(Botanical Museum and Herbarium, Utrecht)

Transcription

1 Wentia 7 (1-116) Pollen morphology of the euphorbiaceae with special reference to taxonomy W. Punt (Botanical Museum and Herbarium, Utrecht) (received December 28th, 1961) CONTENTS Chapter I General Introduction 2 a. Introduction 2 b. Acknowledgements 2 Chapter II History 2 a. Pollen morphology 2 b. Euphorbiaceae 4 Chapter III Material 5 Chapter IV Methods 6 a. Flowers 6 b. Pollen preparations 7 c. Preservation of pollen grains 7 d. Microscopes 8 e. Punched cards 8 f. Drawings 9 Chapter V Some nomenclaturalremarks 9 Chapter VI Pollen morphology 10 Chapter VII Glossary 15 Chapter VIII Results 18 A. Pollen grains of the Euphorbiaceae 18 B. Discussion of the results 20 a. Phyllanthoideae 20 Antidesma configuration 20 Amanoa configuration 32 Phyllanthus nutans configuration 37 Breynia configuration 38 Aristogeitonia configuration 40 b. Crotonoideae 47 Croton configuration 47 Cnesmosa configuration 57 Dysopsis configuration 60 Plukenetia configuration 60 Chiropetalum configuration 65 Cephalomappa configuration 68 Sumbavia configuration 69 Bernardia configuration 73 Mallotus configuration 77 Claoxylon configuration 90 Cladogynos configuration 93 Hippomane configuration 95 Summary 106 References 107 Index 110

2 CHAPTER I GENERAL INTRODUCTION a. Introduction Many investigators have stated (e.g. Lindau 1895, Wodehouse 1935, Erdtman 1952), that pollen morphology can be of great importance for plant taxonomy, while it was also known that in Euphorbiaceae several types of pollen grains exist (e.g. Erdtman 1952). On the suggestion of Professor who Lanjouw, himself has worked on the Euphorbiaceae of Surinam, the author has investigated the pollen grains of this family of that area. From the result it was apparent that in the Surinam Euphorbiaceae many different pollen types could be distinguished. In the present work the study was therefore extended to all the genera in the Euphorbiaceae, from which one or more species have been examined. Besides a description of the pollen grains examined a drawing of most of the pollen types is given. The pollen morphologic groupings found have been compared with the systems of several authors (Bentham and Hooker 1880, Pax and K. Hoffmann 1931, etc.). The taxonomic conclusions, however, can only have a character. provisional Future investigators of the Euphorbiaceae will have to check the taxonomic notes and comments. b. Acknowledgements The investigations have been carried out at the Botanisch Museum en Herbarium, Utrecht. The writer is greatly indebted to the director, Professor Dr. J. Lanjouw, for giving him the opportunity to carry out the investigations at his institute and for his help and guidance in preparing this paper. The author is much indebted very to Professor Dr. F. P. Jonker who advised him in the special problems of pollen morphology. He very gratefully acknowledges the Ministerie van Kunsten en for Onderwijs, Wetenschappen sponsoring the present study. The author wishes to his sincere thanks to the express directors of the herbaria at Brussels, Leiden, Kew and Paris for the hospitality and help during his residence in these places. He is also very grateful to the Miquelfonds which covered the expenses of his stay in London, Paris and Brussels where so much valuablematerial could be examined. CHAPTER II HISTORY a. Pollen morphology The study of pollen morphology started in 1675, when Malpighi observed, that various types of pollen grains are to be found in different plants. He especially saw variations of colour and shape. Grew (1682) was already aware of a differencein structure, too. He distinguished

3 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 3 smooth and spiny pollen grains. In the next two centuries quite a number of botanists enriched our knowledge of pollen morphology.; In 1834 a synopsis of the literature on the subject was published by v. Mohl, which greatly contributed to a better understanding of pollen morphology. He studied representatives of most of the plant families and gave a descriptive classification of their forms, v. Mohl recognized that the apertures of the pollen grains are the most important morphological features. Some later years Fritzsche published his principal work on pollen grains (Ueber den Pollen, 1837). From the descriptions it is clear that, besides a careful study of the shape and apertures of a pollen grain, the structure was accurately analysed. His drawing and description of Jatropha panduraefolia gives an excellent idea of the pollen grain. The description even very accurately reproduces the construction of the croton-pattern (see PI. I, E4 and PI. VIII, 1). In the period after v. Mohl and Fritzsche we come across publications with drawings of pollen grains as additions to analytic; drawings of plants (e.g. in Martius, Flora Brasiliensis). Most botanists, however, added but little to pollen morphology. C. A. H. Fischer was the first scientist to give, in 1890, a comprehensive comparative study of the data then available. From the study of 2200 species in 158 families he was able to draw, the e.g. following conclusions: 1: Pollen grains of related species are generally similar. 2: Some families have more than a single basic form. 3: Sometimes unrelated plants have similar pollen grains. In 1883 Radlkofer made use of pollen morphology to obtain a better classification of the Acanthaceae. It \yas done more profoundly by Lindau (1895) in Die natiirlichen Pfianzenfamilien by Engler and Prantl when he stated: Es ist deshalb nur wenn auf consequent, die Pollenbeschaffenheit die Einteilung der Acanthaceae gegriindet wird, wie dies bereits von Radlkofer angcdeutet wurde und im folgenden ganz streng durchgefiihrt werden soil. It is obvious that, with regard to some families, pollen morpholqgy had become a wellknown and indispensable part of taxonomy at the end of the 19th century. In fact, a renewed general interest in pollen morphology emerged, not from the side of botany, but from the side of geology. It was the v. geologist Post who, in 1919, for the first time published a modern pollen analytic diagram; the primitive pollen analysis had resulted in a refined method in quaternary geology. However, the composition of a pollen diagram requires a sound knowledge of the morphology of pollen grains. Originally a primary acquaintance of the pollen grains of trees and anemophilous plants was considered to be sufficient, but more and more the necessity of a wider knowledge of pollen types became evident. Consequently the number of investigators who made researches into the pollen morphology of special taxonomic groups increased. Potoni in 1934 and Wodehouse in 1935 published works on pollen morphology that formed the basis of our modern terminology.

4 4 w. PUNT Potonie as well as Wodehouse needed a system of well defined terms for the description of pollen grains. The construction and details of a pollen grain were, however, not well enough known to be arranged in a convenient system. Nevertheless, many terms of these investigators are still in use, although sometimes the circumscription is more or less altered. of the terms In 1950 Faegri and Iversen, and Iversen and Troels-Smith published a system on the morphology of pollen grains. Their terminology excelled in exactitude and simplicity. The well-defined terms made it quite possible to determine pollen grains which occur in their determination-key of European pollen types. At the same time Erdtman also worked at a system of pollen terminology. He published several papers on this subject, but his main work (Pollen morphology and plant taxonomy, Angiosperms) appeared in In this book Erdtman proposed an extended system of In pollen terminology. many respects, however, the terminology seems too complicated to be of use. practical Only pollen morphologists with a large experience are able to identify all the terms mentioned in the glossary. Moreover, the descriptions of terms are not always as exact as they should be. What, for instance, is the exact boundary between brevicolpate and brevissimicolpate or foramen and foraminoid? Nevertheless, his book is a source of valuable data providing primary information of the pollen Angiosperms. types of all the families in the b. Euphorbiaceae This family is one of the largest in the Angiosperms. It is a difficult one on account of the many different species, so thatit is not surprising, that only a few botanists possessed a general knowledge of the whole family. Since the present report is mainly a pollen morphologic work, the following lines will only briefly comment on the principal authors on Euphorbiaceae. In the 19th century A. de Jussieu, J. Mueller of Argau and H. Baillon published important papers on the In Euphorbiaceae. the of Journal the Linnean Society (Botany) of 1880 Bentham reviewed the work of Baillon and Mueller Arg. In the same paper Bentham gave extensive notes on his own system. From Pax (partly assisted by Kathe Hoffmann) monographed most of the genera in Euphorbiaceae in Engler s Das Pfianzenreich. In the second edition of Engler and Prantl, Die Natiirlichen Pfianzenfamilien, Band 19c (1931), a revised synopsis of the Euphorbiaceae is given. In the present paper this synopsis of 1931 has been used as the basic classification of the family. After Pax and K. Hoffmann only incidental remarks have been made the on system of the Euphorbiaceae. Most of them were given in papers concerning floras of certain areas (e.g. Leandri, Flora of Madagascar; Gagnepain, Flora of Indo-Chine; Leonard, Flora of Congo). Important improvements have been made by Croizat on

5 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 5 several genera. Schultes amply discussed the Hevea and its genus allies. A very important work is Webster s publication on Phyllanthus. In his monographic study The West Indian species of Phyllanthus" he used as one pollen morphology of the main characters to divide the genus Phyllanthus into subgenera and sections. CHAPTER III MATERIAL Herbarium material obtained from the following herbaria was used in the study; London (BM): British Museum (Natural History). Bruxelles (BR); Jardin Botanique Bruxelles (BRLU): Herbier d Afrique de 1 Etat. de l Universit6 de Bruxelles. Kew (K); The Herbarium, Royal Botanic Gardens. Leiden (L): Rijksherbarium. Paris (P): Museum National d Histoire Naturelle Laboratoire de Phanerogamic. Utrecht (U): Botanical Museum and Herbarium. Wageningen (WAG): Laboratory for Plant Taxonomy and Plant Geography. Generally the use of type-material has been avoided. Only in those cases, where no other specimens were available while flowers were present, type-material has been used for pollen The classification of the Euphorbiaceae as given by investigation. Pax and K. Hoffmann in Engler and Prantl, Die natiirlichen Pflanzenfamilien ed. 2, Band 19c (1931), was the principal system used for the choice of genera. Besides this system attention has been paid to the new genera published after 1931 and listed in the Index Kewensis. As far as possible all known genera of the Euphorbiaceae were examined. In a number of genera it was, for various reasons, impossible to study the pollen grains. They fall in the following groups: 1. Male flowers unknown. Croizatia Lasiochlamys Paradrypetes Phyllanoa Steyermark Pax et K. Hoffmann Kuhlman Croizat 2. The material studied did not contain male flowers, or only a few, or unripe ones. Claoxylopsis Calicopeplus Cometia Corythea Ditta Keyodendron Mettenia Monadenium Leandri Planghon Thouars ex Baillon Watson Grisebach Leandri Grisebach Pax

6 6 w. PUNT Neomphalea Pax et K. Hoffmann Neotrigonostemon Pax et K. Hoffmann Ramelia Baillon Riseleya Hemsley Syndyophyll um Schumacher et Lauterbach Sphyranthera Hooker, J. D. Thelypetalum Gagnepain 3. Genera of which no material has been seen. Acalyphopsis Annesijoa Chlamydojatropha Clarorivinia Gitaria Senefelderopsis Pax et K. Hoffmann Pax et K. Hoffmann Pax et K. Hoffmann Pax et K. Hoffmann Pax et K. Hoffmann Steyermark On account of the large number of genera and species the author has found it impossible to be always sure of the accuracy (both taxonomic and nomenclatural) ofthe names used in the present paper. Future monographic studies will perhaps prove some names to be incorrect. With all the descriptions of pollen grains, however, the collector s name and number and the herbarium, where the specimen is preserved, are In given. those cases when the name has to be corrected the description of the pollen grain can thus always be placed under the right name. In the Rijksherbarium at Leiden a large collection of plants of D Alleizette is present. These plants, however, have not been collected by D Alleizette himself. In the herbarium at Paris D Alleizette obtained duplicates from several collectors. These duplicates were sent to the Rijksherbarium without mention of the names and collectioning numbers of the original collectors. As it lies beyond the of this scope work to search for the names of the original collectors, the name of D Alleizette has been maintained in the present work. CHAPTER IV METHODS a. Flowers From each atleast genus one species and when possible two species were examined. From those genera that can be devided into two or more sections, species were chosen from different sections. Usually the male inflorescenses of generally speaking, no important loss occurs Only the Euphorbiaceae have many flowers, so that, it one or two flowers are dissected. - those flowers of the Euphorbiaceae are useful that have just opened. When the flowers are younger the pollen has not yet ripened, so that it is impossible to give an accurate description of shape and structure of the pollen grains. It is far better to use old, outpollinated anthers than young and undoubtedly unripe ones. In the corners of their thecas the old anthers always have some pollen grains which, by a special micro-method described below, can be isolated.

7 POLLEN MORPHOLOGY OF THE EUPHORBIACEAB 7 b. Pollen prepararions For the study of the morphology of pollen grains it is desirable that cytoplasma and intine are dissolved. The microscopic image becomes obscure and the important cxine remains indistinct if intine and cytoplasma are not dissolved. For that reason Wodehouse s methylene-green method, the lactic-acid method and any other method, which does not completely dissolve intine and cytoplasma are considered less suitable. The treatment of the pollen grains with the acetolysis-method of Erdtman (1943) produces the best results. Everything inside the exine is dissolved, while, moreover, a colouring of the exine takes place. This method, described by Erdtman in 1943 and 1952 for herbarium material, has, in its original form, a great disadvantage. Too much material is required to obtain a usable preparation. In order to reduce this disadvantage a micro-method (see below) has been applied, which leaves the method as such untouched, but requires less material. If necessary even only one anther is sufficient. Micro-method The micro-method is a perfection of the method described by Willrath in a publication of Potonie (1934). A flower is boiled in water till it is free of air. Under a preparing-microscope one anther (or more if the material allows) is dissected and laid in the pit of a hollow slide. Some drops of acetolysis mixture (9 parts anhydric acetic acid : 1 part H2SO4 cone.) are dropped on the anther till the pit has been filled with the mixture. Now the slide is placed on a heating-apparatus as shown in Fig. 1. The end of the heating-apparatus is heated by Bunsenburner. The flame should a not be too high, Fig. 1 but not too low either. The slide furthermore has to be removed so far from the flame that the mixture cannot take fire. A part of the mixture will evaporate. When the slide has become almost dry some more drops are added. This is repeated 3 or 4 times till the remaining mixture is turning dark brown. When the preparation once more is nearly dry, it is quickly cooled down by placing the slide on a stone or iron surface. Under the dissecting microscope the pollen grains are to be seen as brown globules. With anthers which have a solid skin, that cannot easily be dissolved, it is desirable, while heating, to prick this skin in places with a needle to make the acetolyse mixture penetrate deeply enough into the anther. c. Preservation of pollen grains For the preservation of grains the paraffin-comprising method has been used. A small drop of glycerin (diam. 1 mm.) is laid on a slide. With a brush moisted in glycerin the pollen grains are wiped off the slide or fished out of the liquid. In

8 8 w. PUNT the drop of glycerin the pollen grains come off the brush more or less easily. This is repeated till at least 10 grains are present in the glycerin drop. Now little a piece of glycerin-gelatin is added. The gelatin is melted and the grains mixed with it. To are prevent the pollen grains from expandingabnormally the whole should not be heated too much. It is of the utmost importance that the grains can be studied from all sides, so that care should be taken that the coverglass does not flatten the grains. Therefore a granule of clay (no synthetic material) is laid next to the glycerin drop. The cover-glass is now supported by the clay when placed on the preparation and the pollen grains cannot be compressed. By lightly pressing the cover-glass the thickness of the preparation can be regulated. By running melted paraffin under the cover-glass the grains are isolated from the air. Webster (1956) rejects the method of sealing up by paraffin on the ground of the preparations not durable. In Utrecht, however, preparations ten years old are still in excellent condition. It is, however, essential that the layer of paraffin should not be too thin and the drop of glycerin should not approach the edge of the cover-glass (Erdtman and Praolowski 1959). Preparations made by the micro-method have the disadvantage that some acid will enter with the pollen grains. This may cause the colour to fade after some time. After three years, however, form and structure were still unchanged. For that reason the macro-method oferdtman is always preferable to the micro-method, but the latter has the advantage that very little material is sufficient, so that any rigorous sacrifice of herbarium-material may be prevented. In the case of flowers with large anthers it is even possible that part of a stamen suffices (Campanula, Lilium). In describing pollen grains it is of great importance that the method used for preparation should be indicated, since the size of the pollen grain varies with it. Though the expansion of a certain pollen grain is constant in either method, only relative value can be attached to the indication of the size of that particular pollen grain. Only with regard to pollen grains that have been treated in precisely the same way the grain measures retain their value. Moreover, the size is influenced by the grade of ripeness ofthe pollen. When a pollen grain is on the verge of maturing, form and structure may already be fully developed, while the size has not yet reached its maximum. It is, therefore, again emphasized that all absolute numbers which are mentioned in the descriptions have a relative value. For this reason any accurate calculation of sizes has been average relinquished. The figures give an impression of the size in these special circumstances and allow a mutual comparison of the examined pollen grains. The proportionate numbers, however, are more dependable. Proportions remain the same in either method. Some proportionate numbers are, e.g.. Polar Area Index (P.A.I.) and Polar axis: Equatorial axis ( P:E) (see Pollen Morphology p. 12). d. Microscopes The pollen grains were studied with a Bausch and Lomb binocular (objective 97 x, oculair 10 X ), and an Olympus phase contrast microscope was used. The employment of this microscope was found very useful in detecting the configuration of structure elements. Most of the structure patterns also are visible with the ordinary light-microscope, but, especially in the case of small structure elements, not without difficulty. The phase contrast microscope immediately gives clear a image of the structure. e. Punched cards The study of large quantities of material requires special methods of compiling the knowledge obtained. A simple card-index is inadequate when there are many genera to be mutually compared; the great number of characters hampers any efficient arrangement. A solution of this problem was found in the use of punched cards. Punched card systems were made for the pollenmorphologic as well as for the taxonomic characters. Thus any particular botanical object could be classified in two independent The of within systems. those groups genera systems are now mutually comparable so that it became possible to recognize correlations

9 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 9 of characters, which, without the punched cards, would have been found at the cost of much time and labour or not at all. A further advantage is that the investigator is compelled to look over the same characters continually. Even the absence of a special character may be of importance and finds expression in this system. A disadvantage may be that too much reliance is placed upon the characters noted. Often characters that are indicated in the same way are not mutually comparable. If e.g. stamens are connate, this coalescence can take various forms. There be may investigators, who are inclined to distrust this system. Nevertheless the of use punched cards and other mechanic systems should undoubtedly be taken into consideration, especially when studying larger taxa. Surveying and working up the continually increasing numbers of characters in a reasonable time is becoming more and more difficult. Loss of time, chance of errors and the risk of being incomplete are apt to compel the taxonomist to utilize some form of mechanic taxonomy. Who would, at present, venture to revise the whole Solanum genus or Euphorbia? Only through a perfection and especially through an acceleration of our methods the larger taxa will as yet have a chance of being monographically worked on. f. Drawings The pollen grains have been drawn in such a manner that in one drawing as many characters as possible are to be seen at the same time. Each drawing therefore is midway between a scheme and a photographical reproduction. The pollen grains have been drawn without a camera lucida or other drawing instruments. Drawing has the advantage that the most important characters can be indicated distinctly, so that the grains are easier to distinguish from others. A disadvantage is the subjective image which the drawing is apt to give. The other possibility of reproduction is making use of photography. For the following reasons this method has not been applied; 1. In most cases at least 4 photographs of perfect quality are required to obtain a distinct picture of the grain. 2. In a photograph the structures of the pollen grains can never be pursued distinctly, because the photographic image will only be sharp for a certain optical section. 3. Deeper lying layers debase the photographic image and cause an insufficient sharpness of the essential characters. The plates have been drawn to scale. Each picture is 1000 X enlarged, except some of the large pollen grains (e.g. Manihot and Croton matourensis),,, which are less enlarged. In those cases where the enlargement is not 1500 X the scale is given in the text to the plates. CHAPTER V SOME NOMENCLATURAL REMARKS Bischofia Blume Although Blume dedicated this genus to Bischoff he spelled, in his Bijdragen the (1826), name with a single f. Obviously he latinised Bischoff to Bischofius. For this reason the name of the should be written in the genus original spelling. Briedelia Willdenow In his Species Plantarum (4: ) Willdenow deliberately spelled the genus with i e. He commented as follows on the choice of the name: Genus hoc a Cluytia et Rhamno abunde diversum in honorem Clariss. S. E. Briedel nominavi. Although the name of the botanist in question is spelled Bridel it is not clearly demonstrated that the author made an unintentional error. Under the rules of the international code the should be written in the genus original spelling.

10 10 w. PUNT substitute name Anabaenella to this taxon. They did not know, however, that an Cnesmosa Blume In his Bijdragen (1825) Blume spelled this as Cnesmosa. genus In the Flora van Java, Praefatio in adnotation, he changed the name, probably for grammatical reasons, in Cenesmon. This is not allowed under the rules of nomenclature. The original spelling should be maintained. Koilodepas Hasskarl Koilodepas was first published in 1855 in the Greek version (Versl. Med. Akad. Wetenschappen Amsterdam 4: 139). Two later years the name was altered in Coelodepas (Flora 40; ). As Croizat (1942b) and Airy-Shaw (1960) have stated, the validity of the spelling Koilodepas is unquestionable. Omalanthus A. de Jussieu The first spelling ofthis genus appeared in A. de Jussieu s De Euphorbiacearum in In his Conspectus (1828) H. G. L. Reichenbach altered the name in Homalanthus. Of course the first spelling has to be maintained. Carumbium is a synonym of Omalanthus. Reinwardt published this name as a nomennudum in Oken s Isis So Omalanthus is the name first published and therefore valid. Romanoa Trevisan In 1824 A. de Jussieu published the genus Anabaena. De Bory, however, had given the name Anabaina to a new algae genus (Cyanophyta) in 1821, so that the name of A. de Jussieu is a homonym. In 1918 Pax et K. Hoffmann gave the older name was available. In 1848 Trevisan had substituted Anabaena A. de Jussieu by Romanoa (Alghe Coccotale p. 99). If the genus is held separate from Plukenetia, Romanoa is the first correct name for Anabaena A. de Jussieu. CHAPTER VI POLLEN MORPHOLOGY In describing pollen grains the number of terms employed should be as small as possible. This was the starting point of Faegri and Iversen in One of the arguments they advanced was the difficulty of giving an analytic description of a pollen grain in exact terms. The system, however, that resulted from their attempt at PLATE I. A. 1. pori; 2. colpi; 3. prolate pollen grain; 4. oblate pollen grain; Polar axis; E Equatorial axis. B. polar view; 1. triangular; 2. three lobed; 3. circular; 4. convex triangular; S smallest distance between two colpi ends; G greatest breadth; S:G Polar Area Index. C. 1. and porus colpus nudate; 2. p. and c. with a membrana granulata; 3. p. and c. with an operculum; 4. and p. c. with a margo (thickening a or thinning b ofthe ekt.); 4c. costae pori and costae colpi; 5a. porus with a vestibulum; 5b and c. composite aperture with an atrium. D. 1. colpus transversalis; 2. circular colp. transv.; 3. equatorial colp. transv.; m meridional axis of the cop. transv.; e. equatorial axis of the colp. transv.; o os; 4. costa transversalis; 5. costa circularis; 6. costa equatorialis. E. 1. tectum; 2. tectum perforatum; 3. reticulum; 4. croton-pattern. F. 1. psilate; 2. scabrate; 3. verrucate; echinate; 8. pila. gemmate; 5. clavate; 6. baculate;

11 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 11 PLATE I.

12 12 W. PUNT simplicity was inadequate, and amplification was necessary. Especially when the amount of material to be described is considerable many difficulties are liable to arise. The number of characters of a pollen grain is comparatively smalland they often appear in many transitional forms. Yet a precise and exact definition of characters is desirable. Erdtman, who, in his study of the met with Angiosperms (1952), these difficulties, therefore enlarged the He terminology considerably. tried to nominate every transitional form, each nuance, which could be achieved only at the cost of exactitude and convenient arrangement. Nevertheless many pollen morphologists will feel highly attracted towards his method. For the use of pollen morphology as an additional character in taxonomy this system is too complicated. The more obvious characters will suffice. For that reason a compromise seems to be the right course for the description of diverse pollen grains. Erdtman the present author has attempted Without going as far as to use a terminology as simple as was possible without loss of exactitude and completeness. Pollen grains are composed of three layers. The central is the part The cytoplasma. middle layer is the intine; the outer layer the exine. The exine of pollen grains consists of a particularly resistant substance of a lipoid character: sporopollenin. Sporopollenin is even more resistant to the attack by micro-organisms than cutin. The exine substance is also highly resistant to all sorts of chemicals; it can not, or only slowly, be attacked by fluid KOH, concentrated H2SO4 or even fluoric acid (HF). The exine which can be devided in two layers, the endexine and the ektexine, supplies most of the data necessary for pollen description. These data fall into three groups: 1. Shape 2. Apertures 3. Structure 1. Shape If a pollen grain is not spheroidal we can distinguish two axis: a. Polar axis (P) (PI. I, A3 and A4) b. Equatorial axis (E) (PI. I, A3 and A4) In equatorial view pollen grains show a shape which will be determined easily now by calculating the P: E. to this According relation we distinguish the following classes (Erdtman 1952): P : E > 2 : perprolate 1,33-2 : prolate (PI. I, A3) 1,14-1,33: subprolate 1 1,14: prolate spheroidal 1 : spheroidal 1-0,88: oblate spheroidal 0,88-0,75: suboblate 0,75-0,50: oblate (PI. I, A4) < 0,50: peroblate

13 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 13 In polar view the circumference of a pollen grain can be: 1. triangular (PI. I, Bl) 2. convex triangular (PI. I, B4) 3. circular (PI. I, B3) 4. three lobed (PI. I, B2) 2. Apertures In the system of Iversen and Troels-Smith the principal classification is made according to the occurence of apertures: pori (PL I, Al), colpi (PI. I, A2) and colpi transversales (PI. I, Dl, 2, 3) (in composite apertures only). Real pori and real colpi are formed by thinnings in the ektexine. A colpus transversalis, on the other hand, by a thinning in the endexine. For that reason of the endexine and the ektexine apertures have to be described of each other. independently Pori can be distinguished from colpi by calculating the length: breadth. If length: breadth < 2 the is aperture a porus although the shape is not necessarily a pure circle. If length: breadth > 2 the aperture is a colpus. The endexine below the edges of or colpi pori may be thickened. These thickenings are called costae colpi and costae pori (PI. I, C4c). The colpus transversalis (PI. I, D) is an easily recognisable character. It is a thinning of the endexine. Yet this aperture has been inaccurately described Faegri and by many investigators. Iversen confuse (1950) this with aperture a real porus when the colpus transversalis has the shape of a circle. Consequently, when using their key for the determination of European pollen it is types, sometimes difficult to decide if a porus longitudinal elongated or a transversal furrow is present. Erdtman calls colpi + colpi transversales composite apertures. For the term colpus transversalis he introduced the new term os (Pi- ora). If this term is used in that sense the name will have to be rejected as being a synonym. It may, however, be maintained for that part of the colpus transversalis that crosses the colpus (PL I, D3). transversales stand Colpi perpendicular on the colpi. Usually they are boat-shaped. Sometimes they fuse together into one single aperture along the equator of the pollen grain (colpus transversalis equatorialis), but it is also possible that the colpus transversalis becomes isodiametric (colpus transversalis circularis PL I, D2). The medial dimension of the colpus transversalis is rarely larger than the equatorial one ( Pleiostemon type, p. 30). Edges of transversales colpi can also have thickenings; according to the shape of the colpus transversalis they are called: costae circulares, costae transversales and costae equatoriales (PI. I, D4, 5, 6). To express the proportions of the colpus transversalis in exact figures, the m: e quotient is added to the description of this character. In this formula m is the meridional axis and e the axis of the equatorial colpus transversalis (PI. I, Dl). Iversen and Troels-Smith (1950) introduced the term Polar Area Index ( P.A.I. ). The P.A.I. is the ratio between the greatest distance of the ends of two colpi and the greatest breadth of the pollen grain

14 14 W. PUNT u x usually the Equatorial axis. This character is of great value for the description of pollen grains. 3. Structure The ektexine furnishes characters of great diagnostic importance. The description of these characters with meets, however, considerable difficulties, so that it is necessary to use a system as simple as possible and highly desirable to avoid confusing terms (e.g. sculpture, see below). The ektexine is composed of structure elements. Different forms and arrangements of these elements cause several structure types to occur. The most common structure elements occuring on the endexine are pila (PI. I, F8). Each pilum consists of a swollen apex' (caput) cylindrical basal and a part (collum; homologous columella). Pila are strictly found on the endexine. They may occur ordinate or inordinate. Ordinate arranged pila can form regular figures (e.g. reticulum PI. I, E3) or stand in rows (e.g. striae). Frequently capita of are connate or pila fused. If the fusion is and less of the surface only laterally, is covered by the capita, we speak of tegillate pollen grains. Pollen grains are tectate when the fused capita cover 80 % or more of the surface (PI. I, El). In that case the form a second membrane pila outside the endexine. If the tectum is not completely closed (fusion of the capita more than 80 % and less than 100 %) the perforate membrane is a tectum perforatum (PI. I, E2). On the top of the tectum we distinguish shape. structure elements of different According to Faegri and Iversen they are called: baculae, clavae, echinae, gemmae, scabrae and verrucae (PI. I, F2 t/m 7) (for descriptions see glossary). Psilate pollen grains have no structure elements on the tectum (PI. I, FI). Structure elements on the top of the tectum may also be arranged in different structure types (reticulum, striae). If a tectum is present, corresponding structure of types the columellae are called intra-reticulum, intra-striae. Sometimes columellae can give an impression of a reticulum. This is, however, a pseudo-reticulum in which each columella represents a lumen and the space between the columellae, the muri. A real reticulum is always formed by In the present study no difference is sculpture. According to Potonie (1934), form the sculpture of a pollen grain. several structure elements. made between structure and all elements outside the tectum Potonie tried to give sharp definitions of these terms, but admitted at the same time, that structure and sculpture pass into each other. Faegri and Iversen (1950) also stated, that structure and sculpture are to confuse. For that easy reason it seems better to speak of structure elements and structure only.

