscripts, MSS Cairo DJ 709,6, (fols. 29r-53v, copied ca.
1750) and Princeton Garrett (Hitti) 1016 (61 B) (copied ca. 1875), to Nr al-Dn {Al ibn
{Abdallh al-awsh, an individual otherwise known to the modern literature.10 The treatise
contains a poem on the midday shadows for the lunar mansions (fol. 35r of the Cairo
manuscript): the values are a simplified version of those given by Ibn Tmart (see 6.1), varying
linearly between a maximum of 51/2 feet for al-Balda and a minimum of zero between al-Haq{a
and al-Jabha.
4.7 al-Thbit
Muammad ibn {Abd al-Laf al-Thbit was a Syrian who lived in Zabid.11 In 1047 H [= 1637/
38] he compiled a set of tables for timekeeping by the lunar mansions and by shadow lengths,
of which I have examined MS Vatican ar. 962 (fols. 13r-19r). The shadow-tables display the
shadow at midday and the {ar: values to base 61/2 are given to two digits, the second called
bann, fingers, representing twelfths of a shadow digit. The midday shadows are computed
according to a linear zigzag-scheme, with a maximum of 55/12 at the winter solstice and a
maximum of 22/12 (south) at the summer solstice. These extremal values are equivalent to 10
and 4 digits for a gnomon length of 12 digits, which suggests that al-Thbit converted a set
of values from base 12 to base 61/2. In fact, for = 15 and base 12 the extremal values are
10
Ibid., no. 27.
11
ibid., no. 33, and II-12.7.
490 part iii, chapter four
closer to 10 and 2 digits. al-Thbits afternoon shadow lengths are always 61/2 feet more than
the absolute value of his midday shadow lengths.
In MS Cairo DM 899,2 (fols. 9r-9v, copied ca. 1700), appended to an abridged version of
the treatise of al-Aba (4.1) by Ab amad Maf ibn {Abd al-Ramn al-aram, there
is a list of midday shadows to base 61/2 written out in words for days in Syrian calendar,
specifically intended for Zabid. The values are the same as those given by al-Thbit. Values
are also given for the {ar and are all correctly 61/2 feet more than the midday values.
On fol. 97v of MS Cairo DM 948 (see 4.1), amidst some notes on timekeeping, there is a
reference to a commentary (shar) by one Ab Shukayl (?), not otherwise known to me. In
this work the midday shadow is given as zero on Ayyr 1; it has a maximum (southerly) value
2 1/6 on azrn 22, and a maximum (northerly) value 5 1/4 + 1/6 on Knn I 22. No locality
is specified, but the extremal values are the same as those of al-Thbit (see 4.7).
(a) In MS Oxford Marsh 134, copied 855 H [= 1451/1452], of the treatise of folk astronomy
by al-Aba (4.1), there is a shadow-scheme amidst some notes at the end of the treatise (fol.
52v). The scheme is presented in the form of a circle and values of the midday shadow lengths
are given for each of the Syrian months; the corresponding names of the Himyaritic (ancient
South Arabian) months are also given.12 The scheme is simply:
5 4 3 2 1 0 1 2 3 4 5 6,
and is probably very old. One may wonder whether the scheme was known in the Yemen before
the advent of Islam. In the view of what has come to light of Graeco-Roman influence in South
Arabia,13 I suspect that it was indeed known there. The only other Islamic source in which I
have come across such a scheme is from in Kayf (now in Southern Turkey), where one
could presume Byzantine influence see 7.2.
(b) MS Aleppo Awqf 968 (16 pp., copied ca. 1600) contains an anonymous treatise entitled
Kitb al- al-shf bi-l-itqn f ma{rifat al-manzil wa-l-azmn. The author mentions the
Qms of Majd al-Dn (al-Frzbd) so that he must postdate ca. 1400. He also mentions
the Kitb al-Yawqt by al-Janad, otherwise known as al-Aba (4.1). Two shadow-schemes
are presented for the seasonal hours. The shadows (in feet), not shadow increases, are as
follows:
12
A. F. L. Beeston, New Light on the Himyaritic Calendar, Arabian Studies 1 (1974), pp. 1-6; George
Saliba, A Medieval Note on the Himyarite Calendar, JAOS 105 (1985), pp. 715-717, based on MS Oxford
Hunt. 233 of al-Ashrafs Tabira (4.3); and Varisco, Yemeni Almanac, pp. 64-71, etc.
13
See, for example, A. H. Masry, An Introduction to Saudi Arabian Antiquities, Riyadh, 1975; A.R. al-Ansary,
Qaryat al-Fau. A Portrait of Pre-Islamic Civilisation in Saudi Arabia, Riyadh, 1982, and Annie Caubet, Aux
sources du monde Arabe: LArabie avant lIslam Collections du Muse du Louvre, Paris: Institut du Monde
Arabe & Runion des muses nationaux, 1990.
yemeni sources 491
38 18 9 6 3 0
35 15 9 5 3 0
The numbers are displayed in the text in Arabic numerical notation.
(c) In an anonymous Egyptian source MS Cairo Dr al-Kutub K 3769, (fols. 10v-11r, copied
1910), appended to a late Yemeni treatise on timekeeping, the scheme is:
1 2 3 4 5 6 7 8 9 10 11
26(!) 13 7 5 3 - 3 5 7 13 24(!)
4.10 al-Wsi{
Shaykh {Abd al-Wsi{ ibn Yay al-Wsi{ published a set of prayer-tables for the Yemen in
the year 1939/40.14 Amongst the functions tabulated for each day of the year are the shadow
lengths at midday and the {ar given to base 7 in feet and fingers (as with al-Thbit). The
shadow lengths are computed according to a very primitive scheme. From Taurus 8 to Cancer
18 the midday shadows have a constant value 4/12 feet north instead of south, and different
values are given for the vernal equinox and autumnal equinox, as follows:
VE: 2 SS: 4/12 AE: 1 4/12 WS: 5 .
For = 15 and n = 7 these are actually closer to:
110.5/12 15/12 (north!) 110.5/12 57/12 .
14
On al-Wsi{ see King, Astronomy in Yemen, no. 47. His tables were published in Cairo, Maba{at ijz,
1358 H [= 1939], under the title Kanz al-thiqt f {ilm al-mqt. For an analysis see II-12.12.
492 part iii, chapter five
CHAPTER 5
ANDALUS SOURCES
The earliest known almanac for al-Andalus is the Kitb al-Anw} of Abu l-asan {Arb ibn
Sa{d al-Ktib compiled in the late 10th century. This work, known as the Calendar of Cordova,
was translated into Latin and was transmitted to medieval Europe. The Judaeo-Arabic and Latin
texts have been published by Reinhart Dozy and translated into French by Charles Pellat.1
{Arb ibn Sa{d gives information on the lunar mansions, the duration of day and night and
twilight, as well as midday solar altitudes and shadows (expressed in words in terms of the
gnomon length), for each 13, 14, 15 or 16 days. This information has recently been investigated
by Julio Sams.2 The following table displays the values given for the last two quantities (the
equinoxes are at III 16 and IX 18):
Table 5.1
Date H z n Date H z n
I2 292/3 13 / 4 VII 16 732/5 11
/36
I 16 321/2 13 / 5 VIII 1 692/3 3
/8
II 2 365/6 11 / 3 VIII 16 651/4 1
/2
II 15 401/2 11 / 6 IX 1 55 /4 (sic)
1 7
/12
III 2 465/6 15
/16 IX 15 541/2 2
/3
III 18 521/2 3
/4 X1 472/3 11
/12
IV 2 585/6 7
/12 X8 deest 1
IV 17 634/5 1
/2 X 16 4111/12 11 / 8
V2 deest deest XI 1 361/4 11 / 3
V 14 724/5 1
/3 XI 15 322/5 17/12
VI 2 751/3 11
/40 XII 2 292/3 13 / 4
VI 16 761/3 1
/4
VII 1 753/4 1
/4 XII 16 282/3 15 / 6
The midday solar altitudes are clearly computed, and underlying them are to be detected the
parameters:
= 37;30 and = 23;50 .
