Org. Divers. Evol. 2, 107–125 (2002)
© Urban & Fischer Verlag
http://www.urbanfischer.de/journals/ode
Morphological phylogenetics of the sea spiders
(Arthropoda: Pycnogonida)
Claudia P. Arango*
Department of Zoology and Tropical Ecology, James Cook University, Townsville, Australia
Received 1 July 2001 • Accepted 2 April 2002
Abstract
Pycnogonids or sea spiders are a group of marine arthropods whose relations to the chelicerates have been an issue of controversy. Higher-level
phylogenetic relationships among the lineages of sea spiders are investigated using 36 morphological characters from 37 species from all extant
families and a Devonian pycnogonid fossil. This is one of the first attempts to analyze the higher-level relationships of the Pycnogonida using
cladistic techniques. Character homoplasy (implied weights) is taken into account to construct a polytomous, most-parsimonious tree in which
two major clades within Pycnogonida are obtained. Clade A includes Ammotheidae paraphyletic with Colossendeidae, Austrodecidae and Rhynchothoracidae, and clade B is formed by Nymphonidae, Callipallenidae (apparently paraphyletic), Pycnogonidae and Phoxichilidiidae.The analysis
of equally weighted data is presented and helps to identify those characters less consistent. The reduction of the chelifores, palps and ovigers –
shown independently within each of the clades as parallel evolution events – challenges the assumption of a gradual mode of reduction within
the group, according to analysis of unordered vs ordered characters. Most of the phylogenetic affinities proposed here are compatible with traditional classifications. However, traditional taxonomic characters need to be complemented by sets of anatomical, molecular and developmental
data, among others, to produce more robust phylogenetic hypotheses on the higher- and lower-level relationships of the sea spiders.
Key words: Pycnogonida, phylogeny, morphology, character evolution, gradual reduction, sea spiders
See also Electronic Supplement at http://www.senckenberg.de/odes/02-05.htm
Introduction
Pycnogonida or sea spiders are an extraordinary group of
marine arthropods containing more than 1100 species
living in tropical to polar seas, from the shoreline to
abyssal depths (Arnaud & Bamber 1987, Munilla 1999).
The subphylum Pycnogonida (Hedgpeth 1955) is defined
by striking autapomorphies that strongly support monophyly, such as the prominent external proboscis, an extremely reduced abdomen, and the presence of a ventral
pair of appendages (ovigers) on the cephalic segment
(Boudreaux 1979). Pycnogonids have recently been proposed as the sister-taxon of all other extant arthropods
(Giribet et al. 2001). Other studies have carried out a controversy about the status of pycnogonids as either relatives of the Chelicerata (Snodgrass 1938, Weygoldt
1986, Wheeler et al. 1993, Wheeler & Hayashi 1998,
Edgecombe et al. 2000) or basal arthropods (Zrzavy et al.
1997, Giribet & Ribera 2000, Edgecombe et al. 2000,
Giribet et al. 2001). The scarcity of fossil records has not
helped to solve uncertainties about the evolutionary history of pycnogonids. Nonetheless, the Devonian fossil
Palaeoisopus problematicus Broili has been used to infer
ancestral conditions of the group (Stock 1994). Recently,
findings of pycnogonid larvae from the Upper Cambrian
‘Orsten’ have been interpreted as new evidence to relate
pycnogonids to chelicerates (Walossek & Müller 1997,
Walossek & Dunlop in press).
The uncertain position of the sea spiders is also reflected in the lack of understanding of the phylogenetic relationships within the group. Few studies have addressed
the phylogenetic relationships of the pycnogonids at family level, and none of them used explicit cladistic analysis
(but see Lovely 1999). Hedgpeth (1955) proposed a clas-
*Corresponding author: Claudia P. Arango, Department of Zoology and Tropical Ecology / School of Marine Biology and Aquaculture, James
Cook University, Townsville, 4811 Queensland, Australia; e-mail: claudia.arango@jcu.edu.au
1439-6092/02/02/02-107 $ 15.00/0
Org. Divers. Evol. (2002) 2, 107–125
108
Arango
sification of eight families of pycnogonids based on the
presence and complexity of the chelifores, palps and
ovigers. However, he stated that it would be almost impossible to draw a phylogenetic family tree, referring to
the failed attempts in separating some of the families due
to the occurrence of ‘transitional’ genera such as Pallenopsis (Hedgpeth 1947). Hedgpeth also suggested a direction in the evolution of the group based on a gradual
reductive trend. He regarded the Nymphonidae as the
most generalized lineage. Its members have functional
chelae, 10-segmented ovigers, and long palps. Pycnogonidae are considered the most derived due to the absence
of all the head appendages (Hedgpeth 1947). Fry (1978)
examined the morphological similarities among pycnogonids to re-classify the group using a numerical taxonomy approach. However, the unclear outcome of the analysis and the creation of about 20 new families did not receive much support (Arnaud & Bamber 1987, Munilla
1999). More recently, Stock (1994) expressed his ‘personal philosophy’ about the phylogenetic relationships of
pycnogonids based on a comparison of the extant Pycnogonida with the fossil Palaeoisopus problematicus, based
on the assumption that 10-segmented appendages were
the plesiomorphic state (Fig. 1A). Stock reiterated the
ideas of a gradual reduction stated by Hedgpeth (1947),
but he proposed Ammotheidae as the basal lineage of the
pycnogonids with Eurycyde as the possible most primitive extant form (Stock 1994). Munilla (1999) also concluded that pycnogonid phylogeny could be derived from
the assumption of ‘regressive evolution’, or the gradual
loss of structures over evolutionary time (Fig. 1B).
So far, the hypothesis of an evolutionary trend of successive reductions in the number of segments of the appendages has not been examined under current cladistic
techniques, and the validity of the families as monophyletic groups, mainly Ammotheidae and Callipallenidae, needs to be revised. Here, I test the assumption
of a reductive trend in the Pycnogonida examining the
phylogenetic affinities among the main lineages using
quantitative cladistic analysis.
Ordered multistate characters can be useful when assuming trends of reduction series (Wilkinson 1992). For
example, one can assume – by introducing unordered
characters – that all possible changes in the number of segments of the appendages have an equal chance to occur.
Alternatively, it can be assumed that the chelifores, palps
and ovigers of pycnogonids went through a gradual reduction from the maximum number of segments to a complete
loss; then, ordered characters are used. The most parsimonious outcome is chosen after comparing the results of the
‘ordered’ with those of the ‘unordered’ analysis.
When dealing with high-level phylogenies, taxon
sampling can affect resulting topologies (BinindaEmonds et al. 1998). The Pycnogonida are characterized
by great morphological plasticity, and there are excepOrg. Divers. Evol. (2002) 2, 107–125
tions for almost every morphological character state
within families and genera (Thompson 1904). In this
analysis, the exemplar method approach (Yeates 1995)
was attempted. Although not exhaustive because of the
extremely wide variation in some taxa and unavailability
of material, in most cases genera polymorphic for the
characters coded are represented by more than one
species and not by a single hypothetical supraspecific
terminal taxon (see discussion in Prendini 2001).
In this paper, I present a cladistic study of the major
lineages of the subphylum Pycnogonida, based entirely
on morphological characteristics of adults. The main
goals of this study are: 1) to examine phylogenetic
affinities among extant sea spiders under different coding strategies, 2) to quantitatively test the hypothesis of a
gradual reduction trend of the chelifores, palps and
ovigers, 3) to identify taxa of doubtful monophyly and
thus in need of taxonomic revision, and 4) to discuss the
compatibility of the presented cladistic results with existing classifications of Pycnogonida.
Additional information on this study is available from
the Organisms Diversity and Evolution Electronic
Supplement 02–05, Parts 1–3, on the internet at
http://www.senckenberg.de/odes/.
Materials and methods
Taxon sampling
A total of 38 species belonging to 21 genera of extant Pycnogonida and one fossil species, Palaeoisopus problematicus, were
included in the analysis (Appendix 1; Table 1; Electr. Suppl.,
Pt 3). For a matter of convenience, the traditional family assignments of pycnogonid genera have been used throughout
the study while their validity is being tested. Ammotheidae and
Callipallenidae, the most diverse families in terms of morphology and number of genera, are represented here by eight and
five genera, respectively (Pallenopsis Wilson is included in the
Callipallenidae following Child 1979). Colossendeidae is represented by a species of the type genus Colossendeis Jarzinsky
and by a species of Rhopalorhynchus Wood-Mason. The monogeneric Rhynchothoracidae and Pycnogonidae, which are remarkably uniform, are represented here by single species.
Nymphonidae, a cosmopolitan family with a large number of
closely related species belonging to the type genus Nymphon
Fabricius, is represented by three species.
Several factors influenced the selection of species for the
analysis. Firstly, most of the taxa included in the analysis are
part of the collections made by the author, mostly in North
Queensland but also in the Colombian Caribbean (Arango
2000). Additional material was kindly provided by collaborators in Australia and overseas, but for a few species descriptions from the literature had to be used (all material sources
listed in Electr. Suppl., Pt 1). Type genera and those whose
members are abundant and/or of widest distribution were selected from each of the families. The genera for which more
Morphological phylogenetics of sea spiders
109
Fig. 1. Previous hypotheses of phylogenetic relationships among
pycnogonid families. A. Modified after Stock (1994), with basic body
plan in each lineage (illustrations from Hedgpeth 1948; Stock 1989,
1991). B. From Munilla (1999).