15 pi. pi. - pi. CHAPTER VII GLOSSARY Annulus (Iversen and Troels-Smith 1950) (PI. I; G 4a en b) Area which surrounds a porus and is distinguishedby a thickening or a thinning of the ektexine; thus forming a prominent or depressed margin of the latter. Aperture (Faegri and Iversen 1950) Any weak performed part of the ektexine or endexine of a pollen grain (e.gcolpus, colpus transversalis). Atrium (PL I; 5a, b en c) A cavity inside a composite aperture caused by a separation between two layers of the exine. Bacula - pi. Baculae (Iversen and Troels-Smith 1950) (PI. I; F 6) Structure elements with at least one dimension 1 or larger in the shape of rods. Occur on the tectum only. Baculate (Iversen and Troels-Smith 1950) Pollen grains provided with baculae. Caput - Capita (Erdtman 1952) Upper part of a pilum. Clava - Clavae (Iversen and Troels-Smith 1950) (PI. I; F 5) Structure elements with at least one dimension 1 fi or larger in the shape of clubs. Occur on the tectum only. Clavate (Iversen and Troels-Smith 1950) Pollen grains provided with clavae. Collum (Erdtman 1952) Lower part of a pilum; homologous with columella. Columella Columellae (Iversen and Troels-Smith 1950) Lower part of a pilum. Columellae bear the tectum. In 1956 Faegri used the term as a synonym for pilum. Colporate (Erdtman 1945) Pollen grains with composite apertures. Colpus pi. Colpi (Erdtman 1943) (PI. I; A2 and Cl) Longitudinal of apertures the ektexine. Length : breadth > 2. Colpus transversalis (Erdtman 1943) (PI. I; D 1, 2, 3) Aperture of the endexine perpendicular to the colpus. Appears in composite apertures only. Composite aperture (Erdtman 1952) (PI. I; D 1, 2, 3) Ektexine aperture combined with an endexine not congruent). aperture (apertures Costa - pi. Costae (Faegri and Iversen 1950) (PI. I; C 4 and D 4, 5, 6) Thickenings of the endexine. Costae circulares (D5): Thickened edges of circular colpi trans\ ersales.

16 16 w. PUNT pi. pi. pi. Costae colpi (C4c): Thickened endexine below the edge of colpi. Costae equatoriales (D6): Thickened edges of colpi transversales extended along the equator. Costae pori (C4c): Thickened endexine below the edge of pori. Costae transversales (D5): Thickened edges of colpi transversales. Croton-pattern (Erdtman 1952) (PI. I; E4) Structure elements in rings of 5 or 6 around a circular area. Echina - Echinae (Wodehouse 1928) (PI. I; F7) Structure elements with at least one dimension 1 or larger in the shape of spines. Occur on the tectum only. Echinate (Wodehouse 1928) Pollen grains provided with echinae. Ellipsoid Inaperturate pollen grains with a long and a short axis. Endexine (Erdtman 1943) Inner part of the exine. Ektexine (Erdtman 1943) Outer part of the exine. Equatorial axis (Van Zinderen Barker 1953) (PI. I; 3E and 4E) The line perpendicular to the Polar axis in the equatorial plane. Exine (Fritzsche 1837) The outer, usually resistant, layer of the wall of a pollen grain. Gemma Gemmae (Iversen and Troels-Smith 1950) (PI. I; F4) Structure elements with at least one dimension 1 or larger of which the lower parts are constricted. Occur on the tectum only. Gemmate (Iversen and Troels-Smith 1950) Pollen grains provided with gemmae. Inaperturate (Iversen and Troels-Smith 1950) Pollen grains without apertures. Intectate (Iversen and Troels-Smith 1950) Capita of the pila covering less than 80 % of the surface. Intra-reticulate (Faegri and Iversen 1950) Columellae inside the tectum form a network. Intra-striate (Faegri and Iversen 1950) Columellae inside the tectum stand in rows. Lumen - Lumina (Potonie 1934) Spaces between the muri of a reticulum. Margo (Iversen and Troels-Smith 1950) Area which surrounds a colpus and is distinguished by a thickening or thinning of the ektexine; thus forming a prominent or depressed margin of the latter. Membrana granulata (Erdtman 1952) (PI. I; C2) Colpus or porus membrane ( endexine) with some scattered structure elements.

17 pi. pi. pi. pi. pi. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 17 Murus Muri (Potoni6 1934) Ridges separating the lumina of a reticulum. Operculum (Wodehouse 1928; Faeori and Iversen 1950) (PI. I; C3) Isolated part of the ektexine which is separated from the edges of a colpus or porus by a narrow zone in which the ektexine is missing or greatly reduced. Os Ora (PI. I; D3, o) That part of a composite aperture which is formed by the crossing of the colpus and the colpus transversalis. Periporate (Iversen and Troels-Smith 1950) Pori T uniformly distributed over the surface of a pollen grain. Pilum Pila (Erdtman 1952) (PI. I; F8) Structure elements consisting of a swollen apex (caput) and a cylindrical basal part (collum, columella). Occur on the endexine only. Polar Area Index P.A.I. (Iversen and Troels-Smith 1950) (PI. I; B3) The proportion of the greatest distance between the ends of two furrows and the greatest breadth of the pollen grain (usually the Equatorial axis). Polar axis (Erdtman 1943) (PI. I; A3p and A4p) The perpendicular line connecting the poles of a pollen grain. Porus - Pori (Erdtman 1943) (PI. I; A1 and Cl) Circular apertures of the ektexine. Length breadth : < 2. Psilate (Wodehouse 1928) (PI. I; FI) Surface of tectum smooth. Reticulum (Potonie 1934) (PI. I; E3) Structural pattern in the form of a network in which columellae represent the muri of the reticulum. Scabra - pi. scabrae (Iversen and Troels-Smith 1950) (PI. I; F2) Structure elements visible but smaller than 1 /i. Occur on the tectum only. Scabrate (Iversen and Troels-Smith 1950) Pollen grains provided with scabrae. Sculpture (Potonie 1934) A term proposed by Potonie for the joint structure elements outside the tectum (p. 14). Stephanocolpate (Iversen and Troels-Smith 1950) Pollen grains with more than three colpi perpendicular to the equatorial plane. Stephanocolporate (Iversen and Troels-Smith 1950) Pollen grains with more than three composite apertures. Stephanoporate (Iversen and Troels-Smith 1950) Pollen grains with more than three apertures in the equatorial plane. Stria - Striae (Iversen and Troels-Smith 1950) Structure elements standing in rows forming narrow grooves (Jt parallel). Striate (Iversen and Troels-Smith 1950) Pollen grains with striae.

18 pi. 18 w. punt Structure Form and arrangement of the structure elements. Structure elements Individual elements either on the endexine or on the tectum. Syncolpate (Iversen and Troels-Smith 1950) Colpi anastomosing at the poles. Tectate (Iversen and Troels-Smith 1950) (PI. I; El) The capita of pila are fused and cover together 80 % of a pollen grain. or more of the surface Tectum (Iversen and Troels-Smith 1950) (PI. I; El) The membrane outside the endexine formed by fused capita of pila. Tectum perforatum (Iversen and Troels-Smith 1950) (PI. I; E2) The tectum is not completely closed. Small holes are present. Tegillate (Erdtman 1952) The capita of the pila are laterally united or fused and cover less than 80 % of the surface of a pollen grain. Tricolpate (Iversen and Troels-Smith 1950) Pollen grains with three colpi perpendicular to the equatorial plane. Tricolporate (Iversen and Troels-Smith 1950) Pollen grains with three composite apertures. Triporate (Iversen and Troels-Smith 1950) Pollen grains with three pori in the equatorial plane. Verruca - Verrucae (Iversen and Troels-Smith 1950) (PI. I; F3) Structure elements with at least one dimension 1 /i or larger in the shape of warts. Occur on the tectum only. Verrucate (Iversen and Troels-Smith 1950) Pollen grains provided with verrucae. Vestibulum (Iversen and Troels-Smith 1950) (PI. I; C5a, b) Cavity inside a porus, caused by a separation between two layers of the exine. CHAPTER VIII RESULTS A. POLLEN GRAINS OF THE EUPHORBIACEAE Pollen grains of the Euphorbiaceae show a number of types, which can be distinguished more or less clearly. Generally pollen grains of species inside one genus belong to the same pollen type. The dimensions and proportions may differ, but in principle the pollen grains identical. In are some larger genera different pollen types occur in one genus ( Tragia More Phyllanthus). often we find that different, genera possess the same pollen type. Consequently it is not possible, as it is

19 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 19 with the Acanthaceae, to place a plant instantly in a genus by means of its pollen grains. For the greater part the pollen types within the subfamilies Phyllanthoideae and Crotonoideae can be distinguished clearly from each other. Some types, however, show some affinities; e.g. Clutia type (p. 77) with Phyllanthus pentaphyllus subtype (p. 24), and Amanoa type (p. 33) with Sumbavia type (p. 69). According to the system of Iversen and Troels-Smith (1950) the following principal types occur; inaperturate, periporate, triporate, stephanoporate, tricolpate, stephanocolpate, tricolporate, stephanocolporate. Pollen grains with composite apertures occur most frequently. Especially the number of tricolporate pollen grains is large. The shape varies from perprolate (P; E > 2,00) to oblate (P : E 0,50-0,75). In polar view the pollen grains are mostly convex triangular. In some cases the circumference is triangular (dclutia type), circular ( Moultonianthus type) or three lobed ( Omalanthus nutans subtype). The colpi are usually narrow, but sometimes broad ( Amanoa type, Plukenetia type). In many types the colpi are accompanied by costae colpi ( Antidesma type, Mallotus type). The P.A.I. diverge from large (> 0,7) to syncolpate. Syncolpate pollen grains are, however, rarely found in the Euphorbiaceae (Amperea type)- Colpi transversales, if present, can be of great diagnostical value. Shape and size are characteristic for several pollen types (Securinega type, Hippomane type). Pollen grains of Euphorbiaceae are often tectate. Sometimes the columellae supporting the tectum are so small that they are barely visible ( Alchornea type, Ricinus type). If the pollen grains are intectate, they generally possess a reticulum. In a few cases only the pollen grains are intectate, pilate (some Phyllanthus species, Tragia fallax type). On the tectum structure elements of different shape may be found; e.g. clavae, baculae, echinae, etc. The croton-pattern is a structure type which we meet in many genera of the Crotonoideae. Outside the Euphorbiaceae this structure type is found in Buxaceae and Thymelaeaceae (Erdtman 1952). In the following treatment of the investigated material the pollen grains are first divided into the two subfamilies: a. Phyllanthoideae b. Crotonoideae Inside these subfamilies the pollen grains are discussed according to their similarity in type. Different pollen grains are placed in different pollen types. So the type unit has been chosen as the base of the classification. Descriptions of a type are therefore as extensive as possible. If the differences are

20 20 w. PUNT ~~ ofminor importance, the pollen grains are placed in subtypes. Several types can have some characters in common. To express the correspondences, these types are assembledin configurations. After the diagnoses of the pollen groups comment is given on the pollen morphologic characters, mostly in their relation to other Besides pollen groupings. this comment in many cases a taxonomic discussion is added, in which the value of pollen morphology in taxonomic problems is discussed. B. DISCUSSION OF THE RESULTS a. Phyllanthoideae Antidesma configuration Tricolporate or stephanolcolporate (4). Pollen grains with a Polar axis larger than the Equatorial axis. P. rarely shorter than the E. (e.g. Phyllanthus acidus subtype). Colpus transversalis usually with costae. Pollen grains reticulate or not reticulate; never echinate. The Antidesma configuration is characterised by its prolate shape of the pollen grains. The Polar axis is usually larger than the Equatorial axis. In the Phyllanthus acidus subtype (p. 26) only the pollen grains are spheroidal or slightly oblate spheroidal. However, the pollen grains resemble so much the Securinega subtype, that it seems better to place them in the same type. The three main of types the configuration are the Antidesma type, the Securinega type and the Dicoelia type. The three types are closely related and have several characters in common e.g. shape, costae colpi and costae transversales. The other types resemble the main types but lack one of the important characters e.g. costae ( Heywoodia type, p. 30) or have special characters (verrucae in the Zimmermannia type, p. 29). Fig. 2 The differences between the subtypes are rather faint and difficult to define and separate. The most reliable character for determination and separation is found in the colpus transversalis. These apertures are bordered by costae which make them distinctly visible. In some Antidesma subtypes (e.g. subtype) the rims of the costae transversales run parallel and the outer ends of the colpus transversalis are diffuse. Most Phyllanthus species have costae transversales with rounded ends. The outer ends of the colpus transversalis are in that case distinct and frequently even closed.

21 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 21 Antidesma type Tricolporate; perprolate to prolate spheroidal. Colpus transversalis small or narrow elongated; costae. Colpi costae narrow; colpi. Tectate; psilate. Mostly not reticulate; rarely indistinct reticulate. The Antidesma type differs from the Dicoelia type (p. 28) by the shape and dimensions of the colpus transversalis. In other characters the two types cannot be differentiated. From the Securinega the Antidesma differs type type not only by the shape of the colpus transversalis but also by the structure of the exine. Taxonomic discussion The plants in the Antidesma type can be divided into two groups. The first group combines the Antidesma subtype with the Baccaurea subtype and the second includes the two group Phyllanthus subtypes. The first group comprises most of the genera which Pax and K. Hoffmann (1922, 1931) united in the subtribe Antidesminae. The Antidesminae are chiefly characterised by the inflorescense which is catkin-like or in slender spikes. The following characters are of importance too; 1. The male flowers have only few (six stamens round an ovarium rudiment (in Aporosa is small or absent). or less; usually five) the ovarium rudiment 2. Usually a disc is present. 3. In the female flowers a disc may be present; the styles are usually bifid. 4. The plants are nearly almost dioecious. All the genera are found in the Old World except Hyeronima, which is the American representative of Antidesma (Bentham, 1880). Chonocentrum was first recognised by Mueller as a species of Drypetes. Pax and K. Hoffmann (1922, 1931) separated the plant from this genus and placed it in the vicinity of Pollen of the Discocarpus. grains latter genus are however quite different. Chonocentrum differs from Discocarpus by the following characters: 1. Petals are never present. 2. The stamens are not connate but free. Some genera included in the Antidesminae by Pax and K. Hoffmann possess pollen grains without an Antidesma type. As for the genera Dicoelia, Richeria and Richeriella they belong to related types. Martretia and Hymenocardia have totally different pollen grains. The second group of the Antidesma type includes many species of Phyllanthus. Also pollen grains of Reverchonia possess the Phyllanthus pentaphyllus subtype. In many respects the genus is related to Phyllanthus. The branching of the plants, however, is not phyllanthoid. The male disc, always present in' Phyllanthus, is rather indistinct. The two stamens are inserted on a thick disc-like tissue. As in Phyllanthus there is no ovarium rudiment, while in the female flowers a disc is present.

22 22 w. PUNT - Antidesma subtype Pollen grains tricolporate; perprolate prolate. Colpus transversalis long; outer ends indistinct and diffuse. Edges of the costae transversales parallel. The pollen grains are strongly elongated and have narrow elongated colpi transversales. Though always referred to the genus Securinega, Securinega congesta has pollen grains largely different from the other pollen grains in the genus. The large P: E and narrow elongated colpus transversalis place the pollen grains straight into the Antidesma subtype. Antidesma bunius Java [U] PI. II, 2 (Linn.) Spreng. Antidesma subtype; perprolate. 37,5 E n 18,5 P:E n 2,04. Antidesma venosum Tul. m:e < 0,5 P.A.I. 0,25. Antidesma subtype; prolate. Stolz 476 [U] 19 fi E 13 n P:E 1,50. Antidesma agusanense Elmer Elmer [U] Antidesma diandrum (Roxb.) Roth Antidesma subtype. Idem Antidesma Antidesma subtype. venosum. Hohenacker 167 [U] Idem Antidesma venosum. Antidesma edule Merr. Antidesma subtype. Ramos 1083 [U] Idem Antidesma venosum. Antidesma ghesaembilla Gaertn. Hort. Calcutta 1897 [U] Antidesma subtype. Idem Antidesma venosum. Antidesma membranaceum Eels 126 [U] Muell. Arc. Antidesma subtype. Idem Antidesma venosum. Antidesma obovatum J. J. Smith Versteeg 1789 [U] Hyeronima alchorneoides Allem. Krukoff 6110 [U] Hyeronima laxiflora (Tul.) Muell. Arg. Sandwith 544 [U] PI. II, 1 Thecacoris leptobotrya (Muell. Arg.) Antidesma subtype. Idem Antidesma Antidesma subtype. 28,5 fi Antidesma subtype. Antidesma subtype. venosum. E 12 n P:E 2,20. Brennon Zenker 23 [U] Cyathogyne spathulifolia Pax Schlieben [P] Cyathogyne viridis Muell Arg. F.H.I [K] Apodiscus chevalieri Hutch. Pobequin 1909 [P] Protomegabaria macrophylla Hutch. Andoh 5474 [K] PI. II, 3 36 E n 17,5 P;E 2,04. Antidesma subtype. 40 fi m:e± 0,5. Antidesma subtype. E 20,5 fi Too young for measurements. Antidesma subtype. P:E 1, n E 19 fi P:E 1,58. Antidesma subtype. 27,5 E 19 fi n P:E 1,45. Spondianthus preussii Engl. Versuchsanst. 164 [U] Antidesma subtype. 26 n E 18,5 n P;E 1,40. Maesobotrya floribunda Benth. Lebrun 1039 [U] Antidesma subtype. 21 iu E 15 /a P;E 1,40. Maesobotrya dusenii (Pax) Hutch. Antidesma subtype. Zenker 221 [U] Securinega Muell. Arg. congesta Antidesma subtype. Black 2432 [U] 45 fi E 26 n P:E 1,73. m:e < 0,5. Baccaurea subtype Pollen - grains tricolporate; subprolate prolate sphaeroidal. Colpus transversalis small. Outer ends indistinct and diffuse. Edges of the costae transversales parallel.

23 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 23 plate ii. 1. Hyeronima laxiflora; 2. Antidesma bunius; 3. Protomegabaria macrophylla; 4. Aporosa lindleyana; 5. arenaria; Phyllanthus pentaphyllus; 6. Baccaurea sumatrana; 7. Reverchonia 8. Margaritaria nobilis; 9. Phyllanthus niruri.

24 24 W. PUNT In its exine structure this subtype is related to the Antidesma subtype. The pollen grains, however, are less elongated and the colpus transversalis is smaller. Baccaurea sumatrana (Miq,.) Muell. Arg. Hort. Bog. VI C. 185 A [U] PI. II, 6 Baccaurea subtype; subprolate. 19 ft E 15 ft P:E 1,33. P.A.I. 0,35 0,3. Baccaurea javanica (Blume) Muell. Arg. Baccaurea subtype. Hort. Bog. 344 [U] Idem Baccaurea sumatrana. Baccaurea racemosa (Reinw. ex Blume) Baccaurea subtype. Muell. Arg. Hort. Bog. 553 [U] Idem Baccaurea sumatrana. Baccaurea oxycarpa Gagnep. ( Gatnaia annamica Gagnep.) Eberhardt 3042 [P] Aporosa lindleyana (Wight) Baill. Hohenacker 352 [U] PI. II, 4 subtype. Baccaurea 21 ft E 16 n P:E 1,31. subtype. Baccaurea P 19 n E 17,5 ft P;E 1,10. P.A.I. 0,2 0,25 Aporosa dioica (Roxb.) Muell. Arg. Baccaurea subtype. King 298 [U] Aporosa frutescens Blume Herbr Utrecht Idem Aporosa lindleyana. Baccaurea subtype. Idem Aporosa lindleyana. Collector unknown. Phyllanthus pentaphyllus subtype Pollen - grains tricolporate; prolate subprolate. Colpus transversalis small; outer ends almost distinct. Edges of the costae transversales rounded at the end. The difference with the Antidesma subtype and Baccaurea subtype is found in the colpus transversalis. There is also a relation with the Phyllanthus niruri subtype. The latter subtype, however, is 4 Phyllanthus pentaphyllus Wright ex Grisebach Boldingh 7378 [U] PI. II, 5 Phyllanthus sublanatus Schum. et Thonn. de Wit 1010 [WAG] colporate. Phyllanthus pentaphyllus subtype; prolate. 32 /I E 22,5 n P:E 1,43. m:e 0,5 P.A.I. 0,3. Phyllanthus pentaphyllus subtype. 45 'n E 30 n P:E 1,50. P.A.I. 0,45. Phyllanthus stipulatis (rap.) Webster Geyskes 13 [U] Phyllanthus amarus Schum. B. W. 10 [U] de Wit 474 [WAG] Reverchonia arenaria A. Gray Warnock [U] PI. II, 7 Phyllanthus pentaphyllus subtype. 24 n E 17,5 n P:E 1,37. Phyllanthus pentaphyllus subtype. P 35,5V E 27 n P:E 1,32. Phyllanthus pentaphyllus subtype. 37 E' 30 n P:E 1,60. m:e ca. 0,4. Savia erytroxyloides Griseb. Wright 1434 [BM] Phyllanthus pentaphyllus subtype. 32 ii E 26 n P:E 1,24. m:e < 0,5 P.A.I. 0,25. Phyllanthus niruri subtype Pollen grains stephanocolporate (4); prolate. Colpi transversales; costae transversales. Edges of the colpus transversalis rounded at the end. Except for the number of colpi the Phyllanthus niruri type is in all its characters identical with the Phyllanthus pentaphyllus subtype.

25 prolate sphaeroidal POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 25 As in the Phyllanthus pentaphyllus subtype the columellae are sometimes fairly coarse. The surface pattern seems to form a reticulum. This is however a pseudo-reticulum (p. 14). Only in Phyllanthus niruri a real reticulum formed by a regular arrangement of several columellae was seen. This reticulum was only to be seen with a phase-contrast microscope and then very indistinctly. Phyllanthus niruri Linn. Fuertes 89 [U] PI. II, 9 Phyllanthus niruri subtype. 39 n E 21 n P:E 1,86. P.A.I. > 0,5. Indistinct intra-reticulate. Phyllanthus guianensis Klotz. B.W [U] Phyllanthus hyssopifolioides H.B.K. Focke 1278 [U] Phyllanthus urinaria Linn. Lanjouw 246 [U] Phyllanthus nirwri subtype. P 36 /z. Phyllanthus nirwri subtype. 30 E 18 P:E n n 1,67. Phyllanthus niruri subtype. Too young. Chonocentrum cyathoforum (Muell. Arg.) Pierre ex Pax et K. Hoffmann Spruce 3781 [P] Antidesma type-, tricolporate. Colpus transversalis with costae. P:E > 1. Too to young give it a definite place in the Antidesma type. Securinega type Tricolporate or - stephanocolporate (4); subprolate sphaeroidal). (rarely oblate Colpus transversalis small, circular or broad elliptic; costae. Colpi narrow; costae colpi usually developed. Intectate: distinctly reticulate; lumina 1-2 n- Pollen grains small (smaller 30 p, usually μ). This type has small pollen grains with circular or broad elliptic colpi transversales. The reticulum is distinct and usually fairly coarse in comparison with the small pollen grains. Taxonomic discussion The Securinega type comprises, besides the genus Securinega (and Flueggia), many species of Phyllanthus and the small genera Margarilaria and Astrocasia. These genera are all closely related (Webster, 1956). The in genus Aporosella, Pax s system placed close to Aporosa, should be reduced to a section of Phyllanthus to Webster according (1956). One of the principal reasons which led him to this conclusion is the close relation to Phyllanthus acidus. In fact, the pollen grains of Phyllanthus elsiae, belonging to the section Aporosella, are closely related to those of Phyllanthus acidus (p. 26). Securinega subtype Pollen grains subprolate - sphaeroidal. Colpus transversalis circular broad or elliptic. Pollen grains of the Securinega subtype have always a P: E > 1. The shape of the colpus transversalis is circular to broad elliptic. Securinega suffruticosa (Pallas) Rehder Cantonspark (Baarn) Securinega subtype. 22,5 E /a 2\/j. P:E 1,08. P.A.I. 0,25 0,3. Lumina 1-2 fi.

26 26 w.punt oblate Securinega ramiflora (Aiton) Muell. Arg. Linsley Gressitt 229 [U] Securinega neopeltandra (Griseb.) Urban ex Pax et K. Hoffmann Shafer [U] Securinega virosa (Roxb. ex Wiled.) Baill. Reichs-Kolon. 506 [U] PI. Ill, 2 Margaritaria nobilis Linn. f. Soeprata 2D [U] PI. II, 8 Astrocasia tremula (Griseb.) Webster Gaumer et Sens 1261 [K] Securinega subtype. P 23,5 fi 20/ P:E 1,17. Securinega subtype. P:E 1,18 Securinega subtype. Securinega subtype. Lumina < 1 fi. 22,5 n E 19,5 /a m:e > 0,5 Lumina ca. 1 Securinega subtype. P.A.I. ca. 0,25. /a. P:E 1,15. P 29 E 25,5 P:E n 1,14. m:e ca. 1 P.A.I. 0,3. Lumina 1-2 fi. Phyllanthus de Wit 128 Bakhuizen reticulatus Poir. [WAG] Phyllanthus discoideus v. d. Brink 3540 [U] Leeuwenberg 2789 [U] Muell. Arg. Securinega subtype. 18 E n 16,5 P;E ft 1,09. P.A.I. 0,25. Securinega subtype. Lumina ca. 1 fi. P 22,5 E 18 P:E n n 1,25. Costae colpi. P.A.I. 0,3. Lumina 1-2 fi. Phyllanthus capillaris Schum. et Thonn. De Wit 229 [WAG] PI. Ill, 1 Phyllanthus acuminatissima C. B. Robb. Elmer [U] Richeriella gracilis (Merr.) Pax et K. Hoffmann Securinega subtype. Stephanocolporate (4) or tricolporate. P 21,5 n E 20,5 n P:E 1,05. P.A.I. 0,5. Securinega subtype. 22,5 E 18 n n P:E 1,25. P.A.I. Securinega subtype? 0,25. Lumina ca. 1 fi. Too young for reliable measurements, Hyting [K] but colpus transversalis distinctly circular. Phyllanthus acidus subtype Pollen grains tricolporate; sphaeroidal - sphaeroidal. Colpus transversalis circular. Unlike the Securinega subtype the pollen grains of the Phyllanthus acidus subtype are spheroidal or sometimes slightly oblate spheroidal. The colpus transversalis is always circular. The exine structure is in all respects the same as in the Securinega subtype. Phyllanthus acidus (Linn.) Skeels Versteeg 566 [U] Phyllanthus muellerianus (O. Kuntze) Exell Zenker 253 [U] Phyllanthus elsiae Urban Krug and Urban 4789 [L] PI. HI, 3 Phyllanthis acidus subtype. E 17,5 n P:E ca. 1. P.A.I. 0,2 0,3. Phyllanthus acidus subtype. Colpi very short. Too young. Phyllanthus acidus subtype. E 20 n. P.A.I. > 0,5. Richeria type Tricolporate; subprolate. Colpus transversalis narrow; costae. Colpi narrow. Tectate (?); intra-reticulate. Lumina ca. 1 fi. The Richeria type is an intermediate type between the Antidesma type and the Securinega type. The narrow elongated colpus transversalis

27 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 27 plate in. 1. Phyllanthus capillaris; glauca; 2. Securinega virosa; 3. Phyllanthus elsiae; 4. Drypetes 5. Actephila excelsa; 6. Andrachne phyllanthoides.

28 w. PUNT is typical of the Antidesma type, but the distinct, rather coarse reticulum is characteristic of the Securinega type. Taxonomic discussion Pax and K. Hoffmann (1931) as well as Bentham (1880) placed this genus in the Antidesminae. It occurs in the New World. Richeria laurifolia Baill. ex Muell. Arg. Richeria type; prolate. 1,21. Krukoff 8776 [U] P 29 fi E 24 P:E Lumina ca. 1 fx. Dicoelia type Tricolporate; prolate Colpus transversalis Colpi narrow; subprolate. broad and elongated to broadly elliptic; costae. costae colpi. Tectate, psilate; sometimes intra-reticulate. Pollen grains of medium size (ca. The Dicoelia type possesses as of the Securinega 25 fi /). elements of the Antidesma type type. The pollen grains have always a as well Polar axis of at least 25 /u; usually between 30 fi and 55 pt. The colpus transversalis is large; broad and elongated or broadly elliptic, but never circular. Taxonomic discussion Pax and K. Hoffmann (1922, 1931) included the genus Dicoelia in the Antidesminae for its elongated inflorescences. However, the loose, often androgynous clusters along the rachis of the racemes is different from any inflorescence in the Phyllanthoideae 1880). (Bentham Dicoelia subtype Colpus transversalis broad and elongated; m:e < 0,4. Usually not intra-reticulate. Dicoelia affinis J. J. Smith Hallier 1255 [U] Dicoelia subtype. 44 E 32 // P:E 1,37. m:e < 0,4 P.A.I. 0,3. Drypetes glauca Vahl Sintenis 5111 [U] PI. Ill, 4 Indistinct reticulate. Dicoelia subtype. 34 E n 27 P:E n 1,26. m:e < 0,25 P.A.I. 0,25. Drypetes macrophylla (Blume) Pax et K. Hoffmann Bogor VIII F. 48 [U] Dicoelia subtype. P 45 E 40 P:E n n 1,16. m:e < 0,4 P.A.I. 0,35. Intra-reticulate. Drypetes laterifolia (Swartz) Krug et Dicoelia subtype. Urban Fuertes 824 [U] Drypetes variabilis Uitt. B.W [U] Dicoelia subtype. Andrachne subtype Colpus transversalis broadly elliptic; m:e ca. 0,5 or larger. Usually intra-reticulate.