The first was no doubt intended to serve Cordova, although it is not attested in any other known
medieval Islamic sources;3 the accurate value is 37;53. The second is an approximation for
1
On the context of all of this Andalus material see Forcada, Anw} Books in al-Andalus. On the Calendar
of Cordova see Pellat, Le Calendrier de Cordoue, pp. xiii-xiv, and, more recently, J. Martinez Garquez and J.
Sams, Una nueva traduccin latina del calendrio de Crdoba (siglo XIII), in Vernet, ed., Textos y estudios,
A, pp. 9-78, Sams, Sobre los materiales astronmicos en el Calendrio de Crdoba y en su versin latina
del siglo XIII, in Vernet, ed., Textos y estudios, B, pp. 125-138; and Forcada, op. cit. See also the article on
the author by Charles Pellat in EI2, and Sezgin, GAS, VII, pp. 355-356.
2
See Sams, op. cit., esp. pp. 126 ff. for details.
3
Kennedy & Kennedy, Islamic Geographical Coordinates, p. 95.
andalus sources 493
Ptolemys 23;51,20. Likewise computed are the ratios of the shadows to the gnomon. The
tabulation of such quantities derived by computation borders on the mathematical tradition of
Islamic astronomy.
5.2 al-Umaw
MS Escorial r. 941 (38 fols., copied ca. 1300) is a unique but incomplete copy of a treatise
on folk astronomy by al-asan ibn {Al al-Umaw al-Qurub, who was born in Cordova in
514 H [= 1120/21] and died in Seville in 602 H [= 1205/06].4 The work is divided into two
main parts, both incomplete, the first discussing anw} for each of the {ajam (Christian)
months, and the second giving information on a variety of topics including the times of prayer,
the qibla, the winds and the seasons.
Our present concern is with the shadow lengths for the hours given at the end of the sections
for each month. Unfortunately only three sets of values are preserved in the Escorial manuscript
see the following table but these are adequate to reconstruct the entire set for the year.
Table 5.2
Month \ T 1/11 2/10 3/9 4/8 5/7 6 {ar
January 27 17 13 10 9 7 14
February 26 16 12 9 8 6 [13] a
March 25 15 11 8 7 5 12
a
text: 14
The pattern is similar to, but not identical with, that of the Greek tables of type M see 1.4
and also 7.2. We may suppose, for example, that the midday shadows vary linearly between
8 (Dec.) and 2 (June). The increases for each hour are:
2 1 3 4 10
instead of the usual
1 2 3 4 10 ;
Neugebauer has noted similar absurdities in Latin tables.5
In MS Cairo MJ 202,2 (fol. 33v, copied 1681), immediately following an astrological treatise
Kitb al-Daraj by the 13th-century Andulusian mystic Ibn Sab{n,6 there is a rule for finding
the time of day by shadows. Values for the shadow lengths are given in Eastern Arabic
numerical notation, and the case-ending of the word qadam, feet, indicates the the values
given in the text are probably as originally intended. These are:
1 2 3 4 5 6 7 8 9 10 11 12
22 18 7 6 3 * 3 6 7 8 18 28
No base is stated. For the sixth hour the sun is said to be in the pole of the sphere (qub
al-falak), perhaps here meaning the zenith (the scheme is not necessarily of Andalus origin).6
Notice the lack of symmetry and the fact that the compiler could not resist including a value
for the twelfth hour. A similar set of values, albeit symmetrical, is presented by Ibn Tmart
see 6.1.
6
Even before the advent of Islam the Arabs were aware of the celestial pole, which they called qub see
the article al-^ub by Paul Kunitzsch in EI2, repr. in idem, Studies, XVIII.
maghribi sources 495
CHAPTER 6
MAGHRIBI SOURCES
Muammad ibn {Al ibn Tmart al-Andalus was the author of a treatise entitled Kitb Kanz
al-{ulm wa-l-durr al-manm f aq}iq {ilm al-shar{a and extant in several manuscripts.1
The work deals with religion, law and the sciences, and is of considerable interest. It has been
studied by Georges Vajda, who established that the author can not predate the beginning of
the 13th century. I have used MS Cairo J 47,1 (fols. 1v-85r, copied ca. 1750), in which there
is a section on simple timekeeping by shadows on fols. 80r-81v: see Fig. 6.1.
Although Ibn Tmart was writing in al-Andalus or the Maghrib, the first shadow-scheme
that he presents is for the Yemen and is introduced as such. It gives the midday shadows as
the sun progresses through the 28 lunar mansions. For al-Balda (M21) the shadow is 51/2 feet,
decreasing 1/2 foot for each mansion until al-Dabarn (M4) when there is no shadow (yan{adim
al-ill). From al-Haq{a (no. I) till al-Jabha (M10) there is a very small midday shadow (adn
ziyda), and then in al-Zubra (M11, the text has incorrectly al-Dabarn) the shadow is 1/2 foot,
increasing 1/2 foot each mansion until al-Balda (M21). For each mansion the shadow at the {ar
is 61/2 feet more. The maximum midday shadow length of 51/2 corresponds to = 15 for n
= 61/2, but then the shadow should increase to 1 foot south in the mansions close to the summer
solstice.
The above information is contained in the text of Ibn Tmarts treatise and is summarized
in a poem. Our author then presents a shadow-scheme in feet for the hours, which is as follows:
28 18 9 6 3 0 (istiw})
This scheme was also used by al-Amrn / al-Imrn (3.5). Ibn Tmart concludes this section
with a discussion of time-reckoning by night by means of the lunar mansions.
6.2 Ab Miqra{
Ab Miqra{ was an authority on folk astronomy in the Maghrib in the 13th century.2 His
teachings are known to us through a poem called al-Muqni{ fi khtir {ilm Ab Miqra{ written
by the 17th-century Maghribi astronomer Muammad ibn Sa{d al-Ss al-Marghth, and they
1
On Ibn Tmart see the article in EI2 by J. F. P. Hopkins, and Brockelmann, GAL, I, pp. 506, and SI, p.
697, and I, p. 274 and SI, p. 424 (are these references to the same individual?), as well as Cairo ENL Survey,
no. F14. His treatise is investigated in Georges Vajda, Une synthse peu connue de la rvlation et de la
philosophie: le Kanz al-{Ulm de Muammad ibn {Al Ibn Tmart al-Andalus, in Mlanges Louis Massignon,
Damas: Institut Franais de Damas, 1957, vol. III, pp. 359-374.
2
Colin & Renaud, Ab Miqra{; Renaud, Additions et corrections Suter, no. 531, and Cairo ENL Survey,
no. F17.
496 part iii, chapter six
Fig. 6.1: The section of the treatise of Ibn Tmart dealing with time-keeping by shadows. [From MS Cairo J
47,1, fols. 80r-81v, courtesy of the Egyptian National Library.]
maghribi sources 497
are outlined in greater detail in various commentaries, of which I have consulted one called
al-Mumti{ f shar al-Muqni{ by al-Ss.3
In MS Cairo J 201,2 (fols. 7r-12v, copied ca. 1600) of the Mumti{, al-Ss quotes Ab
Miqra{ (fol. 11r) as having advocated the standard (Indian) formula for timekeeping (with n
= 61/2), and the scheme:
9 7 5 3 2 1 1 2 4 ? 8 10
for the midday shadows for the months of the (Christian) calendar. The values are given in
abjad notation, and the value for month X looks like a European 6, but this could be read as
either a 5 or a 6. In MS Cairo DM 415 (55 fols., copied ca. 1700) of the Mumti{ the same
values are given (fol. 41r) but with a y} = 10 for month X; elsewhere (fol. 41v) the sequence
1 2 4 5 is mentioned for months VII to X, and it seems that a 5 was indeed intended for month
X. al-Ss adds that the above values were specifically for latitude 32, Marrakesh (the actual
value is 31;49), and that al-Jdar had proposed the scheme:
10 8 5 3 2 1 1 2 4 6 8 10
for Fez. See further 6.4.