Org. Divers. Evol. (2002) 2, 107–125
110
Arango
than one species were included due to intrageneric ‘polymorphism’ were: Achelia Hodge, Ammothella Verrill, Anoplodactylus Wilson, Ascorhynchus Sars, Austrodecus Hodgson,
Callipallene Flynn, Cilunculus Loman, Nymphon, and
Tanystylum Miers. ‘Transitional’ or problematic taxa such as
Pallenopsis, Tanystylum and Endeis Philippi, whose taxonomic status had been a matter of debate, have been included to
provide a test for such taxonomic hypotheses.
Most of the phylogenetically informative characters within
the Pycnogonida are derived from structures absent in any
other arthropod taxa (e.g. characters of ovigers and proboscis).
For this reason it is very difficult to take outgroup relationships into account. Palaeoisopus problematicus, a well-known
fossil pycnogonid from the Devonian, was introduced to the
analysis in hope of a root for the cladograms. Another two
species of fossil sea spiders from the Lower Devonian are
known, but very few specimens have been examined and their
morphology is not well understood (Bergström et al. 1980).
Characters of P. problematicus were coded according to published descriptions of the fossil specimens based on radiographs (Bergström et al. 1980). In the absence of algorithms
that distinguish between inapplicable characters and missing
data (Lee & Bryant 1999), the use of fossil taxa offers some
difficulties when coding detailed morphological characters
(Kitching et al. 1998). However, it remains the best option to
provide a sister group for the extant Pycnogonida.
Characters
Thirty-six morphological characters of adult sea spiders were
scored across the 38 species (Appendix 1, Table 1). Of the
total number of characters, 20 are binary and 16 were coded as
multistate. For those multistate characters that refer to the
number of segments of the appendages, the actual number of
Table 1. Species and coding of morphological characters.
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
Palaeoisopus problematicus
Eurycyde raphiaster
Achelia assimilis
Achelia australiensis
Ammothella n.sp.
Ammothella biunguiculata
Ammothea hilgendorfi
Tanystylum haswelli
Tanystylum rehderi
Nymphopsis acinacispinata
Ascorhynchus glaberrimus
Ascorhynchus ramipes
Ascorhynchus tenuirostris
Cilunculus armatus
Cilunculus sekiguchi
Nymphon micronesicum
Nymphon molleri
Nymphon surinamense
Colossendeis megalonyx
Rhopalorhynchus tenuissimum
Austrodecus glaciale
Austrodecus gordonae
Rhynchothorax australis
Callipallene novaezealandiae
Callipallene brevirostris
Parapallene famelica
Propallene saengeri
Pseudopallene ambigua
Pallenopsis schmitti
Anoplodactylus n.sp.
Anoplodactylus batangensis
Anoplodactylus tenuicorpus
Anoplodactylus glandulifer
Anoplodactylus insignis
Anoplodactylus n. sp.
Anoplodactylus longiceps
Endeis mollis
Pycnogonum litorale
Org. Divers. Evol. (2002) 2, 000–000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
2
2
1
1
2
2
2
1
2
2
2
2
2
2
2
2
–
–
–
–
–
2
2
2
2
2
2
2
2
2
2
2
2
2
–
–
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
–
–
–
–
–
0
0
0
0
0
0
0
0
0
0
0
0
0
–
–
0
1
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
–
–
–
–
–
0
0
1
0
0
0
0
0
0
0
0
0
0
–
–
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
1
0
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
5
3
1
1
1
0
1
1
4
4
4
0
0
3
3
3
–
–
–
6
–
7
–
–
–
–
–
–
–
–
–
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
5
3
1
1
1
0
1
1
4
4
4
0
0
3
3
5
–
–
–
–
–
6
–
–
–
–
–
–
–
–
–
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
4
5
1
1
1
1
1
1
1
4
4
4
4
4
4
4
3
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
0
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
1
3
4
1
1
1
1
1
1
1
–
–
–
–
–
–
–
–
–
0
1
0
0
0
0
0
0
0
0
1
1
1
0
0
1
1
1
1
1
0
0
1
0
0
1
0
1
0
0
0
0
0
0
0
0
0
1
0
0
1
1
1
1
1
1
1
1
0
0
0
1
1
1
1
0
0
0
1
0
1
1
1
0
0
0
1
1
0
0
1
0
0
1
1
0
?
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
?
0
0
0
0
0
0
0
0
0
1
1
0
0
0
1
1
1
–
–
0
0
0
0
1
1
1
–
0
0
0
1
1
0
0
1
1
–
?
1
1
1
1
1
0
1
1
1
0
0
0
1
1
0
0
0
–
–
1
1
1
0
0
0
1
–
1
1
1
0
1
0
1
0
0
–
?
1
0
0
0
0
0
0
0
0
0
0
2
0
0
2
2
2
–
–
2
2
0
2
2
2
2
–
2
0
0
0
0
0
0
0
1
–
?
2
0
0
0
0
0
0
0
0
3
3
1
1
1
3
3
3
–
–
1
1
0
1
1
1
3
–
1
1
1
1
1
1
1
1
3
–
1
0
2
2
0
0
1
2
2
1
0
0
0
1
1
0
0
0
0
0
1
1
2
0
0
0
0
2
0
1
1
0
1
0
2
0
0
2
0
0
1
1
0
0
0
1
1
2
0
0
0
0
0
0
0
0
2
0
0
0
1
1
0
0
0
0
0
0
0
0
1
0
1
0
0
1
?
2
2
2
2
2
2
2
2
2
2
3
2
2
2
2
2
1
0
0
3
0
3
2
2
–
2
2
2
2
2
2
2
2
2
2
1
3
?
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
0
0
1
1
1
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
1
0
0
0
1
1
1
0
0
2
2
2
0
0
0
0
0
2
2
2
2
0
0
0
0
0
0
0
0
0
1
0
?
2
2
2
1
2
4
2
1
2
2
2
2
4
2
1
1
1
0
0
0
2
3
1
1
3
1
1
1
3
3
3
3
3
3
3
4
1
?
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
?
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
1
0
0
1
–
–
–
–
–
–
–
–
–
0
1
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
–
–
–
–
–
–
–
–
–
1
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0
0
0
–
–
–
–
–
0
0
1
0
1
1
1
1
0
0
1
1
1
–
–
0
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0
0
0
–
–
–
–
–
0
0
0
1
1
1
1
1
1
1
1
1
1
–
–
0
1
1
1
1
1
1
0
1
1
0
0
0
0
0
1
1
1
0
2
0
0
0
1
1
1
0
0
1
1
1
1
1
1
1
1
1
0
2
1
1
1
1
1
2
3
3
2
1
1
1
2
2
0
0
0
0
0
3
3
1
0
0
2
0
1
0
0
3
0
0
0
0
0
0
1
0
1
1
0
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
0
0
0
0
1
1
1
1
1
2
1
1
1
1
2
1
1
1
1
0
0
2
2
2
2
1
0
0
0
0
0
0
1
1
0
0
0
1
1
1
0
2
2
0
0
0
0
1
0
0
0
2
2
2
2
2
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Morphological phylogenetics of sea spiders
segments present was entered as a character state instead of
creating class intervals to reduce the number of states.
Different alternatives exist for the coding of inapplicable
character states when a certain character is absent in some of
the taxa. In the pycnogonids, this is a major problem when
coding external morphological features, specifically those of
the cephalic appendages and the cement glands because these
are absent in several taxa. I use the reductive coding (Strong &
Lipscomb 1999), denoting inapplicable characters (e.g. number of palp segments in those taxa having no palps) with “-”.
There are problems associated with this reductive or separate
coding, such as redundancy of absence states and their partial
independence relative to the presence (Lee & Bryant 1999).
However, when analyzed under unambiguous optimization
settings – that is, not allowing semi-strict trees (default in
Nona and PeeWee packages; Goloboff 1993b, 1997) – reductive coding is the option that best reflects the information content of the data collected when inapplicable values need to be
introduced to the matrix (Strong & Lipscomb 1999).
All sixteen multistate characters were initially treated as
unordered. Even if the ‘unordered’ analysis is justified because
character transformations should not be assumed but tested by
cladistic analysis (Hauser & Presch 1991), different assumptions of character transformations (e.g. ‘unordered’ and ‘ordered’) should be compared (Wilkinson 1992). Disagreement
between the alternative treatments implies that phylogenetic
inferences are sensitive to assumptions of character evolution
(Wilkinson 1992), and a choice then has to be made. Those
multistate characters for which hypotheses of transformation
series could be assumed (e.g. 10-segmented to 9-segmented to
7-segmented to 6-segmented, etc.) were coded both as unordered and ordered (characters 1, 5, 7, 8, 10 in Appendix 1),
and the results of both analyses were compared.
Character evaluation
The analyses were carried out without assigning a priori polarity to characters. Outgroup relationships and polarization of
characters are issues difficult to approach in pycnogonids due
to the lack of reliable sister taxa. P. problematicus has not been
constrained as an outgroup to polarize the characters. It was
expected to provide a root for the tree but is not meant to be
used for a strict outgroup comparison. Assumptions can be
made regarding the ancestral state of some characters, but
there is no reason to assume that because P. problematicus is
extinct and probably older than the extant forms, all its characters are therefore plesiomorphic.
In the following sections a discussion of the most relevant
characters is presented, including assumptions about their evolution, as well as a brief overview of important features of the
different lineages of pycnogonids.