29 POLLEN MORPHOLOGY OF THE EUPHORBLACEAE 29 Andrachne phyllanthoides (Nutt.) Mueix. Arc. Bush [U] PI. Ill, 6 telephioides Linn. Andrachne Dalmatic 1937, 50 [U] Andrachne colchica Fisch. et Meyer Andrachne subtype. 45 E /i 35,5 P:E n 1,27. m:e costae colpi. Andrachne 0,5 P.A.I. 0,2. Indistinct subtype. Intra-reticulate. Andrachne subtype. Cult. Hort. Canton (Baarn) 4884 [U] Andrachne aspera Spreng. Pappi 4892 [U] Actephila excelsa (Dalz.) Muell. Arg. Hort. Bog. VII B. 271 [L] PI. Ill, 5 Poranthera corymbosa Brong. Constable [U] Andrachne subtype. 43 ft E 31 n P:E 1,40. m:e > 0,5. P.A.I. 0,3. Longitudinal striation. Andrachne subtype. 48 n E 37 n P:E 1,36. P.A.I. 0,3. Andrachne 34 n subtype. Costae colpi. E 29,5 /j. m:e ca. 0,5 P.A.I. 0,25. P:E 1,15. Tectum perforatum; lumina of the intra-reticulum 2-3 /i. Poranthera microphylla Brong. Hastings [U] subtype. Andrachne 24 ft E 19,5 n P:E 1,23. Tectate; intra-reticulate; lumina ca. 1 p. Zimmermannia type Tricolporate; prolate spheroidal. Colpus transversalis small; costae. Colpi narrow. Tectate; verrucate. The Zimmermannia type is the only type in the Phyllanthoideae with verrucae. Except for this remarkable character the pollen grains belong without doubt in the Antidesma configuration. Zimmermannia capillepes Pax Zimmermannia type. FI. Usambarica 5862 [BM] PI. IV, 1 P 58 E 51 P:E /e /e 1,12. m:e > 0,5 P.A.I. 0,25. Leptonema type Tricolporate; subprolate. Colpus transversalis broad and large; costae narrow. Colpi narrow; costae colpi absent. Tectate; psilate. The columellae are so small, that they are hardly visible. The Leptonema type has some relation with the Dicoelia type (p. 28). The type misses, however, the costae colpi. It differs from the Antidesma type (p. 21) by the broad and elongated colpus transversalis. Taxonomic discussion Pax and K. Hoffmann (1922, 1931) placed in the subtribe Glochidiinae Leptonema together with Glochidion and Breynia. It seems better to keep Leptonema in the vicinity of Phyllanthus and its allied genera, as stated by Bentham (1880). Leptonema venosa (Poir.) A. de Juss. D Alleizette 6405 Mad. Nov [L] PI. IV, 2 Leptonema type. 27 E ju 22,5 P:E yu 1,21. m:e <0,5 P.A.I. 0,5.

30 30 W. PUNT Pleiostemon type Tricolporate; prolate spheroidal. Colpus transversalis; costae; m:e larger 1; 6-8 m fi. Colpi narrow; costae colpi. P.A.I. small. Pollen grains sometimes syncolpate. Tectate; psilate. The Pleiostemon type belongs undoubtedly to the Antidesma configuration. The colpus transversalis however differs from all other types by its meridional elongation. The m : e is larger than 1. Moreover, the colpi are very long. Sometimes the pollen grains are even syncolpate (colpi anastomosing at the poles). Taxonomic discussion The genera belonging to this type are without doubt related to Phyllanthus and its allied genera. Mueller (1866) reduced the genus Pleiostemon to a section of Phyllanthus and Bentham (1880) referred it to Securinega. In the system of Pax and K. Hoffmann (1931) Pleiostemon is maintained as a genus close to Phyllanthus. On the other hand, Lingelsheimia is found in the subtribe Drypetinae in their system. Corresponding characters are: 1. In the male flowers an extrastaminal disc present and an ovarium rudiment wanting. 2. In the female flowers an annular disc and bifid styles present. 3. Calyx is 6-partite (in Pleiostemon the male flower is sometimes 4-5-partite. Petals are absent. 4. Plants monoecious and only found in Africa and Madagascar (.Dangyodrypetes ). Pleiostemon verrucosus Sonder Pleiostemon type. Cooper 312 [BM] PI. IV, 3 20 n E 19,5 fi P;E 1,03 m:e > 1 P.A.I. < 0,2. Lingelsheimia frutescens Pax Pleiostemon type. Lebrun 3855 [K] PI. IV, 4 Dangyodrypetes ambigua L andri Serv. Eaux et Forests 7675 [P] 22,5 fi P.A.I. 0,1. Pleiostemon type. E 21 n P:E 1,07. P24 E n 22,5 n P:E 1,07. P.A.I. almost syncolpate-syncolpate. Heywoodia type Tricolporate; prolate spheroidal. Colpus transversalis broad; costae absent; m:e ca 1. Colpi narrow; costae colpi absent. Tectate; psilate. Columellae distinct. As the Polar axis is larger than the Equatorial axis the type is placed in the Antidesma configuration. There is, however, a possibility that Heywoodia is wrongly placed in this configuration because of the absence of costae. Heywoodia lucens Sim Bally 5133 [K] PI. IV, 5 Heywoodia type. 27,5 n E 25,5 n P:E 1,08. m:e ca. 1 P.A.I. 0,3.

31 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 31 iv. Zi mmermanni a capillepes; 2. Leptonema venosa; plate Lingelsheimia fruticosa; 5. Heywoodia lucens; 6. Amanoa guianensis; 7. Pseudolachnostylis glauca. Pleiostemon verrucosus;

32 32 w. PUNT Amanoa configuration Tricolporate-stephanocolporate (4-5). Pollen grains usually with a P. shorter than the E. Colpus transversalis usually with costae. Pollen grains reticulate. Rarely not reticulate (some species of Amanoa and the Discocarpus type). The Amanoa configuration includes pollen types with a Polar axis shorter than the Equatorial axis and lesser than six (usually 3 4) colpi. The pollen grains are reticulate except in a few species. These latter species, however, have so many characters in common with related reticulate species, that there is no doubt that they belong to the same type-. As in the Antidesma configuration the types are difficult to separate. Especially the Amanoa type and Savia type are closely related. Taxonomic discussion The genera belonging to this configuration are all closely related. The division of the pollen grains into four types is not found again in the systems of Mueller, Bentham and Pax and K. Hoffmann. Mueller (1866) removed Briedelia and Cleistanthus to a distinct subtribe on account of their valvate calyx (nearly all the other genera in Phyllanthoideae have an imbricate calyx). For this reason Pax and K. Hoffmann (1931) also placed these genera in a special tribe. Bentham (1880) took less account of the imbricate calyx and found, in the remaining characters, much affinity with Amanoa. So he placed Briedelia and Cleistanthus in close proximity of Amanoa. The pollen grains of Cleistanthus cannot be separated from those of Savia. Except for the longitudinal striation the pollen grains of the Briedelia type are also very close related to those of the Savia type. From these results it is clear that pollen morphology supports the statement of Bentham. Bentham (1880) discussed Uapaca in the group of the Antidesminae. Pax and K. Hoffmann (1931) established a special subtribe for this genus. Although the pollen grains belong to the Savia type it seems better to keep this genus out of the subtribes with a Savia type. The involucrum and the absence of a disc in the male flower are characters too divergent for placing Uapaca in the vicinity of any other genus with a Savia type. In the system of Pax and K. Hoffmann (1931) the remaining genera of the Amanoa configuration are divided into four subtribes. 1. Wielandiinae. 2. Amanoinae. 3. Discocarpinae. 4. Pseudolachnostylidinae. In these subtribes the following genera do not have the typical characters of the Amanoa configuration. 1. Astrocasia. According to Webster (1956) the genus is related to Securinega. The pollen grains undoubtedly belong to the Securinega type. 2. Actephila. The is genus classified together with Amanoa by all authors. Actephila excelsa, however, possesses pollen grains of the Dicoelia type. Examination of the pollen grains of other species is necessary.

33 POLLEN MORPHOLOGY OF THE BUPHORBIACEAE Chonocentrum. The pollen grains of Chonocentrum belongs without doubt to the Antidesma type (p. 21). 4. Savia erytroxyloides. Pollen grains of this Savia species belong to the Phyllanthus pentaphyllus subtype. The remaining genera have the following characters in common: 1. The calyx is 5-partite. 2. Frequently petals are present. 3. Male as well as female flowers have a disc. 4. In the male flowers an ovarium rudiment is present. 5. Stamens are often connated into a column. 6. Seeds never have a caruncula. Amanoa type Tricolporate-stephanocolporate (4-5); suboblate-oblate spheroidal. Colpus transversalis broad; ora large; costae usually present. Colpi wide; no costae colpi. Tectate or intectate; reticulate or not reticulate. If the pollen grains are reticulate the lumina are at least 3 fi in diameter. When the pollen grains are not reticulate the structure elements are tall and at least 4 fi. Exine very thick, at least 4 /i. The Amanoa type is characterised by a coarse reticulum and a thick exine. The reticulum of the Amanoa type had to be considered as an intrareticulum since the muri are connected by a thin membrane, which covers the lumina. Some species of Amanoa lack the coarse reticulum. Amanoa guianensis has a tectum provided with long baculae or echinae. The columellae of Amanoa grandiflora in are, optical section, gemma shaped and inordinate In arranged. both Amanoa species the exine is, however, more than 4 fi thick. Amanoa oblongifolia Muell. Arg. Krukoff 7015 [U] PI. V, 1 Amanoa grandiflora Muell. Arg. Murca Pires 835 [U] Amanoa bracteosa Planch. D Alleizette 6392 [L] Amanoa boiviniana Baill. D Alleizette Mad. Nov [L] Amanoa type. 33,5 n. E 43 n P:E 0,78. Colpi wide. P.A.I. 0,25. Lumina up to 11,5 fi. Amanoa type. 39 n P.A.I. 0,5. Not reticulate. Columellae gemma shaped, inordinate arranged. Amanoa type. E 45/i P:E 0,87. E 80 /i. Lumina up to 10 fi. P.A.I. 0,4. Amanoa type. 37,5 E 45 P:E fi n 0,83. P.A.I. 0,4 0,5. Lumina up to 4 fi. Amanoa guianensis Aubl. Bafog S.F [U] Wood Herb. 28 [U] PI. IV, 6 Amanoa type. 46 fi E 52 /i P:E 0,88. P.A.I. 0,3. Costae transversalis. Co- lumellae very short. Baculate-echinate; baculae up to 7 fi. Endexine thick.

34 34 w. PUNT.. _. 35 Pentabrachion reticulatum Jantzon Zenker 2982 [L] (Muell. Arg.) Amanoa type; 3-4 colporate. P 45 j it E 48/i P;E 0,94. Lumina up to 5 ft. P.A.I. 0,4 0,5. Costae transversales. Pseudolachnostylis glauca (Hiern) Hutch. Lanjouw 1181 [U] PI. IV, 7 Amanoa type; 2-3 colporate. 31 n E 34 /i P:E 0,91. Lumina up to 4,5 ft. Savia type Tricolporate-stephanocolporate (4); subolate-prolate spheroidal. Colpus transversalis usually narrow; small. ora Rarely colpus transversalis broad and ora large (Cleistanthus ferrugineus). Colpi narrow; sometimes costae colpi present. Tectate or intectate; reticulate. Lumina of the reticulum 3 Exine is medium thick, lesser then 4 ft but still distinct. ft or smaller. In many respects the Savia type resembles the Amanoa type. The lumina of the reticulum are, however, distinctly smaller. Since the pollen grains are also relatively smaller this difference is perhaps not offundamental importance, but, besides the smaller lumina, the colpi are narrow and the ora small, so it seems better to separate the pollen grains of the Savia type from those of the Amanoa type. Some species with a Savia type (Cluytiandra madagascariensis, Lachnostylis hirta) have a Polar axis larger than the Equatorial axis. By this character the pollen grains should be in the Antidesma placed configuration. The exine structure, however, is quite different from the Antidesma configuration and is, on theother hand, in full agreement with the Savia type. Savia andringitrana Leandri Decary 1926 [L] Savia type. 26 ft E 29 n P:E 0,88. m;e < 0,5. P.A.I. 0,4. Colpi narrow and short. Lumina 3 ft. Costae transversales. Savia sessiliflora (Sw.) Wield. Eggers 1905 [P] Savia type. P 24,5 E 29 fi!x P.A.I. 0,3. P:E 0,85. Blotia oblongifolia (Baill.) D Alleizette [L] Mad. 4 Mai Leandri Savia type. 25,5 n E 28.5 n P:E 0,89. P.A.I. 0,45. Wielandia elegans Baill. D Alleizette Seych. [L] Colpi short and narrow. Lumina 1-2 ft. Savia type. 3 4 colporate. 37 /i E 41,5 n P:E 0,88. P.A.I. 0,65. Lumina 1-2 /a. Lachnostylis hirta Muell. (Linn, f.) Arg. [ Discocarpus hirtus (Linn, f.) Pax et K. Hoffmann] Bolus 2396 [K] Cluytiandra madagascariensis Leandri D Alleizette Mad [L] Colpi narrow and short. Savia type; prolate spheroidal. 37 n- E fi P:E 1,05. P.A.I. 0,3. Colpus transversalis with costae. Lumina 1-2 fi. Savia type; prolate spheroidal. 32 fi E 31 P.A.I. 0,3. n P:E 1,04. Costae transversales. Reticulum indistinct. Lumina 1-2 fi. Cleistanthus winkleri Jabl. Bogor IX C. 165 A. [U] Cleistanthus dichotomus J. J. Smith Gjellerup 142 [U] Savia E type. 32 ft. Costae colpi. Savia type.

35 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 35 oblongifolia; 2. Cleistanthus ferrugineus; Uapaca heudelotii; 1. Amanoa plate v. 3. Discocarpus essequeboensis; monoica; 6. Breynia fruticosa; 7. Breyniopsis Briedelia pierrei; 8. Glochidion sericeum.

36 36 w. PUNT Cleistanthus ferrugineus Muell. Arg. Hort. Bogor. IX C. 43 A. [U] PI. V, 2 Savia type; prolate. P 44 E 33 P:E 1,34. P.A.I. 0,4 0,5. Colpus transversalis isodiametric; no Uapaca heudelotii Baill. Zenker 59 [U] PI. V, 3 costae; ora large. Lumina 1-2 fi. type; oblate spheroidal. Savia 22 fi E 25 n P:E 0,88. m:e < 0,5. P.A.I. 0,5. Tectate. Uapaca kirkiana Stolz 622 [U] Muell. Arg. Savia type. 27 E 29,5 n P:E 0,93. Tecta te. Briedelia type Tricolporate; oblate spheroidal. Colpus transversalis broad; costae. Colpi narrow at the ends, broadened in the equatorial Sometimes part. costae colpi present. Intectate; reticulate. Lumina small and in chains along the colpi (pseudostriate). Muri distinct but not tall. This type is closely related to the Savia type. The greatest difference is found in the longitudinal arrangement of the lumina of the reticulum. The chains oflumina give the pollen grains a striate appearance, which is, however, only a a pseudo-striation : striation not caused by individual columellae but the by lumina of the reticulum which of are, course, composed of several columellae. Briedelia monoica (Lour.) Merr. Endert E 1066 PI. [U] V, 5 Briedelia type. 24 E ju 25,5 fi P:E 0,94. m:e > 0,5. P.A.I. 0,35. Lumina < 1 ft. Briedelia glauca Blume Winckel 865 B [U] Briedelia assamica J. D. Hook. Herb. East Ind. Comp Briedelia stipularis (Linn.) Blume v. Leeuwen-Reynvaan [U] [U] Briedelia type. E 31 /a. Briedelia type. E 31 /t. Lumina 1-2 /i. Briedelia type. E 42 //. m:e < 0,5. P.A.I. 0,25. Godefroya rotundata Gourgaud [P] (Jabl.) Gagnepain Briedelia type. 32 n E 33,5 n P:E 0,96. P.A.I. 0,25. Costae colpi. Discocarpus type Stephanocolporate (4-5); oblate spheroidal. Colpus transversalis small; no costae present. Colpi narrow. No costae colpi present. Tectate, psilate. Not reticulate. The Discocarpus is type an intermediate between the of type types the Amanoa configuration and the types of the Antidesma configuration. The small columellae and the absence of a reticulum are characters in favor of placing the type in the Antidesma configuration. On the other hand the oblate shape, the absence of costae and the 4 5 colpi are more in agreement with the Amanoa configuration. For these latter reasons it seems better to discuss the type in the Amanoa configuration. Discocarpus essequeboensis TC THT7SPH Schomburgk 659 [U] PI. V, 4 Discocarpus type. P 34,5 ft E 37/t P:E 0,94. Ora small; m:e > 0,5. P.A.I. 0,6.

37 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 37 Phyllanthus nutans configuration Pollen grains with an areolate structure. Taxonomic discussion After a proposal of Iversen, Erdtman recommended the term areolatus for the structure described under Phyllanthus nutans type (1947). The definition given by him is: Pollen grains with small areas separated by small grooves forming a negative reticulum. Webster (1956) states that the areolate pollen grains are mostly found in the New World. Some species of Phyllanthus in the Old World, however, have areolate pollen grains that are superficially very similar to those of the New World. Phyllanthus nutans type Periporate; the surface of the pollen grains is divided into or hexagons. The pori are situated in the angular points. pentagons Pori small. Intectate. From each aperture two rows of pila lead to the other apertures. In the centre of the penta- and hexagons the pila form an irregular reticulum. Phyllantus nutans Sw. Hooker, Jamaica [U] PI. VI, 2 Phyllanthus Hulk 21 adianthoides Klotzsch [U] Phyllanthus speciosus Jacq,. Fresh material, Greenhouse Utrecht Phyllanthus nutans type. Longer axis 27,5 /r. Shorter axis 25 /. ca. 25 pori. Phyllanthus nutans type. Diam. 17 /s. Phyllanthus nutans type. Diam. 20 Pori ca. 15. /i. Dendrophyllanthus type Tricolporate (?); oblate. Colpus transversalis circular; costae. Colpi very narrow; operculum consisting of one row of pila. Intectate; not reticulate. Without doubt the pollen grains are related to the Phyllanthus nutans type in their exine structure. The surface of the pollen grains is divided into three regions, separated by strings of pila, the colpi. Each colpus consists of three rows of pila. If the two outer rows represent the borders of the colpus, the inner row mustbe taken as an operculum. However, the space between the rows is so small, that it can hardly be regarded as a weak performed part in the ektexine (see definition aperture: Glossary). For this reason it is possible to defend the opinion, that the pollen grains of Dendrophyllanthus are triporate. In that case the rows of pila would only be an ornamental peculiarity. Nevertheless, the rows look so much like real that in colpi, the author s opinion it is better to maintain the pollen grains in the of the group tricolporate. Dendrophyllanthus ripicolus Guill. Mackee 3629 [L] PI. VI, 3 Dendrophyllanthus type. 19,5 E27 P:E n n 0,72. m:e ca. 1.

38 38 w. PUNT Breynia configuration Stephanocolporate (colpi sometimes with two ora). Colpus transversalis circular rarely somewhat elliptical; costae. Colpi narrow and long. Tectate or intecta te. By its exine structure the Breynia configuration is related to the Antidesma configuration. Most pollen grains, however, have a Polar axis shorter than the Equatorial axis as in the Amanoa configuration. The typical character of the number of colpi with often two circular colpi transversalis per colpus seem to justify this configuration. Breynia type Stephanocolporate (colpi diorate); oblate- subprolate. Usually two circular colpi transversales per colpus, some colpi have only one colpus transversalis; costae. Colpi sometimes costae narrow; colpi. Intectate or tectate; reticulate or not reticulate; lumina small. The Breynia type is easily recognised by the large number of colpi (six or more) and the double circular colpi transversales. In the species of Breynia one of the two colpi transversales is often reduced. The remaining endexine aperture is never situated in the equatorial plane but towards the ends of the colpus. Taxonomic discussion Hitherto the genera belonging to this type have never been placed in one group by any author. Mueller (1866) recognised the relationship of Agyneia with which Sauropus, opinion was shared by Bentham (1880) and Pax and K. Hoffmann (1922, 1931). Breynia and Glochidion, on the other hand, are always distinguished as genera closely related to Phyllanthus. Glochidion is even often reduced to a section of Phyllanthus. Pax and K. Hoffmann keep Glochidion, Breynia and Breyniopsis in the subtribe Glochidiinae. In the author s opinion the above genera should all be placed in one group (either subtribe or other taxon). They have the following corresponding characters: 1. The number of sepals is usually six (i Glochidion sometimes fewer). 2. In the male flowers generally three connated stamens form a column. Connectives are frequently appendiculated. Anthers extrors. There is no ovarium rudiment nor a disc. As Bentham states, the scale-like thickenings at the base of the sepals of Agyneia and Sauropus are not to be regarded as disc glands. 3. In the female flowers disc glands are wanting. 4. The plants are generally monoecious (Some Glochidion species and Arachnodes, however, dioecious) and only occur in the Old World. Breynia subtype Pollen grains oblate spheroidal- oblate. Intectate; reticulate; lumina 1-2 //.

39 prolate POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 39 Breynia fruticosa (Linn.) Benth. J. and M. S. Clemens 3182 [U] Beekler 2 [U] PI. V, 6 Breynia subtype; oblate. 21 E fi 27 P:E /x 0,76. Colpi transversales situated almost at the end of the colpi. Colpi not always provided with two colpi transversales. Breynia racemosa (Blume) Muell. Arg. Bakhuizen v. d. Brink [U] Breynia nivosa (W. G. Smith) Small Collector unknown. Colpi Lumina 1-2 p. Breynia subtype; oblate. 5-6 colpi. Breynia subtype. 7-8 colpi. Agyneia bacciformes (Linn.) A. de Jussieu v. Leeuwen-Reynvaan 1978 [U] Breynia subtype; oblate. 27,5 n E 32,5 /j. P:E 0,85. Sauropus androgynis (Linn.) Merrill Doctors v. Leeuwen, Depok [U] Always two colpi transversales present. Golpi 5-6. Costae colpi. Breynia subtype; oblate spheroidal. 25,5 i u E 28 P:E 0, colpi. Breyniopsis subtype Pollen grains subprolate. Tectate; not reticulate. The pollen grains of Breyniopsis are related to the pollen grains of Breynia by the number of colpi and the double circular colpi transversales. The shape and structure are, however, quite different. Breyniopsis pierrei Beille Poilane [P] PI. V, 7 Breyniopsis subtype; subprolate. 32,5 n E 28,5 n P:E 1,14. Always two colpi transversales colpus. 6-7 colpi. Costae colpi. per Glochidion type Stephanocolporate (4); spheroidal - spheroidal. Colpus transversalis circular or slightly elliptical; costae. Colpi narrow. Intectate or tectate; reticulate; lumina 1-2 p. In many morphological respects the Glochidion type is closely related to the as Breynia type, Erdtman (1952) and Webster (1956) have stated. The pollen grains lack, however, the typical double colpus transversalis. It is also possible to place the in the Antidesma type configuration. Its single circular, sometimes broad elliptic endexine and aperture the distinct reticulum are characters in favour of a place in the Securinega type (p. 25). Glochidion sericeum (Blume) J. U. Hooker Bakhuizen v. d. Brink 1886 [U] PI. V, 8 Glochidion concolor Muell. Arg. Glochidion type; spheroidal. E 24 n. Colpi narrow. P.A.I. 0,3-0,4. Lumina 1-2 /< Intectate. Glochidion type. Yuncker [U] Glochidion littorale Blume Glochidion type. Bakhuizen v. d. Brink 3182 [U] Glochidion obscurum (Willd.) Blume Glochidion type. Buysman 295 [U] Glochidion ramiflorum Yuncker [U] R. et G. Forster Glochidion type. Glochidion superbum Baillon Dumas 1507 [U] Glochidion type.

40 40 W. PUNT Glochidion spec. Teysman 4411 [U] Arachnodes chevalieri Gagnep. Fleury [P] PI. VI, 1 Glochidion type. Glochidion 25,5 n type; prolate spheroidal. P:E 1,06. P.A.I. > 0,5. Tectate; intra-reticuculate. Colpus transversalis slightly elliptical. Aristogeitonia configuration Stephanocolpate, stephanocolporate, stephanoporate, periporate or inaperturate. The pollen grains are spheroidal or have a Polar axis shorter than the Equatorial axis. Tectate; usually echinate, sometimes psilate. The types of the Aristogeitonia configuration form a closely related group which is sharply differentiatedfrom the other configurations in the Phyllanthoideae. They are easily recognised by the number of apertures and structure of the exine. Most types have echinate pollen grains and ca. six small apertures. The shape is spheroidal or oblate to spheroidal oblate. Taxonomic discussion The Aristogeitonia configuration includes a typical group of genera which are, in several respects, different from other Phyllanthoideae. Mueller (1866) was acquainted with too few genera of this group to recognise it. Bentham (1880), however, enumerated, in the fourth group of his Phyllantheae, all the genera of the Aristogeitonia configuration then known. Only the genus Bischofia of Bentham s group does not belong to the configuration. All the genera belonging to the configuration are found in one of the three into which Pax groups divides the Phyllanthoideae (Pax 1924). This group includes his subtribes Drypetinae, Petalostigmatinae, Toxicodendrinae ( Hyaenanchinae), Dissiliarinae, Paivaeusinae and the tribus Caletieae of the Stenolobeae. Only Drypetes, Lingelsheimia and Heywoodia of the Drypetinae differ pollen-morphologically. The genera are small. The largest genus Longetia does not come up to more than 10 species. The male flower lacks an extra staminal disc. The flowers always are while the apetal, number of stamens is seldom constant; there are mostly more than six. In the female flower the styles are undivided, at best emarginate, but never incised. The leaves of this group are often opposite in or whorls. Some genera have, moreover, composite leaves. Aristogeitonia type Stephanocolpate, stephanocolporate or stephanoporate; oblate spherioidal - suboblate. Colpi, if present, small and very short; costae colpi. Pori, if present, with costae. Colpus transversalis, if present, with costae. Tectate; echinate or psilate. In the Aristogeitonia type the apertures are always situated in the

41 35,5 pollen morphologyof the euphorbiaceae 41 equatorial plane. The apertures can have the shape of pori, colpi or composite apertures, but they are always small. Aristogeitonia subtype Pollen grains echinate. Aristogeitonia limoniifolia Prain Herb. Kew. PI. VI, 5 Oldfieldia africana Benth. Small 621 [K] Paivaeusa dactylophylla Welw. et Oliver Shabana 9 [K] Piranhea trifoliata Baill. Glaziou [K] PI. VI, 4 Aristogeitonia subtype; stephanocolpate. P 30,5 [i Colpi ca. 4,5 /i long. Echinae E 35 P:E 0,86. ca. 2,5 jx. Aristogeitonia subtype; stephanocolpate. P - fi E 43 n P:E 0,83. Aristogeitonia subtype; stephanocolpate. 43 ft E 46,5 n P:E 0,92. Aristogeitonia subtype; stephanoporate. 27 fi E 32 n P:E 0,85. Diam. pori 3 fx. Tetracoccus ilicifolius Coville et Gilmay Gilmay Mai. 2A [L] Aristogeitonia subtype; stephanocolporate. 35,5 fi E 40 P:E 0,91. Colpi 6-7. Echinae 3,5 /x. Hyaenanche globosa (Gaertn.) Vahl Splitgerber [L] PI. VI, 6 Mischodon zeylanicus Thwait. Lamb, et Hort Bogor. IX A [U] Paragelonium perrieri Leandri Perrier de la Bathie 1178 [P] Aristogeitonia subtype; stephanocolporate. 39 E 46 P:E ti n 0,85. Colpi 6-7. Echinae 2 ft. Aristogeitonia subtype; stephanocolpate. P 32 E n 36 P:E 0,89. n Colpi 5-6. Echinae 3,5 fi. Aristogeitonia subtype; stephanoporate. 32 n E 35 P:E v 0,92. Diam. pori 3 fi. Echinae 2,5-3,5 /x. Petalostigma quadriloculare F. v. Mueller Queensland [U] Aristogeitonia subtype; stephanoporate. 28,5 n E 32 «P:E 0,89. Pori 4-6; not always exactly in the equatorial plane; diam. 3 fi. Echinae Longetia malayana (Benth.) Pax et K. Hoffmann D Alleizette, Penang 6421 [L] ca. 1 fi. Aristogeitonia subtype; stephanoporate. 47 E p 52 P:E ft 0,90. Pori slightly ellpitical diam. ca. 5 fi. Echinae up to 6,5 fi. Longetia carunculata (Baill.) Pax et Aristogeitonia subtype; stephanoporate. K. Hoffmann Pori 5 (-6). Mackee 4230 [L] Longetia buxoides subtype Pollen grains psilate. Longetia buxoides Baillon D Alleizette, Nw. Caled [L] PI. VI, 7 Dissiliaria tricornis Benth. Brass 5755 [L] PI. VI, 9 Longetia buxoides subtype;stephanocolpate. V 22 n E27 n P:E 0,81. Colpi (5-) 6 (-7); small and short with thick costae. longetia buxoides subtype; stephanoporate. 19 E /x. 22,5 P;E [x 0,85. Pori 6; diam. 2 ft. Stachystemon type Periporate or inaperturate; spheroidal or ellipsoid (with a longer and a shorter axis). Pori small; costae pori. Tectate; echinate. In the Stachystemon type the apertures are never situated in the equatorial plane, but scattered over the surface.

42 42 W. PUNT plate vt. 1. Arachnodes chevalieri; 2. Phyllanthus nutans; 3. Dendrophyllanthus ripicolus; 4. Piranhea trifoliata; buxoides; 5. Aristogeitonia limoniifolia; 6. Hyaenanche globosa; 7. Longetia 8. Phyllanthus acuminatus; 9. Dissiliaria tricornis.