Ibn al-Bann} was a Morroccan astronomer and mathematician of some renown who worked
in Marrakesh about 1300 and whose major work was a zj, extant in several copies.4 An almanac
entitled Risla fi l-Anw} is also attributed to him and has been published by Henri-Paul-Joseph
Renaud.5 Along with diverse information on agriculture, meteorology, and astronomical
folklore given for each day of the year, this almanac displays the lengths of day and night for
each 15 days of the solar year based on a latitude of about 38, that is, Cordova, as well as
the lengths of the gnomon shadow in feet (aqdm) at the uhr and {ar for each month. Since
Ibn al-Bann} worked in Marrakesh ( 32) there are problems with the attribution of this
work to him, but I shall not pursue these here.
Renaud used five manuscripts (S, P1, P2, A1 and A26) for his translation of the almanac
and noted the variants in the shadow lengths; there are clearly three main traditions see Table
6.3a. Neugebauer, perhaps overeagerly, restored some of the values in the scheme represented
by MS P2 to produce the primitive scheme of Greek origin shown in Table 6.3b.7
3
Renaud, op. cit., no. 540.
4
On Ibn al-Bann} see Suter, MAA, no. 399; Renaud, op. cit., no. 399; idem, Ibn al-Bann}; Cairo ENL
Survey, no. F23; and the article by J. Vernet in DSB.
5
Renaud, Calendrier dIbn al-Bann}, esp. pp. 22-23. See now Forcada, Les sources andalouses du
Calendrier dIbn al-Bann.
6
These stand for: S Sal (private owner); P1 Paris BNF ar. 4764; P2 Paris BNF ar. 6020; A1 Algiers
941; A2 Algiers 1468 (see Renaud, op. cit., p. 26).
7
Neugebauer, HAMA, II, pp. 743-744.
498 part iii, chapter six
Table 6.3a
Table 6.3b
Modifications to the scheme in MS P2 made by Neugebauer are underlined.
uhr 7 6 5 4 3 2 3 4 5 6 7 8
{ar 14 13 11 10 9 8 9 10 12 13 14 15
z 7 7 6 6 6 6 6 6 7 7 7 7
The uhr entries for months XI and XII, and perhaps also month X, are garbled. Virtually all
of the {ar shadows are garbled since, at least according to the standard definition of the {ar,
they should each be in excess of the corresponding midday entries by 7. (Even in Neugebauers
reconstruction the increases are either 6 or 7.) The uhr shadows are clearly all midday shadows,
with perhaps one exception (see below), rather than midday shadows augmented by one-quarter
of the length of the gnomon, which was the standard definition of the time of the uhr in al-
Andalus and the Maghrib (see also 6.4).8
The anomalous but well-attested value 91/5 for the uhr in month XI appears to represent
an attempt to add one-fifth of the gnomon length to the gnomon shadow. This corresponds
to a time one seasonal hour after midday using the standard Indian formula (see 2.1 and 3.2).9
The value of the gnomon length varies between 6 and 7 feet in the table, but the value
91/5 could be derived from a midday shadow of 8 augmented by one-fifth of a gnomon length
of six digits.
MS Oxford Pococke 249 (63 fols., copied ca. 1400) is a unique copy of a treatise on folk
astronomy by a Maghrib astronomer named Ab {Abdallh Muammad ibn {Abdallh ibn
{Amr al-Mu}addib (the teacher), not otherwise known to the modern literature. On fols. 46v-
47r there is a table displaying the shadow lengths in digits (n = 12) at each seasonal hour and
at the uhr and {ar for each date of the solar year when the shadow has an integral value
see Fig. 6.4 and Table 6.4 for extracts.
8
On the motivation for the value z = 1/4 n see IV-4.4.
9
See further King, Tunisian Sundial, p. 195, and now IV-2.
maghribi sources
Fig. 6.4: A table for timekeeping by Ab {Abdallh al-Mu}addib. [From MS Oxford Pococke 249, fols. 46v-47r, courtesy of the Bodleian Library.]
499
500 part iii, chapter six
Table 6.4
Datea | Hours 6 5 4 3 2 1 uhr {ar
VI 17 2 2/ 5 b
5- c 81 / 2 141/2 261/2 621/2 51 / 2 141/2
VII 9 3 52 / 5 9 15 27 63 6 15
VII 11 d 4 62 / 5 10 16 28 64 7 16
VIII 14 - 25 .....
IX 4 - 13 - 22 - 29 .....
X 7 - 13 - 19 - 25 - 31 .....
XI 6 - 13 - 20 .....
XI 30 19 212/5 25 31 43 79 22 31
XII 16 20;7 e 222/5 26 32 44 80 23 32
I 2 19 212/5 25 31 43 79 22 31
I 12 - 19 - 26 .....
II 1 - 7 - 13 - 19 - 25 .....
III 3 - 10 - 19 - 28 .....
IV 7 - 27 f .....
V4 .....
V 16(?) 3 52 / 5 9 15 27 63 6 15
V 27(?) 2 /5
2
*g
a
the first date is followed by the phrase: yakn ill al-wuqf ... , meaning: the midday shadow will be ... . The
remaining dates are preceded by the phrase: thumma yazd / yanqu fa-yakn il ... , meaning: then it increases
/ decreases until it becomes (amount) until (date). b iba{n wa-khumsn. c fraction illegible; one could read
a khumsayn (= 2/5) but there is more. d read 21? e {ishrn wa-sab{at ajz}. f read 17? g wa-raja{ (read: wa-rja{
?) il awwal al-jadwal {al ... (??) in sh}a llh ta{l, meaning: go back to the beginning of the table ... ...
(??) if God wills.
6.5 al-Jdar
{Abd al-Ramn ibn Ab Ghlib Muammad al-Jdar was famous as the author of a short
poem on folk astronomy compiled in 794 H [= 1391/92] and entitled Rawat al-azhr, which
enjoyed several commentaries at the hands of later Maghribi astronomers.10 In this poem there
is no shadow-scheme, but one specifically for Fez is attributed to al-Jdar by the 17th-century
Maghribi astronomer al-Ss. This scheme is:
10 8 5 3 2 1 1 2 4 6 8 10
and is presented, for example, in MS Cairo DM 415, fol. 41v, of al-Sss treatise entitled al-
Mumti{. See further 6.2.
The only other known work of al-Jdar is an almanac for Fez, extant in MS Istanbul Lalel
3748,10. In this work values of the solar altitude at midday and the corresponding shadows
to base 62/3 are given for each few days of the year, and I have not investigated these. However,
the scheme
9 7 5 3 2 1 1 2 4 5 8 10
for the shadows at the uhr beginning with January with corresponding values increased by
7 for the {ar is found in a note on fols. 131v, immediately following al-Jdars almanac.
6.6 al-afqus
{Al al-Sharaf of Sfax compiled a navigational atlas ca. 1579.11 I have examined two of the
four available copies, namely, MSS Paris BNF ar. 2273 and Oxford Marsh 294 (both completed
ca. 1600); I label these A and B, respectively.
Both copies contain a table displaying the shadow lengths at significant times of day
see Table 6.6. The table in A displays shadow lengths for the uhr and the {ar, with difference
six units: clearly the length of the gnomon is supposed to be 6. The scheme is not attested
in any other known source. The table in B displays values for midday, the uhr and the {ar.