Chelifores
The term ‘chelifores’ makes reference to the first pair of appendages frontally on the cephalon. They are believed to be homologues to the chelicerae in Arachnida. The presence of chelifores in adult pycnogonids (character 0, see Appendix 1) is considered a plesiomorphic state based on outgroup comparison
(comparing them to the chelicerae) and on the ontogenetic criterion, since all pycnogonid larvae and juvenile forms have cheli-
111
fores. Chelifores are completely functional in adults of all members of the families Nymphonidae, Phoxichilidiidae (excluding
Endeis after Child 1992), and most species of Callipallenidae.
When present, chelifores of extant taxa have one or two basal
segments forming the scape. The fossil specimens have three
segments (character 1). Among modern species, and following
the idea of a reductive trend, a two-segmented scape has been
assumed as a primitive condition (Stock 1994, Munilla 1999).
Different degrees of reduction of chelifores are found within the
Ammotheidae, ranging from a few fully chelate species to
species having no chelifores but just a single short segment on
the front of the cephalon. The apomorphic state of complete absence of these appendages is characteristic of Pycnogonidae,
Austrodecidae, Rhynchothoracidae, and the genus Endeis. The
presence of spines on the chelifores (character 3) and teeth on
the chelae (character 28) might have phylogenetic importance
and these characters have been coded among those forms bearing chelifores. The former is an apparently stable feature within
some genera of Ammotheidae (Nymphopsis Haswell, Ammothella, Achelia), the latter is useful for genera and species of Callipallenidae and Nymphonidae, respectively. The fossil P. problematicus has neither spines on the chelifores nor teeth on the
chelae, and it is possible that their occurrence might be a derived feature of certain extant taxa. However, I refrain from accepting the absence of spines and teeth on chelifores as a plesiomorphic state based on knowledge of a single fossil species.
Families of pycnogonids have been classified based on the presence-absence of chelifores, but it is unlikely, or at least not yet
proven, that all the forms whose adults lack chelifores are phylogenetically related.
Palps
Considered homologues to pedipalps in arachnids. As with the
chelifores, the absence of palps is a feature used to classify genera into families. When present, their pattern of segmentation
shows a wide variation, from more than ten segments to a single
segment (characters 5 and 7). Longer and more segmented
palps have been assumed as the plesiomorphic condition (Stock
1994, Munilla 1999). The palps of the known fossils cannot be
evaluated with complete certainty, but apparently they have
nine segments (Bergström et al. 1980). Ten-segmented palps are
coded for some Ammotheidae (e.g. Eurycyde Schiodte and Ascorhynchus) and Colossendeidae, although discrepancies exist
in the literature regarding the counting of the basal portion of
the palps of Colossendeidae as a segment. Palps of nine, eight
or six segments are found in Ammotheidae, Austrodecidae,
Rhynchothoracidae and some callipallenids. The number of
palpal segments is polymorphic within some of the ammotheid
genera coded (e.g. Tanystylum, Ascorhynchus). Although males
and females are mostly similar in regard to the palps, presence/absence and number of segments have been coded separately for each sex to include differences observed in the callipallenid Propallene Schimkewitsch (characters 4–7). A transformation series of the palps towards reduction is tested coding
the characters of number of segments as ordered.
Ovigers
Unlike the chelifores and the palps, there is no counterpart or
structure homologous with the ovigers of pycnogonids in any
Org. Divers. Evol. (2002) 2, 107–125
112
Arango
other arthropod group. They can be present or absent, and
show different degrees of reduction and patterns of segmentation. These characters vary between males and females.
Eleven-segmented ovigers are assumed here to occur in males
and females of P. problematicus, since sex cannot be distinguished in the radiographs of the fossils. Presence of ovigers
and number of segments were coded separately for males and
females because they are sexually dimorphic. Females of Pycnogonidae, Phoxichilidiidae and Endeis lack ovigers completely. The terminal claw (character 11) of nymphonids and
some ammotheids and callipallenids could be derived from the
eleventh segment of the oviger observed in the fossil. However, an ancestral condition within the group could be the presence of a terminal claw as a remnant of the main claw of the
propodus, retained during the modification of the oviger, if the
latter is assumed to be a modified leg (Arnaud & Bamber
1987). Thus, the loss of the terminal claw could be seen as an
apomorphic condition. Different types of spines can be present
on the terminal segments of the ovigers (character 25), or
spines can be completely absent, as in members of Phoxichilidiidae and Pycnogonidae, which is seen as a reversal (Stock
1994). Nymphonids, callipallenids, colossendeids and members of Ammotheidae share compound or denticulate spines.
Denticulate spines are generally arranged in a single row on
the last four segments. However, multiple rows of spines are
present in colossendeids and two ammotheid genera, Ascorhynchus and Eurycyde (character 27). It is not possible to
define either a plesiomorphic or an apomorphic condition for
this character.
Legs
Characteristics of the propodus are useful to segregate genera
and species of pycnogonids. The presence of auxiliary claws
(character 12), or ‘ungues’ when compared to the pretarsal
structure of spiders in Snodgrass (1952), has been an important character to recognize genera of ammotheid and callipallenid affinities, and species within Nymphonidae and Phoxichilidiidae. They are not evident in the fossil, and homology
has not been established with similar chelicerate structures
such as tridactyl claws believed to be ancestral (e.g. in Nothrus
sp. (Acari); van der Hammen, 1986). The presence of heel
spines is included to examine its phylogenetic informativeness
(character 22). The absence of heel spines in divergent taxa appears as a parallel event of secondary loss.
Cement gland openings on the femora of males (in Propallene occurring on tibiae as well) are present in most pycnogonid taxa (character 13), suggesting this as the plesiomorphic
state despite the character being uncodeable for the fossil P.
problematicus. Absence of these structures in the unrelated
taxa Colossendeidae, Pycnogonidae and Pseudopallene Wilson can be assumed to be due to loss. Cement glands are present as single or multiple openings, the former appearing as
characteristic of more basal taxa. A clear pattern of the distribution of the type of cement glands cannot be distinguished
among the families. Both pores and conspicuous tubes occur
within Ammotheidae, Callipallenidae and Phoxichilidiidae
(character 15). The situation is similar regarding cement gland
position with respect to the femora, but a mid-dorsal position
seems to be the general state (character 16 and 17).
Org. Divers. Evol. (2002) 2, 107–125
Genital pores or gonopores are located ventrally on the second coxae of one, two, three, or all pairs of legs (characters 20
and 21). Multiple openings of the gonads are assumed to be a
plesiomorphic condition when compared to chelicerates and
euarthropods in general (Boudreaux 1979). Most of the female
sea spiders have gonopores on all leg pairs, but females in
Rhynchothorax Costa and Pycnogonum Brunnich, for instance,
possess a single pair of gonopores. Within the Pycnogonida this
state might be assumed as a secondary loss occurring independently in different lineages. In some members of Ammotheidae
and Phoxichilidiidae, the genital pores of males are present on
prominent ventral spurs of the coxae (character 32), that appear
as an independent specialization of the reproductive outlets.
Trunk
The shape of the body (character 18) is estimated by the distance
between the lateral processes of each of the segments. Elongate,
slender forms appear to be more common in the group; they are
characteristic of Nymphonidae, Endeis, and most Colossendeidae. Many tenuous forms are also found in Anoplodactylus
(Phoxichilidiidae), and some Ammotheidae which also include
discoid-shaped forms (e.g. Achelia and Tanystylum). A small and
compact body characterizes Rhynchothoracidae and Pycnogonidae. The general appearance of the body of sea spiders has
been related to factors of the physical environment, more elongate forms being common on deeper soft-bottom substrata, and
medium and compact forms generally found in shallow waters
exposed to strong wave action (Arnaud & Bamber 1987). However, this has not been found to be a reliable rule, and any form
can occur in any type of habitat. The segmentation of the trunk
can be clearly distinguished by marked dorsal lines, which is the
general state and presumably ancestral, but many species show
partial or complete absence of segmentation lines. Lack of segmentation is more common in the compact forms although
Colossendeis species, many with well-separated lateral processes, show no signs of trunk segmentation.
The position of the ocular tubercle with respect to the
cephalic segment is explored as a phylogenetic character
(character 23). Its posterior position is a synapomorphy of
Nymphonidae and Callipallenidae, believed to be a specialized state with no biological implications discovered so far.
The shape of the tubercle, although diverse within the group,
is not useful as a phylogenetic character due to frequent intraspecific variation (examples shown in King 1973). An anterior cephalic hood in which the proboscis is embedded occurs
in the ammotheid Cilunculus and has been coded as an autapomorphy for the genus.
The position of the abdomen of pycnogonids is rather consistent within genera. A horizontal position is assumed to be
the plesiomorphic form when compared to other arthropod
groups, it is also the state observed in the fossil P. problematicus. The ancestral condition is present in some taxa, but the
significance of an erect abdomen has yet to be explained. Different degrees of abdominal inclination have been observed
but are all coded as erect if not colinear with the trunk.
Proboscis
The proboscis of pycnogonids has been considered a homologue of the proboscis in polychaetes (Henry 1953 cited in
Morphological phylogenetics of sea spiders
Hedgpeth 1954, Sharov 1966), provoking suggested placement
of the sea spiders closer to the basal arthropodan stock than to
chelicerates (Sharov 1966). So far, there is no evidence that the
proboscis is anything other than the elongated acron
(Boudreaux 1979), a unique specialization within arthropods.
Fry & Hedgpeth (1969) tried to code the different shapes of the
proboscis using a system of geometrical shapes and coordinates. The coding presented in this study is based on the five
main types of proboscis shapes these authors proposed (character 31), using the geometrical criteria but not the system of coordinates (see Electr. Suppl., Pt 2). The particular shape and
length of the proboscis can usually define families and genera.