43 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 43 vermicularis Planch. Stachystemon Pritzel 270 [L] PI. VII, 2 Micrantheum ericoides. Desf. Buysman 1873; 347 [U] Stachystemon type. Diam. 36 pc. Pori ca. 12; diam. pori 2,5 Echinae pc. 1,5 ju long. Stachystemon type. Longer axis 45 fi. Pori ca. 25; diam. 2 Echinae fi. usually in groups of three around the pori; 3 n Micrantheum hexandrum J. Tasmania [U] D. Hook long. Stachystemon type. Diam. 45 pc. Pori small; more than 40. Pseudanthus pimeleoides Sieb. ex Spreng. Buysman 1890 [U] PI. VII, 7 Pseudanthus nematophorus F. v. Muell. Murchinson [U] Neoroepera banksii Benth. Banks and Solander 1770 [K] Longetia gynotricha Guile. Mackee 4191 [L] Androstachys johnsonii Prain Pole-Evans, [K] PI. VII, 2 Stachystemon Diam. 20 type. pc. Pori ca. 7; diam. 2 pc. Echinae 1 pc long. Stachystemon type. Diam 20 pc. Stachystemon type. Diam. 40 pc. Diam. pori 1,5 pc; ca. 30 pori. Echinae 2 pc long. Stachystemon type. Diam. 34 pc. Pori ca. 16. Echinae 2 pc. Stachystemon type. Longer axis 46 fi. Shorter axis 40 pc. Echinae ca. 1 pc. Pollen types not placed in one of the above configurations Some types cannot readily be placed in one of the discussed configurations. Since most types have only one representative they are discussed separately and not in put any configuration. Phyllanthus acuminatus type Tricolporate (colpi diorate); spheroidal. Colpi transversales (two in each colpus) circular; costae. Colpi short. Intectate; pilate. Pila ordinate arranged but not in a reticulum. The presence of two ora in the colpi is also an important character of the The Breynia type (p. 38). structure of the exine, however, is quite different. Its remaining characters cannot be compared with any other configuration. Phyllanthus acuminatus Vahl Phyllanthus acuminatus type. Hostman 413 [U] PI. VI, 8 P E 27 n. P.A.I. 0,5. Bischofia type Tricolporate; spheroidal, P:E is 1. Colpus transversalis small; costae. Edges of the costae transversales rounded at the ends. Colpi narrow and long; no costae colpi. Tectate; psilate. Columellae short. This type should be best placed in the Antidesma configuration but differs from other in types this configuration by its shape (P: E 1) and the absence of costae The columellae colpi. are moreover very short.

44 44 W. PUNT - Taxonomic discussion Bentham (1880) discussed Bischofia together with Oldfieldia and Piranhea because of their digitately compound leaves. Pax and K. Hoffmann (1931) thought it better to put the genus apart in a subtribe. As the pollen grains are certainly not similar to those of Oldfieldia and Piranhea and moreover difficult to place in any other configuration, it seems right to maintain the separate subtribe. Bischofia javanica Blume Hohenacker 1573 [U] PI. VII, 5 Bischofia type. E 25 ju. P:E ca. 1, m:e < 0,5. P.A.I. 0,2. Putranjiva type Tricolporate; spheroidal. P:E is 1. Colpus transversalis large; costae. Colpi costae narrow; colpi. Tectate;psilate. Exine thick. Columellae short but endexine and tectum thick. The Putranjiva (p. 28). There is type has some resemblances with the Dicoelia... a large colpus transversalis with costae, type and also costae colpi. The type differs, however, by its shape (P: E 1) and the remarkably thick endexine. In a specimen of Putranjiva (Stocks, India), examined by Erdtman (1952), the pollen grains are subprolate, reticulate. The specimens Reorders 2156 and Broadway 9249 differ from that considerably description. The typical character of the thick exine is not mentioned by Erdtman and was probably not in present his pollen grains. Taxonomic discussion In the system of Pax and K. Hoffmann (1931) Putranjiva is found in the subtribe Glochidiinae. The genus is put in relatiosnhip with Breynia and Glochidion. Bentham (1880) gives it a place next to Drypetes. It must be admitted, that the pollen grains of Putranjiva are far more similar to those of Drypetes than to the Breynia type. Putranjiva roxburghii Wall. Koorders 2156/3 [U] Braodway 9249 [U] PI. VII, 3 Putranjiva type. E 40 fi. P:E ca. 1. m:e<0,5. P.A.I. 0,3. Exine thick (4 fi). Hymenocardia type Triporate (-diporate); oblate. Costae pori. Tectate; psilate. Columellae short. The type differs completely from all the other pollen grains in the Phyllanthoideae. Taxonomic discussion On account of its catkin-like male inflorescenses Pax and K. Hoffmann (1931) in placed Hymenocardia the Antidesminae. The most striking character of the genus is the compressed two-cocced fruit with lateral wings. Hymenocardia differs from the other genera in the Antidesminae in some characters:

45 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 45 plate vh. I. Androstachysjohnsonii; 2. Stachystemon vermicularis; 4. Martretia quadricornis; 5. Bischofia javanica; 6. pimeleoides. Hymenocardia ulmoides; 3. Putranjiva roxburghii; 7. Pseudanthus

46 46 w. PUNT 1. Disc wanting in male as well as in female flowers. 2. The styles are long and undivided. Hymenocardia ulmoides Oliv. Koechlin 2371 [U] PI. VII, 6 Hymenocardia acida Tul. Bequaert 7142 (BR) Duvigneaud 2422 (BRLU) Hymenocardia type. 19 E fi 22,5 n P:E 0,85. Hymenocardia type. 28,5 /j E 31 fi P:E 0,92. Diam. pori 3 fi. Martretia type Tricolporate; oblate spheroidal. Colpus transversalis; costae. Edges of the costae transversales rounded at the ends. Atrium present. Colpi narrow; no costae colpi present. Tectate; psilate. The pollen grains of the Martretia type are oblate spheroidal but do not have any other character in common with the Amanoa or Aristogeitonia configuration. The colpus transversalis has rounded ends. This character is present in some types of the Antidesma configuration. The oblate shape and absence of costae colpi keeps the Martretia type out of the Antidesma configuration. A typical character is the atrium, which is not found in any other type. Taxonomic discussion Martretia has only one species. In many respects this species is related to the Antidesminae (e.g. in its inflorescense). There are, however, some differences. 1. No disc is present in the male flower nor in the female one. 2. The seeds are provided with a caruncula. No other genus in the Antidesminae has carunculate seeds. 3. The only genus in the whole family with a false partition in the fruits (Pax and K. Hoffmann, 1922). Martretia quadricornis Beille Adames 251 [K] PI. VII, 4 Martretia type. 28,5 n E 31 n P;E 0,92. m:e 0,3 0,4. P.A.I. 0,25. Taxonomic comment on the Phyllanthoideae The Phyllanthoideae can be divided into three groups of pollen types. These groups, broadly outlined, agree with the three groups Pax (1924) could already distinguish on other grounds, though without nominating them. The first group, to which, for instance, the subtribes Antidesmineae and Phyllanthinae belong, has pollen grains which mostly have a prolate shape and distinct costae transversales ( Antidesma configuration, p. 20). The second group has pollen grains which are mostly reticulate and generally have an oblate shape (, Amanoa configuration, p. 32). The subtribes Amanoinae, Sarropodinae, etc., belong to this group. The third group comprises those genera that are to be found in the Aristogeitonia configuration (p. 40).

47 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 47 For the Phyllanthoideae (Pax and K. Hoffmann, 1931) composed a highly differentiated system of many subtribes. This system, however, leaves much to be desired. Some subtribes (e.g. Glochidiinae) comprise genera which decidedly do not show any relationship with each other. On the other contrary, genera in different subtribes resemble each other so much, that it would seem desirable to unite them (Wielandiinae with Amanoinae and Pseudolachnostylidinae). In the Phyllanthoideae pollen morphology is obviously exceedingly useful to arrive at a natural classification. b. Crotonoideae Croton configuration Croton-pattern present. Structure elements on the tectum are usually clavae but can also be echinae, baculae etc. Sometimes the structure elements are located on ridges (e.g. Croton matourensis, Manihot saxicola). Tectate. The Croton configuration includes all pollen grains with a croton- In the pattern. Euphorbiaceae the croton-pattern is found only in the subfamily of the Crotonoideae. Erdtman (1952) states, that, outside the Euphorbiaceae, pollen grains with a have been found croton-pattern in the Buxaceae and Thymelaeaceae. Most pollen grains with a croton-pattern belong to the Croton type. The Manihot type, Klaineanthus type and Hevea type occur less commonly. Taxonomic discussion In the subfamily of the Crotonoideae a great number of genera possess pollen grains with a croton-pattern. In Pax and K. Hoffmann s system (1931) these occur genera in the following tribes: Grotoneae, Ghrozophoreae, Joannesieae, Cluytieae, Manihoteae, Celonieae and Ricinocarpeae. Further also Neoboutonia as the only genus out ofthe It Acalypheae. follows from this enumeration, that the genera possessing a croton-pattern are to be found in widely divergent groupings. In classifying the Crotonoideaemost authors considered the bursting open of the male calyx to be a factor of primary importance for arranging the genera. Thus Mueller (1866) and Pax and K.Hoffmann arrived at a classification, in which genera with an obvious relationship are kept widely apart, e.g. Joannesia and Micrandra (Baldwin et Schultes 1947; Schultes 1955). Bentham (1880) already stated that within the genus Croton the male calyx opens both valvate and imbricate, which is in why, his system, this character plays a less In important part. Bentham s the system genera with a croton-pattern are placed close together. They are to be found in his subtribes Jatrophineae, Eucrotoninae and Ghrozophorinae. In comparing all the taxa that possess a croton-pattern it is seen that several characters correlate with this pollen character. With hardly an exception the male and the female flower contain a disc, while an ovarium rudiment is absent. A less constant but striking

48 48 W. PUNT character is the presence of petals in many genera. Genera which form an exception Sagotia. to these characters are: The disc is wanting in both sexes. The genus, however, is related to Dodecastigma, Sandwithia and Garcia (Croizat 1948), which, on the other hand, have a disc. Bertya. This Australian genus has no disc in both sexes either. There is, however, a distinct relation to two other Australian genera, viz. Beyeria and Ricinocarpos, which certainly have a disc. Tritaxis has Adenocline and no disc in the female flower. Tetrorchidium have no disc in the male flower. The genera Pantadenia and Oligoceras, described by Gagnepain, were said to possess an ovarium rudiment in their male flowers. Pantadenia does have a little conical projection in the centre ofits receptaculum, but whether this projection represents an ovarium rudiment is not certain. in Oligoceras possesses, its male flower, a column which, according to a drawing by Gagnepain in the Flore General de ITndochine (1926), was said to be an ovarium rudiment. In studying the material type in Paris, however, the author concluded, that the drawing gives an erroneous picture of the The column is reality. not bottle-shaped but cylindrical. The five stamens are not placed at the foot of the column, but are grown together with it at a little distance of the base. The projections which stand at the end of the column and which, in the drawing, look like stigmas, are three staminodies. These staminodies even contain rudimentary pollen sacs. Therefore there are two whorls of stamens, the topmost of which is sterile. A more or less distinct ovarium rudiment is found in: Klaineanthus, Cladogelonium, Micrandra and Endospermum, which all belong to the Klaineanthus type. Genera not possessing a corolla are: Neoboutonia, Benoistia, Elateriospermum, Baliospermum, Eremocarpus, Micrandra ( Cunuria ), Hevea, Bertya, Beyeria, Adenocline and Tetrorchidium. Suregada mostly has no corolla (Leonard 1958). The type occuring most frequently is the Croton type. Less common are the Manihot the Klaineanthus Hevea type, type, type and Adenocline type. Croton type Inaperturate; spheroidal or ellipsoid i.e. with a longer and a shorter axis. Tectate; endexine thin. Croton-pattern. Clavate or echinate. The Croton has type no apertures and the shape is fairly uniform. There may be a longer and a shorter axis, but usually the pollen grains are spheroidal. The structure elements have more variations. The differences in these elements and the lumina enclosed by the elements, however, are slight and many transitions exist, so that it seems better to omit the formation of groups (subtypes). The pollen grains of Sagotia racemosa show two shapes. The specimens Y. Mexia and Woob Herb, are echinate, while the specimen For. Dep.

49 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 49 plate vm. 1. Croton cuneatus; 2. Croton matourensis (x 800).

50 50 w. PUNT is clavate. Further examination will be necessary to ascertain whether these two shapes of Sagotia racemosa can be distinguished by other characters besides. According to Schultes (1952) Cunuria cannot be distinguished from Micrandra. The pollen grains of the Cunuria specimen Ducke 1087 belong to the Croton type, and not to the Micrandra type. In another Cunuria specimen (Ducke 24874) Erdtman (1952) described the pollen grains as being 5-colporate. According to Schultes (1952) the two Ducke specimens both belong to Cunuria spruceana var. bracteosa. Micrandra siphonoides and Micrandra brownsbergensis stand pollenmorphologically between the two specimens of Cunuria spruceana var. bracteosa. Croton cuneatus Klotzsch J. et P. Florschutz 1113 [U] PI. VIII, 1 Croton bahamensis Millsp. Brace 4064 [U] Croton longiradiatus Lanj. B.W [U] Croton hirtus L Heritier Versteeg 504 [U] matourensis Aubl. Croton B [U] PI. VIII, 2 Croton type. Diam. longer axis 75 /. Clavate; diam. clavae 3,5 fi. Clavae on ridges. Croton type. Diam. 55 /x. D. cl. 2,8 fi. Croton Diam. 75 type. fi. D. cl. 2,2 Clavae /x. distinctly on ridges. Croton type. Croton type. Clavae acuminate; distinctly on ridges. Diam. 115 fi. D. cl. 6,5 fx. Croton pullei Lanj. Rombouts 654 [U] Croton type. Diam. 115 D. cl. 6,5 fi. Clavae distinctly on ridges. Julocroton triqueter Baillon Hassler 6016 [U] Croton type. Diam. D. cl. 3,2 fx. 120 /x. Julocroton argenteus (Linn.) Didrichs C. L. Schulz 1225 [U] Crotonopsis elliptica WlIXD. Delzie Demaree 7066 [U] Eremocarpus setigerus (Hook.) Benth. L. S. Rose [U] Grossera Pax major Zenker 469 [U] Croton type. Diam. 120 /x. D. cl. 3,2 fx. Croton type. Diam. 60 fx. D. cl. 1,5 fx. Croton Diam. 60 type. /x. D. cl. 2 /x. Croton Diam. 48 type. fx. D. cl. 1,8 /x. Holstia tenuifolia Pax Drummond et Hemsley 3480 Cyrtogonone argentea (Pax) Prain Zenker 561 [U] Crotonogyne Pax preussii Zenker 4666 [L] _ Crotonogyne parvifolia Prain D AlTeizette 6845 [L] [K] Croton type. Diam. 45 fx. D. cl. 2 fx. Croton type. Diam. 60 /x. D. cl. 3,5 fx. Croton type. Diam. 60 /x. D. cl. 2 /x. Croton type. Diam. 50 fx. D. cl. 2,8 /(. Manniophyton fulvum Muell. Arg. Zenker 1404 [L] Aleurites moluccana (Linn.) Willd. Croton Diam. type. 60 /t. D. cl. 3,5 /x. Croton type. Diam. 70 fx. Bunnesmeyer 8060 [U] Deutzianthus tonkinensis Gagnep. Balansa 3149 [P] D. cl. 2,0 /x. Croton type. Diam. 65 /x. D. cl. 2,0 /x. Oligoceras eberhardtii Gagnep. Eberhardt [P] Alphandia furfuracea Baillon Balansa 3439 [P] Croton type. Diam. 50 /(. D. cl. 1,8 fx. Croton type. Diam. 40 /x. D. cl. 2,0 /.i. Garcia nutans Vahl Croton type. Diam. 60 /(. Lundell [U] D. cl. 2,0 fx.

51 D. cl. 1,5 ft. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 51 Joannesia princeps Velloso T. W. Brown 35 [K] Croton type. Diam. 70 ft. D. cl. 3,0 ft. Clavae on ridges. Neoboutonia macrocalyx Pax Croton type. Diam. 40 ft. Stolz 2043 [U] Pantadenia adenanthera Gagnep. D. cl. 1,5 fi. Croton type. Diam. 45 ft. British Museum D. cl. 1,0 fi. Baloghia lucida Endlicher Croton type. Diam. 40 fi. Beckler, Hastings River [U] D. cl. 1,0 ft. Ostodes paniculata Blume Croton type. Diam. 65 ft. Griffith 4790 [U] D. cl. 3,5 ft. Ostodespendulus (Hassk.) Meeuse et Croton type. Diam. 53 ft. ADIPT RFRT D. cl. 2,8 ft. Sandakan For. Dep. A [U] Codiaeum variegatum (Linn.) Blume Croton type. Diam. 50 ft. Fresh material from greenhouse at D. cl. 2 ft. Utrecht. PI. IX, 3 Codiaeum stellingianum Ware. Docters v. Leeuwen 1567 _[U] Croton type. Diam. D. cl. 2,0 ft. 50 ft. Fontainea pancheri (Baillon) Heckel Croton type. Diam. 60 ft. D Alleizette 6474 Dimorphocalyx murina [L] Elmer D. cl. 2,8 ft. Croton type. Diam. 80 ft. Elmer [U] D. cl. 3,5 ft. Benoistia perrieri Leandri D Alleizette Madag. Dec [L] Croton type. D. cl. 1,8 ft. Diam. 35 ft. Tritaxis gaudichaudi Baill. Croton type. Diam. 45 ft. Clemens 4351 [U] Strophioblachia fimbricalyx Boerl. Sort. Bog. VIS F. 43 [U] Blachia umbellata (Willd.) Baill. D Alleizette Austr. Nov [I.] Echinate. Diam. ech. Croton Diam. 60 type. ft. D. cl. 1,8 ft. Croton Diam. 45 type. ft. D. cl. 1,8 ft. 1,8 ft. Sagotia racemosa Baillon For. Dep [U] Croton type. Diam. 50 ft. D. cl. 2,0. Clavae sharp pointed. Mexia 6050 [U] Wood Herb. 311 [U] Jatropha multifida Linn. Groll-Meyer, St. Eustatius [U] Jatropha integerrima Jacq,. Docters v. Leeuwen [U] Mildbraedia paniculata Pax Leeuwenberg 3004 [U] Echinate. Diam. 40 D. ech. ft. 1,8 ft. Diam. 95 Croton type. ft. D. cl. 4,0 ft. Clavae on ridges. Croton type. Idem. Jatropha multifida Croton type. Diam. 50 ft. D. cl. 2,5 ft. Acidocroton adelioides Crises. Croton type. Diam. 45 ft. Ekman [K] Pausandra densiflora Lanj. Krukoff 5464 [U] D. cl. 1,5 ft. Croton type. Diam. D. cl. 2,0 ft. 50 ft. Pausandra morisiana (Casar.) Radlk. Croton type. Diam. 75 ft. Ducke et Kuhlmann [U] D. cl. 3,5 ft. Givotia rottleriformis Griffith Croton type. Diam. 48 ft. Wight 2638 [L] Ricinodendron heudelotii (Baill.) Pierre D. cl. 2,8 ft, Croton type. Diam. 50 ft. D. cl. 2,5 ft. Zenker 487 [U] Elateriospermum tapos Blume Hort. Bog. IX. A. 129 [U] Croton type. Diam. 70 ft. D. cl. 3,2 ft. Clavae on ridges. Baliospermum motanum (WlLLD.) MUELL. Croton type. Diam. 43 ft. Bengal, J. D. Hooker [U] Cavacoa aurea (Cavaco) Leonard Croton type. Diam. 75 ft. de Winter 2109b [K] Anomalocalyx uleanus (Pax et K. Hoffm.) OlTPIfP D. cl. 2,0 ft. Croton Diam. 55 type. ft. D. cl. 3,2 ft. Ducke [U]

52 52 W. PUNT plate rx. 1. Manihot saxicola (X 600); 2. Tannodia cordifolia; 3. Codiaeum variegatum.

53 POLLEN MORPHOLOGY OF THE EUPHORBLAGEAE 53 Sandwithia guyanensis Lanj. Croton type. Diam. 55 fi. For. Dep [U] D. cl. 3,5 fi. Dodecastigma amazonicum Ducke Croton type. Diam. 80 ft. Ducke [U] Cunuria spruceana Baillon var. bracteosa D. cl. 4,5 ft. Croton type. Diam. 50 fi. (Ducke) R. E. Schultes D. cl. 1,8 fi. Ducke 1087 [K] Vaupesia cataractarum R. E. Schultes Croton type. Diam. 70 fi. Schultes [K] Ricinocarpos pinifolius Desf. Austr. felix [U] D. cl. 3,0 ft. Croton type. Diam. 60 fi. D. cl. 1,8 ft. Beyeria leschenaultii (Decand.) Baill. Croton type. Diam. 43 fi. Austr. felix [U] D. cl. 1,5 u. Clavae sharp pointed. Bertya gummifera Planch. Johnson et Constable [U] Croton type. Diam. 55 fi. Croton-pattern indistinct. Domohinea perrieri L andri Croton type. Diam. 50 ft. D Alleizette 6479 Dec [L] PI. X, 1 Croton-pattern on ridges. Echinate; diam. ech. 1,5 fi. Tannodia cordifolia Baillon D Alleizette, Comoren 6484 [L] PI. IX, 2 Croton type. Diam. 35 fi. Echinate; diam. ech. 1-1,5 fi. Manihot type Periporate; spheroidal. Pori large, borders indistinct. Tectate. Croton-pattern. The type is closely related to the Croton type. The pollen grains, however, have large with indistinct rims. The number of apertures pori varies in the different species. Taxonomic discussion According to Croizat (1943b) Cnidiscolus does not have a calyx but a corolla. This is also the case with Manihot. In a very definite sense Cnidiscolus connects Jatropha, sensu strictu, with Manihot. Both the sorts of Cnidiscolus examined had pollen grains of the Manihot type, sothat at least in this respect they are closer related to Manihot than to Jatropha. The pollen-morphological similarity between Manihot and Suregada is probably accidental and has to be considered a parallel development. Manihot saxicola Lanj. Manihot type. Proefst. Buitenz. 922 Boom 8 PI. IX, 1 [U] Diam. 160 fi. Croton-pattern on a distinct reticulum. Pori ca. 25. Diam. clavae 6,8 fi. Manihot esculenta Crantz Rombouts 757 [U] Manihot type. Idem Manihot saxicola. B.W [U] Pori ca. 25. Cnidiscolus urens (Linn.) Arthur Boldingh 4829 [U] Manihot type. Diam. 75 fi. D. cl. 2,5 Pori 3^1. fi. Cnidiscolus stimulosus (Michaux) Gray Manihot type. Pori ca. 6. Killip [U] Suregada glomerulata (Blume) Baill. Griffith 4772 [U] Suregada subglomerulata (Elmer) Croizat Elmer [U] Manihot Diam. 40 type. ft. D. cl. 2,5 Pori fi Manihot type. Diam. 45 /i. Pori 5-6.

54 54 W. PUNT Klaineanthus type Tricolpate; oblate spheroidal to oblate. Colpi wide or narrow; no operculum present. Tectate; croton-pattern. The Klaineanthus type, too, is closely related to the Croton type and the Manihot type. Pollen grains of this type have colpi without an operculum. The colpus membrane is frequently ruptured. Klaineanthus subtype Boarders of colpi distinct. Glavate. The rims of the colpi of this subtype are distinct. The colpus membrane is frequently totally In ruptured. Cladogelonium the colpus membrane is not ruptured but granulate. Sometimes a rudiment of the colpus membrane remains at the outer ends of the colpi. Klaineanthus gaboniae Pierre ex Prain Zenker 583 [U] PL X, 3 Cladogelonium madagascariense L6andri Perrier de la Bathie 9696 [P] PI. X, 4 Endospermum moluccanum (Thysm. et Binnend.) Beccari Kornasse 462 [U] Glycydendron amazonicum Ducke Ducke [U] Klaineanthus subtype; suboblate. 29,5 /i 37/1 P:E 0,80. Colpi wide. P.A.I. 0,4. Diam. clavae 1,8 /t. Klaineanthus subtype; spheroidal. 30/i E 30,5 P:E fx 0,98. Colpus membrane granulate. P.A.I. 0,4. Klaineanthus subtype; suboblate. 33,5 E n 44,5 P:E /i 0,80. Colpus membrane absent. P.A.I. 0,4. Diam. clavae 1,8 /i. Klaineanthus subtype; oblate. 29 /«E 41,5/t P:E 0,69. Colpus membrane absent. P.A.I. O, Diam. clavae 1,5 fi. Micrandra brownsbergensis Lanj. Klaineanthus subtype. v. Emden 9 [U] Micrandra siphonioides Benth. Colpus membrane Klaineanthus subtype. absent. Lopez 9625 [U] PI. X, 5 41,5 fi E 46 /< P:E 0,92. P. 0,3. Croton-pattern indistinct. Clavae candelabrum-shaped, in-ordinate arranged. Adenocline subtype Borders of the colpi not sharply cut but irrigularly ruptured. Baculate. Baculae small and crowded. The baculae on the tectum are small and crowded. With low magnification it is not directly obvious that the structure elements form a croton-pattern and not a reticulum. Taxonomic discussion The genus Adenocline was placed in the vicinity of Seidelia and Leidesia in the Acalypheae by Bentham (1880). Pax and K. Hoffmann (1931) could not find the same relation and put the genus in a separate subtribe in the tribe Gelonieae. This tribe also comprises the genus Klaineanthus. The two genera are, however, quite different in other respects.

55 POLLEN MORPHOLOGY OF THe'eUPHORBIACEAE 55 plate x. 1. Domohinea perrieri; 2. Suregada subglomerulata; 3. Klaineanthus gaboniae; 4. Cladogelonium madagascariense; 5. Micrandra siphonioides; 6. Hevea brasiliensis.

56 56 W. PUNT _ Adenocline acuta (Thunb.) Baiix. [ A. mercurialis Turcz.] Lanjouw 613 [U] PI. XI, 2 Adenocline subtype; oblate spheroidal. 36 /a E 37,5 n P:E 0,96. P.A.I. 0,4. Hevea type Tricolpate; oblate spheroidal. Colpi wide; operculum present. Tectate; baculate. The baculae are so small and crowded, that the structure is hardly recognised as a croton-pattern. As in the Adenocline subtype the baculae of the Hevea type are small in diameter and crowded. With low magnification the structure elements seems to form a reticulum, but with high magnification the arrangement of the structure elements is clearly a croton-pattern. The operculum distinguishes this type from the Klaineanthus type. Hevea brasiliensis (Willd. ex A. Jussieu) Mueix. Arc. Schultes 8050 [U] B.W. 752 [U] PI. X, 6 Hevea guianensis Aublet B.W. 41 fu] Krukoff 5800 [U] Hevea type; oblate spheroidal. 33,5 E 35 P:E fi n 0,96. P.A.I. 0,3. Hevea type. 35 i E 38 n P:E 0,95. Tetrorchidium type Tricolporate; oblate spheroidal to prolate spheroidal. Colpus transversalis exceptionally broad and elongated. No costae present. Colpi narrow; rims irregular. Intectate; pilate. Pila coarse. The Tetrorchidium type is tricolporate and differs by this character from the tricolpate Klaineanthus type. The extremely broad colpus transversalis is elongated to a colpus transversalis equatorialis. The tectum of this type is indistinct and the croton-pattern is not clear in all the pollen grains. In Tetrorchidium didymostemon the structure elements on the tectum are certainly not arranged in a croton-pattern. Tetrorchidium rubrivenium shows the croton-pattern more readily. Although the most important characters of the Croton configuration are not always present, the structure of the pollen grains and the shape of the colpi are much like those in the Klaineanthus very type. The author therefore prefers to keep the Tetrorchidium in type the vicinity of the Klaineanthus type. Taxonomic discussion Pax and K. Hoffmann (1931) v, referred Tetrorchidium into the subtribe Tetrorchidiinae. This subtribe was placed in the tribe Gelonieae, to which also Klaineanthus belongs. Although in habit and character Adenocline and Tetrorchidium differ widely, the flowers have some characters in common: 1. In the male flowers disc and ovarium rudiment are absent. 2. In the female flowers the disc consists of three free glands.

57 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 57 Tetrorchidium rubrivenium Poepp. Harris and Britton [K] Tetrorchidium didymostemon (Baillon) Pax et K. Hoffmann Zenker 4590 [L] PI. XI, 3 Tetrorchidium type; oblate spheroidal. P 36 n 40/1 P:E 0,90. P.A.I. 0,35. Bacculate-verruca te. Tetrorchidium type; prolate spheroidal. 28 /i E 27 n P:E 1,04. P.A.I. 0,35. Cnesmosa configuration Inaperturate; no croton-pattern present. Tectate or intectate. Like the Croton type the Cnesmosa configuration is inaperturate, but a croton-pattern is always absent. Cnesmosa type Inaperturate. Pollen grains ellipsoid, i.e. with a longer and a shorter axis, Tectate; psilate. No reticulum. The pollen grains of this type are closely related to the Tragia fallax type and the Platygyne type. In the Cnesmosa type the pollen grains are tectate. Taxonomic discussion The genera belonging to this type are closely related and not easy to separate. Croizat (1941a) discussed the relationship between some genera of the Plukenetiinae and came to the following conclusions: 1. Cenesmon Gagnepain cannot be separated from Cnesmosa Blume. 2. Megistostigma malaccense, reduced to Sphaerostylis by Pax and K. Hoffmann, is the taxonomic type of the genus Megistostigma, which is congeneric with Clavistylus J. J. Smith. 3. Tragiella Pax and K. Hoffmann should be treated as a synonym of Sphaerostylis, ofwhich S. tulasneana is the taxonomic type (p. 63). The pollen-morphological results are in full agreement with the statements of Croizat. Cnesmosa javanica Blume Cnesmosa type; ellipsoid. Elmer, Borneo [U] Longer axis 55 p. Java 1855 [L] Megistostigmapeltatus (J.J. Smith) Croizat Cnesmosa i type; ellipsoid [L] PI. XI, 6 Longer axis 56 fi ; Steenis shorter axis 53 p. v. Megistostigma malaccense J. D. Hook. Sinclair [U] PI. XI, 4 Cnesmosa type; ellipsoid. Longer axis 42 //; shorter axis 37 /i. Columellae in groups. Acidoton urens Swartz Jamaica, Hooker [U] Cnesmosa type; ellipsoid. Longer axis 37,5 fi\ shorter axis 35,5 ft. Tragia fallax type Pollen grains inaperturate; spheroidal or ellipsoid. Intectate, pilate. Pila not in a reticulum. Endexine very thin. This type differs from the Cnesmosa type by its intectate structure. The pila are not arranged in a reticulum.