The difference between the first and second is 2 units. The difference between the first and
the third is 7 units: clearly this is the length of the gnomon. The same scheme is found in other
Maghribi sources (see 6.4 and 6.6). A marginal note states that the shadow increase for the
uhr could be two units but that the value of three digits is precautionary (itiy). These
increases of 2 and 3 units probably represent approximations for 13/4 and 31/2 units, that is,
one-quarter and one-half the length of the gnomon 7. In A, but not in B, the lengths of day
and night for each month are also tabulated but they do not concern us here. The entries in
both sources are displayed in the following table:
10
On al-Jdar see Suter, MAA, no. 424a; Renaud, Additions et corrections Suter, no. 424a; Brockelmann,
GAL, SII, pp. 217-218; and Cairo ENL Survey, no. F26.
11
On al-afqus and his work see Nallino, Scritti, V, pp. 533-548; and Brockelmann, GAL, SI, p. 710. A
new study of these manuscripts is currently in preparation by Mnica Herrera Casais (La Laguna, Tenerife, and
Frankfurt).
502 part iii, chapter six
Table 6.6
MS A MS B
Month day night uhr {ar midday uhr {ar
III 101/3 132/3 7 13 5 8 12
IV 122/3 111/3 4 9 3 6 10
V 134/5 101/5 2 8 2 5 9
VI 142/3 91 / 3 1 7 1 4 8
VII 141/2 91 / 2 11 / 3 71 / 3 1 4 8
VIII 134/5 91 / 5 21 / 3 81 / 3 2 5 9
IX 132/3 111/3 4 10 4 7 11
X 111/3 122/3 5 11 5 8 12
XI 101/3 132/3 8 14 8 11 15
XII 92 / 3 141/3 10 16 10 13 17
I 101/4 133/4 9 15 9 11 16
II 101/3 132/3 7 13 7 10 14
(1) The standard approximate rule (with n = 7 and n = 12) is mentioned in an anonymous
Maghribi treatise on the astrolabe extant in the unique MS Cairo DM 1169,6 (fols. 45v-47v,
copied 1158 H [= 1745/46])12 (see esp. fols. 47r-47v).
(2) Hoest (1781), Delphin (1891), Renaud (1948), and Roche (1965) have noted the use of
the mnemonic
9 7 5 3 2 1 1 2 4 5 8 10
in the Maghrib.13 This differs from the Egyptian version (see 9.7a) only in that the shadow
for month X is 5, not 6. Hoest records the text of an almanac in which these lengths occur.
Also given in this source, but not the others, is the length of day and night for each month:
9;30 10+ 12 13 14 15;40 14;30 13;20 12 10;30 9;20 8;55
14;30 13+ 12 11 10 8;20 9;30 10;40 12 13;30 14;40 15; 5
Hoest reported that his informants had told him that they thought that the almanac had come
to the Maghrib from Istanbul. The winter cold in the Maghrib was not so severe as implied
by the almanac, and the shadows there were a little shorter.
12
See IV, Arabic text 2.8.
13
G. Hoest, Nachrichten von Marokos und Fes im Lande selbst gesammelt, in den Jahren 1760 bis 1768,
Copenhagen: Christian Gottlob Prost, 1781, pp. 252-257; Delphin, Lastronomie au Maroc, pp. 199-200;
Renaud, Calendrier dIbn al-Bann}, p. 23; and P. Roche, Lirrigation et le statut juridique des eaux au Maroc,
Revue juridique et politique dindpendance et coopration 19 (1965), pp. 55-120 and 537-561, esp. p. 85, n.
68.
{ir@q sources 503
CHAPTER 7
{IR@Q SOURCES
{Abd al-Qdir ibn {Abdallh al-Kln was a celebrated f, and founder of the Qdiriyya
order.1 He was born in 1078 in Gilan, studied in Baghdad, and died in 1167. His major work
al-Ghunya contains some passages on reckoning time by shadow lengths, which I came across
in MS Cairo MJ 180,17 (fols. 146v + 148r-149r, copied ca. 1750).2
al-Kln first presents the midday shadow lengths for the Syrian months, beginning with
Ayll (= September), as follows:
IX X XI XII I II III IV V VI VII VIII
5 6 7 8 7 6 5 4 3 2 3 4.
He then presents another opinion, which is that the midday shadow of a person is 3 feet
on @dhr (= III) 19, decreasing to 1/2 by azrn (= VI) 19, then increasing at a rate of 1 foot
each 36 days to 3 feet on Ayll (= IX) 19 and then at 1 foot each 14 days to 71/2 on Knn
I (= XII) 19.
Yet another scheme is presented next. The shadows at midday and the {ar for the months
beginning with azrn are:
VI VII VIII IX X XI XII I II III IV V
3 4 5 6 7 8 101/2 9 71 / 2 6 41/ 2 31 / 2
91/2 101/2 111/2 121/2 131/2 141/2 17 15 141/2 121/2 11 10
This scheme is clearly a combination of two linear schemes. The midday shadow increases
from 3 in June to 8 in November at a rate of 1 foot each month and decreases from 101/2 in
December to 41/2 in May at a rate of 11/2 feet each month. No tables displaying this feature
are recorded by Neugebauer. The values for the {ar are in general, but not always, 61/2 units
more than those for the uhr. Since for n = 6;30 the minimum shadow 3 corresponds to a solar
altitude of 65 and the maximum shadow 101/2 corresponds to a solar altitude of 32, even
the limiting values of the shadow are inconsistent with reality for any latitude. The same scheme
is recorded by Ibn Qudma see 8.1.
7.2 al-ikaf
MS Paris BNF ar. 2540,2 (fols. 16r-27v, copied ca. 1450) contains a set of planetary tables
compiled about the year 1400 by Nir al-Dn ibn {s al-abb al-ikaf (from in Kayf
1
Article {Abd al-^dir al-Djln by W. Braune in EI2, and Brockelmann, GAL, I, pp. 560-563, and SI,
pp. 777-779.
2
The passages extracted in this manuscript correspond to {Abd al-Qdir al-Kln, al-Ghunya, 2 vols., Cairo,
1304 H [= 1886/87], II, pp. 89-90.
504 part iii, chapter seven
Fig. 7.2: A table for the seasonal hours based on a Hellenistic model and preserved in the planetary tables of
al-ikaf. [From MS Paris BNF ar. 2540, fol. 27v, courtesy of the Bibliothque Nationale de France.]
on the Tigris).3 Another copy of this work is in MS Vatican ar. 1499 (copied 1576/77). I label
these two sources A and B.
Both manuscripts contain (A: fol. 27v, B: fol. 13r) a table of the shadows at each seasonal
hour from 1 to 11 for each month of the Syrian year starting with Nsn (= April). See Fig.
7.2 for the table as it appears in A. The values given in the table for hours 1 to 6 are displayed
in Table 7.2a; those for 1 to 5 are confirmed by the identical entries for 11 to 7. The underlying
scheme for midday is simply:
5 4 3 2 1 0 1 2 3 4 5 6.
but the entries for the hours are in some cases, particularly for the first and third, corrupt. Table
7.2b shows a matrix of values based on the above midday scheme and increases at the hours
of:
1 2 3 4 10 (individual) or 1 3 6 10 20 (cumulative).
3
al-ikaf and the Paris copy of his tables are listed in Brockelmann, GAL, SI, p. 869.