In Colossendeidae and Austrodecidae, the proboscis is longer
than the trunk (character 33). This is not expected to be a
synapomorphy for these two lineages, but probably an independently attained specialization. A ventral position of the proboscis is described in the fossil species (Bergström et al. 1980).
This position resembles that observed in the basal ammotheid
genera (e.g. Eurycyde, Ascorhynchus and Cilunculus) but also
in the callipallenid form Pseudopallene. Fry (1965) pointed out
the possible phylogenetic relevance of characteristics of the
musculature and internal structure of the proboscis. There is information on only six species from five distinct genera, thus this
character could not be defined in the present analysis. Morphological adaptations to preferred prey, as shown for Austrodecus,
Rhynchothorax and Pycnogonum (Fry 1965), could also be further investigated for evolutionary implications.
Cladistic analysis
A parsimony analysis under an ‘a posteriori weighting’ approach, using the implied weights of the package PeeWee
(Goloboff 1993b), was carried out to produce a phylogeny of
the Pycnogonida based on morphological characters. As a preliminary exploration and for the aim of comparison with the implied weights analysis, similar analyses were done in Nona and
PAUP* version 4.0b.4a (Swofford 2000), entering the characters as equally weighted or ‘unweighted’. These days it is commonly considered that an equally weighted analysis is a preliminary estimate of the relative value of the data (Kitching et al.
1998). Based on Farris’ ideas on character weighting (Farris
1969), Goloboff proposed a non-iterative method that uses evidence on homoplasy to estimate character reliability (Goloboff
1993a). It does not depend on initial estimations of weights, and
produces trees of maximum fit F = ∑ fi, which implies the characters to be maximally reliable (Goloboff 1993a, 1995). The fit
of the character i is measured with fi = k /(k + es), where k is a
constant that changes the concavity of the fitting function to
allow homoplastic characters to have more or less influence,
and es is the number of extra steps. No theoretical justification
exists for selecting a particular k value (Turner & Zandee 1995;
Prendini 2000). However, extreme values of k are not recommended since very mild concavity functions (lowest value of k)
do not differ much from analysis with equal weights, and very
strong functions cannot be justified (Goloboff 1993b). The concavity or k value in this analysis was set to 5, weighting less
strongly against characters with homoplasy (Goloboff 1993a).
When k values of 4 and below were introduced, the analysis resulted in 21 most-parsimonious trees and a decrease of 6–23%
in total fitness compared to the results with k = 5. When the ex-
113
treme value of k = 6 was used, a slight increase in the total fit occurred (1%), but the topologies remained the same as with k = 5.
Prior fits or weights of the characters were scaled to 10 as recommended by Goloboff (1993a, b).
Heuristic searches were run in PeeWee using the commands “hold500; hold/20; mult*50”, i.e. hold 500 trees in
memory; keep 20 starting trees in each replication; perform
Tree-Bisection-Reconnection (TBR) swapping on 50 random
addition replicates. The command “jump” was used for additional swapping among multiple ‘islands’ of trees (Goloboff
1995). This same analysis can be run in PAUP* 4.0 by selecting unambiguous optimization under the parsimony settings
and entering the “Goloboff”, “GPeewee” and “GK” options.
The analysis with equally weighted characters was run in
Nona (under the same commands as PeeWee) and PAUP*.
Heuristic searches employed 20 iterations of random stepwise
addition of the taxa, tree space was sampled using 100 random
addition sequence replicates with three trees sampled per iteration (nchuck = 3, chuckscore = 1) in PAUP*. The resulting
trees were branch-swapped using TBR checking for shorter
resolutions and to fill out tree space (as in Edgecombe et al.
2000). This procedure is computationally less demanding than
the default options for search in PAUP*. Under a strict or unambiguous optimization, only those trees are shown on which
all minimum branch lengths are greater than zero. This is the
default procedure in Nona and PeeWee, but the same conditions of analysis can also be implemented in PAUP* (pset collapse = minbrlen) (G. F. Wilson pers. comm.).
The synapomorphies present in all the dichotomous trees
yielded by PeeWee (total set of trees found using the command
“poly-”) are found using the command “apo” in PeeWee
(Goloboff 1997). Relative degree of support for each node was
also examined by means of the branch support indices (Bremer 1994). Bremer support values up to ten extra steps were
calculated in PeeWee using the command “hold 1000; bsupport10”, i.e. hold a maximum of 1000 trees and branch support
indices up to 10 extra steps.
The results of the implied weights analysis, obtained using
unordered characters, were compared to those obtained with
five multistate characters coded as ordered (characters 1, 5, 7,
8, 10 in Appendix 1) and representing the number of segments
of chelifores, palps and ovigers in males and females. Hypotheses of pycnogonid evolution were investigated by constraining
clades proposed by other authors and comparing them with the
present results using the commands “ref”, “swap”, “mv”, and
“cmp” in PeeWee (as in Szumik 1996, Prendini 2000).
Results
A single most-parsimonious polytomous tree of maximum fit (F = 2391.9 [53%], L = 180) was found using
unordered characters and implied weights (Fig. 2). The
two polytomies – one for Achelia species, the other for
Anoplodactylus species – when uncollapsed produce
nine equally most-parsimonious dichotomous trees. The
strict consensus of the nine trees with these two nodes
collapsed is the basis for the discussion of the pycnogonid relationships presented (Fig. 2).
Org. Divers. Evol. (2002) 2, 107–125
114
Arango
Fig. 2. Single most-parsimonious polytomous tree (F = 2391.9 [53%]; length = 180; ci = 39; ri = 67) obtained from the analysis of unordered
characters and implied weights and presented as the preferred hypothesis of the pycnogonid phylogeny. This is the strict consensus tree of nine
dichotomous resolutions. Solid circles represent synapomorphies present in all nine trees. Numbers above circles identify characters, numbers
below circles the character states (Appendix 1). Nodes indicated by solid squares show a maximum Bremer support value of 3.
Org. Divers. Evol. (2002) 2, 107–125
Morphological phylogenetics of sea spiders
Two major lineages of pycnogonids are obtained: ammotheids grouped with Colossendeidae, Austrodecidae
and Rhynchothoracidae (clade A in Fig. 2), and nymphonids grouped with callipallenids, Pycnogonidae,
Phoxichilidiidae (including Endeis), with Pallenopsis as
a sister-taxon (clade B). Clade A is supported by two
synapomorphies: absence of chelae in the adults (character 2), and size of the proboscis relative to the trunk
(character 33). However, character 33 changes twice
from state 1 to state 2 in Ascorhynchus ramipes and Austrodecus. Clade B is supported by the shape and position
of the proboscis (characters 31 and 34), although these
characters are homoplastic, changing in Anoplodactylus
batangensis and Endeis mollis, respectively. The support
for the monophyly of living pycnogonids, A+B, is given
mainly by the presence of fewer segments in chelifores
and ovigers (characters 1, 8 and 10) in the living taxa
than in the fossil Palaeoisopus.
The ‘unweighted’ analysis resulted in a single mostparsimonious tree (L = 174) (Fig. 3). This tree differs
from the ‘weighted’ tree (Fig. 2) in the basal positions of
Cilunculus and Ammothea, the derived position of Nymphonidae, and by Pycnogonum and Pseudopallene being
placed within one clade.
After converting a subset of five multistates from unordered to ordered (characters 1, 5, 7, 8, 10 in Appendix
1), another set of searches in PeeWee yielded 100 mostparsimonious dichotomous trees. These are summarized
in the strict consensus tree (Fig. 4) showing five collapsed nodes (F = 2361.2 [49%], L = 195). Fit decreased
for eleven characters and increased for three when ordered characters were introduced (Table 2), resulting in
a decrease in total fit of 4%.
Although most of the shallow clades were rather similar between the unordered and ordered analyses, the
deep divergence of the two main clades is not obtained
in the latter. Instead, a chain-like cladogram joins both
major groupings (Fig. 4). A decreased resolution when
the five multistates are coded as ordered (excepting
character 10, which is uninformative as unordered)
makes the alternative treatment of an ‘unordered analysis’ a better choice to represent their possible evolution.
Characters 1, 5, 7, 8 and 10 are traced onto the initial
proposed phylogeny (Fig. 5) to visualize the possible
evolution of these characters according to the resulting
topology in Fig. 2.
Discussion
Goloboff’s method for estimating character weights during tree search (implied weights) is used here to construct a high-level phylogeny of Pycnogonida. Since the
reliability of the characters is a logical implication of the
trees being examined, this approach supports the notion
115
that there is no necessity to estimate weights prior to the
analysis in a parsimony analysis. On the other hand, a
parsimony analysis of equally weighted or ‘unweighted’
data can be regarded as a preliminary estimate of phylogeny and could only be defended with a claim that all
the characters provide equally strong evidence. It is
widely known that such a claim is usually rejected, in
cladistic analyses some characters show a lot of homoplasy while others are perfectly hierarchical (Goloboff
1993a). Topological differences between the weighted
and ‘unweighted’ analyses presented here are due to less
weight given to the homoplastic characters by the implied weights method. A comparison of the consistency
and retention indexes for each character between the implied weighted (iw) and the equally weighted (ew) reflects this action (Table 2).
The cladogram yielded by the implied weights analysis of unordered characters (Fig. 2) is the basis for the
pycnogonid phylogeny discussed below. The internal relationships within each of the two major clades are described according to the conventional families of pycnogonids, which are used throughout the study while their
validity is being tested.