58 58 W. PUNT Taxonomic discussion The examined species belong to the section Bia of Tragia. Klotzsch and Baillon (1858) separated this section from Tragia. Pax and K. Hoffmann (1919a) also stated: Die mannliche Bliite weckt den Eindruck, als ob es um eine eigene Gattung handelt Mfall the species of the section Bia have the inaperturate pollen grains, it is perhaps better to treat this section as a distinct genus, although a great affinity with Tragia cannot be denied. Tragia fallax Muell. Arc. Klug 4207 [U] PI. XI, 5 Tragia sellowiana (Klotzsch) Muell. Arc. Smith 3558 [U] Tragia fallax - yp e ; Diam. 50 u. Pila 1,5 /J. Tragia fallax type. Diam. 55 /a. Platygyne type Inaperturate. Pollen grains ellipsoid. Tectate; intra-reticulate. Endexine with circular thickenings. The Platygyne type is distinguished from the Cnesmosa type by its distinct intra-reticulum. Small circular thickenings occur on the endexine. These peculiar costae are distributed irregularly over the endexine and certainly not accompagnied ofthe by thinnings ektexine. Platygyne hexandra (Jacq.) Muell. Arg. Platygyne type; ellipsoid. Rutten-Pekelharing 346 [U] PI. XI, 1 Longer axis 40 fx. Shorter axis 36 u. Lumina 2-3 fi. Trigonostemon verrucosus type Inaperturate; spheroidal or ellipsoid. Tectate. Gemmate. Trigonostemon verrucosus J. J. Smith Cult. Hort. Bog. VIII. E-16 [U] PI. XII, 3 Trigonostemon fungii Merrill Linsley Grishitt 1089 [BM] Trigonostemon verrucosus type; spheroidal. Longer axis 60 /(. Diam. gemmae 2 jx. Gemmae not in a croton-pattern. Trigonostemon verrucosus type. Diam. 60 Diam. fi. gemmae 1,5-2 /i. Trigonostemon redioides type Inaperturate; spheroidal to elliptical. Intectate; reticulate. Both Trigonostemon types are inaperturate without a croton-pattern. They belong for these reasons to the Cnesmosa configuration. The Trigonostemon verrucosa type with the gemmate structure elements on its tectum is perhaps related to the Croton type (p. 48). The gemmaeare, however, certainly not arranged in a croton-pattern. Except for the shape and lack of apertures the Trigonostemon redioides type does not have any character in common with the Croton configuration.

59 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 59 plate xi. 1. Platygyne hexandra; 2. Adenocline mercurialis; 3. Tetrorchidium didymostemon; 4. Megistostigma malaccense; 5. Tragia fallax; 6. Megistostigma peltatus.

60 ..., 60 w. PUNT Taxonomic discussion Although the pollen grains of the Trigonostemon type with a croton-pattern, the genus are not provided Trigonostemon has great affinity with genera in the Croton configuration. Corresponding characters are: 1. In the male as well in the female flowers a disc is found and an ovarium rudiment is absent. 2. Sepals are present. Trigonostemon redioides (Kurz) Craib Coudre [BM] PI. XII, 5 Trigonostemon longifolius Wall, ex Baillon Achmad 142 [U] Prosartema gaudichaudii Gagnep. Poilane [K] Trigonostemon redioides type. Longer axis 34 Shorter axis 30 Lumina 1 fi. ft. -2 ft. Trigonostemon redioides type. Longer axis 38 p\ snorter axis 35 fi. Lumina ca. 2 fi. Actephilopsis malayana Ridl. Ridley 2300 [K] Trigonostemon redioides Trigonostemon redioides type- Diam. 55 Lumina 3-5 ft. fi. type- Longer axis 45 fi. Lumina small, ca. 1 /i. Dysopsis configuration Inaperturate. Pollen grains quasi tricolpate, distinctly three-lobed. Tectate; no croton-pattern present. Dysopsis type Inaperturate; quasi tricolpate; oblate spheroidal. Tectate; psilate. Columellae fairly coarse. The most striking character is found in the three-lobed shape of the pollen grains. The declination between the lobes is distinct, but there is not the slightest indication of any part of the exine being weakened. According to the definition (see glossary) the pollen grains are inaperturate. Taxonomic discussion Mueller (1866) as well as Pax and K. Hoffmann (1914, 1931) kept Dysopsis in close relationship with Leidesia and Seidelia. These latter African genera are, however, geographically far removed from the Andine genus Dysopsis. The three genera belong undoubtedly to the Acalypheae. The pollen grains of Dysopsis have little affinity with those of Leidesia and Seidelia. Perhaps it should be better to separate Dysopsis in a different subtribe. Dysopsis glechomoides(rich.) Muell. Arg. Dysopsis type.. E 22,5 i f P:E 0,93. Lechler 956 [K] PI. XII, 6 P 21 /i Plukenetia configuration Tricolpate or triporate. Shape of the pollen grains oblate spheroidal to oblate (rarely prolate spheroidal). Apertures broad; costae absent. Colpus or porus membrane ruptured.

61 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 61 Plukenetia verrucosa; 2. Fragariopsis scandens; Trigonostemon verrucosus; plate xn Plukenetia volubilis; Trigonostemon redioides; Dysopsis glechomoides.

62 62 W. PUNT The types of the Plukenetia configuration have simple apertures. In contrast with the Chiropetalum configuration the apertures are broad and have a ruptured colpus membrane. In some species a small piece of this membrane persists at the outer ends of the colpi. Plukenetia type Tricolpate; oblate spheroidal to oblate (rarely spheroidal or prolatespheroidal). Colpi broad; colpus membrane ruptured. At the outer ends of the apertures a small piece of the colpus membrane (i.e. endexine) persists. The ruptured part is sometimes bordered by a margo. Tectate or intectate. Reticulate or not reticulate. Pollen grains belonging to this type can have very different exine structures. They may be intectate, tectate or tectum perforate, while the columellae are inordinate arranged or form a reticulum. The various subtypes are based on the structure patterns of the exine. Taxonomic discussion From the genus Plukenetia new genera have been separated by several authors. 1 j ax and K. Hoffmann maintain (1919a) this division ~ of Plukenetia. Bentham (1880), however, has pointed out that Fragariopsis and Romanoa differ only slightly from the typical Plukenetia species. Other authors, too, see little difference. Uroizat (1941a): JNo characters are available to separate Pterococcus and Tetracarpidium from Plukenetia Macbride (1951): Apodandra and Fragariopsis could readily be included in Plukenetia". Certainly the pollen grains of the above genera are not easily distinguished from those of Plukenetia. In South-America three tree-shaped, non-climbing genera are to be found. Sandwith (1950) says of them: Astrococcus and Angostyles, Haematostemon three are closely allied genera of the Plukenetiinae. The pollen is indeed related. They all belong to theplukenetia volubilis subtype and possess a tectum perforatum. Astrococcus and Haematostemon have moreover a margo. On account of the imbricate calyx Bentham and Pax and K. Hoffmann placed Omphalea outside the Plukenetiinae. Croizat (1942a, 1948), however, gives as his opinion that the genus shows strong affinity to the Plukenetiinae. The pollen of Omphalea also appears to resemble the Plukenetia volubilis subtype very closely. Plukenetia volubilis subtype Tectate; psilate. Sometimes the tectum is a tectum perforatum. Plukenetia volubilis Linn. Baker 153 [U] PI. XII, 4 Plukenetia abutifolia (Ducke) K. Hoffm. Ducke [U] Eleutherostigma lehmannianum K. Hoffm. Pax et Pax et Lehmann 5158 [K] PI. XIII, I Plukenetia volubilis subtype; oblate. 36 E 56 ft P:E 0,64. P.A.I. 0,35. Plukenetia volubilis subtype. 50 n E 62 n P;E 0,80. Tectum perforatum. Plukenetia volubilis subtype. 37 E fi 47,5 P;E n 0,79. P.A.I. 0,35.

63 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 63 Tetracarpidium conophorum (Muell. Arg.) Pax et K. Hoffm. [ Plukenetiaconophorum Muell. Arg.] Mann 1739 [U] Plukenetia volubilis subtype. 30,5 E 44 P;E n n 0,69. P.A.I. 0,15 0,2. Margo. Versuchsanst. 9 [U] Pterococcus corniculatus (E. Pax et Smith) K. Hoffm. [ Plukenetia corniculatus E. Smith] Koorders B [U] Plukenetia volubilis subtype. P 32 n E 44 P:E n 0,66. P.A.I. 0,2. Anabaenella tamnoides K. Hoffm. (Juts.) Pax et Plukenetia volubilis subtype. 43 n E 49,5 fi P:E 0,87. nom. illeg. Romanoa [ ] Trevisan Plukenetia tamnoides (Juss.) Muell. P.A.I. 0,2. Margo. Lutz [U] Angostyles longifolia Benth. Ducke [U] Astrococcus cornutus Benth. Spruce 2050 [K] PI. XIII, 3 Plukenetia volubilis subtype. P45 Ju P.A.I. 0,45. Plukenetia volubilis subtype. E 60 /< P:E 0,75. Tectum perforatum. P 36,5 n E 50 P:E n 0,73. P.A.I. 0,4. Margo; tectum perforatum. Haematostemon guianensis Sandwith Fanshawe 6016 [U] Plukenetia volubilis subtype. 29 /j. E40/r P:E 0,74. Omphalea diandra Linn. P.A.I. 0,25. Margo; tectum perforatum indistinct. Plukenetia volubilis subtype. Frees (Krukoff) 1893 [U] PI. XIII, 2 Lanjouw and Lindeman 3164 [U] Tragia capensis Thunb. Hooker Herb. [K] 21,5 n E 30,5 n P.A.I. 0,3. P;E Plukenetia volubilissubtype; spheroidal. E 17 / a. Plukenetia verrucosa subtype Intectate (?); reticulate. Plukenetia verrucosa Smith B.W. 993 [U] PI. XII, 1 Plukenetia verrucosa subtype; suboblate. 28 /x E 35 // P:E 0,80. Apodandra buchtienii (Pax) Pax [ Plukenetia buchtienii Pax] Krukoff [U] Fragariopsis scandens St.-Hill. D Alleizette 6594 [L] PI. XII, 2 P.A.I. 0,3. Lumina 2-3 fi- Plukenetia verrucosa subtype. Plukenetia verrucosa subtype. 33,5 n E 41,5 P:E n 0,81. P.A.I. 0,45. Tragia stolziana Pax et K. Hoffm. Plukanetia verrucosa subtype. Stolz 1775 [U] PI. XIII, 4 P29/t E 32 P:E 0,90. Reticulum consisting of separate pila, lumina 1-2 (i. P.A.I. 0,6. Sphaerostylis natalensis (Sond.) Croizat v. Someren 2409 [K] Plukenetia verrucosa subtype. P 38,5 ti E43/t P:E 0,89. P.A.I. 0,6. Lumina 1-2 /i. Sphaerostylis tulasneana Baill. D Alleizette, Nossibe. Nov [L] Plukenetia verrucosa subtype. 35 /«E 44 /I P:E 0,84. P.A.I. 0,3. Lumina 1-2 /i. Tragia tristis subtype Intectate; pilate. Pila not in a reticulum. Tragia tristis subtype. Tragia tristis Muell. Arg. Hassler 6645 [BM] PI. XIII, 5 27 E 40 n fi P;E 0,68. P.A.I. 0,3.

64 64 W. PUNT plate xin. 1. cornutus; Eleutherostigma lehmannianum; Tragia stolziana; 5. Tragia tristis; Omphalea 6. diandra; Pachystilidium 3. Astrococcus hirsutum.

65 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 65 Tragia geraniifolia Baill. Herter 1006 A. [U] Tragia tristis subtype. P 31,5 /i E 41,5 ti P:E 0,76. Tragia volubilis Linn. Focke 992 [U] Tragia ramosa Torr. Rogerson [U] P.A.I. 0,3. Pila ca. 1 fi. Tragia tristis subtype. Tragia tristis subtype. Pachystylidium type Triporate; spheroidal. Apertures are circular in shape and sunk down. The borders are indistinct. Tecta te; psilate. The Pachystilidium has circular of which the rims type apertures, are indistinct. It is, however, not clear if the type belongs with certainty to the Plukenetia configuration. The ektexine is only very indistinctly weakened and it is therefore possible to call the type inaperturate. The endexine, however, is clearly thinned. In the author s opinion it is better to keep the type in theplukenetia configuration and not in the Cnesmosa configuration. Pachystilidium hirsutum (Blume) Pax et Pachystylidium type.., K. Hoffmann E 29 fi. P.A.I. 0,6-0,7. Elbert 3340 [L] PI. XIII, 6 Chiropetalum configuration Tricolpate or stephanocolpate (4); shape variable. No Colpi long. costae colpi. Colpus membrane (endexine) persistent. [Rarely tricolporate, but then pollen grains prolate in shape and an operculum present.] This configuration is distinguished from the Plukenetia configuration by the colpus membrane, which is never ruptured. Taxonomic discussion The genera of the Chiropetalum type, Ditaxis type and Argythamnia type are closely related. Mueller and (1866) Bentham united them in one taxon, the genus Argythamnia (except Leucocroton). Pax and K. Hoffmann (1912c, 1931) kept the genera separated, and according to pollen morphology this seems the right opinion. Chiropetalum type Tricolpate (or rarely tricolporate); spheroidal to prolate spehroidal. Colpi wide; operculum present; costae colpi absent. Tectate or intectate; finely reticulate. [Rarely colpus transversalis present.] The Chiropetalum type and the Lasiocroton type are related by the presence of an operculum. Chiropetalum and its allies, however, are finely reticulate. From the Ekman 16425a sheet Erdtman (1952) described Leucocroton flavicans as colpate. The Ekman sheet as examined by the

66 66 w. PUNT " author from the Kew herbarium has These tricolporate pollen grains. pollen grains should be placed in a different configuration, although other characters (e.g. the operculum and fine reticulum) are consistent with the Chiropetalum type. Noth withstanding the colpus transversalis the author prefers to place the pollen grains of Leucocroton flavicans in the Chiropetalum type. Chiropetalum lanceolatum A. Jussieu Wedermann 34 [U] PI. XIV, 2 Chiropetalum type; spheroidal. E 37,5 fi P: E ca. 1. P.A.I. 0,3. Lumina ca. 1 /J,. Intectate. Chiropetalum canescens Phil. Wedermann 773 [U] Aonikena patagonica Spego. [ Chiropetalum patagonica (Spegg.) O Donn. et Lourteig] Eyerden, Beetle and Grond [K] PI. XIV, 4 Leucocroton flavicans Muell. Arg. Ekman [K] PI. XIV, 5 Chiropetalum type; prolate spheroidal. Lumina 1-2 fi. Intectate. Chiropetalum type; prolate spheroidal. 34,5 n E 30,5 n P:E 1,12. P.A.I. 0,3. Lumina 1-2 fi. Chiropetalum type; subprolate. 30,5 n E 25,5 P:E 1,19. Colpus transversalis without costae. P.A.I. 0,4. Lumina ca. 1 fi. Ditaxis type Tricolpate; oblate to suboblate. The pollen grains are bi-laterally symmetric. One lobe of the three-lobed pollen grain is smaller than the other two. Colpi long and not very narrow. Tectate; psilate; not reticulate. This type has bi-symmetric pollen grains. Of the three lobes one is smaller than the other two. This curious shape was observed in six species of Ditaxis. Argythamnia candicans does not possess these bisymetric pollen grains. Taxonomic discussion It could not be investigated whether other species of Argythamnia perhaps do show the Ditaxis type. Croizat (1945) did not see any reason to separate Ditaxis and Argythamnia. The difference in pollen type, however, supports the opinion of Pax and K. Hoffmann who (1931), maintain Ditaxis next to Argythamnia. Ditaxis fendleri (Muell. Arg.) Pax et K. Hoffm. Staffers 1703 [U] PI. XIV,_1 Ditaxis catamarcensis (Griseb.) Pax Brizuela 759 [U] Ditaxis diversiflora Glokey Clokey 7576 [U] Ditaxis fasciculata Vahl ex A. de Juss. Boldingh 2733 A [U] Ditaxis lancifolia Schlechtend. Broadway 9257 [U] Ditaxis salina Pax et K. Hoffm. Ditaxis type; oblate. 30,5 n E 51 n P:E 0,60. P.A.I. 0,15. Ditaxis type. 33 n E 42 n P:E 0,79. Ditaxis type. Some Ditaxis type. pollen grains 4 colpate. 43 /i E 50 n P:E 0,86. P.A.I. 0,2 0,25. Ditaxis type. P.A.I. 0,2 Ditaxis type. 0,25. Pedersen 2638 [U]

67 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 67 plate xiv. 1. Ditaxis fendleri; 2. Chiropetalum lanceolatum; 3. Argythamnia candicans; 4. Aonikena patagonica; 5. Leucocroton flavicans; fi. Lasiocroton macrophyllus.

68 68 w. PUNT Argythamnia type Stephanocolpate; oblate spheroidal. Colpi long, not very narrow. Tectate; on the of the top tectum a reticulum is superimposed. Although there are some similarities with the Ditaxis type, shape as well as structure differ too much for the pollen grains to be placed in the same type. Argythamnia candicans Sw. Boldingh 3401 B. [U] PI. XIV, 3 Argythamnia type; 4 (-3)-colpate. E45/t P:E 42 n 0,93. P.A.I. 0,5. Lasiocroton type Tricolpate or stephanocolpate; spheroidal to oblate spheroidal. Colpi wide; operculum present; no costae present. Tectate; psilate; no reticulum present. The difference between the Chiropetalum type and Argythamnia type is found in the absence of a reticulum. According to Erdtman (1952) the pollen grains macrophyllus are reticulate. With the aid of a phase of Lasiocroton contrast microscope it was possible to conclude, that the of pollen grains the Harris specimen are certainly not reticulate. Taxonomic discussion Lasiocroton and Adelia, together with Leucocroton (p. 66), formed the series Adeliiformes (Pax and K. Hoffmann 1914, in 1931) the subtribe Mercurialinae. These three genera are certainly related by the following characters: 1. Disc present in the male as well as in the female flowers. This disc is usually annular. 2. Ovarium rudiment in the male flower absent. 3. Petals absent. Number of sepals usually Plants usually dioecious and only growing in the New World. Lasiocroton macrophyllus (Swartz) Griseb. Lasiocroton type; oblate spheroidal. Harris [K] PI. XIV, 6 18,5 n E P.A.I. 0,3. 19,5 fi P:E 0,95. Adelia ricinella Linn. Rutten-Pekelharing [U] Fuertes 833 [U] Lasiocroton type; spheroidal. (3-) 4-coIpate. P E 33 [i. Operculum narrow. P.A.I. 0,4-0,5. Cephalomappa configuration Tricolpate. Colpi very short; costae colpi present. Cephalomappa type Tricolpate; oblate spheroidal. Colpi very short; costae colpi present. Tectate; intra-reticulum coarse. The Cephalomappa type differs completely from all other types in

69 - stephanocolporate; POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 69 the Crotonoideae. Besides the very coarse reticulum the colpi are short and bordered by distinct costae colpi. Taxonomic discussion The genus Cephalomappa is certainly related to those in the Cladogynos type (p. 93). Cephalomappa malloticarpa J. J. Smith Herb. Lugd. Batav [L] Cephalomappa type... 30,5 fi E 32,5 P:E p 0,94. PI. XV, 5 Colpi length ca. 4 / 1. Lumina 2-3 fi. SUMBAVIA CONFIGURATION Tricolporate usually Polar axis shorter than Equatorial axis. Colpi wide or narrow. Tectate or intectate; reticulate. Reticulum coarse. The most important character of this configuration is the coarse reticulum. Some species in the Cladogynos configuration (p. 93) also have a fairly coarse reticulum but inthat case they have a Polar axis larger than the Equatorial axis. In the Sumbavia configuration the pollen grains usually have a Polar axis shorter than the Equatorial axis. Taxonomic discussion Most genera in this configuration belong to the Chrozophoreae (Bentham, 1880; Pax and K. Hoffmann 1912c, 1931). These authors split up this tribe in Bentham s third groups. group is comparable with the subtribe Regulares of Pax and K. in Hoffmann, which Sumbavia, Chrozophora and Caperonia are found. Sumbavia type Tricolporate; oblate spheroidal to oblate. Colpus transversalis large; broad or long and narrow. Colpi long; costae colpi present or absent. Intectate; reticulate. Reticulum coarse. The pollen grains of the Sumbavia type are always tricolporate and have a distinct reticulum. They differ from the Chrozophora type and Caperonia subtype by the number of colpi and from the Philyra subtype by the coarse reticulum and elongated colpus transversalis. In 1952 Erdtman described pollen grains of Monotaxis grandiflora which are subprolate and have a membrana granulata. This discription is not in accordance with the data found by the author. The pollen grains examined were suboblate and had no membrana granulata. The exine structure, however, is quite the same. Taxonomic discussion In the of system Pax and K. Hoffmann (1931) Thyrsanthera and Sumbaviopsis are placed next to Sumbavia. This is in with the agreement close relationship of the pollen grains.

70 . _ 70 W. PUNT Monotaxis differs from the other genera in the Sumbavia type by the absence of stellate hairs and the presence of a disc and an ovarium rudiment in the male flowers. The flowers have, however, also five petals as. in most genera of the Sumbavia. typein many respects Melanolepis resembles Sumbavia and Sumbaviopsis. As with these genera there is no disc or ovarium rudiment in the male flower. On the other hand the female flower possesses a disc. The plants have stellate hairs, they are monoecious and show similarity in the formof the leaves. Melanolepis, however, has no petals. Baillon (1858) therefore calls Melanolepis a Sumbavia without petals. Mueller (1866), Bentham (1880) and Pax and K. Hoffmann see Mallotus as nearest relation of Melanolepis. An accurate comparison of the corresponding characters will have to decide whether Melanolepis is closer related to Mallotus than to Sumbavia and Sumbaviopsis. Sumbavia rottleroides Baill. Merrill Sp. Blancoanae PI. XV, [L] Sumbavia type; oblate spheroidal. 43ju E 46 P:E 0,93. m:e 0,6. P.A.I. 0,35. Costae colpi absent. Lumina 3-4 /x. Sumbaviopsis albicans J. J. Smith Backer [L] Sumbavia type; suboblate. P 32 n E 38,5 fi P:E 0,80. P.A.I. 0,3. Lumina 1-2 /x. Costae colpi. Thyrsanthera suborbicularis Pierre ex Gagnepain Pierre 512 [K] PI. XV, 3 Melanolepis multiglandulosa (Reinw. ex Blume) Reich.f. et Zollinger Hort. Bogor [U] PI. XV, 1 Monotaxis grandiflora Endl. Pritzel XI, 1900 [U] PI. XV, 4 Sumbavia type; suboblate. 43,5 E 52 P:E i n 0,84. P.A.I. 0,3. Lumina 1-2 /x. Costae colpi absent. Sumbavia type; suboblate. 32 / 38,5/1 P:E 0,82. P.A.I. 0,3 0,35. Lumina 1-3 /x. Costae transversales. Sumbavia 30,5 type; /j. oblate. E 36 P:E 0,84. m:e < 0,5- P.A.I. 0,15. Colpus transversalis long and narrow. Lumina 1-2 fi. Chrozophora type Stephanocolporate; suboblate - oblate. Colpus transversalis broad, isodiametric; costae. Colpi short; costae colpi. Intectate reticulate. Reticulum (?); coarse. The lumina of the reticulum are much larger in the Chrozophora type than in any other type of the Sumbavia Besides configuration. this character the colpi are short and the number of colpi is large (at least six). In the subfamily of the Crotonoideae pollen grains with more than three or four are colpi rather rare. the related Only Caperonia type has also six colpi. Chrozophora plicata (Vahl) A. de Juss. Feinbrun etc A. [U] PI. XV, 6 Chrozophora type; (7-) 8 (-9)-colporate; suboblate oblate. 60 n E 80 P:E ft 0,75. m:e ca. 1. P.A.I. > 0,5. Lumina 4-6 fx. Chozophora tinctoria (Linn.) A. de Juss. Faure A. [U] Chrozophora type. 6-colporate.

71 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 71 plate xv. 1. Melanolepis multiglandulosa; 2. Sumbavia rotlleroides; 3. Thyrsanthera suborbiculatus; 4. Monotaxis grandiflora; 5. Cephalomappa malloticarpa; 6. Chrozophora plicata.

72 broad prolate 72 W. PUNT Caperonia type Stephanocolporate or tricolporate; oblate spheroidal Colpus transversalis broad, circular - - elliptic. spheroidal. Colpi long and narrow; costae colpi absent. Intectate; reticulate. Lumina fairly small. Columellae elongated. This resembles the type Chrozophora in its number of type colpi (usually six). The colpi are, however, longer, and the lumina of the reticulum much smaller. The Philyra subtype differs from the Sumbavia type by its large circular colpus transversalis. Taxonomic discussion Mueller (1866) and Bentham (1880) included Philyra in the genus Pax Argythamnia. and K. Hoffmann (1912c, 1931) separated Philyra from this genus and referred it close to Caperonia. Pollen-morphologically the genus is more related to Caperonia than to Argythamnia and its allies. Caperonia Pollen subtype grains stephanocolporate. Caperonia palustris (Linn.) St. Hill. Lanjouw 1044 [U] PI. XVI, 4 Caperonia serrata Presl de Wit 429 [WAG] Caperonia corchoroides Rombouts 358 [U] Muell. Arc. Caperonia subtype; oblate spheroidal. 32 n E 33 n P:E 0,94. m:e ca. 1. P.A.I. 0,3. Lumina ca. 1 fi. Caperonia subtype; spheroidal. E 39 n. P.A.I. 0,5. Caperonia subtype. Philyra subtype Pollen grains tricolporate. Philyra brasiliensis Klotzsch Hassler [K] PI. XVI, 2 Philyra subtype; prolate spheroidal. P 38,5 E 35 fi n P;E 1,10. m: e 0,7. P.A.I. 0,4. Lumina 1-2 ft. Caryodendron type Tricolporate; oblate suboblate. spheroidal- In polar view pollen grains convex triangular. Colpus transversalis broad; no costae present. Colpi narrow; no costae colpi. Tectate, tectum perforatum; psilate. Intra-reticulum but lumina coarse, small. The Caryodendron type is closely related to the Sumbavia type. The tectate pollen grains are, however, and the lumina of the intrareticulum are fairly small. The tectum, moreover, is perforate. The holes in the tectum are very small and only visible with high magnification. Taxonomic discussion Caryodendron is not closely related to the genera in the Sumbavia type. It lacks a stellate indument and petals. Bentham (1880) as well as Pax and K. Hoffmann (1914, 1931) of a thought relationship

73 POLLEN MORPHOLOGY OF THE EUPHORBTACEAE 73 with Alchornea. Caryodendron possesses, however, an annular disc in the female flower, a character never to be found in the genera of the Alchornea type, but in the always present genera of the Sumbavia type. This character is also in present genera of the Bernardia type and On Speranskia the other hand the authors cited above type. stated, that in their opinion Adenophaedra is closely related to Caryodendron. Adenophaedra belongs pollen-morphologically to the Bernardia type (p. 75). The pollen grains of Caryodendron show indeed some resemblance with this type, but the differences are too many for a combination of Adenophaedra and Caryodendron in the same type. Caryodendron orinocense Krukoff 5553 [U] Karst. PI. XVI, 3 Caryodendron type; suboblate. 29 fi E 34,5 n P:E 0,84. m:e 0,4. P.A.I. 0,3. Lumina ca. 1 /i; muri thick. Caryodendron grandifolius Muell. Arc. Kuhlmann et Ducke [U] Caryodendron type; oblate spheroidal. P 31,5 n E 33 p P:E 0,95. P.A.I. 0,2. Nealchornea type Tricolporate; oblate spheroidal. In polar view the pollen grains circular. Colpus transversalis broad; ora large. Colpi short and wide. Intectate; reticulate. Reticulum coarse. The pollen grains have a Polar axis shorter than the Equatorial axis and there is also a reticulum with large lumina. For these reasons the type is placed in the Sumbavia configuration. It differs from the other types by the three short colpi. Taxonomic discussion Nealchornea is not related to any genus discussed in other types of the Sumbavia configuration. Its nearest allies are perhaps Hamilcoa, and Plagiostyles Pimeleodendron (p. 103). Striking corresponding characters with these genera are the short to sessile anthers and very thickened pedicels. Pax and K. Hoffmann (1931) put the genus in the subtribe Cluytiinae with e.g. Trigonostemon, Moultonianthus and Trigonopleura. The latter three genera are all placed in different pollen types and even in different configurations. Nealchornea yapurensis Huber Krukoff 4967 [U] PI. XVI, 1 Nealchornea type. 27 fi E 30 / P:E 0,90. m:e ca. 1. Lumina 3-4 fi. Bernardia configuration Tricolporate or stephanolcoporate (except Erismanthus type). Usually Polar axis larger than Equatorial axis. Colpi narrow; costae colpi. Tectate or in tecta te; reticulate. Reticulum fine; lumina <2 // (Cyttaranthus without reticulum). The types of the Bernardia configuration are finely reticulate and have usually a Polar axis larger than the Equatorial axis. Cyttaranthus

74 74 W. PUNT plate xvi. 1. Nealchornea yapurensis; 2. Philyra brasiliensis; 3. Caryodendron orinocensis; 4. Caperoniapalustris; 5. Pseudagrostistachys africana; 6. Heterocalyx laoticus; 7. Cyttaranthus congolensis; 8. Bernardia pulchella.