{ir@q sources 505
The entries in Table 7.2a different from these are italicized. The entries in Table 7.2b are
less by 2 than the corresponding ones in the standard classical tables of type M (see 1.4)
reproduced in Table 7.2c.4
Table 7.2a
Month \ T 1 2 3 4 5 6
IV 23 13 9 5 3 2
V 24 11 7 4 2 1
VI 20 a 10 6 3 1 0b
VII 24 11 7 4 2 1
VIII 22 11 9 5 3 2
IX 23 13 10 6 4 3
X 24 14 12 7 5 4
XI 25 15 13 8 6 5
XII 26 16 15 9 7 6
I 25 15 13 8 6 5
II 24 14 12 7 5 4
III 23 13 10 6 4 3
a
thus in both A and B; b
A: 2 (sic)
Table 7.2b
22 12 8 5 3 2
21 11 7 4 2 1
20 10 6 3 1 0
21 11 7 4 2 1
22 12 8 5 3 2
23 13 9 6 4 3
24 14 10 7 5 4
25 15 11 8 6 5
26 16 12 9 7 6
25 15 11 8 6 5
24 14 10 7 5 4
23 13 9 6 4 3
Table 7.2c
Month | T 1 2 3 4 5 6
I 24 14 10 7 5 4
II 23 13 9 6 4 3
III 22 12 8 5 3 2
IV 23 13 9 6 4 3
V 24 14 10 7 5 4
VI 25 15 11 8 6 5
VII 26 16 12 9 7 6
VIII 27 17 13 10 8 7
IX 28 18 14 11 9 8
X 27 17 13 10 8 7
XI 26 16 12 9 7 6
XII 25 15 11 8 6 5
Note that the midday June shadow of zero serves the tropic of Cancer, not the upper reaches
of the Tigris. Since it is unlikely that the table was intended to display increases of the shadows
4
Neugebauer, HAMA, II, p. 738.
506 part iii, chapter seven
over the summer midday minimum, we are left with the conclusion that it is simply the work
of an incompetent.5
By the side of the main table in both copies is another in which the midday shadows for
the Syrian months beginning with Knn II (= January) are given as:
9 7 5 3 2 1 1 2 4 6 8 10
This scheme is identical to the one attested in the Coptic and Arabic sources described in 9.7a.
5
Compare the Ethiopic tables (based on limits 0 and 9 for Z) in idem, Ethiopic Astronomy and Computus,
pp. 212-213.
syrian sources 507
CHAPTER 8
SYRIAN SOURCES
MS Damascus hiriyya 10732 (fols. 2r-4v, copied ca. 1600) contains a short treatise on the
zodiacal signs, lunar mansions and simple timekeeping by Muwaffaq al-Dn Ab Muammad
{Abdallh ibn Amad ibn Muammad ibn Qudma al-Maqdis (1147-1223), a scholar of law
and theology who worked mainly in Jerusalem and Damascus.1
The tract includes some prayer-tables (the underlying latitude is 30, which corresponds
more happily to Cairo),2 and also a section on the shadows at the uhr and the {ar for each
month of the Syrian year beginning with azrn (VI). The values given are the same as those
in the third scheme of al-Kln (see 7.1) with midday values missing for months X and V
and an {ar value 151/2 for month I.
In his major legal treatise, al-Mughn, Ibn Qudma presented a different scheme.3 This is
described as an approximation (taqrb) and is attributed to Abu l-{Abbs al--n-j-y (?), an
individual otherwise unknown to me. For the middle of each of the Syrian months, the shadow
lengths in qadams are as follows:
VI VII/V VIII/IV IX/III X/II XI/I XII
1 1
1 / 3 1 + / 2 + /31 3 1
4 /2 1
6 /2 9 10 1/6
These are said to be valid for the provinces of Iraq and Syria and others with similar latitudes.
The extremal values are roughly consistent with a latitude of ca. 34 for n = 61/2.
Ibn al-Shir was the head of the astronomers at the Umayyad Mosque in Damascus in the
mid-14th century, and he is well known to the history of science for his work in planetary
astronomy, spherical astronomy and timekeeping.4
The introduction to a set of prayer-tables for latitude 34 by Ibn al-Shir survives in the
unique MS Leiden Or. 1111 (fols. 108r-113r, copied 1601), and the prayer-tables themselves
survive in MS Cairo DM 1170,2 (fols. 11r-22v, copied ca. 1650). In the introduction to the
1
On Ibn Qudma see the article Ibn ^udma by George Makdisi in EI2. Only one scientific work is
(incorrectly) attributed to him by Brockelmann (SI, p. 689, no. 20), namely, a work on comets which is actually
by a later Egyptian writer also called {Abdallh ibn Amad al-Maqdis (on whom see Brockelmann, GAL, II,
p. 470, and SII, p. 486, and Cairo ENL Survey, no. D45).
2
These are analyzed in II-6.2.
3
Ibn Qudma, al-Mughn, 12 vols., Cairo, 1341-48 H [= 1922-30], I, p. 270.
4
On Ibn al-Shir (II-9.3) see my article in the DSB and the bibliography there cited (to which add Kennedy
& Ghanem, eds., Ibn al-Shir), and also King, Astronomy of the Mamluks, pp. 538-539 and 547-548.
508 part iii, chapter eight
tables, which themselves are quite sophisticated and based on accurate formulae (II-9.3), Ibn
al-Shir mentions some approximate methods for timekeeping, including (fol. 111v) the
standard approximate formula for n = 7.5
8.3 Anonymous
MS Princeton Yahuda 1168 (= Mach 4872), copied ca. 1800 (?)), contains a short text on the
shadows at midday for the Syrian months, accompanied by a table: see Fig. 8.3. The text states
that shadows in feet (qadams, base 7) can be measured by ones feet lengthwise, and shadows
in digits (iba{s, base 12) can be measured by ones feet widthwise (it is not completely clear
what is intended here).
Fig. 8.3: The entire text of a short anonymous Syrian treatise on timekeeping by shadow-lengths. [From MS
Princeton Yahuda 1168, courtesy of the Special Collections, Princeton University Library.]
5
IV, Arabic text 2.9.
syrian sources 509
For each month in the Syrian calendar values are given in qadams for the shadows at midday
and the {ar. Those values of the midday shadow serve the first, middle and last ten days of
each month.
Table 8.3
Months
Days 1-10 10-20 20-30
Midday {ar [{ar - 7]
I 5 4 31 / 2 10 3
II 3 21 / 2 2 9 2
III 13 / 4 12 / 3 11 / 2 8 1
IV 12 / 3 11 / 2 12 / 3 8 1
V 2 21 / 6 21 / 2 9 2
VI 3 31 / 2 41 / 6 10 3
VII 5 51 / 6 6 13 6
VIII 71 / 3 71 / 2 8 14 7
IX 9 91 / 2 10 14 7
X 12 11 101/2 14 7
XI 11 91 / 2 8 13 6
XII 71 / 2 71 / 4 51 / 2 12 5
These midday shadows range between 11/2 and 12, which correspond quite well to the latitude
of Northern Syria ( 36). The values for the {ar are unrelated to these midday values. If
we subtract 7 from each value we obtain a series of values for midday ranging from 1 to 7;
these correspond to no terrestrial latitude.
510 part iii, chapter nine
CHAPTER 9
EGYPTIAN SOURCES
9.4 al-Drn
The early-13th-century Egyptian mystic al-Drn wrote a treatise on folk astronomy, partly in
verse, which is extant in several manuscripts (II-2.6), including MS Cairo DM 651,5 (fols.
41r-58r, copied 828 H [= 1424/25]). al-Drn presents the following shadow-scheme in feet
(aqdm) for the zodiacal signs:
/ / / /
/
2/ 1 2 3 1/ 5 1/ + 1/ 7 1/ 8 1/
3 2 4 8 4 8
These values correspond roughly to latitude 30 with base 6. al-Drn also presents the daily
shadow increases for each sign in minutes (daq}iq) thus:
1 1/ 3 4 6 71 / 2 71/ 2 31/ 2
but these values correspond to base 12!