Ammotheidae+Colossendeidae+Rhynchothoracidae+Austrodecidae
Two main groupings are found in clade A indicating the
Ammotheidae as paraphyletic. The ammotheids Eurycyde and Ascorhynchus are grouped basally with
Colossendeidae, supported by the presence of a terminal
claw (character11), multiple rows of spines on the ovigers
(character 27), the mid-position of the ocular tubercle
(character 23), and the shape of the proboscis (character
31), the latter two being reversals. Munilla (1999) showed
the presence of multiple rows of spines on the ovigers as
an autapomorphy of Colossendeidae, also noting – without further discussion – the curious presence of the multiple rows of spines in the ovigers of Ascorhynchus.
Colossendeidae is a highly specialized family as indicated by the eight synapomorphies grouping its two most
conspicuous genera, Colossendeis and Rhopalorhynchus.
It is worth noting that the clade for ammotheids and
colossendeids is characterized by species known from
deeper waters, with the exception of a few Eurycyde and
Rhopalorhynchus species. Eurycyde and Cilunculus
have been considered primitive forms among extant pycnogonids, based on sutures on the proboscis and abdomen believed to be remains of an ancestral segmentation pattern (Stock 1994). Ammothea is presented here
as sister taxon of ((Ammothella+Nymphopsis) (Achelia+
Rhynchothorax + (Tanystylum +Austrodecus))), in contrast to the unresolved position of the genus presented by
Lovely (1999).
At least 40 genera have been placed within Ammotheidae at one time or another. Of these, about 30 are generalOrg. Divers. Evol. (2002) 2, 107–125
116
Arango
Fig. 3. Single most-parsimonious tree (length = 174; ci = 40; ri = 69) obtained with unordered and equally weighted characters. Branch
lengths are shown.
Org. Divers. Evol. (2002) 2, 107–125
Morphological phylogenetics of sea spiders
117
Fig. 4. Single most-parsimonious polytomous tree (F = 2361.2 [49%]; length = 195; ci = 36; ri = 69) obtained with five multistate ordered
characters (1, 5, 7, 8, 10) under implied weights. This is the strict consensus tree of 100 most-parsimonious dichotomous resolutions.
Org. Divers. Evol. (2002) 2, 107–125
118
Arango
Fig. 5. Possible evolution of cephalic appendage segment numbers. A. Chelifores. B. Palps in males. C. Palps in females. D. Ovigers in males. E. Ovigers in
females. Congeneric species collapsed if
not exhibiting different character states.
Org. Divers. Evol. (2002) 2, 107–125
Morphological phylogenetics of sea spiders
119
Table 2. Character statistics for ‘unordered’ and ‘ordered’ analyses. Characters coded as ordered in bold. Consistency index (ci) and retention
index (ri) are given for the analyses with implied weights (iw) and with equal weights (ew). Prior fits or weights of all characters scaled to 10.
Fit values were obtained from the ‘implied weights’ analysis.
Char
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Unordered analysis
-----------------------------------------------------------------------------------------------------------fit
ci (iw)
ri (iw)
ci (ew)
ri (ew)
Ordered analysis
------------------------------------------------------------------------------------------------------–––––-------fit
ci (iw)
ri (iw)
ci (ew)
ri (ew)
55.5
83.3
100.0
71.4
100.0
100.0
100.0
100.0
83.3
100.0
20
66
100
33
100
100
100
100
83
100
55.5
41.6
71.4
62.5
50.0
55.5
55.5
35.7
45.4
50.0
83.3
50.0
55.5
41.6
100.0
62.5
100.0
71.4
71.4
41.6
41.6
55.5
45.4
55.5
100.0
20
12
33
25
16
33
42
18
25
37
50
16
33
36
100
25
100
33
33
22
30
20
25
33
100
55.5
83.3
83.3
71.4
100.0
71.4
100.0
83.3
62.5
100.0
83.3
55.5
38.4
71.4
62.5
50.0
55.5
45.4
33.3
41.6
50.0
83.3
55.5
55.5
41.6
100.0
55.5
100.0
71.4
71.4
41.6
38.4
55.5
38.4
55.5
100.0
33
66
100
33
100
100
100
100
85
100
Uninformative
66
56
33
72
61
63
71
40
33
16
0
61
63
68
100
50
100
71
66
41
63
55
64
50
100
25
66
50
33
50
100
50
85
83
50
50
66
92
33
91
100
92
87
85
87
20
12
50
20
20
33
42
28
25
42
50
20
40
40
50
20
100
33
33
33
33
20
22
40
50
66
56
66
63
69
63
71
66
33
33
0
69
72
72
87
33
100
71
66
66
68
55
58
62
0
ly accepted as valid taxa (Fry & Hedgpeth 1969, Child
1998). The great morphological diversity within the family has led taxonomists to propose some genera as separate
families, e.g. in the case of the Tanystylidae (Schimkewitsch 1913). In the present study, Tanystylum appears as
an ammotheid genus closely related to Achelia, Rhynchothorax and Austrodecus. Some authors fail to recognize a common origin for Achelia and Tanystylum, arguing that there is a lower number of palp segments in
Tanystylum (Hedgpeth 1954, Clark 1977, Munilla 1999).
The results of the present study differ from that position
and agree with Stock’s classification in showing Achelia
20
66
50
33
100
77
100
85
62
100
50
20
11
33
25
16
33
33
16
22
37
50
20
33
36
100
20
100
33
33
22
27
20
20
33
100
33
75
92
33
100
93
100
96
88
100
66
66
50
33
72
61
63
57
33
22
16
0
69
63
68
100
33
100
71
66
41
57
55
52
50
100
25
50
50
33
50
77
50
66
55
50
80
20
12
50
20
20
33
37
25
22
42
50
20
40
40
50
20
100
33
33
28
33
20
20
40
50
50
50
92
33
91
93
92
86
85
87
66
66
56
66
63
69
63
64
60
22
33
0
69
72
72
87
33
100
71
66
58
68
55
52
62
0
and Tanystylum as closely related taxa (Stock 1954).
However, no synapomorphies were found to hold T.
haswelli Child and T. rehderi Child in a single clade. A
similar case is presented by the Ammothella species: Ammothella sp. appendiculata-group is grouped together
with Nymphopsis, solely by the presence of spines in the
chelifores. A single character grouping the two species together is not considered to be strong evidence to modify
the status of Ammothella and Nymphopsis.
The affiliations of Rhynchothorax have long been uncertain. The genus was associated to ammotheid forms
(e.g. Bouvier 1923), considered a genus of the family
Org. Divers. Evol. (2002) 2, 107–125
120
Arango
Tanystylidae (Hedgpeth 1955), or to belong to the
Colossendeidae (see Arnaud & Krapp 1990). More recently, Stock (1994) suggested its affinity to Austrodecus.
Thompson (1904) had given it family status, creating
Rhynchothoracidae which has been widely accepted and
redefined more recently (Arnaud & Bamber 1987, Arnaud & Krapp 1990). According to the present results,
Rhynchothorax appears closely related to the ammotheids
Achelia and Tanystylum. Characteristics of the trunk
(characters 18 and 19), number of palp segments (characters 5 and 7), and the position of the abdomen (character
30) bring Rhynchothorax closer to the small ammotheid
forms. Munilla (1999) had suggested Rhynchothoracidae
to be the sister taxon of Pycnogonidae, based on the presence of a single genital pore in the females of both taxa,
assuming it to be a product of ‘regressive’ evolution. That
pattern of relationship is not obtained in this analysis.
Austrodecidae is a compact and homogeneous family
with highly specialized characteristics, the most remarkable being the pipette-like proboscis and the extreme reduction of the ovigers. Austrodecus species used to be
considered members of Tanystylidae (Hedgpeth 1947),
before Stock created the family Austrodecidae (Stock
1954). In this study the Tanystylidae are rendered paraphyletic by the Austrodecidae. This close relationship is
based on the trend observed in Tanystylum towards a
proboscis tapered downward (character 31), and on the
sharing of simple spines on the ovigers of both taxa
(character 26). The presence of six-segmented palps in
the females of T. rehderi and the species of Austrodecus
shows these taxa to be closely related (character 7), segregating them from T. haswelli and Rhynchothorax in
which females possess four-segmented palps.
A major revision of the relationships of the genera in
Ammotheidae is long overdue (Fry & Hedgpeth 1969). A
revision of the family is a difficult task, mainly because
of the large number of species to be examined and the
scarcity of type material available for rare genera. The relationships presented here are a first impression of what
might have been the course of evolution in this lineage.
(Nymphonidae+Callipallenidae+Pycnogonidae+
Phoxichilidiidae)+Pallenopsis
The shape and the frontal position of the proboscis
(characters 31 and 34) are synapomorphies that segregate clade B from the rest of the Pycnogonida. Nymphonidae appears as a monophyletic group based on the
presence of a terminal claw on the ovigers (character 11)
in the three Nymphon species. This family has been considered a relatively homogeneous lineage of numerous
(ca. 240) closely related species. Nymphon surinamense
exhibits a different position of the male genital pores
(character 20) and was selected as representative of the
species complex with no auxiliary claws (character 12),
contrasting with N. molleri and N. micronesicum.
Org. Divers. Evol. (2002) 2, 107–125
The position of the ocular tubercle (character 23) and
the presence of teeth on the chelae (character 28) relate
Nymphonidae and callipallenid genera. According to
Stock (1994), these two lineages are closely related.