75 prolate subprolate. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 75 only is not reticulate, but in other respects the pollen grains of this genus closely resemble the Speranskia type. Although placed in this configuration by their fine reticulum and prolate shape, the Moultonianthus type and the Clutia type differ largely from the Bernardia type and the Speranskia type. Bernardia type Tricolporate; spheroidal prolate - Pollen grains in polar view distinctly three-lobed. Colpus transversalis small; costae. Colpi narrow; costae colpi. P.A.I. small < 0,25. Intectate; finely reticulate; lumina ca. 1 fi. The Bernardia type differs from the Speranskia type by its distinctly three-lobed pollen grains. Taxonomic discussion Mueller originally described Adenophaedra as a section of Bernardia. In later publications he raised it to a separate genus. Bentham (1880) and Pax and K. Hoffmann (1931) maintained the two genera and even placed them in a far-away group to which Alchornea belongs. Pollen-morphological study of the two genera showed a close relation, so it seems better to maintain Adenophaedra in the vicinity of Bernardia. Both genera are restricted to the New World. Bernardia pulchella (Baillon) Muell. Tessman 6078 [U] PI. XVI, 8 Bernardia type; prolate spheroidal. 25,5 E ft 24/i P:E 1,07. m:e ca. 0,5. P.A.I. 0,15. Bernardia Schulz lorentzii Muell. Arc [U] Bernardia type. 29 n E 25 n P:E 1,17. P.A.I. 0,2. Bernardia corensis (Jacq..) Muell. Arg. Bernardia type. Boldingh 665 B. [U] 31 iu E 25 /x P;E 1,24. P.A.I. 0,2. Adenophaedra minor Ducke Herb. Rio [U] Bernardia type. P 28 n E 27 /i P:E 1,04. Colpus transversalis narrow; m:e < 0,5. P.A.I. 0,1-0,15. Speranskia type Tricolporate; spheroidal convex triangular. spheroidal. Pollen grains in polar view Colpus transversalis large; costae present or absent. Colpi narrow; costae colpi. Sometimes costae indistinct. Tectate; intra-reticulum present or rarely absent (Cyttaranthus). This type looks somewhat like the Bernardia type but differs by its polar view and the shape of the colpus transversalis. Cyttaranthus belongs undoubtedly to this type but lacks the fine reticulum. Taxonomic discussion Genera with a Speranskia type are only found in the Old World. They have several characters in common. 1. In the female flowers always a disc present. 2. Anthers usually pendulous; connectives frequently appendiculated.

76 76 w. PUNT 3. In the male flowers an ovarium rudiment is wanting except in some Agrostistachys species. 4. The Asiatic genera have petals and an extrastaminal disc. The African representatives frequently lacks petals, while glands are always present on the receptaculum. 5. Stellate hairs never present. Speranskiapekinensis Pax et K. Hoffmann Bullock 10 Mai [K] PI. XVII, 8 Pseudagrostistachys africana (Muell. Arg.) Pax et K. Hoffm. Moller and Quintus 16 [U] PI. XVI, 5 Cyttaranthus congolensis Leonard Gollens 3541 [K] PI. XVI, 7 Agrostistachys indica Dalz. Stocks [U] Agrostistachys sessilifolia (Kurz) Pax et K. Hoffmann East Ind. Comp [U] Speranskia type; subprolate. 28,5 n E24/t P:E 1,16. m:e 0,5-1. P.A.I. 0,10. Lumina < 1 /s. Speranskia type; subprolate. 33 E /i 26,5 P;E / 1,24. m:e < 0,5. P.A.I. 0,3. Lumina ca. 1 fi. Speranskia type; prolate spheroidal. 30 E n 28,5 /i P.A.I. 0,35. Speranskia type; Not reticulate! P:E 1,05. oblate spheroidal. P 31,5 E fi 33,5 P:E n 0,94. m:e < 0,5. P.A.I. 0,25. Lumina 1-2 ft.margo (thinning ofthe ektexine). Speranskia type; subprolate. 26,5 E /j. 23,5 P:E fi 1,12. P.A.I. 0,2-0,25. Costae transversales. of the ektexine). Margo (thinning Heterocalyx laoticus Gagnep. Kerr [K] PI. XVI, 6 Argomuellera macrophylla Pax de Wit 32 [WAG] PI. XVII, 3 Speranskia type; spheroidal. P E 30 P.A.I. /i. 0,25. Lumina 1-2 /(. Margo (thinning ofthe ektexine). Speranskia type; spheroidal. 25,5 E n 25 P;E /i 1,02. m:e 0,7-0,8. P.A.I. 0,4. Lumina ca. 1 absent. /i. Costae transversales Swynnerton 119 A [BM] Discoglypremna caloneura (Pax) Prain Zenker 2643 [L] 30 fi E 28,5 n P:E 1,08. In other characters as above. Speranskia type; prolate spheroidal. P27 ii E 25,m P:E 1,09. P.A.I. 0,25-0,3. Lumina ca. 1 fi. Moultonianthus type Tricolporate; spheroidal. In polar view the pollen grains circular. Colpus transversalis narrow and small; costae. Colpi narrow and long; costae colpi. Intectate; reticulum fine. The pollen grains of the Moultonianthus type are finely but distinctly reticulate and are therefore placed in the Bernardia configuration. The shape, polar view and colpus transversalis of this type differs greatly from the other types in this configuration. Taxonomic discussion According to v. Steenis Moultonianthus is related (1948) to Erismanthus in several characters. Unfortunately pollen grains of Erismanthus indochinensis do not agree with those of Moultonianthus leem-

77 54 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 77 bruggianus. Perhaps other species of Erismanthus will show more relationship in their pollen-morphology. Moultonianthus leembruggianus (Boerl. et Koord.) v. Steenis Sarawak 464 [L] PI. XVII, 1 Moultonianthus type. P E 23,5 fi. m;e < 0,5. P.A.I. 0,4. Lumina 1-2 n- Erismanthus type Tricolpate; oblate spheroidal. In polar view pollen grains circular. Colpi Rim of the narrow. colpi irregularly ruptured. Tectate; psilate. Columellae short. Pollen grains of Erismanthus indochinensis distinctly differ from the Moultonianthus-type. There are no composite apertures and the grains are tectate. Erdtman (1952) examined Erismanthus oblique and on the did contrary find in composite this apertures specimen. Erismanthus indochinensis Gagnep. D Alleizette [L] PI. XVII, 4 Erismanthus type. P 24 n E 25,5 P:E n 0,93. P.A.I. 0,4. Clutia type Tricolporate; subprolate. In polar view the pollen grains triangular. Colpus transversalis small; costae. Colpi narrow; costae colpi. Margo present (thinning of the ektexine). Tectate; intra-reticulum indistinct. The Clutia type strongly recalls the Phyllanthus pentaphyllus su btype in the subfamily of the Phyllanthoideae. The type differs, however, by the shape of the colpus transversalis. The Clutia type differs from the Bernardia type by its polar view. The Bernardia is type three-lobed, the Clutia type triangular. Taxonomical discussion This African genus stands alone in the Euphorbiaceae. The inflorescence and habit are those of Phyllanthus; but the essential characters are those of the Chrozophorae (Bentham, 1880). Clutia paxii Knauf Crookewit 198 [WAG] PI. XVII, 2 Clutia type. 35 /i E 27 n P:E 1,30. Clutia natalensis Bernh. m:e < 0,5. Clutia type. Wylie Jan [U] Clutia abyssinica Jaub. et Spach Stolz 1876 [U] Clutia rubricaulis Eckl. ex Sond. [ C. alaternoides Linn.] Lanjouw 165 [U] 43 E n 33 P:E n 1,30. Reticulum indistinct. Clutia type. 52 E 40 P:E fi n 1,30. Reticulum indistinct. Clutia type. P fi E 41 fi P:E 1,30. Reticulum indistinct. Mallotus configuration Tricolporate - stephanocolporate (4). Polar axis usually shorter than Equatorial axis. Colpi narrow. Tectate (tectum perforate in Blumeodendron type); usually not reticulate; columellae short, capita indistinct.

78 78 W. PUNT plate xvu. 1. Moultonianthus leembruggianus; Argomuellera macrophylla; 4. Erismanthus indochinensis; spinosa; Mallotusalbus; 6. Cleidionjavanicum; 7. Speranskia pekinensis; Clutia paxii; 9. Trewia nudiflora. Macaranga

79 - - stephanocolporate suboblate. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 79 The most important characters of the Mallotus configuration are the short columellae and the indistinct capita. Besides these important characters the pollen grains are usually not reticulate. Mallotus type, Blumeodendron type, Acalypha type, Alchornea type, Discocleidion type and Mareya type are closely related and differ only in minor points. The Ricinus type and Pera type are connected by the very narrow colpi and nearly visible columellae. They certainly belong to the Mallotus configuration although the shape of the pollen grains is often prolate. The Seidelia type and Leidesia type are related to each especially other, in their exine structure and polar view. The Microdesmis type and Cheilosa type are in placed the Mallotus configuration because of their shape and short columellae. The differences with the other types in this configuration, however, are considerable. Mallotus type Tricolporate - (4); oblate spheroidal - Sometimes spheroidal. Colpus transversalis usually small; costae. Colpi narrow; costae colpi. Tectate; psilate or scabrate. Rarely intra-reticulate (Adriana). The pollen grains are easily distinguished from the other types. The colpi, are short (P.A.I. is at least and 0,4) the shape is usually oblate spheroidal. The columellae are so short, that only with difficulty the tectum can be distinguished from the endexine. Costae colpi distinct. Taxonomic discussion In the system of Mueller (1866) genera of the Mallotus type are all found in his subtribe Euacalypheae. Bentham (1880) placed them in the subtribe Acalypheae and Pax and K. Hoffmann (1914, 1931) in the subtribe Mercurialinae of the tribe In Acalypheae. all three systems the genera are mixed up with those, closely allied to them in the Acalypha Alchornea Blumeodendron Discocleidion type, type, type, type and Mareya type. Corresponding characters are: 1. A disc is wanting in the male as well as in the female flowers. 2. An ovarium rudiment is wanting. 3. Usually numerous stamens are inserted on a convex receptaculum (Macaranga with few stamens has no Most receptaculum). times the receptaculum is glabrous. Sometimes it is provided with glands (some species of Mallotus and Pycnocoma) or hairs ( Wetria). 4. The majority of the in the Old genera grows World; only Avellanita is from Chile. Mallotus subtype Pollen grains psilate.

80 80 w. PUNT Mallotus albus (Roxb.) Mueix. Arg. Thomson [U] PI. XVII, 5 Mallotus claoxyloides (F. v. Mueller) Muell. Arg. Buysman 1874 [U] Mallotus miquelianus (Scheff.) Pax et K. Hoffm. Rutten-Kooistra 27 [U] Mallotus subtype; oblate spheroidal. 30 E n 30,5 n P:E 0,98. m:e < 0,5. P.A.I. 0,4. Mallotus subtype. 30 E n 32,5 P:E /u 0,95. P.A.I. 0,4. Mallotus subtype. 20 // E 21 /u P:E 0,95. P.A.I. 0,4. Mallotus repandus (Willd.) Muell. Arg. Mallotus subtype. Too young. Griffith 4760 [U] Coelodiscus lanceolatus D Alleizette Jan. Trewia nudiflora Ltnn. Gagnep [L] Hohenacker 453 [U] PI. XVII, 9 Avellanita bustillosii Philippi Philippi [BM] Cordemoya integrifolia (Willd.) Baillon Gover. Pires to Kew Nov [K] Coccoceras borneënse J. J. Smith F. H. Endert 2078 [L] Deuteromallotus acuminatus (Baillon) Pax et K. Hoffm. D Alleizette Mad. Nov [L] Wetria macrophylla (Blume) J. J. Smith Cult. Hort. Bog. IXA. 124 [U] Adriana quadripartita (Labile.) Gaudich. F. v. Mueller, Freemantle 1861 [U] Cleidion javanicum Blume Cult. Hort. Bog. IX A. 38 [U] PI. XVII, 6 Cleidion angustifolium Pax et K. Hoffm. Le Rat 468 [P] Macaranga spinosa Muell. Arg. Steiner 1292 [U] PI. XVII, 7 Macaranga harveyana Muell. Arg. Yuncker [U] Macaranga densiflora Wars. Gjellerup 236 [U] (Muell. Arg.) Mallotus subtype. P 25,5 /I E 27 /i P:E 0,94. P.A.I. 0,4. Mallotus subtype. 27/r E 30,5 n P:E 0,89. Colpi short. very P.A.I. > 0,75. Mallotus subtype. 25,5 E n 26 P:E n 0,98. Costae colpi indistinct. P.A.I. 0,45. Mallotus subtupe. 20 E n 21,5 P:E 0,93. P.A.I. 0,7. Mallotus subtype; 3-4 colporate. 37,5 n E 39 P: E 0,96. Colpi very short. P.A.I. 0,7. Mallotus subtype. 20 i u E 22,5 P:E fi 0,88. Colpi very short. P.A.I. > 0,7. Mallotus subtype; 4- (3) colporate. E 40 P.A.I. n 0,5. Mallotus subtype. P 33,5 ft 35// P;E 0,96. P.A.I. > 0,5. Mallotus subtype. Intra-reticulate. P 25,5 n E 27 n P:E 0,95. P.A.I. > 0,7. Mallotus subtype. 19 fi E 18 /< P:E 0,95. P.A.I. > 0,7. Mallotus subtype. E 14,5 P.A.I. > 0,5. Mallotus type. fi P:E ca n E 17 n P:E ca. 1. Mallotus subtype; E 17 p. Pycnocoma Pollen subtype grains scabrate. This subtype differs from the Mallotus subtype by the scabrae on the tectum. In Pycnocoma other pollen grains cornuta the colpi are longer than those in of the Mallotus type. macrophylla, however, are as short as the others. The colpi of Pycnocoma Pycnocoma cornuta Muell. Arg. Pycnocoma subtype. Leeuwenberg 3003 [U] PI. XVI11, 1 33,5 E p 35,5 p P;E 0,94. m:e 0,4. P.A.I. 0,3.

81 . POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 81 Pycnocoma macrophylla Bentham Zenker 1251 [L] Pycnocoma 30,5 fi subtype. P.A.I. > 0,5. E 32 fi P:E 0,95. Blumeodendron type Tricolporate; oblate spheroidal. Colpus transversalis small and narrow; costae. Colpi very short and costae narrow; colpi. Tectum perforatum that may be regarded as a coarse reticulum. This type is closely related with the Mallotus type. The tectum, however, is perforated. The holes are rather wide and it is perhaps better to speak of a coarse reticulum. The muri of the reticulum are thick and consist of many small columellae. Taxonomic discussion The affinities of Blumeodendron with Botryophora are extensively explained by Airy-Shaw (I960). Without doubt Ptychopyxis has several characters in common with Blumeodendron and Botryophora. 1. In the male flowers a disc and an ovarium rudiment are absent. Many stamens inserted on a receptaculum. 2. In the female flowers a disc is present. (Airy-Shaw 1960, states, that in Ptychopyxis at least some species certainly have a disc.) Blumeodendron tokbrai (Blume) Kurz Achmad 216 [L] PI. XVIII, 6 Ptychopyxis costata Miquel Kings collector 6740 [L] Ptychopyxis kingii Ridley de Haviland et Hose 3673K [K] Botryophora geniculata (Miq,.) Beumee ex Airy-Shaw Docters v. Leeuwen 7707 [L] Blumeodendron type. 21/t E 23 P:E 0,91. Blumeodendron type. 18,5 n E 21 P:E n 0,88. Blumeodendron type. 19,5 /j. E 21 /i P:E 0,83. Blumeodendron type. 22 E n 23,5 P:E n 0,93. P.A.I. 0,5-0,6. Colpi somewhat longer. Acalypha type Tricolporate or stephanocolporate; oblate to suboblate. spheroidal The pollen grains are quasiporate. With high magnification the are apertures undoubtedly composite apertures. Colpus transversalis and colpus of the same size; costae. Tectate; psilate. The Acalypha is type undoubtedly related to the Mallotus.... type. The dimensions of the colpi and the colpi transversales, however, are so small that the composite apertures resemble pori. Acalypha indica Linn. Dewan, [U] PI. XVIII, 4 Acalypha aronioides Ellenberg 1064 Pax et K. Hoffm. [U] Acalypha racemosa Wallich ex Baillon Versuchsanst. 410 [U] Acalypha Muell. Arg. platyphylla Killip et Garcia [U] Acalypha scandens Bentham Y. Mexia 6386 [U] PI. XVIII, 5 Acalypha diversifolia Jacq. Versteeg 466 [U] Acalypha type- 14,5 E 16 /t [ji P:E 0,90. Acalypha type; 4 5 (col)-porate. P" 14 E pc 16,5 P:E f i 0,85. Acalypha type; 3 (col)-porate. P' 12 E n 13,5 P:E /x 0,89. Acalypha type; 3 (col)-porate. 13,5 /i E 14,5 fi P:E 0,93. Acalypha type; 4 5 (col) porate. P' 17,5 E 19,5 n P:E 0,90. Acalypha type. Too young.

82 82 W. PUNT Alchornea type Tricolporate; spheroidal to suboblate. Colpus transversalis with costae. Colpi narrow; costae colpi. Operculum present. Tectate; psilate. In shape and structure the Alchornea is related type to the Mallotus type. The pollen grains differ from the latter type by the presence of an operculum. Taxonomic discussion Genera of this type have much resemblance with those of the Mallotus type. In the systems of Mueller (1866), Bentham (1880) and Pax and K. Hoffmann (1931) they are mixed up Indeed the characters of this group correspond with them. well with those described in the Mallotus type (p. 79). There are only a few indistinct differences. 1. Number of stamens is often less than 20. Neotrewia, Gavarretia and Polyandra have up to 60 stamens (In the Mallotus type up to several hundreds). 2. Receptaculum usually flat and peltate (not in species with many stamens). As in the Mallotus type glands (Alchornea and Neotrewia ). are absent or scarce Coelebogyne and Lepidoturus are genera, which nowadays are considered as belonging to Alchornea. Originally Croizat placed Adenophaedra prealtum (Croizat) Croizat in the genus Cleidion (1943c). This genus has its widest spread in Asia and was said to have some representatives in America. Afterwards Croizat s opinion was that all American Cleidion species belong to Adenophaedra (1946a). Pollen-morphological investigation shows, that pollen grains of the species Adenophaedra prealtum on no account resemble those of Adenophaedra minor Ducke. The diagnosis of A deno phaedra prealtum however,, completely fits into the new genus Polyandra described by Leal (1951). Leal moreover published a photograph of the pollen grains of this genus, from which it is clear that the pollen grains belong to the Alchornea like those of type Adenophaedra prealtum. Alchornea villosa (Benth.) Muell. Arg. Bosch Proefst. F [U] PI. XVIII, Alchornea rugosa (Lour.) Elmer [U] Muell. Arg. Alchornea type; oblate spheroidal E 26,5 P:E 0,94. m:e 0,6. P.A.I. 0,4. Alchornea type. P24 J u E26,5 n P:E 0,92. P.A.I. > 0,5. Alchornea schomburgkii Klotzsch Hilleris-Lambers 1922 Alchornea triplinervia (Spreng.) Muell. Arg. Regnell III a [U] Lepidoturus laxiflorus Benth. [ Alchornea laxiflora (Benth.) Pax et K. Hoffmann] Alchornea type. P 24,5 fi _E 24,5ji Alchornea Alchornea type. type. E 27 /i. P:E 1. 20,5 n E 22 /M P:E 0,93. P.A.I. 0,35-0,4. Faulkner 531 [K]

83 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 83 Coelebogyne ilicifolia J. Smith [ Alchornea ilicifolia (J. Smith) Muell. Arg.] F. v. Muell., Rockhampton 1873 Alchornea type. 20 p E 23 n P:E 0,88. P.A.I. 0,4-0,45. [K] PI. XVIII, 7 Neotrewia cumingii (Muell. Arg.) Pax et K. Hoffm. Elmer [U] Gavarretia terminalis Baill. Ducke [U] Conceveiba simulata Steyermark Krukoff 6643 [U] Conceveibastrum martianum (Baill.) Pax et K. Hoffm. I.B.V [U] Veconcibea latifolia (Benth.) Pax et K. Hoffm. [ Conceveiba latifolia Bentham] Spruce 2826 [BM] Aparisthmium cordatum (Juss.) Baill. Y. Mexia 6273 [U] PI. XVIII, 3 Alchornea type. 21 /x E 22,5 fi P:E 0,93. P.A.I. 0,4. Operculum only one row of pila. Alchornea type. 27, 5 fi P.A.I. 0,3. Alchornea type. consists of E 30,5 P;E 0,91. 19,5 n E 22,5 n P:E 0,88. P.A.I. 0,4. Scabrate. Alchornea type. 21,5 /i P.A.I. 0,4-0,45. Alchornea type. E 22,5 fi P:E 0, /u E27 l x P:E 0,92. P.A.I. 0,4. Alchornea type. 24 E 30 / P:E 0,80. P.A.I. 0,4. Lautembergia multispicata Baill. D Alleizette Madag. Nov [L] Alchornea type. 21,5 /x E 25 n P:E 0,86. P.A.I. 0,35. Lautembergia coriacea (Baill.) Pax Alchornea type; sometimes 4 colporate. D Alleizette Madag. Oct [L] Bocquillonia sessiliflora Balansa 274 [L] Baill. Alchornea type. 21,5 n E 24,5 n P:E 0,88. P.A.I. 0,35. Adenophaedraprealtum (Croizat) Croizat [ Polyandra L al?] Krukoff 6602 [U] Alchornea type. 25,5 n E 29 n P:E 0,88. P.A.I. 0,45. Discocleidion type Tricolporate; spheroidal to suboblate. Colpus transversalis; costae as broad as the colpi. Colpi narrow; broad margo present (thickening of the ektexine). P.A.I. small, smaller than 0,2. Tectate; from psilate to coarsely warted. The ornamentation differs strongly. The pollen type shows some resemblance to the Mareya type as well as to the Alchornea type. It differs from both types by its curious ornamentation. This ornamentation varies in the individual pollen It grains. is not certain whether other specimens of Discocleidion rufescens will show the same variation of ornamentation. The type differs from the Mareya type by the presence of costae and from the Alchornea type by the absence of an operculum. Taxonomic discussion Discocleidion differs from the genera in the Alchornea type by the presence of a disc in the female flowers and many glands inserted on thereceptaculum of the male flowers. These characters place the genus next to the genera of the Mareya type. Mueller first described Discocleidion as a section of Cleidion. Pax

84 84 W. PUNT plate xvra. 1. Pycnocoma cornuta; 2. Alchornea villosa; 3. Aparisthmium cordatum; Acalypha indica; Acalypha scandens; 8. Mareya brevipes; 9. Discocleidion rufescens; 6. Blumeodendron 10. tokbrai; 7. Coelebogyne ilicifolia; Kunstlerodendron sublanceolatum.

85 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 85 and K. Hoffmann (1931) thought it better to raise the section to genus level which in they place the Bernardiiformes close to Bernardia and Mareya with their allies. Hurusawa (1954) at last reduced the genus again to a section, this time of Alchornea. Discocleidion rufescens (Franch.) K. Hoffmann Pax et Discocleidion type. 24 E27 n P:E 0,89. D Alleizette China Mai [L] m:e 0,7. P.A.I. 0,1-0,15. PI. XVIII, 9 Mareya type Tricolporate; oblate spheroidal to prolate spheroidal. In polar view pollen grains convex triangular. Colpus transversalis elongated; no costae. Colpi narrow; no costae colpi present. Tectate; psilate. The pollen grains are small. The largest axis is smaller than 30 fx. The Mareya type is distinguished from the Mallotus type by the absence of costae. Taxonomic discussion In Pax and K. Hoffmann s system (1914, 1931) genera of this type belong to the Bernardiiformes. By the following characters the genera differ from those in the Mallotus type and Alchornea type: 1. In the female flowers a disc is present. 2. The male flowers have a receptaculum often set with glands. A disc and an ovarium rudiment are wanting. (Mareyopsis is anomalous by the presence of extra-staminal disc glands and the absence of glands in the receptaculum.) 3. Anthers often have appendiculate connectives. 4. Plants usually grow in the Old World. (,Alchorneopsis in the New World.) The American genus Alchorneopsis differs from the above characters by having an ovarium rudiment and the presence of hairs on the receptaculum instead of glands. Mareya bervipes Pax Zenker 1794 [L] PI. XVIII, 8 Mareya type; oblate spheroidal. 21 v E 23 P:E 0,91. m:e < 0,5. P.A.I. 0,3. Colpus membrane granulate. Crotonogynopsis usambarica Eggeling 3826 [K] Pax Mareya type. Too young to give reliable measures. Chondrostylis bancana Boerl. Cult. Hort. Bogor. II. I 8. [L] Mareya type; spheroidal, p E 25 ji. P.A.I. 0,3. Kunstlerodendron sublanceolatum Ridl. Mareya type; oblate spheroidal. Sinclair [L] PI. XVIII, 10 20,5 n E 21 // P:E 0,98. P.A.I. 0,3. Mareyopsis longifolia (Pax) Pax et Mareya type; oblate spheroidal. K. Hoffmann 18 /i E 19 n P:E 0,96. Zenker 4228 [K] Neopalissya castaneifolia (Baill.) Pax PerviI [P] P.A I. 0,2. Mareya type; prolate spheroidal. 22 E 21 P:E n n 1,05. P.A.I. 0,25-0,3.

86 86 W. PUNT Necepsia afzelii Prain Leeuwenberg 2211 [U] Amyrea humberti Leandri D Alleizette Dec [L] Mareya type; prolate spheroidal. 21,5 /i E21 ft P:E 1,02. P.A.I. 0,3. Mareya type; prolate spheroidal. P 24 /< E21,5yu P;E 1,13. P.A.I. 0,3. Alchorneopsisfloribunda (Benth.) Muell. Mareya type; oblate spheroidal. 18 n E 19 P:E 0,95. For. Dept [U] Alchorneopsis trimera Lanj. Lanjouw 1926; 49 [U] P.A.I. 0,2. Mareya type; oblate spheroidal. 21/t E 22 P:E fi 0,96. P.A.I. 0,25. Ricinus type differs from Tricolporate; shape suboblate to subprolate. Colpus transversalis narrow; no costae. Colpi very narrow; no costae colpi present. Tectate; psilate. Columellae very short, nearly visible in some species. The pollen grains resemble the Pera type, but the Ricinus type misses costae as in the Mareya type. It differs from the latter type by narrow colpi and nearly visible columellae. Taxonomic discussion The genera in this type fall apart in two groups. The first group includes Ricinus, Homonoia and Lasiococca. These genera are remarkable for the large number of anthers crowded on repeatedly branched filaments. The second group comprises Koilodepas only, which genus occupies an isolated position in the family (Airy-Shaw, 1960). Croizat (1942b) as well as Airy-Shaw reduced Nephrostylus Gagnepain and Calpigyne Blume to Koilodepas. According to Croizat, Calpigyne hainanense and Nephrostylus poilanei are synonyms. In the opinion of Airy-Shaw, Nephrostylus poilanei belongs to the species Koilodepas longifolium. Pollen grains of both species are practically identical and show great structural resemblance with Koilodepas bantamense. Ricinus communis Linn. Heerdt 531 v. [U] Ricinus type; spheroidal. 32 p E 32 p P:E ca. 1. Versteeg 243 [U] PI. XIX, 3 Lasiococca symphylliifolia (Kurz) J. D. Hook. Hort. Bogor. Cult [L] Homonoia javensis (Blume) Muell. Arg. m: e < 0,5. P.A.I. 0,2. Ricinus type; spheroidal, p E 23 fi. P.A.I. 0,45. Ricinus type; prolate spheroidal. Hort. Bogor. IX A. 60 [U] Homonoia riparia Lour. Heifer 4746 [U] PI. XIX, 2 P 22,5 n E 2l/x P:E 1,08. P.A.I. 0,5. Ricinus type; Costae transversales? oblate spheroidal-suboblate. V 2bn E 29 P:E 0,88. Koilodepas bantamense Hassk. Ricinus type; spheroidal-suboblate. Hort. Bogor. IX C. 36 [U] Pi. XIX, 4 20 fi E 24 n P:E 0,84. m:e < 0,5. P.A.I. 0,3. Koilodepas hainanense (Merr.) Croizat [ Calpigyne hainanense Merr.] S. K. Lau 3542 [P] [? Nephrostylus poilanei Gagnep. D Alleizette, Annam [L] Ricinus type; subprolate. 21 /x E 17,5 ft P:E 1,18. PA.I ,5 /< E 17,5 P:E /x 1,17. P.A.I. 0,3.

87 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 87 plate xix. 1. Leidesia procumbens; 2. Homonoia javensis; Ricinus communis; Koilodepas bantamense; Chaetocarpus castanocarpus; Pera bicolor; Seidelia triandra; Microdesmis casearifolia.

88 88 w. PUNT Pera type Tricolporate; prolate spheroidal. Colpus transversalis small; costae. Colpi narrow; costae colpi narrow but distinct. Tectate; psilate. Columellae very short. The Pera type has the same very narrow colpi and nearly visible columellae as the Ricinus type, but there are distinct costae colpi and costae transversales present. Pera bicolor (Klotzsch) Mueix. Arg. Ule 8409 [U] PI. XIX, 6 Pera glabrata (Schott) Baill. Reitz 4548 [U] Pera type. P 25,5 E 24 P:E n yu 1,07. m;e < 0,5. P.A.I. 0,3. Pera type. 19 n E 18 n P:E 1,06. P.A.I. 0,5. Seidelia type Tricolporate; prolate spheroidal. In polar view the pollen grains distinctiy three-lobed. Colpus transversalis small; costae. Colpi narrow; no costae colpi. Tectate; psilate. Columellae short. The pollen grains of this type are distinctly three-lobed and have costae transversales. Seidelia triandra (E. Meyer) Pax MacGregor Mus. 1254[K] PI. XIX, 7 Seidelia type. 17,5 n E 16,5 /s P:E 1,06. m:e 0,6. P.A.I. 0,25. Leidesia type Tricolporate; oblate spheroidal. In polar view the pollen grains distinctly three-lobed. Colpus transversalis narrow and long; no costae. Colpi narrow; no costae colpi. Narrow operculum. Tectate; psilate. Columellae short. The Leidesia type resembles the Seidelia type in its polar view and exine structure. It differs, however, by the absence of costae and the addition of a narrow operculum. Taxonomic discussion Leidesia and Seidelia are Mercurialis-like plants which Bentham (1880) as well as Pax and K. Hoffmann (1914, 1931) enumerated close together with Mercurialis. This relationship, however, is not confirmed by pollen-morphological results. Leidesia procumbens (Linn.) Prain Leidesia type. Splitgerber [L] PI. XIX, 1 19 E ft 20,5 n P:E 0,92. m:e < 0,5. P.A.I. 0,3. Microdesmis type Tricolporate; spheroidal to oblate speroidal. In polar view pollen grains convex triangular. Colpus transversalis isodiametric; m:e 1; ora large. Colpi narrow; costae colpi broad and indistinct or absent. Tectate; psilate. Columellae short.

