Content uploaded by Wayne Maddison
Author content
All content in this area was uploaded by Wayne Maddison on May 30, 2014
Content may be subject to copyright.
1255
Accepted by N. Scharff: 12 Jun 2006; published: 10 Jul. 2006 37
ZOOTAXA
ISSN 1175-5326 (print edition)
ISSN 1175-5334 (online edition)
Copyright © 2006 Magnolia Press
Zootaxa 1255: 37–55 (2006)
www.mapress.com/zootaxa/
Lapsiines and hisponines as phylogenetically basal salticid spiders
(Araneae: Salticidae)
WAYNE P. MADDISON1 & KAREN M. NEEDHAM2
1Departments of Zoology and Botany and Centre for Biodiversity Research, University of British Columbia,
6270 University Boulevard, Vancouver, British Columbia, V6T 1Z4, Canada.
E-mail: wmaddisn@interchange.ubc.ca
2Spencer Entomological Museum, Department of Zoology, University of British Columbia, 6270 University
Boulevard, Vancouver, British Columbia, V6T 1Z4, Canada. E-mail: needham@zoology.ubc.ca
Abstract
Increased phylogenetic resolution of the basal lineages of salticid spiders will help us understand
their early evolution and provide better outgroups for phylogenetic studies within the major clades.
We gathered sequences of nuclear and mitochondrial gene regions (28S, 18S, Histone 3, 16S-ND1,
CO1) and used them to reconstruct salticid phylogeny by parsimony, likelihood and Bayesian
methods. Our results confirm that lapsiines and hisponines are among the basal salticids, i.e. outside
the major clade Salticoida. The lapsiines are resolved as sister group to the spartaeines. The precise
placement of hisponines is unclear, but they may represent a deep-branching lineage independent
from the spartaeines.
Key words: Araneae, Salticidae, Thrandina, Galianora, Hispo, Massagris, Tomocyrba, Goleba,
lapsiines, Hisponinae, Spartaeinae, Lyssomaninae, jumping spider, basal groups, phylogeny
Introduction
Morphological and molecular data have begun to resolve the basal phylogenetic structure
of salticid spiders (Wanless, 1980, 1982, 1984, Rodrigo & Jackson, 1992, Maddison, 1988,
1996, Wijesinghe, 1992, 1997, Maddison & Hedin, 2003). One of the best corroborated
clades is the Salticoida (Maddison & Hedin, 2003), within which falls the vast majority of
salticids, about 95% of the approximately 5000 described species (Platnick, 2005).
Excluded from the Salticoida are three much smaller groups: the lyssomanines, the
spartaeines, and the Cocalodes group. Six extant Old World and 2 New World genera are
placed in the Lyssomaninae (Wanless, 1980, Logunov, 2004); 15 genera, entirely from the
Old World, are placed in the Spartaeinae (Wanless, 1984, Wijesinghe, 1992, abka &
MADDISON & NEEDHAM
38 © 2006 Magnolia Press
1255
ZOOTAXA Kovac, 1996); 4 Old World genera are in the Cocalodes group (Wanless, 1982, 1985). We
will refer to these non-salticoid groups as "basal salticids", not to imply that they have
predominantly primitive characteristics, merely to indicate that they fall outside of the
speciose clade Salticoida.
Basal salticids offer special insight into early salticid evolution for two reasons. First,
if the basal salticids do not form a single clade, but diverged successively from the line
leading to the Salticoida, then they will have a strong influence over inference of the
family's ancestral states, whether parsimony or likelihood methods are used. Second, even
if basal salticids do form a single clade, which therefore would stand equal to the
Salticoida as an indicator of ancestral states, each sampled basal species would have more
influence on ancestral state inference than each sampled salticoid species. For these
reasons the unusual predatory behaviour (Jackson & Pollard, 1996, Li, 2000) and eye
anatomy (Blest et al., 1990) of basal salticids are of particular interest in understanding the
origins of salticid diversity. By recognizing what species are among the basal groups and
how they are related, we will be able to characterize better the early evolution of salticids.
In addition, we will have more complete outgroup information for reconstructing
phylogeny within the Salticoida.