However, an electrophoretical study had shown Nymphonidae as a basal clade of the Pycnogonida distant
from the Callipallenidae (Munilla & De Haro 1981).
This plesiomorphic condition has also been supported
by the idea that Nymphonidae species show a generalized plan of the Pycnogonida closer to that of an arachnid (Hedgpeth 1947). From the information currently
available, it is not possible to ascertain the ancestral conditions of the extant Pycnogonida, until more fossil evidence becomes available. Nevertheless, in this study
Nymphonidae appears as a fairly basal group of pycnogonids at the same time related to the callipallenids.
Males from both Nymphonidae and Callipallenidae
generally possess one cement gland (character 14), except
for some species of Callipallene such as C. brevirostris
with more than one, and the genus Pseudopallene that
shares the absence of cement glands with Pycnogonum
and Colossendeis. The Callipallenidae appear as a paraphyletic group, containing the Phoxichilidiidae, Endeididae and Pycnogonidae, all related by the absence of palps
in the females (character 6). Pseudopallene links the callipallenids to (Pycnogonum+Endeis+Anoplodactylus)
based on the anterior position of the ocular tubercle and
the ventral orientation of the chelae (characters 23 and
29, respectively). The absence of ovigers in the females
(character 9) and the glabrous condition of the ovigers in
males bring together the Pycnogonidae and Phoxichilidiidae. The position of Pycnogonidae as derived from a
callipallenid ancestor reflects the same relationships proposed by Stock (1994) in his diagram (Fig. 1A).
The genus Endeis has been considered by some specialists as a distinct entity that needs to be placed as a
separate family, Endeidae (Hedgpeth 1947, King 1973)
or Endeididae (Child 1992), but Stock (1965) preferred
to include it in the Phoxichilidiidae. The present cladogram shows Endeis as closely related to the Phoxichilidiidae based on the absence of a terminal claw on the
ovigers (character 11). E. mollis appears as sister to
Anoplodactylus longiceps which could be taken as representative of Anoplodactylus species with auxiliary
claws (character 12) and a relatively long proboscis
(character 33). Although E. mollis is grouped with A. longiceps by these two characters, Endeis has clear autapomorphies (Fig. 2) that support it as a taxon separate from
Anoplodactylus. The inclusion of Phoxichilidium, the
most probable sister taxon to Anoplodactylus, in future
analyses might be of help to test the proximity of
Anoplodactylus and Endeis.
The paraphyly of Anoplodactylus might be explained
by the enormous variability within the genus in the characteristics of the cement glands, shape and segmentation
Morphological phylogenetics of sea spiders
of the trunk, size and shape of the proboscis, and a number of other characters. The absence of palps, of ovigers
in females, and the number of segments of the male
ovigers are quite stable among the species, but the internal relationships of the diverse genus Anoplodactylus
are worthy of further detailed studies.
Pallenopsis has been considered a transitional genus
between Callipallenidae and Phoxichilidiidae (Hedgpeth
1947), classified as a callipallenid (see Hedgpeth 1948,
Child 1979) and as a phoxichilidiid (Stock 1978). In the
present analysis Pallenopsis is the basal taxon of clade
B, which node is supported by the presence of a single
cement gland (character 14), the middle position of the
ocular tubercle (character 23), and the absence of teeth
on the chelae (character 28). However, Pallenopsis
shares these characters with taxa from one or the other
family, and a certain variation is also known within the
genus (e.g. toothed chelae in P. mascula Bamber). Child
(1992) has suggested Pallenopsis might deserve familial
rank, this particular data set is agnostic and does not give
indication whether to relate Pallenopsis to Callipallenidae or Phoxichilidiidae.
Character evolution
Presence/absence and features of the head appendages
are the most commonly used characters relating families
and genera of pycnogonids. Since it has been argued that
a gradual reduction of the appendages might have occurred within the group, the assumption of an order in
the evolution from highest to lowest number of appendage segments should show the most-parsimonious
resolution of pycnogonid phylogeny. The present results
do not show support for this argument. The most parsimonious trees were obtained when the number of segments of chelifores, palps and ovigers were coded as unordered. This suggests that there has not been a strict
gradual reduction of the appendages throughout the evolution of the group. The mapping of the characters onto
the proposed phylogeny (Fig. 5 A–E) shows that a trend
of reduction and loss of the appendages occurs independently in each of the two major clades.
Regarding the evolution of the chelifores, according
to the cladogram proposed here, a complete loss of the
chelifores in adults has independently occurred on five
occasions (Fig. 5A). Chelate larval stages and juveniles
are known for most of the taxa in which chelifores are
lost in the adult stage. The functional importance of the
chelae in larvae and juveniles is believed to be related to
their parasitic habits (King 1973, Staples & Watson
1987), but the absence of chelae in adults has not been
discussed in functional or ecological terms.
The absence of palps (characters 4 and 5) appears as
an apomorphic condition relating Pycnogonum, Phoxichilidiidae and callipallenid taxa, although males of a
121
few genera of Callipallenidae (e.g. Propallene) have 1or 2-segmented palps. It could be argued that it was within the callipallenids that sea spiders lacking palps began
to diversify. A reduction in the number of palp segments
is also evident within the ammotheid clade (Achelia+Rhynchothorax+Tanystylum), but complete loss does not
occur (Fig. 5B, C). The reduction or absence of palps
cannot yet be explained in terms of their functional or
ecological significance. The relevance of the palps as
sexually dimorphic features, in some callipallenid taxa
such as Propallene, remains to be studied. Setae and
glands observed on the palps of Nymphon and ammotheid species are believed to be sensory structures used for
the recognition of prey (in King 1973, Arnaud & Bamber
1987). However, for those taxa in which palps are completely absent alternative sensory structures are yet to be
recognized. Detailed developmental and physiological
studies might help to form hypotheses on the significance of the loss of palps, which seems to have occurred
first in the females of callipallenid forms, suggesting also
that sexual dimorphism might be involved.
A reduced number of oviger segments in males is
common to Pycnogonidae and phoxichilidiids; in Austrodecidae not only the numbers of segments are lower
(Fig. 5D), but ovigers are also extremely reduced in
overall size (Stock 1957). A total absence of ovigers has
been observed in species of Pycnogonum (Stock 1968,
Child 1998), as well as in a few Austrodecus (Stock
1991). However, the process of reduction seems to be
different since the size of the ovigers of Pycnogonum
males is not as extreme as in Austrodecus. The number
of oviger segments in females appears as an uninformative character in the analysis (character 10, Fig. 5E). Although it is not possible to identify a common origin for
the reduction of the ovigers in the females, their complete absence (character 9) clearly defines the clade for
Pycnogonidae and Phoxichilidiidae, including Endeis
(Fig. 2). Ecological or functional differences between
species with conspicuous, long ovigers and those with
reduced or absent ovigers are not well studied. However,
males of species lacking ovigers have been observed
carrying the eggs cemented to the ventral side of the
trunk (Child 1998). Again, as well as for the palps, the
reduction of the ovigers especially in the males is shown
as a parallel event in the two major clades of the Pycnogonida.
Previous classifications
The fit and length of the trees obtained in this study were
evaluated against previous classifications. Constraints
were forced onto the tree according to relationships previously suggested by Stock (1994) and by Munilla
(1999), both summarizing the most accepted traditional
classification of the Pycnogonida.
Org. Divers. Evol. (2002) 2, 107–125
122
Arango
Table 3. Clades enforced according to previous classifications, and their fit compared to the phylogeny proposed in this study.
Prior fits or weights of all characters scaled to 1.
Tree
Fit
difference
Characters with better fit
Characters with worse fit
Colossendeidae+Rhynchothoracidae+
Austrodecidae (Stock, 1994)
–4.6
0(2/1.6); 20 (1/0.5); 22(2/1.2);
30(1/0.4)
5(1/1.7); 7(2/2.9); 12(1/0.3);
15(1/0.5); 18(1/0.2); 26(1/0.7);
27(1/1.7); 31(1/0.3)
Rhynchothoracidae+Austrodecidae
(Stock, 1994)
–1.0
0(1/0.7); 20(1/0.5); 22(1/0.5)
7(1/1.7); 26(1/0.7); 31(1/0.3)
Pallenopsis as a phoxichilidiid
(Stock, 1978)
–6.3
22(1/0.5); 29(1/1.2)
4(1/1.7); 6(1/1.7); 9(1/1.7);
14(1/0.7); 15(1/0.5); 25(1/1.7)
Pallenopsis as a callipallenid
(Hedgpeth, 1948; Child, 1979)
–2.3
22(1/0.5)
14(1/0.7); 23(1/0.5); 28(1/0.9)
Pycnogonidae+Rhynchothoracidae
(Munilla, 1999)
–8.5
0(1/0.7); 11(1/0.7); 20(1/0.5);
21(1/1.7); 22(1/0.5)
4(1/1.7); 5(1/1.7); 6(1/1.7);
7(1/1.7); 9(1/1.7); 12(1/0.3);
14(1/0.7); 15(1/0.5); 17(1/0.5);
25(1/1.7); 33(1/0.4)
Endeis+Phoxichilidiidae (Stock, 1994)
–0.7
No better fit for any character
12(1/0.3); 33(1/0.4)
According to his diagram (Fig. 1A), Stock seemingly
proposed a clade Colossendeidae+Rhynchothoracidae+Austrodecidae (Stock 1994). When this group was
enforced, an overall decrease in total fit of 4.6 was observed, eight characters decreasing and four characters
increasing in fit (Table 3). A clade formed by these three
taxa does not explain the data set well. However, when a
clade for Rhynchothorax and Austrodecus was constrained (according to Stock’s diagram), the total fit of
the proposed phylogeny was just slightly affected (Table
3). The absence of chelifores (character 0) is the main
character with a better fit, but the type of spines in the
ovigers (character 26) and the shape of the proboscis
(character 31) are not synapomorphies of Rhynchothorax+Austrodecus (Table 3).