89 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 89 The Microdesmis type is characterised by the isodiametric colpus transversalis and the large ora. Taxonomic discussion Bentham (1880) discussed the genera of this in two type tribes, Galearieae and Gelonieae. Microdesmis, Galearia and Pogonophora occurring in the Galearieae have characters both of the Phyllanthoideae and Crotonoideae. Pollen grains of these genera are, however, not at all comparable with those in the Phyllanthoideae. Pierre s description being unknown to Pax (1899), he described Centroplacus glaucinus Pierre in Microdesmis as Microdesmis paniculata. Pax, however, did not have female flowers at his disposal. Pierre (1899) and Gilg (1908), on the other hand, studied female material and found 2 ovules per locule, for which reason the had to be plant placed in the subfamily of the Phyllanthoideae. The structure of the male flowers, however, resembles that of Microdesmis to such an extent, that it is by no means strange that Pax reckoned this species to Microdesmis. Since it is evident that the pollen grains of Centroplacus cannot be distinguished from those of Microdesmis, the correctness of placing the genus in the subtribe Antidesminae seems (Pax 1931) Two questionable. conclusions are possible: 1. Centroplacus forms a link between the sub-family Phyllanthoideae and Crotonoideae. On one side we find the characters that suggest a close relationship to Microdesmis, while the Phyllanthoideaecharacter of the two ovula per locule in the ovary stands on the other side. 2. Centroplacus has been described on the basis of material belonging to two species. The plant is dioecious, so that the possibility of an erroneous combination of plants should not be overlooked. A renewed accurate examination of the known material, will possibly remove the doubt. Microdesmis subtype Pollen grains small. Smaller than 20 fi. Costae colpi absent. casearifolia Planch. Microdesmis Thorenaar T. 50 [U] PI. XIX, 8 Microdesmis zenkeri Pax de Wit 4 [WAG] Centroplacus glaucinus Pierre Tester 1922 [K] Zenker 1761 [K] Galearia dognaiensis Pierre ex Gagnep. Clemens 3528 [U] Galearia filiformis (Blume) Pax Blume, Java [U] Microdesmis subtype; oblate spheroidal. 15,5 E // 17,5 P:E ft 0,89. P.A.I. 0,3. Microdesmis subtype. 18 E 19 P:E n // 0,95. P.A.I. 0,5. Microdesmis subtype. 14 E 15,5 n P:E 0,90. P.A.I. 0,3. Microdesmis subtype; suboblate. 14,5 n E 17,5 P:E /u 0,83. P.A.I. 0,35. Microdesmis subtype. P 16,5 E 19 P:E ft 0,89. fi P.A.I. 0,3.

90 90 w. PUNT ' Chaetocarpus subtype Pollen grains large. Larger than 30 //. Costae colpi broad but indistinct. Chaetocarpus castanocarpus (Roxb.) Th WAITES Elmer, Borneo [U] PI. XIX, 5 Chaetocarpus schomburgkianus (O. Kuntze) Pax et K. Hoffm. For. Dept [U] Trigonopleura malayana J. D. Hook. Bosch Proefst. 83 E. IP [L] PI. XX, 1 Pogonophora schomburgkiana Miers B. W [U] Chaetocarpus subtype; spheroidal. PE29/< P.A.I. 0,3. Chaetocarpus subtype. Chaetocarpus subtype; m:e ca. 1. oblate spheroidal. 35 E 38 P:E n n 0,92. P.A.I. 0,2. Chaetocarpus subtype; oblate spheroidal. 42 fi E 45 n P:E 0,93. Krukoff 8645 [U] P.A.I. 0,4 0,45. Tectum perforatum. Cheilosa type Tricolporate; suboblate to oblate spheroidal. In polar view pollen grains convex triangular. Colpus transversalis; costae. Colpi narrow; indistinct costae colpi. Tectate; echinate or micro-echinate. Columellae short. The Cheilosa type possesses the most important characters of the Mallotus configuration and is therefore placed in this configuration. The type differs, however, distinctly by the spines fn Cheilosa these spines are larger than 1 ' pt (echinate); on the tectum. ' in Neoscortechinia smaller than 1 /j.. According to the nomenclature of Faegri and Iversen the pollen grains of Neoscortechinia are scabrate. The author would like to call the small spines microechinate according to Erdtman (1952). Echinate pollen grains are rarely found in the subfamily of the Crotonoideae, only in the Croton configuration. Taxonomic discussion Bentham (1880) discussed Cheilosa in the Acalypheae, although the male calyx is divided into slightly imbricate lobes. In the system of Pax and K. Hoffmann (1931) the genus is found in the tribe Gelonieae, to which also Neoscortechinia belongs. The pollen grains of both genera are undoubtedly related. Unfortunately this is not confirmed by other characters. Cheilosa montana Blume Cheilosa type; suboblate. Korthals, Java [L] PI. XX, 9 22,5 n E 29/ P:E 0,78. m;e 0,5. P.A.I. 0,5-0,6. Echinate; echinae 1-2 ft. Neoscortechinia arborea K. Hoffm. (Elmer) Pax et Cheilosa type; oblate spheroidal. 21y«E 22,5 P:E n 0,93. Hort. Bogor. IX A. 6a. [U] PI. XX, 8 P.A.I. 0,3. Scabrate (microechinate). Claoxylon configuration Tricolporate or stephanocolporate (4, rarely 5). Colpi narrow. Tectate; capita of the columellae distinct. Pollen grains not reticulate.

91 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 91 Trigonopleura malayana; 2. plate xx. 1. Athroandra mannii; 4. Claoxylon 5. cuneatum; Cladogynos orientalis; xiphoclada; 8. Neoscortechinia arborea; 6. Cephalocrotonopsis socotrana; 7. Amperea 9. Cheilosa 3. Mercurialis annua; montana.

92 92 W. PUNT E 21 The Claoxylon configuration is related to the Mallotus configuration. The capita of the columellae, however, are distinctly visible. Claoxylon type Tricolporate or stephanocolporate; oblate spheroidal-prolate spheroidal. Colpus transversalis small; costae but sometimes indistinct. present Colpi narrow; costae colpi absent or indistinct. Tectate; psilate or scabrate. Columellae coarse, capita distinct. This type is in some characters comparable with types of the Mallotus configuration. The small spheroidal pollen grains resemble, in their shape, those of the As in the Mallotus Mareya type (p. 85). type (p. 79) more than three composite apertures can be present. Placing the type in the Bernardia configuration (p. 77) is out of question as a reticulum is always absent. The Claoxylon ' is type distinguished by the distinct capita of the columellae, which form a coarse and inordinate pattern. Taxonomic discussion Genera belonging to this type are all readily recognised by the shape and attachment of the anthers. Other corresponding characters are: 1. A disc present in the female flowers. 2. In the male flowers an ovarium rudiment is always absent. Usually an extrastaminal disc is also absent, but glands are inserted on the receptaculum. In Discoclaoxylon receptaculum glands fails, but in this genus an extrastaminal disc is present. 3. Plants are usually dioecious and grow only in the Old World. Erythrococca stolziana Pax et K. Hoffm. Stolz 1556 [U] Claoxylon type; spheroidal. E 22 n- m:e< 0,5. Athroandra mannii (J. D. Hook.) K. Hoffmann Pax et P.A.I. > 0,5. Claoxylon. type; spheroidal-suboblate. oblate 5 ( 4) oblate colporate. [ Erythrococca mannii (J. D. Hook.) Prain] Mann 260 [U] PI. XX 2 ; Athroandra africana (Baill.) Pax et K. Hoffmann [ Erythrococca africana (Baill.) Prain] Zenker 453 [U] Claoxylon cuneatum J. J. Smith Versteeg 1752 [U] PI. XX, 4 Claoxylon tumidum J. J. Smith Versteeg 1708 [U] 30 n E 33,5 n P:E 0,89. Claoxylon type; prolate spheroidal. 22,5 E21 P:E ft ft 1,07. Claoxylon type; spheroidal. E fi. Claoxylon type; spheroidal. 3 (4) colporate. E 19 /t. P.A.I. 0,4. Psilate to scabrate. Micrococca mercurialis (Linn.) Benth. D Alleizette Mad [L] Claoxylon type; prolate spheroidal. 3- (4) colporate. 27 n E 25 n P:E 1,08. P.A.I. 0,3. Discoclaoxylon hexandrum (Muell. Arg.) Pax et K. Hoffm. Claoxylon type; spheroidal. P - 21 n- P.A.I. 0,3. Zenker 35 [U]

93 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 93 Mercurialis type Tricolporate; subprolate to prolate. Colpus transversalis small; costae as long as the costae colpi. Colpi narrow; of only costae colpi. Operculum present one row of columellae. Tectate; psilate. but very narrow, consisting Columellae distinct. There is a certain relation between the Mercurialis type and the Claoxylon type. The two types have distinct capita; both are not reticulate and possess small colpi transversales. The Mercurialis differs by the presence of an operculum, which is type narrow and indistinct, and by the distinctly prolate shape of the pollen grains. Taxonomic discussion Shape and attachement of the anthers show much similarity with the of genera the For Claoxylon group. that reason Bentham (1880) placed Mercurialis in the immediate vicinity of Claoxylon, Erythrococca, etc. Although the Mercurialis differs somewhat from the type Cloxylon type, there is Bentham. enough affinity for the author to share the opinion of Mercurialis annua Linn. Fresh material, Z.-Limb. PI. XX, 3 Mercurialis reverchonii Faure [U] Rouy Mercurialis tomentosa Linn. Leeuwenberg 1317 [U] Mercurialis type; subprolate. P 25,5 y E 22,5 y P:E 1,16. m:e < 0,5. P.A.I. 0,4. Operculum indistinct, consisting of some scattered pila on the endexine (membrana granulata). Mercurialis type. 28,5 E /j. 22 P:E /j. 1,30. Mercurialis type. 40 E n 28,8 /i P:E 1,40. Cladogynos configuration Tricolporate. Colpi costae narrow; colpi. Tectate; columellae distinct. Exine thick. The Cladogynos configuration includes those pollen grains with distinct columellae that bear always a tectum or tectum perforatum. These columellae can be arranged in a reticulum but more frequently they do not form a regular The pattern. pollen grains are usually large; the largest axis is at least 25 and /i often up to 50 (x. Cladogynos type Tricolporate; spheroidal prolate (Epiprinus oblate spheroidal). Colpus transversalis; costae. Colpi narrow; costae colpi. Tectate or tectum perforatum. Columellae distinct. Exine thick. Some species intra-reticulate. The most striking character of this type is the distinct, sometimes even tall, columellae. The pollen grains are large and generally they have a P:E larger than 1.

94 94 W. PUN'I Taxonomic discussion By Pax and K. Hoffmann (1914, 1931) most genera of this type are placed in the series Cladogyniformes of their subtribe Mercurialinae, to which also Mallotus and Alchornea The belong. group can be distinguished as follows: 1. An ovarium rudiment present in the male flowers while a disc is absent. The few stamens are not inserted on a receptacle. 2. Male flowers usually in a capitule. 3. Anthers pendulous. 4. Stellate hairs always present. 5. Plants are monoecious and grow in the Old World. Cladogynos orientalis Zipp ex Spanoghe Pierre 6213 [L] PI. XX, 5 Cladogynos type; prolate spheroidal. 42/i E 40 /r P:E 1,05. Colpus transversalis narrow. Tectum perforatum. P.A.I. 0,3. Intra-reticulate; lumina 1-2 /s. Cephalocroton cordofanus Hochst. D Alleizette Nubie, Aug [L] Cladogynos type; subprolate. 49 n E 42 p P;E 1,17. Colpus transversalis broad; m:e >0,5. P.A.I. 0,3. Intra-reticulate. Cephalocrotonopsis socotrana (Bale, f.) Pax Schweinfurth 594 [K] PI. XX, 6 Symphyllia siletiana Baill. Hort. Bog. Cult. [L] Adenochlaena leucocephala Baill. Hildebrandt 3258b [L] Cladogynos type; prolate spheroidal. 46 E 42,5 fi P:E 1,09. P.A.I. 0,25. Intra-reticulate. Cladogynos type; prolate spheroidal. 27 ii E 25 n P;E 1,07. P.A.I. 0,25-0,3. Not reticulate. Cladogynos type; prolate spheroidal. 53 /u E48/< P:E 1,10. m:e 0,4. P.A.I. 0,3. Not reticulate. Epiprinns poilanei Gagnep. Clemens 3058 [U] Cladogynos type; oblate spheroidal. 32 ft E 35 n P:E P.A.I. 0,2-0,25. Not reticulate. 0,91. Amperea type Dicolporate or tricolporate. Shape differs with the number of colpi from oblate spheroidal (dicolporate) to prolate spheroidal (tricolporate). Colpus transversalis narrow; costae. Colpi narrow and long; costae colpi. P.A.I. small to 0. Tectate; psilate. Exine thick. Intrareticulate or not reticulate. The Amperea-type shows much similarity with the Cladogynos t >T )e - Especially the thick exine and the tall columellae are striking. The T-. P.A.I., however, is much smaller. Some pollen grains are even syncolpate. Taxonomic discussion Amperea belongs to those Australian genera, that were kept separate on account of their narrow cotyles (Stenolobae). Should this group as such be abolished, then, according to the statement of Pax (1924), this genus has to be placed in the Mercurialinae. This group also comprises the genera of the Cladogynos type.

95 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 95 Amperea xiphoclada (Sieb. et Spreno.) Druce Constable [U] PI. XX, 7 Amperea type. 2- P:E 0, P:E 1, E ft fi. P.A.I. 0,15-0. m:e < 0,5. Not reticulate. Amperea spartioides Brongn. Tasmania [U] Amperea type; spheroidal. E 26 ft. Intra-reticulate. Pseudocroton type Tricolporate; prolate spheroidal to subprolate. Colpus transversalis; costae? Colpi costae narrow; col pi? Tectate; psilate. Not reticulate. The pollen grains of the sheet examined were not mature, so it is difficult to put them in the proper type. There is some resemblance with the Speranskia type (p. 75). A reticulum, however, is absent. As distinct columellae are present, it seems best to place this type in the Cladogynos configuration. Taxonomic discussion Bentham (1880) as well as Pax et K. Hoffmann (1931) discuss Pseudocroton in the Chrozophoreae. They accept relationship with Argythamnia and its allies. The pollen grains have some affinity with the Speranskia type. Although the genera of this type to belong the Chrozophoreae, they are only found in the Old World, while Pseudocroton grows in Central America. Affinities with genera of the Cladogynos configuration are few. Pseudocroton tinctorius Muell. Arg. O. Kuntze 1808 [K] Pseudocroton type. Too young to give reliable measures. Hippomane configuration Tricolporate; pollen grains in polar view circular or three-lobed. Never convex triangular or triangular. Colpus transversalis frequently short; sometimes narrow elongated. Colpi narrow; usually long. A margo is nearly always present. Intectate or more often tectate. Columellae distinct, frequendy large. It is not possible to define this configuration with one or two characters. exceptionless The pollen grains are, however, easily recognised by a combination of several typical characters. The most important ones are: three-lobed polar view; margo; distinct, sometimes large, columellae and short colpi transversales. If one of these striking characters is absent the combination of the others still places the pollen grains immediately in this configuration. Taxonomic discussion The genera of this configuration can be divided into two groups. The first group comprises genera of the tribe Hippomaneae of Pax and K. Hoffmann (1912b, 1931). In 1880, Bentham stated of his subtribe

96 96 W. PUNT ~ Hippomaneae: This is one of the most natural subtribes of Crotoneae. Indeed there is not any doubt of a close relationship genera. between the Pachystroma is added to the first group. Evidently on technical grounds Bentham added Pachystroma to the ofthe group Adrianeae with Manihot, Adriana and Cephalocroton, three genera now placed in three different groups of plants. Pax and K. Hoffmann (1919d, 1931) separated the monotypic genus into the tribe Pachystromateae, although they were aware of some affinity with the Hippomaneae. The welldeveloped calyx of the male flower, which is never present in the Hippomaneae, led them to the decision to separate the genus. There are, however, more characters in favour of a fusion than of a separation. Pollen morphology is one of the data in favour of fusing. The second group includes all the genera of the Euphorbieae the addition of Hura. It is with remarkable that the typical Euphorbieae, so well characterised by their cyathium, possess pollen grains which can hardly separated from those of the Hippomaneae. Pax (1924) when discussing the phylogenetic relations of the already to this relationship. He noticed, e.g., the unbranched laticiferous vessels which both taxa possess. Euphorbiaceae, pointed Pollen morphology is another character to support the opinion of Pax. Hippomane type Tricolporate; shape differs from subprolate to suboblate. Pollen grains in polar view circular or three-lobed. Colpus transversalis small (rarely elongated); costae. Colpi narrow and long; costae colpi. Usually margo present. (smaller 0,3) except in some species of Mabea. P.A.I. is small Tectate; psilate. Columellae sometimes tall. The Hippomane type differs from the Euphorbia hirta type by its narrower margo. Usually the margo is narrower than the colpus transversalis. In other respects both types can hardly be separated. Pollen grains of the Omalanthus nutans subtype can be separated from the Hippomane subtype by the deeply three-lobed pollen grains. Some species in the Hippomane type have an intra-reticulum. They do not belong altogether to the Dichostema type (p. 100) as the upper part of the columellaeis fused in all directions, thus forming a tectum. Hippomane subtype Pollen grains in polar view circular or slightly three-lobed. Hippomane mancinella Linn. Buysman 1856 [U] PI. XXI, 3 Omalanthus populneus (Geisel) Pax Docters v. Leeuwen 1716 [U] Hippomane subtype; prolate spheroidal. 32,5 E /j. 30,5 n P:E 1,06. m:e < 0,5. P.A.I. 0,25. Hippomane subtype. 40 fi E34/i P:E 1,17. P.A.I. 0,15-0,20. Mabea piriri Aube. B. W [U] Hippomane subtype. 50 ju E 44 n P:E 1,14. P.A.I. 0,20-0,25. Not reticulate. Mabea taquari Aubl. Pulle 67 [U] Hippomane subtype. 43,5 fi E 41 n P:E 1,05. P.A.I. 0,4. Intra-reticulate.

97 ' POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 97 Mabea caudata Pax et K. Hoffm. v. Emden [U] Hippomane subtype. E 70 p. P.A.I. 0,4-0,3. Not reticulate. Mabea fistulifera Mart. Y. Mexia 4473 [U] Hippomane subtype. E 70 /i. P.A.I. 0,3. Indistinct intra-reticulate. Tectum perforatum. Actinostemon lanceolatus Saldanha Y. Mexia 4987 [U] Gymnanthes lucida Swartz Jamaica, Hohsle [U] Hippomane subtype; oblate spheroidal. 33,5 n E 35 n P:E 0,94. P.A.I. 0,25. Margo absent. indistinct. Costae Hippomane subtype; oblate spheroidal. 25 n E 26 n P:E 0,96. P.A.I. 0,25. Margo absent. Colu- mellae short. Glyphostylus laoticus Gagnep. Lakshnaka 1352 [K] Spirostachys venenifera (Pax) Pax [ Excoecaria venenifera Pax] Graham 1464 [K] PI. XXIII, 3 Excoecaria bicolor Hassk. Hort. Bogor. IX C. 77 [U] PI. XXII, 4 Hippomane subtype. P 36,5 n E 34,5 /x P:E 1,06. P.A.I. 0,25. Colpus transversalis elongated. Columellae tall. Hippomane subtype. 22 n E 23 n P:E 0,96. P.A.I. 0,25. Hippomane subtype. Margo absent. 38 n E 36 ft P:E 1,05. P.A.I. 0,25. Stillingia Pax et K. Hoffm. gymnogyna Wiggins 1351 [U] Sapium longifolium (Muell. Arg.) Huber Pedersen 72 [U] Hippomane subtype. P 35,5 /u E 33 P:E fi 1,08. P.A.I. 0,2. Hippomane subtype. 55 n E 43 n P:E 1,26. P.A.I. 0,25. Colpus transversalis elongated. Costae equatoriales or almost equatorial. Colliguaya integerrima Gill, et Hook. Donat 167 [U] Hippomane subtype. 35,5 fi E 34 P:E n 1,04. P.A.I. 0,2. Margo indistinct; narrow. Omalanthus nutans subtype Pollen grains in polar view distinctly three-lobed. Omalanthus nutans (Forst.) Pax Yuncker [U] PI. XXI, 8 Mabea indorum Sp. Moore Krukoff 1084 [U] Actinostemon concolor (Spreng.) Muell. Omalanthus nutans subtype. 35 n E 31,5 fi P:E 1,11. P.A.I. 0,15. Omalanthus nutans subtype. 45 E n 36,5 n P:E 1,22. P.A.I. 0,2-0,25. Omalanthus nutans subtype. 30 n E 34 /t P:E 0,88. Bornmueller 536 [U] PI. XXII, 3 Gymnanthes jamaicensis (Britten) Urban Harris [U] Sebastianiachamaelea (Linn.) Muell. Arg. Columellae short. Omalanthus nutans subtype. E 22 ft. Margo absent. Omalanthus nutans subtype. Hohenacker 830 [U] 33/< P;E 1,14. P.A.I. 0,2. Sebastiania corniculata(vahl) Frees [U] Muell. Arg. Omalanthus nutans subtype. E 40 n- P.A.I. 0,15-0,2. Macguire and Stahel [U] Sebastiania scottiana Muell. Arg. Omalanthus nutans subtype. Bertoni 1714 [U] E 27 n. P.A.I. 0,2.

98 98 W. PUNT plate xxi. 1. Stenadenium spinescens; 2. Elaeophorbia drupifera; 3. Hippomane mancinella; 4. Dichostema glaucescens; 5. Pachystroma longifolium; Tetraplandra gibbosa; 8. Omalanthus nutans. Senefeldera macrophylla;

99 pollen morphology of the euphorbiaceae 99 Duvigneaudia inopinata (Prain) Leonard Zenker 217 [U] Omalanthus nutans subtype. 23 /j. E 20 /i P:E 1,16. P.A.I. 0,2. Maprounea brasiliensis St.-Hill. Williamson-Assis 8047 [U] Omalanthus nutans subtype. 33,5 E /j. 30,5 P:E n 1,11. P.A.I. 0,2. Maprounea guyanensis Aubl. Omalanthus nutans subtype. B. W [U] 24 n E 20 n P:E 1,20. P.A.I. 0,2. Excoecaria philippinensis Merrill Elmer, Palawan [U] Omalanthus nutans subtype. 32/ E 32/i. P.A.I. 0,2-0,25. Sapium ellipticum (Hochst.) Pax Omalanthus nutans subtype. Croockewit 728 Stolz 621 [U] [WAG] E 30 n. P.A.I. 0,2. 40 E 34 n P:E 1,18. Sapium sebiferum (Linn.) Roxb. Lilnee, Japan [U] Sapium aubletianum (Muell. Arg.) Huber B. W [U] Sapium montanum Lanj. B. W [U] Sapium klotzschianum (Muell. Arg.) Huber Omalanthus nutans subtype. 46 p. Omalanthus nutans subtype. P 46 p. Omalanthus nutans subtype. P 52 p. Omalanthus nutans subtype. 54 p. Slabel, Oct [U] Grimmeodendroneglandulosum (A. Rich.) Urban Curtiss 190 [K] Omalanthus nutans subtype. 28 fi E 26 n P:E 1,08. P.A.I. 0,2. Colpus transversalis Bonania cubana A. Rich. narrow elongated. Omalanthus nutans subtype (?). Howard 5774 [U] Bonania domingensis (Urb.) Urb. 27 E /z 25,5 p P.A.I. 0,2. Omalanthus nutans subtype. P:E 1,06. Fuertes 813 [U] B 22 p E 22 n. P.A.I. 0,25. Columellae small. Adenopeltis colliguaya Bert, ex Juss. Omalanthus nutans subtype. Andreas 834 [U] Dalembertia populifolia Baill. Hinton 4353 [K] 34,5 n E 33,5 n P:E 1,03. P.A.I. 0,15. Omalanthus nutans subtype. Intra-reticulate. 41 E 36 n P:E 1,12. P.A.I. 0,15-0,2. Intra-reticulate. Tetraplandra gibbosa Pax et K. Hoffm. Omalanthus nutans subtype. Glaziou 5613 [K] PI. XXI, 7 30 n E 27 n P:E 1,11. P.A.I. 0,15. Ophthalmoblapton macrophyllum F. Allem. Glaziou [K] Omalanthus nutans subtype. 42 n E 39 /i P:E 1,08. P.A.I. 0,2. Algernonia brasiliensis Baill. H. F. Hance 1887 [K] Omalanthus nutans subtype. Young. Pachystroma type Tricolporate; spheroidal. Pollen grains slightly three-lobed. Colpus transversalis broad elliptic; costae thick. Colpi narrow and short; costae colpi. No margo present. Tectate; psilate. Pollen grains of this type are strongly related to those of the Hippomane The type. colpus transversalis, however, is broadly elliptic and the P.A.I. is much larger.

100 100 w. PUNT Pachystroma longifolium (Nees) I. M. Johnston Krug and Urban 1367 [L] PI. XXI, 5 Pachystroma type. P 41 n E 41 ft P:E ca. 1. Length of colpi differs; mostly short, sometimes long. Costae thick. Euphorbia hirta type Tricolporate; subprolate to oblate spheroidal. Pollen three-lobed. grains in polar view Colpus transversalis; costae. Colpi narrow; margo broad. Sometimes an operculum is present. Tectate; psilate. Columellae distinct. The pollen grains belonging to this type can hardly be separated from the Hippomane type. The margo is broader; mostly as broad as the colpus transversalis. The colpi are narrow. If in the future more Euphorbiaceae species are examined pollen-morphologically, it will perhaps be necessary to combine the two types. Euphorbia hirta subtype Pollen grains without an operculum. Euphorbia hirta Linn. Soeprata 20 H [U] Euphorbia peplus Linn. Fresh material, Utrecht. Euphorbia palustris Linn. Fresh material, Hortus Utrecht. Euphorbia hirta subtype; subprolate. 30 /x E 23 n P:E 1,30. P.A.I. 0,10. Euphorbia hirta subtype. Oblate spheroidal. 26,5 p E 28,5 p P:E 0,93. P.A.I. 0,25. Euphorbia hirta subtype. Prolate spheroidal. 50 n E 45 n P:E 1,10. P.A.I. 0,3. Diplocyathium capitata (Reichb.) H. Schmidt Baldacci, Aug [BR] Euphorbia hirtasubtype; oblate spheroidal. 31,5 fi E 34,5 fi P:E 0,91. P.A.I. 0,2. Euphorbia esula subtype Pollen grains with an operculum. This subtype differs from the other pollen grains in the Euphorbia hirta type by its opercolum. Euphorbia esula Linn. Euphorbia esula subtype. Oblate spheroidal. Fresh material, Utrecht. 43 E 40 P:E n n 0,93. P.A.I. 0,2. Dichostema type Tricolporate; oblate spheroidal. Pollen grains in polar view circular to threelobed. Colpus transversalis; costae. Colpi narrow; small or absent. margo Intectate; reticulate. Columellae large. Exine thick. The Dichostema type is differentiated from the other types by its distinct reticulum and narrow margo. The columellae are fused laterally in their upper part and form a tigillate reticulum. The pollen grains have a thick exine. glaucescens Pierre type; spheroidal-suboblate. Dichostema oblate Klaine 1129 [L] PI. XXI, 4 0,87. /a P 31 E 36 P;E m:e0,5. P.A.I.0,3. Lumina 1-3 ft.

101 Lumina 1-2 u. 0,25. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 101 Anthostema aubryanum A. de Juss. Dichostema type; oblate spheroidal. Spuitebanto 1074 [U] 29 E n 31,5 n P:E 0,92. Senefeldera macrophylla Ducke For. Dept [U] PI. XXI, 6 Dichostema type; spheroidal, p E 36 ft. P.A.I. 0,2-0,25. Lumina 2-3 fx. Euphorbia cotinoides Ellenberg 1635 [U] Miquel Dichostema type; prolate spheroidal. 28 /i E 26 n P:E 1,08. Stahel and Gongryp 816 [U] m:e 0,5. P.A.I. - Margo absent; Lumina 2-3 /x. Stenadenium type Tricolporate; subprolate. Pollen grains large. Colpus transversal is broad and elongated; costae. Colpi narrow and long; costae colpi indistinct. Margo absent. Tectate; psilate; not reticulate. Columellae distinct but small in diameter. Exine relatively thin. Pollen grains of this type are large (larger than 40 /a) and have a relatively thin exine. The columellae are slender; their diameter is small. In polar view the pollen grains are not distinctly three-lobed. A margo is absent. The pollen grains of Hura crepitans have a colpus transversalis shorter than those of Stenadenium and Synadenium. The shape and structure, however, are similar. Stenadenium spinescens Pax Goetze 1099 [BM] PI. XXI, 1 Elaeophorbia drupifera (Thonn.) Staff Leeuwenberg 2314 [U] PI. XXI, 2 Hura crepitans Linn. B. W. 694 [U] B.B.S. 502 [U] PI. XXII, 2 Stenadenium type; subprolate. 59 n E 47 n P:E 1,26. Stenadenium type. 58 fx E 50 n P:E 1,16. P.A.I. 0,2. Tectum perforatum. Stenadenium type. 75 n E 63 fi P:E 1,20. P.A.I. 0,4. Synadenium arborescens Boiss. D Alleizette, Medley 8492 [L] Stenadenium Y oung. type. Pedilanthus type Tricolpate; oblate spheroidal to suboblate. Pollen grains slightly three-lobed. Colpi broad. Margo narrow. granulata. Membrana Intectate (tectum perforatum?); psilate. Columellae tall. Pedilanthus differs from other types in the Hippomane configuration by the absence of a colpus transversalis. Also the P.A.I. is somewhat larger. Pedilanthus palmeri Millsp. Hinton [U] PI. XXII, 1 Pedilanthus type. 72 n E 83 n P:E 0,87. P.A.I. 0,3. Pollen grains not placed in one of the Crotonoideae configurations Dalechampia type Tricolporate; Pollen grains large, usually prolate to prolate spheroidal. Colpus transversalis elongated. Sometimes costae quatoriales. Colpi short and narrow. No costae colpi present. Tectate; intra-reticulate. Reticulum very coarse. Between the muri of the reticulum a membrane is present, covering the lumina.