In this paper we present molecular data to examine whether two little-studied groups
of salticids might belong with the lyssomanines and spartaeines as basal salticids: the
lapsiines (Maddison, 2006) and hisponines (Wanless, 1981, Prószy ski & abka, 1983,
Weso owska, 1993). Salticid systematists have made few comments on where these groups
belong. Simon (1901) apparently considered lapsiines and hisponines relatively primitive,
but he also considered various salticoids as equally primitive. Simon conferred suggestive
names on one lapsiine ("Lapsias cyrboides") and one hisponine ("Tomocyrba") that hint to
similarities with a spartaeine, Cyrba Simon. Prószy ski & abka (1983) placed
Tomocyrba Simon within the Euophryinae on the basis of the spiraled embolus.
The neotropical lapsiines include Lapsias Simon and two recently discovered genera
(Maddison, 2006). Their basal phylogenetic placement is suggested by the presence of a
tarsal claw on the female palpus, loss of which is considered a synapomorphy of the
salticoids (Maddison & Hedin, 2003). In addition, the male palp has an extra sclerite
associated with the tegulum, presumably homologous to the median apophysis of
Cocalodes Pocock and Holcolaetis Simon (Wanless, 1982, 1985). The salticoids and
spartaeines are characterized by the loss of this sclerite. If the lapsiines are confirmed as
basal salticids, it would show that the New World has a previously unrecognized radiation
of basal salticids.
Hisponines are Old World salticids primarily from Africa. The three extant genera—
Hispo Simon (Wanless, 1981), Massagris Simon (Weso owska, 1993) and Tomocyrba
Simon (Prószy ski & abka, 1983)—are distinctive for a constriction on the carapace just
behind the posterior median eyes. Our attention was drawn to hisponines by two
observations. First, as in the lapsiines, the tegulum in many species has a small sclerite that
© 2006 Magnolia Press 39
SALTICIDAE
1255
ZOOTAXA
may be homologous to the median apophysis of other basal salticids. Second, Baltic amber
salticids are dominated by two body forms, the first with the small eyes unusually large
(placing them among the basal salticids), the second with a distinctive constriction behind
the small eyes. As implied by Prószy ski & abka (1983), this constriction suggests that
the Baltic amber genera Gorgopsina Petrunkevitch and Prolinus Petrunkevitch are
hisponines. If Baltic amber has hisponines but not salticoids, hisponines may have a
relatively ancient divergence from other salticids.
We therefore obtained molecular data from hisponines and lapsiines to determine
whether, as hinted by their morphology, they lie outside the Salticoida. We also include
data from two nuclear genes not previously used in salticids (18S, Histone 3) and for
several other genera not previously studied.
Material and methods
Taxon sampling
In addition to sequences obtained previously (Maddison & Hedin, 2003), we analyzed
sequences obtained from 26 species (Appendix 1). The primary targets of the study were
lapsiines of the genera Thrandina Maddison and Galianora Maddison (Maddison, 2006)
and hisponines of the genera Hispo, Massagris, and Tomocyrba. In addition, we sequenced
two genera of lyssomanines (Lyssomanes Hentz, Goleba Wanless), two genera of
spartaeines (Portia Karsch, Phaeacius Simon), Holcolaetis, ten salticoids, and four
outgroups. The sampled species of Thrandina, Galianora and Lyssomanes include the type
species of their genera; those of the other hisponine, lyssomanine and spartaeine genera
are closely similar to the respective type species except Massagris schisma Maddison &
Zhang and Tomocyrba andasibe Maddison & Zhang. The salticoids were chosen to
include representatives of the major groups recognized by Maddison & Hedin (2003):
marpissoids (Dendryphantes C.L.Koch, Ghelna Maddison), euophryines (Euophrys
C.L.Koch, Zenodorus Peckham & Peckham), amycoids (Sitticus Simon), plexippoids
(Evarcha Simon, Pellenes Simon), and two unplaced genera (Salticus Latreille, Orthrus
Simon). Whether or not the salticoid species sampled are type species of their higher taxa
or close thereto is not vital as long as they are well scattered among the salticoids. This is
demonstrated in our results (e.g., Figure 10). Additional sequences from Maddison &
Hedin (2003) were added to most analyses. Authors of species names are listed in
Appendix 1. Voucher specimens are accompanied by labels with their voucher numbers
(as shown in the second column of Appendix 1); the label is marked "WPM voucher
DNA".