When Pallenopsis was constrained to Phoxichilidiidae, as proposed by Stock (1978), there was a decrease
in total fit of 3% with nine characters having a worse fit.
Then, enforcing Pallenopsis as a callipallenid taxon, the
fitness was just two units lower than the unconstrained
cladogram, but only one character performed better
(Table 3). The basal position of Pallenopsis in clade B
suggests it could be proposed as a higher taxon, but it
might also indicate lack of sufficient informative characters to attach the genus to any of the known taxa.
When the clade Rhynchothoracidae+Pycnogonidae
proposed by Munilla (1999) was enforced, the overall
fitness decreased consistently, with 11 characters showing a worse fit under this constraint. These results show
that the grouping of the two families is far from being
the most explanatory for the current data set. The conOrg. Divers. Evol. (2002) 2, 107–125
straint of Endeis as a sister taxon of the Phoxichilidiidae
revealed a very slight decrease in total fit, but none of
the characters showed a better fit (Table 3). The close relationship between Anoplodactylus and Endeis has been
shown here, and I propose that Endeis be left within the
Phoxichilidiidae until further evidence becomes available to decide whether the two genera are within a single
family or whether the two families should be grouped
within a higher taxon.
A clade for callipallenids, phoxichilidiids (including
Endeis) and Pycnogonidae, as proposed by both Stock
(1994) and Munilla (1999), is supported in this analysis.
However, the monophyly of Callipallenidae and the status of Pallenopsis and Endeis need to be clarified.
Conclusions
The phylogeny proposed shows two main lineages of
extant sea spiders. Ammotheidae is shown as a paraphyletic group including Rhynchothorax, Austrodecus
and Colossendeidae. These are segregated from a clade
combining Callipallenidae, also a paraphyletic group,
with Nymphonidae, Pycnogonidae and Phoxichilidiidae.
The absence of chelae in adults is a main feature supporting the divergence of these two main clades. Unordered, weighted characters provided the most parsimonious and consistent resolution of a pycnogonid phylogeny on the morphological data set used. A strict gradual reduction of the appendages in a manner of ladderlike evolution is not supported. Instead, a trend of reduc-
Morphological phylogenetics of sea spiders
tion of the palps and ovigers has occurred independently
in each of the two major clades.
In general terms, the outcome of this study agrees
with previous classifications, especially the one proposed by Stock (1994), except for his ideas of Pallenopsis as a phoxichilidiid genus, and of a clade consisting of
Colossendeidae, Austrodecidae and Rhynchothoracidae.
Pycnogonid phylogeny is yet not clear. Little work has
been done based on a set of traditional taxonomic characters, and there is a strong need for additional characters from different sources to either contrast the clades
proposed or give a more robust support. The lack of appropriate outgroups for comparison and polarization remains a major problem in attempts to relate pycnogonids
to other arthropod taxa, and to recognize ancestral character states.
Acknowledgments
I thank J. Collins, J. J. Cruz, G. Díaz-Pulido, O. Floerl, A. Lee,
W. Lee-Long, J. Otto, and L. Turner for help in the collection
of specimens, K. Miyazaki, T. Munilla, and K. H. Tomaschko
for donations of specimens and helpful advice, F. Krapp for
discussion and comments. Thanks to G. F. Wilson and D. K.
Yeates for comments and advice in the analysis of data. Special thanks to A. Child for his taxonomic expertise, and to R.
Rowe, D. Blair and the editors and reviewers, whose comments on earlier versions of this manuscript resulted in considerable improvements. I am grateful to the School of Tropical
Biology, James Cook University, for financial support.
References
Arango, C. P. (2000): Three species of sea spiders (Pycnogonida) from Santa Marta, Colombian Caribbean. Bol. Invest.
Mar. Cost. 29: 59–66.
Arnaud, F. & Bamber, R. N. (1987): The biology of Pycnogonida. Adv. Mar. Biol. 24: 1–95.
Arnaud, F. & Krapp, F. (1990): The genus Rhynchothorax (Pycnogonida) in the Mediterranean Sea. Beaufortia 41: 1–5.
Bergström, J., Stürmer, W. & Winter, G. (1980): Palaeoisopus,
Palaeopantopus and Palaeothea, pycnogonid arthropods
from the Lower Devonian Hunsrück Slate, West Germany.
Paläont. Z. 54: 7–54.
Bininda-Emonds, O. R. P., Bryant, H. N. & Rusell, A. P.
(1998): Supraspecific taxa as terminals in cladistic analysis:
implicit assumptions of monophyly and a comparison of
methods. Biol. J. Linn. Soc. (London) 64: 101–133.
Boudreaux, H. B. (1979): Arthropod Phylogeny with Special
Reference to Insects. 320 pp., John Wiley & Sons, New York.
Bouvier, E.-L. (1923): Faune de France. 7. Pycnogonides. Office Central de Faunistique, Paris.
Bremer, K. (1994): Branch support and tree stability. Cladistics 10: 295–304.
Child, A. C. (1979): Shallow water Pycnogonida of the Isthmus of Panama and the coasts of Middle America. Smiths.
Contr. Zool. 23: 1–86.
123
Child, A. C. (1992): Shallow water Pycnogonida of the Gulf of
Mexico. Mem. Hourglass Cruises 9: 1–86.
Child, A. C. (1998): Pycnogonida (sea spiders). NIWA Biodiversity 109: 1–46.
Clark, W. C. (1977): The Genus Tanystylum Miers, 1879 in the
southern oceans. J. R. Soc. NZ 7: 313–338.
Edgecombe, G. D., Wilson, G. D. F., Colgan, D. J., Gray, M.
R. & Cassis, G. (2000): Arthropod cladistics: combined
analysis of histone H3 and U2 snRNA sequences and morphology. Cladistics 16: 155–203.
Farris, J. S. (1969): A successive approximations approach to
character weighting. Cladistics 18: 374–385.
Fry, W. G. (1965): The feeding mechanisms and preferred
foods of three species of Pycnogonida. Bull. Brit. Mus. Nat.
Hist., Zool. 12: 195–233.
Fry, W. G. (1978): A classification within the pycnogonids.
Zool. J. Linn. Soc. 63: 35–38.
Fry, W. G. & Hedgpeth, J. W. (1969): Pycnogonida: Colossendeidae, Pycnogonidae, Endeidae, Ammotheidae. NZ DSIR
Bull., Part 7: 1–139.
Giribet, G., Edgecombe, G. D. & Wheeler, W. C. (2001):
Arthropod phylogeny based on eight molecular loci and
morphology. Nature 413: 157–161.
Giribet, G. & Ribera, C. (2000): A review of arthropod phylogeny: new data based on ribosomal DNA sequences and
direct character optimization. Cladistics 16: 204–231.
Goloboff, P. A. (1993a): Estimating character weights during
tree search. Cladistics 9: 83–91.
Goloboff, P. A. (1993b): Pee-Wee 3.0. Computer software and
documentation distributed by the American Museum of
Natural History, New York.
Goloboff, P. A. (1995): Parsimony and weighting: a reply to
Turner and Zandee. Cladistics 11: 91–104.
Goloboff, P. A. (1997): Nona 2.0. Computer software and documentation distributed by the American Museum of Natural
History, New York.
Hammen, L. van der (1986): On some aspects of parallel evolution in Chelicerata. Acta Biotheor. 35: 15–37.
Hauser, D. L. & Presch, W. (1991): The effect of ordered characters on phylogenetic reconstruction. Cladistics 7: 243–265.
Hedgpeth, J. W. (1947): On the evolutionary significance of
the Pycnogonida. Smiths. Misc. Coll. 106: 1–53.
Hedgpeth, J. W. (1948): The Pycnogonida of the western North
Atlantic and the Caribbean. Proc. U.S. Natl Mus. 98: 157–342.
Hedgpeth, J. W. (1954): On the phylogeny of the Pycnogonida. Act. Zool. 35: 193–213.
Hedgpeth, J. W. (1955): Pycnogonida. Pp. 163–170 in: Moore, R.
C. (ed.) Treatise on Invertebrate Palaeontology. Geological Society of America and University of Kansas Press, Lawrence.
King, P. E. (1973): Pycnogonids. 144 pp., Hutchinson & Co.,
London.
Kitching, I. J., Forey, P. L., Humphries, C. J. & Williams, D.
M. (1998): Cladistics. The Theory and Practice of Parsimony Analysis. 228 pp., Oxford Science Publications, Oxford.
Lee, D.-C. & Bryant, H. (1999): A reconsideration of the coding of inapplicable characters: assumptions and problems.
Cladistics 15: 373–378.
Lovely, E. C. (1999): Evolution of Pycnogonid Life History
Traits. PhD dissertation, University of New Hampshire,
Durham, NH, U.S.A.
Org. Divers. Evol. (2002) 2, 107–125
124
Arango
Munilla, T. (1999): Evolución y filogenia de los picnogónidos.