102 102 W. PUNT plate xxii. 1. Pedilanthus palmeri; Excoecaria 2. Hura crepitans; 3. Actinostemon concolor; bicolor. 4.

103 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 103 The type is characterised by its large pollen grains and wide lumened reticulum. The pollen grains are so typical, that no other type can be compared with them. Taxonomic discussion Dalechampia is easily distinguished from other genera by its typical involucre. For this reason Mueller (1866) placed it in the vicinity of the Euphorbieae. Bentham (1880) pointed to the similar characters in the Plukenetieae : e.g. climbing growth, stinging hairs and thickened style in the female flowers. Pax and K. Hoffmann (1919) thought it better to separate the genus into an apart tribe. Croizat (1940a) and Hurusawa (1954) once again placed the genus near and its allies. ~ Euphorbia Pollen grains show a typical structure and are neither comparable with the types in the Plukenetinae nor with those in the Euphorbieae. Pax and K. Hoffmann s classification maintaining the genus in a separate tribus not too far from the Plukenetinae seems the best opinion. Dalechampia dioscoreifolia Endlich. Poepp. et Dalechampia type; prolate. 105 n E 70 n P:E 1,50. G. Klug 4186 [U] PI. XXIII, 6 Dalechampia affinis Muell. Arg. B. W [U] Lumina up to 12 ft. Costae equatoriales. Dalechampia type; prolate spheroidal. 105 fi E 97 fi P;E 1,08. Costae equatoriales. Dalechampia brasiliensis Lam. Y. Mexia 5270 [U] Dalechampia type; prolate spheroidal. 57 fi E 55 n P:E 1,04. Costae equatoriales. scandens Linn. Dalechampia Serv. Forest [U] Dalechampia type; oblate spheroidal. 105 n E 75 n P:E 1,40. Costae equatoriales. Dalechampia spathulata (Scheidw.) Baill. Dalechampia type; oblate spheroidal. Fresh material, Hort. Utrecht. PI. XXIII, 5 6 3yu E65 < P:E 0,97. Costae transversales; m:e 0,5. Hamilcoa type Tricolporate; suboblate. convex triangular. In polar view the pollen grains circular to slightly Colpus transversalis narrow. Colpi narrow and short; indistinct costae colpi. Tectate (?); intra-reticulumfine or not reticulate; columellaedistinct but small. The Hamilcoa type cannot be placed in one of the discussed configurations. The pollen grains have columellae with small but distinct capita and are for that reason excepted from the Mallotus configuration (p. 77). The structure of Hamilcoa would be in better agreement with the Claoxylon configuration, (p. 90), but the related pollen grains of Plagiostyles and Pimeleodendron are reticulate; moreover, the polar view differs. Perhaps the Moultonianthus type in the Bernardia configuration (p. 73) is most related to the Hamilcoa type because of its circular polar view and fine reticulum present in some species. As the position of this type is still uncertain the author prefers to keep it apart from all other configurations.

104 104 W. PUNT plate xxiii. I. Plagiostyles africana; venenifera; 4. Hamilcoa zenkeri; 5. Dalechampia spathulata; 2. Pimeleodendron zoanthogyne; 3. Spirostachys 6. Dalechampia dioscoreifolia.

105 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 105 Taxonomic discussion Prain (1913) thought Hamilcoa was nearest related to Plagiostyles Pierre and Pimeleodendron Hasskarl. Pax and K. Hoffmann (1912b, 1931) agreed with the relationship of Plagiostyles and Pimeleodendron, which they placed both in the Hippomaneae. Hamilcoa,... however, was transferred to the Gelonieae. In doing so they rejected the affinity between Hamilcoa and Plagiostyles. The three genera have several characters in common: 1. Disc absent in male as well as in female flowers. 2. In male flowers an ovarium rudiment is absent. Filaments are very short or wanting. 3. Pedicels are thickened. 4. Styles are undivided. 5. Plants dioecious (In Hippomaneae usually monoecious). Perhaps Nealchornea, in the present paper placed in the Sumbavia configuration (p. 69), is also related to Hamilcoa and its allies. Hamilcoa zenkeri (Pax) Prain Zenker 4765 [L] PI. XXIII, 4 Pimeleodendron zoanthogyne J. J. Smith Hallier 152a [U] PI. XXIII, 2 Hamilcoa spheroidal. type; suboblate to oblate P 23,5 n E 27 P:E 0,87. m: e< 0,5. P.A.I. > 0,5. Lumina ca. 1 /i. Hamilcoa type; spheroidal. E 24 fi. m:e ca. 1. P.A.I. 0,5-0,6. Not reticulate. Plagiostyles africana (Muell. Arg.) Prain Zenker 94 [U] PI. XXIII, I Hamilcoa type; oblate spheroidal. P27 E / 29yit P:E 0,90. m:eca. 1. P.A.I. 0,4. Costae colpi. Taxonomic comment on the Crotonoideae The subfamily of the Crotonoideae includes a large number of pollen types. The types with a croton-pattern attract most attention. When in the future a revision of the Crotonoideae will be undertaken, the placing of all genera with a croton-pattern into one group will be advisable. This will benefit a natural classification of the subfamily. The relationship of, e.g., Joannesia with Cunuria and Garcia with Sagotia will, in this way, show to better advantage than it does at present in Pax and K. Hoffmann s system. Two other groups of well-distinguished pollen grains, the Cnesmosa configuration and the Plukenetia configuration, are found in the subtribe Plukenetiinae. Omphalea is added to the Plukenetia configuration, which agrees with the opinion of Croizat (p. 62). Mueller Arg., Bentham and Pax and K. Hoffmann include the tribe Chrozophoreae in their systems. This tribe falls into two groups of pollen types. The first group with e.g. Crotonogyne, Aleurites, etc. possesses pollen grains with a croton-pattern. As stated previously these should be genera separated from the other ones and transferred to the group of which Croton is the principal genus. The second group with, e.g., Sumbavia, Chrozophora, Agrostistachys, has aperturate pollen

106 106 W. PUNT grains, which belong to several pollen types. The present author would like to retain this second group as the subtribe Chrozophorinae. The tribe Acalypheae of Pax and K. Hoffmann includes most genera of the Crotonoideae. There are, however, comparatively few types which are, moreover, closely related ( Mallotus configuration, Bernardia configuration, etc.). The tribes Gelonieae and Cluytieae, too, include genera of which the pollen grains are provided with a croton-pattern mixed with up of which the genera pollen grains lack this structure. These latter genera are divisible into several types. tribes is badly needed. A reclassification of these Although Hippomaneae and Euphorbieae are both homogeneous taxa which are taxonomically well-defined, the morphology of the pollen grains shows but little difference between the two groups. Pax (1924) already pointed out the relationship of these groups on other grounds. Since the Hippomaneae are treated by all authors as a tribe of the Crotonoideae, the Euphorbieae should also be treated as a tribe of the Crotonoideae close to the Hippomaneae. SUMMARY In the present study pollen morphology of the Euphorbeaceae is treated as an additional character in taxonomy. Besides the greaterpart of the genera occurring in the ofpax system and K. Hoffmann (1931), most of the genera published after 1931 are studied. The pollen grains have been described with the aid of a terminology as simple as possible. In principle the terminology of Iversen and Troels-Smith has been followed, although in addition, many improvements of Erdtman have been used. One of the simplifications is the rejection of Potoni6 s term sculpture. All elements occurring on the endexine are called structure elements; all structure elements together form the structure of a pollen grain. For the sake of consequence endexine apertures and extexine apertures are discussed separately. Different pollen grains are placed in different pollen types. If the differences are of minor importance, the pollen grains are placed in subtypes. Several types can have some characters in common. To express the correspondences, these types are assembled in configurations. As the pollen types in Phyllanthoideae and Crotonoideae differ distinctly, the division of the Euphorbiaceae in these subfamilies is maintained m the discussion ofthe results. The Phyllanthodieae can be separated in three large groups of pollen types (Antidesma configuration, Amanoa configuration and Aristogeitonia configuration), which agrees with the grouping of Pax in The remaining small configurations belong in taxonomic respect to the genera of the Antidesma configuration. In the Crotonoideae many genera possess pollen grains with a croton-pattern. These genera should be treated as a single group. Besides this natural group, the Plukenetiinae possess pollen grains which are clearly distinguished from other genera in the Crotonoideae. Pollen grains of Omphalea are similar to those in the Plukenetia configuration. This pollen-morphological result agrees with the opinion of Croizat.

107 a a b c b. POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 107 The remaining pollen grains in the Crotonoideae are less easy to differentiate in groups. One of the largest configurations is the Mallotus configuration, which includes most genera of the Acalypheae and several genera or other tribes. The Hippomane configuration is another large one. This configuration comprises the tribes Hippomaneae and Euphorbieae. The pollen grains of both tribes are very similar. The genus Pachyslroma is pollen-morphologically as well as taxonomically related to the tribe Hippomaneae. Per a. treated as a separate tribeby Pax and K. Hoffmann, is related by its pollen grains to some genera in the Acalypheae. Dalechampia is habitually related to the genera in the Plukenetiinae. Pollenmorphological data, however, do not support this relation. The pollen grains of Dalechampia are not similar to any other pollen type. The morphology of the pollen grains of the Stenolobeae is in agreement with the opinion of Pax, that any separation of these Australian genera is an artificial one. REFERENCES Airy-Shaw, H. K Note on the Euphorbiaceous genus Botryophora. Kew Bull. 3: Notes on Malaysian Euphorbiaceae. Kew Bull. 14: , Baillon, H Etude du groupe des Paris. 684 pp Histoires Plantes 5: des Baldwin, J. T. and R. E. Schultes A conspectus of the genus Cunuria. Bot. Mus. Leaf!. Harv. Univ. 12: Bentham, G Notes on Euphorbiaceae. Journ. Linn. Soc. Bot. 17: and J. D. Hooker Genera Plantarum 3: Blume, C. L Bijdragen tot de flora van Ned. Ind. Batavia pp. Campo-Duplan, M. v. and N. Hall Bull, de ITnst. Franc. d Afrique Noire 21 ser. A: Croizat, L. 1940a. Novelties in Chinese Phyllanthus. Lignan Sc. Journ. 20: b. Thirty-five new species of American Croton. Journ. Arnold Arb. 21: c. New and critical Euphorbiaceae from Eastern Tropical Asia. Journ. Arnold Arb. 21: a. The tribe Plukenetiinae ofthe Euphorbiaceae in Eastern Tropical Asia. Journ. Arnold Arb. 22 : b. Notes the on Euphorbiaceae 2. Bull. Bot. Gard. Buitenzorg ser : Notes on the Euphorbiaceae 3. Bull. Bot. Gard. Buitenzorg ser : b. On certain Euphorbiaceae from the Tropical Far East. Journ. Arnold Arb. 23: c. A study of Manihot in North-America. Journ. Arnold Arb. 23: Notes on American Euphorbiaceae. Journ. 33: Wash. Acad. Sci. New and critical Euphorbiaceae of Brazil. Tropical Woods 76: Novelties in American Euphorbiaceae. Arnold Journ. Arb. 24: New or critical Euphorbiaceae of the Americas. Journ. Arnold Arb. 26: a. Nomenclatoral transfers and corrections in the Euphorbiaceae. Tropical Woods 88 : Novelties in American Euphorbiaceae. Journ Arnold Arb. 27:

108 108 W. PUNT Plant Explorations in 1944, chiefly to the Tafelberg and the Kaietur Plateau; Euphorbiaceae; Erdtman, G Ueber die Verwendung von Essigsaeureanhydrid bei Pollenuntersuchungen. Svensk Bot. Tidskr. 28: An introduction to Pollen Analysis. Waltham, Mass. 239 pp Pollen morphology and plant taxonomy. Svensk Bot. Tidskr. 39: Classification of pollen grains and spores. Svensk Bot. Tidskr. 41: Pollen morphology and Plant Taxonomy, Angiosperms. Uppsala. 539 pp. and J. R. Praglowski On the swelling of pollen grains in glycerine jelly. Bot. Notiser 112: Faegri, K Recent trends in Palynology. Bot. Review 22: and J. Iversen Text-book ofmodern Pollen Analysis. Copenhagen. 168 pp. Fischer, H Beitrage zur vergleichende Morphologic der Pollenkorner. Berlin. 72 pp. Fritzsche, J Ueber den Pollen. Mem. Acad. Sci. (Sav. Etrang.) 3. Gagnepain, F. and L. Beille In Lecomte, Flore de 1 Indochine. 5; Gilg, E Beitrage zur Flora von Africa, Centroplacus. Bot. Jahrb. 40 : Grew, N The Anatomy of Plants. London. 304 pp. Gruening, G In Engler, das Pflanzenreich IV Euphorbiaceae; Porantheroideae and Ricinocarpodeae. 97 pp. Hasskarl, K Brief den Heer Hasskarl. van Versl. en Meded. Kon. Akad. Wet. (Natuurk.) Amsterdam 4: Hurusawa, I Eine nochmalige Durchsicht des herkommlichen Systems der Euphorbiaceae in weiteren Sinne. Journ. Fac. Sci. Univ. Tokyo Bot. 6 ; Hutchinson,J. and J. M. Dalziel Ed. 2. Revised by R. W. J. Keay. London. Euphorbiaceae; Iversen, J. and J. Troels-Smith Pollenmorphologischen Definitionen und Typen. Damn. Geol. Unders. ser (8) : 54 pp. Jablonszky, E In Engler, Das Pflanzenreich IV. 147 (8): Euphorbiaceae- Phyllanthoideae-Bridelieae. 98 pp. Jussieu, Adr. de De Euphorbiacearum Generibus Medisque Earundum Viribus Tentamen. Paris. 118 pp. Lanjouw, J The Euphorbiaceae of Surinam. Thesis. Utrecht. 195 pp. Leal, C Contribucao a estudo du familia Euphorbiaceae. Arq. Jard. Bot. Rio de Janeiro 11: Leandri, J Flore de Madagascar, Euphorbiac6es 1: Leonard, J Euphorbiaceae Africaneae Novae. Bull. Jard. Bot. de PEtat Bruxelles 25: Notes sur les genres Suregada Roxb. ex Rottl. et Gelonium Roxb. ex Willd. Bull. Jard. Bot. de PEtat Bruxelles 28; Notes sur les espfeces africaines continentales des genres Sapium P. Browne et Excoecaria Linn. Bull. Jard. Bot. de PEtat Bruxelles 29: Lindau, G Acanthaceae in Engler and Prantl, Die Natuerlichen Pflanzenfamilien 4 (3b): Macbride, J. F Flora of Peru, Euphorbiaceae. Field Mus. Nat. Hist. Bot. 13 (3a.l) : Malpighi, M Anatome Plantarum. Londini. p. 49, t Mohl, H. v Sur la structure et les formes des graines de pollen. Ann. Sci. Nat. ser. 2. 3: , , Mueller, J. (Arg.) Nachschrift zu meiner systematischen Arbeit fiber die Euphorbiaceae. Bot. Ztg. 24: Replik auf Dr. Baillon s Nouvelles observations sur les Euphorbias. Bot. Ztg. 33: ,

109 and. 1910a b a b c POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 109 and E. Boissier Euphorbiaceae in Decandolle, Prodromus 15 (2) : 1286 pp. O Donnell, C. A. and A. Lourteig Chrozophorae Argentinae. Lilloa 8: Pax, F Euphorbiaceae Africanae 5, Microdesmis paniculata. Bot. Jahrb. 28: 25. In Engler. Das Pflanzenreich IV. 147 (1): Euphorbiaceae- Jatropheae. 148 pp. In Engler, Das Pflanzenreich IV. 147 (2): Euphorbiaceae- Adrianeae. 110 pp Die Phylogenie der Euphorbiaceae. Bot. Jahrb. 59: and K. Hoffmann In Engler, Das Pflanzenreich IV. 147 (3): Euphorbiaceae-Cluytieae. 124 pp.. In Engler, Das Pflanzenreich IV. 147 (4): Euphorbiaceae-Gelonieae. 41 pp. and In Engler, Das Pflanzenreich IV. 147 (5): Euphorbiaceae-Hippomaneae. 319 pp. and In Engler, Das Pflanzenreich IV. 147 (6): Euphorbiaceae-Acalypheae-Chrozophorinae. 142 pp. and In Engler, Das Pflanzenreich IV. 147 (7): Euphorbiaceae-Acalypheae-Mercurialinae. 473 pp. and 1919a. In Engler, Das Pflanzenreich IV. 147 (9-11): Euphorbiaceae-Acalypheae-Plukenetinae, Epiprininae, Ricineae. 134 pp. and. 1919b. In Engler, Das Pflanzenreich IV. 147 (12): Euphorbiaceae-Dalechampieae. 59 pp. and 1919c. In Engler, Das Pflanzenreich IV (13): Euphorbiaceae-Pereae. 13 pp. and 1 919d. In. Engler, Das Pflanzenreich IV. 147 (14): Euphorbiaceae-Additamentum. 231 pp. and In -. Engler, Das Pflanzenreich IV. 147 (15): Euphorbiaceae-Phyllanthoideae-Phyllantheae. 349 pp. and In Engler, Das Pflanzenreich IV. 147 (16): and Euphorbiaceae-Crotonoideae-Acalypheae-Acalyphinae. 231 pp. In Engler and Harms, Die Naturlichen Pflanzenfamilien.. 2. ed. 19c: Phillips, E Some changes in nomenclature. Journ. S. Air. Bot. 9: Pierre, L Centroplacus glaucinus. Bull. Soc. Linn. Paris ser. 2: 114. Post, L. v Skogstradspollen i sydsvenska torfmosselagerfoldjer. Forhandl. 16. Skand. Naturforskermote Potonie, R Zur Morphologic der fossilen Pollen und Sporen. Arb. Inst. Palaeobot. und Petrogr. Brennsteine 4: 5-25, 109. Sandwith, N. Y On two Euphorbiaceae of British Guiana. Kew Bull. 1950: Schultes, R. E Studies in the genus Micrandra 1. Bot. Mus. Leaf!. Harv. Univ. 15: A new generic concept in the Euphorbiaceae. Bot. Mus. Leaf!. Harv. Univ. 17: Smith, J. J In S. H. Reorders en Th. Valeton, Boomsoorten van Java 12: Steenis, C. G. G. J. v Note on the genus Moultonianthus Merr. (Euph.) : Bull. Bot. Card. Buitenzorg ser. Trevisan, V. B. A Alghe Cocotale. Padova, p. 99. Webster, G. L A monographic study of the West Indian species of Phyllanthus. Journ. Arnold Arb. 37, 38 and 39. Wodehouse, R. P The phylogenetic value of pollen-grain characters. Ann. Bot. 42: Pollen grains. New York. 574 pp. Zinderen Barker, E. M. van South African pollen grains and spores I. Amsterdam (Cape Town) 88 pp.

110 INDEX (Asterisks indicate illustrations) Acalypha type, 79, 81 Acalypha aronioides, 81 diversifolia, 81 indica, 81 racemosa, 81 platyphylla, 81 scandens, 81 Acalyphopsis, 6 Acidocroton adelioides, 51 Acidoton urens, 57 Acephila excelsa, 29, 32 Actephilopsis malayana, 60., Actinostemon concolor, 97 lanceolatus, 97 Adelia ricinella, 68 Adenochlaena leucocephala,. 94 Adenocline subtype, 48, 54 xiphoclada, 95 Amyrea humberti, 86 Anabaena, 10 Anabaenella, 10 tamnoides, 63 Anabaina, 10 Andrachne subtype, 28 Andrachne aspera, 29 colchica, 29 phyllanthoides, 29 telephioides. 29 Androstachys johnsonii, 43 Angostyles longifolia, 63 Annesijoa, 6 Anomalocalyx uleanus, 51 Anthostema aubryanum, 101 Antidesma configuration, 20, 30, 32, Adonocline, 48, acuta, mercurialis, 56 Adenopeltis colliguaya, 99 34, 39, 36, 43, 46 Antidesma type, 19, 20, 21, 26, 28 Antidesma subtype, 20, 21, 22, 24 Antidesma agusanense, 22 Adenophaedra, 73 minor, 75, 82 prealtum, 82 Adriana, 96 quadripartita, 80 Agrostistachys, 105 indica, 76 sessilifolia, 76 bunius, 22 diandrum, 22 edule, 22 ghesaembilla, 22 membranaceum, 22 obovatum, 22 venosum, 22 Aonikena patagonica, 66 Agyneia bacciformis, 39 Alchornea type, 19, 73, 79, 82, 83, 85 Alchornea, 73, 75, 85, 94 ilicifolia, 83 laxiflora, rugosa, schomburgkii, 82 triplinervia, 82 villosa, 82 Alchorneopsis, 85 floribunda, 86 trimera, 86 Aleurites, 105 moluccana, 50 Algernonia brasiliensis, 99 Alphandia furfuracea,.50 Amanoa configuration, 32, 36, 38, 46 Amanoa type, 32, 33, 34 Amanoa, 19, 32, 33 boiviniana, 33 bracteosa, 33 grandiflora, 33 guianensis, 33 oblongifolia, 33 Amperea type, 19, 94 Amperea spartioides, 95 Aparisthmium cordatum, 83 Apodandra, 62 buchtienii, 63 Apodiscus chevalieri, 22 Aporosa, 25 dioica, 24 frutescens, 24 lindleyana, 24 Aporosella, 25 Arachnodes, 38 chevalieri, 40 Argythamnia 68 type, Argythamnia, 66 candicans, 266, 68 Argomuellera macrophylla, 76 Aristogeitonia configuration, 40, 46 Aristogeitonia 40 type, Aristogeitonia subtype, 41 Aristogeitonia limoniifolia, 41 Astrocasia, 25, 32 tremula, 26 Astrococcus, 62 cornutus, 63 Athroandra africana, 92 mannii, 92 Avellanita bustillosii, 80

111 POLLEN MORPHOLOGY OF THE EUPHORBIACEAE 111 Baccaurea subtype, 21, 22, 24 Cenesmon, 10, 57 Baccaurea javanica, 24 Centroplacus glaucinus, 89 oxyxarpa, 24 Cephalocroton, 96 racemosa, 24 cordofanus, 94 sumatrana, 24 Cephalocrolonopsis socotrana,, 94 Baliospermum, 48 Cephalomappa configuration, 68 montanum, 51 Cephalomappa type, 68 Baloghia lucida, 51 Cephalomappa malloticarpa, 69 Benoistia, 48 Chaetocarpus subtype, 90 perrieri, 51 Chaetocarpus castanocarpus, 90 Bemardia configuration, 73, 92, 103, schomburgkianus, Cheilosa type, 79, 90 Bernardia type, 73, 75, 76, 77 Cheilosa montana, 90 Bernardia, 85 Chiropetalum configuration, 62, 65 corensis, 75 Chiropetalum type, 65, 66, 68 lorentzii, 75 Chiropetalum canescens, 66 pulchella, 75 lanceolatum, 66 Bertya, 48 patagonica, 66 gummifera, 53 Chlamydojatropha, 6 Beyeria, 48 Chondrostylis bancana, 85 leschenaultii, 53 Chonocentrum, 21, 33 Bischofia type, 43 cyathoforum, 25 Bischofia, 9, 40, 44 Chrozophora type, 69, 70, 72 javanica, 44 Chrozophora, 105 Blachia umbellata, 51 plicata, 70 Blotia oblongifolia, 34 tinctoria, 70 Blumeodendron type, 79, 81 Cladogelonium, 48, 54 Blumeodendron tokbrai, 81 Bocquillonia sessiliflora, 83 Bonania cubana, 99 domingensis, 99 Botryophora geniculata, 81 madagascariense, 54 Cladogynos configuration, 69, 93 Cladogynos type, 69, 93, 94 Cladogynos orientalis, 94 Breynia configuration, 38 Breynia type, 38, 39 Breynia subtype, 38 Breynia, 29, 38, 39, 44 fruticosa, 39 Claoxylon configuration, 90, 103 Claoxylon type, 92, 93 Claoxylon cuneatum, 92 tumidum, 92 Claoxylopsis, 5 nivosa, 39 Clarorivinia, 6 racemosa, 39 Cleidion, 82, 83 Breyniopsis subtype, 39 Breyniopsis Pierrei, 39 angustifolium, 80 javanicum, 80 Briedelia type, 32, 36 Cleistanthus, 32 Briedelia, 9, 32 dichotomus, 34 assamica, 36 ferruginous, 36 glauca, 36 winkleri, 34 monoica, 36 Clutia type, 19, 75, 77 stipularis, 36 Clutia abyssinica, 77 alaternoides, 77 Calpigyne hainanense, 86 natalensis, 77 Calicopeplus, 5 paxii, 77 Caperonia type, 72 rubricaulis, 77 Caperonia subtype, 72 Cluytiandra madagascariensis, 34 Caperonia, 69 Cnesmosa configuration, 57, 58, 105 corchoroides, 72 Cnesmosa type, 57, 58 palustris, 72 Cnesmosa, 10 serrata, 72 javanica, 57 Carumbium, 10 Caryodendron type, 72 Cnidiscolus stimulosis, 53 urens, 53 Caryodendron grandifolius, 73 orinocense, 73 Cavacoa aurea, 51 Coccoceras borneënse, 80 Codiaeum variegatum, 51 stellingianum, 51

112 112 W. PUNT Coelebogyne ilicifolia, 83 Coelodepas, 10 Coelodiscus lanceolatus, 80 Colliguaya integerrima, 97 Cometia, 5 Conveiba lutifolia, 83 simulata, 83 Conceveibastrum martianum, 83 Cordemoya integrifolia, 80 Ditaxis catamarcensis, 66 diversiflora, 66 fasciculata, 66 fendleri, 66 lancifolia, 66 salina, 66 Ditta, 5 Dodecastigma, 48 amazonicum, 53 Corythea, 5 Domohinea perrieri, 53 Croizatia, 5 Crotonoideae, 47 Croton configuration, 47, 58, 90 Croton type, 48, 53, 54, 57, 58 Croton bahamensis, 50 Drypetes, 21, 44 glauca, 28 laterifolia, 28 macrophylla, 28 variabilis, 28 cuneatus, 50 Duvigneaudia inopinata, 99 hirtus, 50 Dysopsis configuration, 60 lingiradiatus, 50 Dysopsis type, 60 matourensis, 47, 50 Dysopsis glechomoides, 60 pullei, 50 Crotonogyne, 105 Elaeophorbia drupifera, 101 parvifolia, 50 Elateriospermum, 48 preussii, 50 tapos, 51 Crotonogynopsis usambarica, 85 Crotonopsis elliptica, 50 Cunuria, 49 Eleuterostigma lehmannianum, 62 Endospermum, 48 moluccanum, 54 spruceana var. bracteosa, 50, 53 Epiprinus poilanei, 94 Cyathogyne spathulifolia, 22 viridis, 22 Cyrtogonone 50 argentea, Cyttaranthus congolensis, 75, 76 Eremocarpus, 48 setigerus, 50 Erismanthus 77 type, Erismanthus indochinensis, 77 oblique, 77 Dalechampia 101 type, Dalechampia affinis, 103 brasiliensis, 103 dioscoriifolia, 103 scandens, 103 Erythrococca africana, 92 mannii, 92 stolziana, 92 Euphorbia hirta type, 96, 100 Euphorbia esula subtype, 100 spathulata, 103 Euphorbia hirta subtype, 100 Dalembertia populifolia, 99 Danguyodrypetes ambigua, 30 Dendrophyllanthus type, 37 Dendrophyllanthus ripicolus 37, Deuteromallotus acuminatus, 80 Deutzianthus tonkinensis, 50 Dichostema type, 96, 100 Euphorbia, 9, 103 cotinoides, 101 esula, 100 hirta, 100 palustris, 100 peplus, 100 Excoecaria bicolor, 97 Dichostema glaucescens, 100 Dicoelia type, 20, 21, 28, 29, 32, 44 phillippinensis, 99 venenifera, 97 Dicoelia subtype, 28 Dicoelia affinis, 28 Dimorphocalyx murina, 51 Diplocyathium capitata, 100 Discocarpus type, 32, 36 Discocarpus, 21 Flueggia, 25 Fontainea pancheri, 51 Fragariopsis, 62 scandens, 63» essequeboensis, 36 hirtus, 34 Discoclaoxylon hexandrum, 92 Discocleidion type, 79, 83 Discocleidion rufescens, 85 Galearia dognaiensis, 89 filiformis, 89 Garcia, 48, 105 nutans, 50 Gatnaia annamica, 24 Discoglypremna caloneura, 76 Gavarretia terminalis, 83 Dissiliaria tricornis, 41 Ditaxis type, 65, 66, 68 Gitaria, 6 Givotia rottleriformis, 51