We were able to identify specimens to species with a few exceptions. Several are
described by Maddison (2006) and Maddison & Zhang (2006). Our Hispo (called here
Hispo cf. frenata) and one of the Massagris (called here Massagris cf. honesta) may
MADDISON & NEEDHAM
40 © 2006 Magnolia Press
1255
ZOOTAXA represent undescribed species, but without a thorough taxonomic review we are reluctant
to describe them. Instead we offer figures of the genitalia (Figs. 11–12). The specimen of
Holcolaetis is an immature female. Its markings are similar to adults of H. zuluensis, found
in the same region. The specimen of Orthrus is one of the two used by Maddison & Hedin
(2003). We resequenced its 28S because of concern that there may have been errors in the
previous Orthrus sequence, suggested by its instability in their phylogenetic analyses. The
"unidentified spartaeine" studied by Maddison & Hedin (2003) has since been identified as
Spartaeus uplandicus Barrion & Litsinger.
Sequencing
Specimens were preserved in 95% or 100% ethanol. From most specimens DNA was
extracted from legs, although for some, other body parts were used. Otherwise-intact
spiders are preserved as voucher specimens in the Spencer Entomological Museum, UBC,
except for specimens borrowed from and returned to the California Academy of Sciences
(see Appendix 1). Genomic DNAs were extracted from tissues using the QIAGEN
DNeasy extraction kit.
Five gene regions were amplified by PCR and sequenced: the nuclear 28S (primers
28SO and 28SC from Hedin & Maddison, 2001), 18S (amplified and sequenced in three
pieces by the primer pairs 1F-5R, 3F-7R, 4F-9R, from Giribet et al., 1996), Histone 3
(primers H3AF and H3AR from Colgan et al., 1998), and the mitochondrial 16S-ND1
(primers 12261 and 13398 from Hedin & Maddison, 2001) and CO1 (primers 1718 and
2776 from Hedin & Maddison, 2001). Histone 3 will hereafter be referred to as "H3". The
16S-ND1 fragment includes a transfer RNA in addition to parts of 16S and ND1 (Hedin &
Maddison, 2001). PCR involved an initial 95°C denaturation followed by 35 cycles of 45 s
at 95°C, 45 s at either 48–50°C (28S, CO1, H3), 44–48°C (16S-ND1), or 48°C (18S), 60 s
at 72°C per cycle, with a final 10-min extension at 72°C.
Sequencing was done either by the University of British Columbia NAPS facility
(ABI 377 sequencer) or by Macrogen, Inc. (ABI 3730 sequencer). The primers used in
PCR were also used for sequencing. Three additional primers were used for Galianora
sacha 28S to bypass a region that caused the sequencing reaction to stutter ("28S-G1F
Forward" 5'-CGA AGG CAG TGC CTC ACG CCT G-3'; "28S-G2F Forward" 5'-CGC
ACA CGT TGG GAC CCG AAA G-3'; "28S-G1R Reverse" 5'-CAG GCG TGA GGC
ACT GCC TTC G-3'). Most base calls in all sequences were confirmed by reads in both
directions except 16S-ND1 in Thrandina parocula, Galianora bryicola, and Massagris cf.
honesta. For these three sequences only a single, although high quality, read was obtained
and used.
Sequences were obtained from the chromatogram files using phred (Ewing & Green,
1998, Ewing et al., 1998, Green & Ewing, 2002) and phrap (Green, 1999) as operated via
the chromaseq package (D. Maddison & W. Maddison, in prep.) for Mesquite (W.
© 2006 Magnolia Press 41
SALTICIDAE
1255
ZOOTAXA
Maddison & D. Maddison, 2005). Phred was run with default options to read bases and
assign quality scores; phrap was used with options "-qual_show 20 -vector_bound 0" to
assemble the reads into contigs. Sequence ends were trimmed by chromaseq using a
moving window analysis: the first window of 10 bases within which at least 6 were above
quality score 20 was used as the start or end of the sequence. If a site had a secondary peak
at least 0.3 the height of the primary peak, it was treated as ambiguous. Following this
automatic processing and contig assembly, the resulting sequences were compared against
the chromatograms by eye using chromaseq, to reconsider ambiguities and trimming. This
resulted in few changes to the sequences, primarily adjustments of the extent of trimmed
ends.
Proofread sequences were imported into ClustalX version 1.83 (Thompson et al.,
1997) for alignment. Multiple alignments were carried out with gap opening/gap extension
costs set to 24/6 following Maddison & Hedin (2003). Following automatic alignment,
minor editing of the alignments for the ribosomal sequences was done manually using
MacClade (D. Maddison & W. Maddison, 2005). Manual editing was restricted to
adjusting a few obviously misaligned regions near the sequence ends. Final sequences
were submitted to GenBank (accession numbers in Appendix 1). The aligned all-genes
matrix used in the Bayesian analysis for Figure 2 is deposited in TreeBASE (treebase.org).