The same set of values together with daily rates of increase occurs on the inside cover
of MS Cairo MM 168, which contains some miscellaneous tables from the 13th-century
Egyptian Muala Zj and some anonymous material on folk astronomy, copied ca. 1400-1500
probably in Egypt but possibly in the Yemen.1
MS Cairo M 127 (59 fols., copied 1255 H [= 1839/40]) is a unique copy of a treatise entitled
Mab al-anwr wa-maft al-asrr f a{ml al-layl wa-l-nahr on folk astronomy and
simple timekeeping. The work is attributed to an individual named simply Sirj al-Dn wa-
l-Duny.2 It is indirectly related to a treatise on the same subject preserved in MS Princeton
Yahuda 4657 (Mach 4983, fols. 40v-80r, copied ca. 1350), which according to a spurious
note at the beginning of the treatise purports to be an abridgement of the Mab al-anwr
by Sirj.
The Cairo manuscript contains (fols. 50r-50v) a section on reckoning the time of day by
shadows in which the schemes:
9 7 5 3 2 1 1 2 4 6 8 10
and
40 20 10 6 3 2
are advocated for the midday shadows for each month of the Syrian year and shadow increases
for the seasonal hours, respectively. I suspect that this treatise is of Egyptian provenance,
perhaps from the early 13th century; the same is the case for the treatise in the Princeton
manuscript.
Of particular interest is a remark by the author of the treatise in the Cairo manuscript that
the author of a book entitled Kitb all al-mushkilt f ma{rifat al-daq}iq wa-l-daraj wa-
l-s{t this title is new to the modern literature devised a sundial called al-basa al-
1
Cairo ENL Survey, no. C12.
2
The author on whom see Cairo ENL Survey, no. C4 is the celebrated anaf legal scholar Sirj al-
Dn al-Sajwand (fl. ca. 1200 see Brockelmann, GAL, I, pp. 470-471, and SI, pp. 650-651, and the article
al-Sadjwand by Rudolf Sellheim in EI2. See further IV-3.3.
512 part iii, chapter nine
sab{niyya because it bore seventy (sab{n) lines for regulating time of day according to these
two rules. I do not understand the text of this passage, which merits further investigation. (See
10.3 on a sundial based on another simple shadow-scheme.) Sirj al-Dn points out that the
author of this treatise, whom he calls al-ra}s al-fil,3 used n = 7 and mentions that the longest
shadow represented on the sundial was at one hour before sunset in the 4th Coptic month of
Kayhak, namely, 10 + 60 = 85 (??). This suggests that the author was using a scheme for the
hours other than the
40 20 10 6 3 2
scheme advocated by Sirj al-Dn. There is some confusion here, not least because z = 60
at T = 1 corresponds to a gnomon length of n = 12.
9.6 al-iqill
MS Dublin CB 4538 (ca. 30 fols., copied ca. 1300) is a unique copy of a work on timekeeping
by the lunar mansions. The title page is barely legible. I read Anw} ... ibn Hrn (? written
Hrwn) al-iqill, that is, (Book on) the periods of the year arranged according to the lunar
mansions (naw}) by ... ibn Hrn al-iqill. The name of this author is new to the modern
literature. Although he or his family hailed from Sicily, his work was clearly compiled for use
in Egypt. Perhaps he is identical with one Abu l-Qsim ibn {Abdallh ibn {Abd al-Ramn
ibn asan al-Qurash al-iqill, recently identified by Julio Sams.4 This mans date is
uncertain, but he is clearly early enough to be a candidate. He is known to have made serious
twilight measurements in Cairo and Alexandria and to have authored a book entitled al-Intiq}
min al-yawqt f {ilm al-yawqt, otherwise unknown to us but obviously dealing with more
than timekeeping by night with the lunar mansions.
Each pair of pages is devoted to one lunar mansion: see Fig. 9.6 for an example. Tables
display the configuration of the lunar mansions with respect to the horizon and meridian at
each of thirteen times during the night. Dates corresponding to each of the days of the naw}
are given in the Coptic, Syrian and Western (Andalus) calendars. Solar meridian altitudes are
also given for each day. This information is accompanied by a diagram of the stars constituting
the mansion and various additional items of information, such as the shadow lengths for the
naw} in question. Values are given in digits (iba{) and sixtieth divisions (daqqa) thereof.
From the extremal values of the solar altitude, which are given to the nearest degree, namely,
83 and 36, I conclude that these are truncated from values based on
= 30;0 (Cairo) and = 23;35 ,
although they might also be computed with = 24. The values of the shadows are as follows:
3
This is a common title given to the early-11th-century scholar Ibn Sn, on whom see the article by A.
M. Goichon in EI2. However, it is somewhat unlikely that Ibn Sn would have used the Coptic calendar.
4
Sams, Astronomical Observations in the Maghrib, p. 174.
egyptian sources 513
Fig. 9.6: An extract, one double page out of 28, from al-iqills tables for timekeeping by the lunar mansions
with additional information about midday shadow lengths throughout the year. [From MS Dublin CB 4538,
courtesy of the Chester Beatty Library.]
Table 9.6
M Z M Z M Z M Z
1 3;?5 8 2;13 15 10;7 22 14;36
2 3;1 9 3;1 16 12;12 23 12;53
3 2;17 10 3;53 17 deest 24 11;4
4 1;15 11 5; 5 18 deest 25 9;20
5 1;35 12 6;20 19 16;50 26 7;43
6 1;33 13 7;45 20 16;49 27 6;16
7 1;49 14 9;23 21 16;1 28 5; 2
They were apparently computed (rather than observed), but extremal values 1;15 and 16;50
do not correspond properly to those for the above parameters, namely, 1;21 and 16;16. The
corresponding values for = 24 are 1;16 and 16;31. Thus al-iqills values are not fully
explained.
514 part iii, chapter nine
9.7a Miscellaneous
Neugebauer has already drawn attention to a midday shadow-scheme for the months occurring
in an Arabic text written for the use of the Copts and ascribed to the 12th patriarch of Alexandria,
Demetrius (d. A.D. 230) but copied only in 1768.5 The shadows for the twelve months
beginning with ba (V) are:
9 7 5 3 2 1 1 2 4 6 8 10
As Neugebauer has pointed out, there are two arithmetical progressions underlying this
primitive scheme. For latitude = 30 and = 24 and base n = 7 we should expect maximum
and minimum shadows of:
0;44 and 9;38 .
Neugebauer labelled the scheme as badly distorted, but its simplistic nature did not detract
from its popularity in medieval Egypt, as the following new sources show. See also 6.7 on
a related scheme in various Maghribi sources and 7.2 on a Syrian version.
First I should mention that I have not found any such scheme in the one Coptic astronomical
handbook from the medieval period which I have inspected, namely the work of the 13th-century
Copt al-As{ad ibn al-{Assl,6 extant in MS Cairo DM 910 (86 fols., copied ca. 1805). This
work, which is of considerable interest and remains to be studied, contains solar and lunar tables
and tables for calendar conversion and determining the feasts. However, it does not contain
any material relating to timekeeping.
The same scheme for the Coptic months starting with ba that was noted by Neugebauer,
now usually formulated as a mnemonic:
5
G. Sobhy (Bey), The Coptic Calendrical Computation and the System of Epacts known as The Epact
Computation Ascribed to Abba Demetrius the XIIth Patriarch of Alexandria, Bulletin de la Socit
dArchologie Copte 8 (1942), pp. 169-199, and 9 (1943), pp. 237-252, p. 187 (French), p. 250 (Arabic), cited
in Neugebauer, HAMA, II, p. 743.
6
On Ibn al-{Assl see the article by A. Atiya in EI2, and Cairo ENL Survey, no. C10.
7
On al-Suy see Suter, MAA, no. 449; Brockelmann, GAL, II, pp. 180-204, and SII, pp. 178-198; Cairo
ENL Survey, no. D103; the article by E. Geoffroy in EI2, and also n. IV-8:26.
8
Brockelmann, GAL, II, pp. 478-479, and SII, pp. 492-493, and Cairo ENL Survey, no. D30.
egyptian sources 515
Fig. 9.7a: Midday shadow lengths for the Coptic months in a late Egyptian source. [From MS Cairo MM 33,1,
courtesy of the Egyptian National Library.]