Pp. 273–279 in: Melic, A., De Haro, J. J., Méndez, M. &
Ribera, I. (eds) Evolución y Filogenia de Arthropoda. Sociedad Entomológica Aragonesa (SEA), Zaragoza.
Munilla, T. & De Haro, A. (1981): An electrophoretical and
immunological study of Pycnogonida with phylogenetic
considerations. Bijdr. Dierkd. 51: 191–198.
Prendini, L. (2000): Phylogeny and classification of the superfamily Scorpionoidea Latreille, 1802 (Chelicerata, Scorpiones): An exemplar approach. Cladistics 16: 1–78.
Prendini, L. (2001): Species or supraspecific taxa as terminals
in cladistic analysis? Groundplans versus exemplars revisited. Syst. Biol. 50: 290–300.
Schimkewitsch, W. (1913): Ueber Regeneration bei Pantopoden. Bull. Acad. Imp. Sci. St.Petersburg 1913: 1147–1156.
Sharov, A. G. (1966): Basic Arthropodan Stock: with Special
Reference to Insects. 271 pp., Pergamon Press, Glasgow.
Snodgrass, R. E. (1938): Evolution of the Annelida, Onychophora and Arthropoda. Smiths. Misc. Coll. 97: 1–159.
Snodgrass, R. E. (1952): The Pycnogonida. Pp. 48–58 in:
Snodgrass, R. E. (ed.) A Textbook of Arthropod Anatomy.
Hafner Publishing Company, New York.
Staples, D. A. & Watson, J. E. (1987): Associations between
pycnogonids and hydroids. Pp. 215–226 in: Bouillion, J.
(ed.) Modern Trends in the Systematics, Ecology and Evolution of Hydroids and Hydromedusae. Oxford University
Press, Oxford.
Stock, J. H. (1954): Papers from Dr. Mortensen’s Pacific
Exped. 1914–1916 LXXVII. Pycnogonida from the IndoWest Pacific, Australian, New Zealand waters. Vidensk.
Medd. fra Dansk Naturh. Foren. 116: 1–168.
Stock, J. H. (1957): The pycnogonid family Austrodecidae.
Beaufortia 6: 1–81.
Stock, J. H. (1965): Pycnogonida from the southwestern Indian Ocean. Beaufortia 151: 13–33.
Stock, J. H. (1968): Pycnogonida collected by the Galathea
and Anton Bruun in the Indian and Pacific Oceans. Vidensk.
Medd. fra Dansk naturh. Foren. 137: 7–65.
Stock, J. H. (1978): Abyssal Pycnogonida from the north-eastern Atlantic Basin, part 1. Cah. Biol. Mar. 19: 189–219.
Stock, J. H. (1989): Pycnogonida collected in interstitia of
coral sand and rubble in the Netherlands Antilles. Bijdr.
Dierkd. 59: 87–96.
Org. Divers. Evol. (2002) 2, 107–125
Stock, J. H. (1991): Deep-water Pycnogonida from the surroundings of New Caledonia. Mem. Mus. Natl Hist. Nat.
(A) 151: 125–212.
Stock, J. H. (1994): Indo-west pacific Pycnogonida collected by
some major oceanographic expeditions. Beaufortia 1: 17–77.
Strong, E. & Lipscomb, D. (1999): Character coding and inapplicable data. Cladistics 15: 363–371.
Swofford, D. L. (2000): PAUP*: Phylogenetic Analysis Using
Parsimony (*and other methods), version 4. Computer software and documentation distributed by the Illinois Natural
History Survey. Sinauer Associates, Sunderland, Massachusetts.
Szumik, C. (1996): The higher classification of the order Embioptera: A cladistic analysis. Syst. Biol. 12: 41–64.
Thompson, D. W. (1904): Pycnogonida. Pp. 501–542 in:
Harmer, S. F. & Shypley, A. E. (eds) The Cambridge Natural History. Wheldon & Wesley, Ltd., England.
Turner, S. J. & Zandee, R. (1995): The Behaviour of Goloboff’s tree fitness measure F. Cladistics 11: 57–72.
Walossek, D. & Dunlop, J. A. (in press): A larval sea spider
(Arthropoda: Pycnogonida) from the upper Cambrian
‘Orsten’ of Sweden, and the phylogenetic position of pycnogonids. Palaeontology.
Walossek, D. & Müller, K. J. (1997): Cambrian ‘Orsten’-type
arthropods and the phylogeny of Crustacea. Pp. 139–153 in:
Fortey, R. A. & Thomas, R. H. (eds) Arthropod Relationships. Chapman & Hall, London.
Weygoldt, P. (1986): Arthropod interrelationships – the phylogenetic systematic approach. Z. Zool. Syst. Evolut. 24: 19–35.
Wheeler, W. C., Cartwright, P. & Hayashi, C. (1993): Arthropod phylogeny: a combined approach. Cladistics 9: 1–39.
Wheeler, W. C. & Hayashi, C. (1998): The phylogeny of extant chelicerate orders. Cladistics 14: 173–192.
Wilkinson, M. (1992): Ordered versus unordered characters.
Cladistics 8: 375–385.
Yeates, D. K. (1995): Groundplans and exemplars: paths to the
tree of life. Cladistics 11: 343–357.
Zrzavy, J., Hypsa, V. & Vlásková, M. (1997): Arthropod phylogeny: taxonomic congruence, total evidence and conditional combination approaches to morphological and
molecular data sets. Pp. 97–107 in: Fortey, R. A. & Thomas,
R. H. (eds) Arthropod Relationships. Chapman & Hall,
London.
Morphological phylogenetics of sea spiders
125
Appendix 1
Morphological characters, states, and scores for
phylogenetic analysis
Characters referred to in the list as “Ordered” were considered
as such only when indicated in the analysis.
0. Chelifores: present (0); absent (1).
1. (Where applicable.) Number of segments of the chelifore
scape: 3-segmented (0); 2-segmented (1); 1-segmented
(2). Ordered.
2. Chelae: present (0); absent (1).
3. (Where applicable.) Chelifores, dorsal spines on scape:
absent (0); present (1).
4. Palps, in males: present (0); absent (1).
5. Palps, number of segments in males: 10-seg. (0); 9- seg.
(1); 8- seg. (2); 6- seg. (3); 5- seg. (4); 4- seg. (5); 2- seg.
(6); 1- seg. (7). Ordered.
6. Palps, in females: present (0); absent (1).
7. Palps, number of segments in females: 10- seg. (0); 9- seg.
(1); 8- seg. (2); 6- seg. (3); 5- seg. (4); 4- seg. (5); 1- seg.
(6). Ordered.
8. Ovigers, number of segments in males: 11- seg. (0); 10seg. (1); 9- seg. (2); 7- seg. (3); 6- seg. (4); 4- seg. (5). Ordered.
9. Ovigers, in females: present (0); absent (1).
10. Ovigers, number of segments in females. 11- seg. (0); 10seg. (1); 9- seg. (2); 6- seg. (3); 4- seg. (4). Ordered.
11. Ovigers, terminal claw: absent (0); present (1).
12. Propodi, auxiliary claws: absent (0); present (1).
13. Cement gland(s): present on femora or other leg segments
(0); completely absent (1).
14. Cement gland(s), on femora: one on each femur (0); multiple on each femur (1).
15. Cement gland shape: pore(s) or slit on the cuticle (0);
tube(s) or protuberances (1).
16. Cement gland position on legs: dorsal (0); lateral (1); ventral (2).
17. Cement gland(s), on femora: located distally (0); at midpoint (1); proximal (2); distributed all along femora (3).
18. Trunk: elongate shape, crurigers or lateral processes separated by distance at least equal to their own diameter (0);
intermediate shape: crurigers separated by less than their
own diameter, but never touching (1); compact shape:
crurigers touching (2).
19. Trunk: distinctly segmented, the three lines of segmentation dorsally visible (0); partially segmented, only one or
two lines visible (1); lines of segmentation not distinct (2).
20. Genital pores in males: present on all four pairs of legs
(0); present on second, third and fourth pairs of legs (1);
present on third and four pairs of legs (2); present on
fourth pair of legs only (3).
21. Genital pores in females: present on all four pairs of legs
(0); present on fourth pair of legs only (1).
22. Propodi, heel spines: present (0); absent (1).
23. Ocular tubercle: anterior on cephalon (0); equidistant to anterior and posterior margins (1); posterior on cephalon (2).
24. Ovigers, largest segment: sixth (0); fifth (1); fourth (2);
third (3); second (4).
25. Ovigers, spines on last segments: present (strigilis) (0);
absent (1).
26. Ovigers, spines: compound or denticulate (0); simple (1).
27. Ovigers, spine arrangement: in single row (0); in multiple
rows (1).
28. Chelae, teeth: present (0); absent (1).
29. Chelae, orientation: opposing each other (0); pointing
downwards, in front of tip of proboscis (1).
30. Abdomen: horizontal in same direction as trunk (0); erect
diagonally or pointing upwards (1).
31. Proboscis, shape (see Electr. Suppl., Pt 2): A = straight (0);
B = inflated proximally, acute distally (1); C = inflated
distally (2); D = tapering or pipette-like (3).
32 Second coxae of last pairs of legs in males, ventral spurs:
absent (0); present (1).
33. Proboscis, length: less than half length of trunk (0); equal
(± 1mm) to half length of trunk (1); at least equal to trunk
length (2).
34. Proboscis, position: frontal and fixed (0); in angle and
movable (1); ventral and highly movable (2)
35. Cephalic hood: present (0); absent (1).
Org. Divers. Evol. (2002) 2, 107–125