Our sequences were supplemented with those of Maddison & Hedin (2003) for 9 of
our taxa. Maddison & Hedin's sequences of 28S, 16S-ND1, and CO1 were used for
Lyssomanes and each of the outgroups; their sequence of 16S-ND1 was used for Salticus;
and their sequences of CO1 were used for Evarcha and Orthrus. Their sequences of 16S-
ND1 and CO1 were used for Pellenes, although this created a chimeric taxon (our
sequences are from P. peninsularis, theirs from the closely related P. shoshonensis
Gertsch). None of the other taxa are species chimeras. (The Evarcha used by Maddison &
Hedin, reported as E. hoyi, was almost certainly E. pr szy skii as is ours, although their
voucher specimen has not been relocated to confirm this.)
Phylogenetic analysis
Phylogenetic analyses were performed on the gene regions separately and combined, and
using various criteria (parsimony, likelihood, Bayesian posterior probability).
Phylogenetic analyses were done on both a large and a small taxon sample. The small
taxon sample included 6 additional taxa from the data of Maddison & Hedin (2003):
Portia labiata Thorell, Spartaeus spinimanus Thorell, S. uplandicus, Helvetia cf. zonata
Simon, Naphrys pulex Hentz, and Terralonus mylothrus Chamberlin. The latter three
salticoids were included because they belong in subfamilies for which some genes would
be otherwise unrepresented. For instance, our Heliophanus sequences are 28S, 18S, and
H3; the heliophanine Helvetia from Maddison & Hedin provides in addition 16S-ND1 and
CO1, and also has 28S to provide the "glue" that would join the two heliophanine taxa
MADDISON & NEEDHAM
42 © 2006 Magnolia Press
1255
ZOOTAXA together in the course of analyses combining genes. We performed small taxon sample
analyses for 28S, 18S, H3, 16S-ND1, and CO1 gene regions separately and together in
various combinations. 16S-ND1 was used as a single partition not because of biological
homogeneity (it is a mix of protein coding and non-coding regions) but because it is
amplified and sequenced as a piece, and we wished to explore its value as a region to
sequence in the future. We also performed analyses on a fused ND1-CO1 matrix translated
to protein assuming the Drosophila mitochondrial genetic code. In analyses combining
more than one gene region a taxon was deleted if sequences were lacking in some genes so
as to give it 75% or more missing data.
The advantage of including few salticoids in the small taxon sample was that we could
focus our efforts on sequencing more gene regions; the disadvantage was that even if we
showed that the lapsiines and hisponines were outside the clade of sampled salticoids, they
might still lie inside the clade of all salticoids. We mitigated this by choosing a broad
sample of salticoids, but we also performed analyses on a much larger sample of taxa,
though for fewer genes. The large sample included all remaining sequences of Maddison
& Hedin (2003) for 28S, 16S-ND1, and CO1, with the exception of 28S in Evarcha,
Orthrus, and Salticus and 16S-ND1 in Evarcha for which we used sequences presented
here.
Bayesian analyses were done using MrBayes 3.1.1 (Huelsenbeck & Ronquist, 2001,
Ronquist & Huelsenbeck, 2003). The GTR invariant-gamma model was used throughout
(nst=6 rates=invgamma). Models were permitted to vary among data partitions, with the
partitions defined as follows: 28S; 18S; H3; 16S; ND1 first, second, and third codon
positions; CO1 first, second, and third codon positions. When ND1 and CO1 were
included together in an analysis, they were united to yield 3 partitions (ND1+CO1 first,
second, and third codon positions) instead of 6. Each analysis was run for 10 million
generations via the command "mcmcp ngen= 10000000 printfreq=1000 samplefreq=1000
nchains=4 savebrlens=yes;", except for the individual analyses for CO1 and H3 which ran
for only 5 million generations. Results are summarized by a majority rules consensus tree
of sampled trees, having discarded generations up to 100,000 (by which point the posterior
probabilities had stabilized), or up to 50,000 for the shorter CO1 and H3 runs.
Likelihood analyses were done using both RAxML-VI (Stamatakis et al., 2005,
Stamatakis 2005) and PAUP* (Swofford, 2002). RAxML was used in general, with PAUP
used in addition for the two All Genes analyses. With RAxML, most analyses were done
by 20 separate runs with the default standard hill climbing, choosing the run with the best
likelihood. For the All Genes analyses, however, 100 separate standard runs were done,
and also simulated annealing was used with a time limit of 40 hours; the run with the best
likelihood was chosen. The model GTRCAT was used for nucleotide matrices, JTTCAT
for amino acid matrices. With PAUP*, we used the same procedure as Maddison & Hedin
(2003), with model parameters estimated on preliminary trees followed by 5 random
addition sequence searches.