(6) a legal treatise by al-Bjr, who was Shaykh al-Azhar in Cairo ca. 1850. This source was
noted by Eilhard Wiedemann, but the mnemonic as interpreted by him is garbled.9
A variation of this scheme is given in MS Cairo DJ 323,9 (fol. 39r, copied ca. 1900) in
a poem attributed to one Ab Bakr al-Mqt, on whom I have no further information. The
midday shadows for the Coptic months beginning with Tt (I) are stated to be:
3 5 8 10 9 8 5 3 2 1 1 2,
these numbers being given in abjad notation thus:
In MSS Cairo DM 119,4 (fol. 9v, copied 1109 H [= 1697/98]) see Fig. 9.7b and 1112
(fol. 10v, copied 1310 H [= 1892/93]), both manuscripts of Egyptian provenance, there are
notes (f}ida) on the shadows increases for each hour to be added to the midday shadow, as
follows:
T 1 2 3 4 5
z 30 12 6 3 1
For n = 6 the Indian formula yields the values:
30 12 6 3 11/5 ,
so the rule is explained. The use of the base 6 is uncommon in post-Abbasid Islamic astronomy.
The scheme
40 20 10 6 3 2
is presented in an anonymous Egyptian almanac containing mainly material relating to folk
astronomy, preserved in MS Cairo DM 187 (recto of penultimate folio, copied ca. 1850).
9
Wiedemann, Aufstze, II, p. 781. On al-Bjr see the article al-Bdjr by Th. W. Juynboll in EI2.
516
part iii, chapter nine
Fig. 9.7b: In these notes appended to a copy of a treatise on a special type of astrolabic quadrant by the late-15th-century Cairo astronomer Sib
al-Mridn, we find a reference to the sundial of the Caliph {Umar ibn {Abd al-{Azz (IV-7.1), a poem on the prayer-times attributed to the
Imm al-Shfi{, elsewhere attributed to his descendant, the Fatimid astronomer Ibn Ynus (II-2.1), and a simple shadow-scheme for
timereckoning. [From MS Cairo DM 119,2-4, fols. 8v-9v, courtesy of the Egyptian National Library.]
egyptian sources 517
Another mnemonic is attested in various Egyptian sources for the shadow lengths at the
seasonal hours. This is:
The following procedure for reckoning time of day with the hand, using the thumb as gnomon,
is outlined in MS Cairo DJ 323,7 (fol. 34r, copied ca. 1900). The text occurs amidst various
astronomical treatise of Egyptian provenance see Fig. 9.7c.11
Translation: A note on finding the hours with the shadow of the thumb (ibhm). You
raise the thumb straight up with the shadow (falling) on the index finger (sabbba).
As long as the shadow has not reached the tip (unmula) of the index finger you are
in the first hour. If the shadow falls on the tip of the index finger then you are in the
second hour. If it has reached the first joint ({uqda) you are in the third hour, and if
it has reached the second joint you are in the fourth hour. If it has reached the third
mark (azza) you are in the fifth hour, and if it has reached the fourth mark you are
in the sixth hour. Then (use) the opposite (procedure) until the end of the day.
On an adult hand this corresponds approximately to a shadow-scheme
T 1 2 3 4 5
z 5 4 3 2 1 (n = 21/2 inches)
I know of no other attestation of such a procedure in the Islamic sources. al-Brn, it is true,
10
Neugebauer, HAMA, II, p. 738.
11
Cairo ENL Survey, no. D228.
518 part iii, chapter nine
Fig. 9.7c: A scheme for reckoning the hours with the thumb and fingers.
[From MS Cairo DJ 323,7, courtesy of the Egyptian National Library.]
Fig. 9.7d: The scheme for timekeeping using the hand in a treatise by Jacob
Koebel entitled Eyn knstliche sonn Uhr inn eynes jeden menschen Lincken
handt, printed in Mainz in 1532. [From Gockerell, Telling Time without
a Clock, p. 133.]
mentions in the Shadows that he had observed the Indians using their middle finger extended
vertically as a gnomon to cast shadows along their forearms held horizontally, measuring the
shadows in finger lengths with the other hand.12 But, as he says regretfully, he neglected to
ask them the details. A method similar, but not identical, to that described in the Cairo
manuscript is attested in a 16th-century German text (see Fig. 9.7d),13 and no doubt such
procedures were more widespread both in the Near East and in Europe than we could ever
know from surviving textual evidence.
In the course of preparing this study I came across the following shadow-schemes for the
monthly midday shadows in my notes on Ibn al-Bann}s almanac, the first marked Egypt.
I have no recollection in which Cairo manuscript I originally found these schemes, which are
as follows:
9 7 5 3 2 1 1 4 8 8 10 10
7 6 4 3 2 1 1 3 3 4 8 8
12
al-Brn, Shadows, pp. 33-34, and Kennedy, al-Brns Shadows, I, pp. 73-74, and II, p. 28.
13
Gockerell, Telling Time without a Clock, p. 133.
miscellaneous 519
CHAPTER 10
MISCELLANEOUS
MS Cairo DM 570 (182 fols., ca. 1875?) is a unique copy of a collection of texts, tables and
diagrams relating to navigation. The work was apparently compiled ca. 1860, and it awaits
1
Neugebauer, HAMA, II, pp. 745-746. For the table of Pseudo-Bede see also Ginzel, Chronologie, III, p.
88. On yet simpler forms of reckoning the hours in medieval Europe see Gockerell, Telling Time without a
Clock (n. 9:13).
520 part iii, chapter ten
detailed study.2 Of concern to the present investigation is a table (fol. 17v) of shadow lengths
attributed to al-shaykh {Abdallh ibn {Umar B Makhrama, a Yemeni scholar who flourished
ca. 1500: see further 4.5.
A horizontal sundial constructed in 1722 in the town of Gresik in Java (see Fig. 10.3) was
published by the Batavian savant Tjondro Negoro in 1882.3 The instrument measures ca. 60
60 cms., and its surface is covered with texts in Javanese, which were reproduced and
rendered into Dutch by Negoro. It has since been discussed by various scholars of ethnography
and ethnoastronomy interested mainly in the associated calendar, which is defined by means
of midday shadows.4
The sundial, called a bencet, consists of a pair of perpendicular scales shaped like a T.
The shorter scale measures shadows at midday. It bears markings for each unit between 2 units
north and 4 units south: the underlying gnomon length must be 7, as we see from the longer
perpendicular scale. This measures shadows at the {ar (Javanese: asar) based on the standard
definition z = n = 7, and it is marked for each unit from 1 up to 11, although it is only the
markings from 7 to 11 which are relevant to the function of the scale. In order to use the scale
for the {ar, the sundial should ideally be rotated away from the meridian so that the afternoon
shadow can be compared with the markings on the scale, but we are not dealing with a precision
instrument.
The year is divided in twelve periods called mangsas, beginning with the summer solstice.
These are defined as the intervals during which the shadow changes by 1 unit. The mangsas
vary in length between 23 and 43 days, an unusual feature for any calendar, and this has
disturbed modern investigators. The inspiration for the calendar is clearly the fact that (for this
latitude and gnomon length) the midsummer shadow is 4 units south and the midwinter one
2 units north, the total change in shadow length thus being 6 units. The midday shadows at
the beginning of the mangsas (mi) are thus:
m1 m2/m12 m3/m11 m4/m10 m5/m9 m6/m8 m7
4S 3S 2S 1S 0 1N 2N
The limits 4 S and 2 N correspond reasonably well to the latitude of Gresik, namely ca. -7
(actually 4.12 S and 2.07 N for n = 7). The Dutch savant van der Stok, unaware of standard
2
Cairo ENL Survey, no. Z41.
3
See R. M. A. A. Tjondro Negoro, De koperen zonnewijzer van Gresik, Tijdschrift van het Bataviaasch
Genootschap van Kunsten en Wetenschappen (Batavia/Jakarta) 27 (1882), pp. 47-68; and also F. van den Bosch,
Der javanische Mangsa-Kalendar, Bijdragen tot de Taal, Land en Volkenkunde 126 (1980), pp. 248-282 and
figs. 13-17, espeically pp. 251-253.
4
See, for example, N. Daldjoeni, Pranatamangsa, the Javanese Agricultural Calendar, The Environmentalist
4 (1984), supp. no. 7, pp. 15-17; idem and B. Hidayat, Astronomical Aspects of Pranotomongso of the 19th-
Century Central Java, in New Delhi 1985 Colloquium Proceedings, pp. 249-252; G. Ammarell, Sky Calendars
of the Indo-Malay Archipelago, ibid., pp. 241-248; as well as A. Maas, Sternkunde und Sterndeuterei im
malaiischen Archipel, Tijdschrift voor Taal, Land en Volkenkunde (Koninklijke Bataviaasch Genootschaap van
Kunsten en Wetenschappen, Batavia / Jakarta), 64 (1924).
miscellaneous 521
Fig. 10.4: A sundial for regulating water distribution from Muscat and Oman. I have no information on its precise
location or the way in which it was used. The divisions are neither equal nor uniform, which may have made
some falln rave wildly if, as a result of this, they were deprived of their fair share of water. [From the archives
of Alain Brieux, Paris.]
522 part iii, chapter ten
medieval Islamic practice, determined the most appropriate gnomon length as 7.11 units (but
then the limiting shadows are 4.19 S and 2.1 N).5
The development of a calendar out of a shadow-scheme is not, as far as I know, attested
in other societies. See, however, 6.4 for a first step in this direction.
From both Northern and Southern Arabia, as well as Egypt, we have accounts of the regulation
of water for irrigation by means of folk astronomical techniques.6 I do not doubt that similar
practices were, and indeed in some cases still are, in vogue elsewhere in the rural areas of the
Islamic world.
In the Oasis of al-{Ul} in North-Western Arabia, for example, the length of daylight is
divided into 24 asds (plural of suds, lit. sixth), and the passage of the asds is regulated by
observing the shadow of a 2-metre high wall pass over sets of stone markers for each suds.7
There are three sets, two for the solstices and one for the equinoxes. The positions of the stones
for the solstices, as recorded recently, approximate rather well to the hyperbolae of a horizontal
sundial.
In the Buraimi Oasis of Northern Oman, the length of daylight is likewise divided into 24
asds, but their passage is regulated by observing changes in the length of a mans shadow.8
This is said to shorten by 40 qadams during the first suds (that is, after the shadow becomes
clearly identifiable), 20 in the second, 5 in the third, 3 in each of the fourth and fifth, 2 in
the sixth and seventh, and 1 in each of the remaining asds until midday. (Note that the clear
shadow is thus 79, or around 80 qadams long.)
In 1917 W. J. H. King published an account of a most curious procedure for regulating the
distribution of water in the Dakhla Oasis to the west of the Nile.9 The water supply for an entire
day from sunrise to sunset is divided into 31 qadams. Two men are involved in reckoning the
passage of the qadams. The meridian is marked and one man stands on it with his back to
the sun and bottom joint of his forefinger against the tip of his nose. The second man measures
with his feet the distance from the shadow of the finger to the meridian. The qadams are then
regulated according to the information recorded in the following table.
5
J. P. van der Stok, Opmerkingen omtrent bovengemelden zonnewijzer, Tijdschrift van het Bataviaasch
Genootschap van Kunsten en Wetenschappen (Batavia / Jakarta) 27 (1882), pp. 62-68, esp. p. 66.
6
On irrigation in the Islamic world see the articles M} in EI2. A new study of links between irrigation
and simple timekeeping is Thomas F. Glick and Simonne Teixeira, Azaira, alhetma: Two medieval Arabisms
Reflecting the Allocation of Irrigation Water, Suhayl 3 (2002-03), pp. 213-219. The duration of the uhr in
n. 3 on p. 214 is incorrectly cited.
7
See A. A. Nasif, Qanats at al-{Ul}, Proceedings of the Seminar for Arabian Studies 10 (1980), pp. 75-
80, esp. pp. 76-77, and idem, A Historical and Archaeological Survey of al-{Ul} with Special Reference to its
Irrigation System, 2 vols., unpublished Ph.D. thesis, Victoria University of Manchester, 1981, I, pp. 246-252,
and II, pl. CCLIII.
8
J. C. Wilkinson, Water and Tribal Settlement in South-East Arabia: A Study of the Aflj of Oman, Oxford:
Clarendon Press, 1977, pp. 106-110.
9
See W. J. Harding King, Irrigation in the Dakhla Oasis, Geographical Journal 50 (1917), pp. 358-364,
for details.
miscellaneous 523
Table 10.4
Morning Perpendicular Afternoon
(shadow west) in qadams (shadow east)
1 26-26 31
2 26-22 30
3 22-18 29
4 18-14 28
5 14-12 27
6 12-10 26
7 10-9 25
8 9-8 24
9 8-7 23
10 7-6 22
11 6-5 21
12 5-4 20
13 4-3 19
14 3-2 18
15 2-1 17
16 1W-1E 16
As King noted, the 16th qadam provides a rather generous allowance, but he obtained his
information from an old fell who raved wildly on the beauties of the system in such a vile
patois that even his village headman had some difficulty in following. Even in the 1910s the
headman possessed a watch, and the old methods are most probably no longer used nowadays.
King also described some methods of regulating water distribution by night using the moon.
Fig. 10.4 shows a sundial related to this kind of procedure from Muscat and Oman.
E. Stack, in his account of a journey through Persia in the late 19th century, presented a
detailed account of contemporary irrigation practices,10 and then alas, without further
information added:
In some villages of Faraidan I found the water-distribution by day regulated by the
length of a mans shadow, measured in feet; and at night by the stars. Elsewhere I
believe a dial is used by day.
The almanac entitled al-Taqwm al-Qaar which is published annually in the Arabian Gulf
state of Qaar (latitude ca. 25) contains a set of prayer tables and agricultural information for
each zodiacal sign and lunar mansion rising at dawn. I have examined a copy published in
the 1980s. Occasional midday shadow lengths in qadams are given and they were clearly lifted
from some earlier work in a rather haphazard fashion. The following values are given for
various signs (S) and mansions (M):
M10: 71/2 M2: 1/2 S11(): 6
M10: 3 S12(): 5 M12: 51/2
M26: 21/2 M14: 6 M28: 1
This information is, for all practical purposes, quite useless.
10
E. Stack, Six Months in Persia, 2 vols., London, 1882, II, p. 269 (cited in EI2, V, p. 870a).
524 part iii, chapter ten
For the sake of future researchers I include a reference to a shadow scheme for the solar months
in Sindi verse that was in circulation in Sind in the 18th century.11 It came to the attention of
the jurist Makhdm Muammad Hshim of Tatta (d. 1760), who pointed out that the shadows
had been calculated for some region other than Sind. He conducted his own observations in
1719-20 on a weekly / bi-weekly basis to produce to a table of the midday shadows in feet
(per) for each two weeks of the solar year. Values are said to be valid for Lower Sind and
are given to the nearest half a foot.
11
See Baloch, Measurement of Space and Time in Lower Sind, pp. 185-186, where the table is reproduced.
part iii, indexes of shadow-schemes 525
See also the schemes for Aden and Taiz cited in 4.5.
part iii, manuscripts consulted 527
Manuscripts consulted
N.b.: An asterisk denotes that I have not been able to consult the manuscript or obtain a copy thereof.
On the manuscripts of the almanac of Ibn al-Bann} which I have not consulted, see n. 6:6.