© 2006 Magnolia Press 43
SALTICIDAE
1255
ZOOTAXA
Parsimony analyses were done using PAUP* (Swofford, 2002), treating character
states as unordered (one step for any state change). Gaps were treated as missing data. For
the small taxon sample analyses, initial searches consisted of 5,000 random addition
sequence replicates, each saving at most 5 trees with each replicate in order to narrow the
search (D. Maddison, 1991), using TBR branch swapping (Swofford, 2002). Only most
parsimonious trees found were saved. These were used as input trees into a second round
of TBR branch swapping that was not constrained except by MAXTREES of 100,000. For
the large taxon sample analysis combining 28S + 16S-ND1 + CO1, the search was the
same except that the initial phase consisted of 20,000 random addition sequence replicates.
For the combined All Genes data matrix with the small taxon sample, replicability of
clades was assessed by a PAUP* non-parametric bootstrap analysis (Felsenstein, 1985)
with 1000 replicates, on each of which 20 random addition sequence replicates obtained
starting trees for TBR branch swapping holding no more than 1000 trees.
Results
Seventy nine sequences were obtained from 26 taxa and five gene regions; they are listed
in Appendix 1. Our new Orthrus sequence for 28S is different at many sites from that
obtained by Maddison & Hedin (2003) from a different specimen from the same locality
(both vouchers reexamined, males, and confirmed as same species). We doubt that the
differences represent polymorphism within the population; we have no explanation for
this. Because of technological improvements we have more confidence in our current
sequence and will use that.
The results of the phylogenetic analyses on the small taxon sample are summarized in
Figure 1, which indicates support from the varied analytical methods and data partitions.
We consider twelve clades resolved with reasonable confidence, and mark them by circled
numbers. These twelve clades are recovered in the Bayesian analyses of All Genes
combined (Fig. 2), and most are recovered in the parsimony analysis (Fig. 3).
For the All Genes likelihood analyses, PAUP* and RAxML gave trees that were
identical in the basal regions of the tree, differing only in placements within the Salticoida.
Differences are not surprising because the two programs calculated likelihood slightly
differently (e.g., gamma rate variation versus CAT) and also did not use identical search
strategies. The models of substitution and rate variation used in PAUP* likelihood
searches, inferred from the data, were GTR plus gamma rate variation with a proportion
invariant (small taxon sample, rmatrix=(0.89729 4.04401 5.58411 2.27794 5.73636)
shape=0.768635 pinvar = 0.508922; large taxon sample, rmatrix=(1.43352 5.93829
6.45299 3.25682 8.16426) shape=0.777875 pinvar = 0.313501).
The individual gene regions analyzed separately provide independent support for
many of these clades (Figs. 4–8). For instance, the exclusion of the lapsiines from the
Salticoida is supported by 18S, 16S-ND1, and CO1. It is also supported by 28S except for
MADDISON & NEEDHAM
44 © 2006 Magnolia Press
1255
ZOOTAXA the placement of Sitticus. If 28S and 18S are combined into a single matrix, or if H3 +
16S-ND1 + CO1 are combined, this placement of lapsiines is also supported (there is no
separate figure for these data combinations, although their results are indicated via the oval
decorations in Fig. 1). The lapsiine placement is also supported in a matrix of CO1 and
ND1 translated to amino acids (Fig. 9). The exclusion of the hisponines from the
Salticoida is supported in the analysis of 18S alone, 28S alone (with the exception of
Sitticus), the combined matrix 28S + 18S, the combined matrix H3 + 16S-ND1 + CO1,
and with All Genes combined. In general there is strong support from varied analyses and
data partitions for the primary conclusion of this paper, that the lapsiines and hisponines
fall outside the Salticoida.
Results for the large taxon sample are shown in Figure 10. They confirm that the small
taxon sample had been chosen well, as it includes representatives spanning the Salticoida.
The Salticoida is resolved as monophyletic with the lapsiines and hisponines excluded. For
legibility of the basal taxa, interrelationships within the Salticoida are de-emphasized in
this figure; they will be treated in a subsequent paper using additional data. The
relationships among the basal lineages matches in general those recovered from the small
taxon sample analyses except for Hispo, Tomocyrba, and Goleba. Hispo fell with the
miturgid in the likelihood and Bayesian analyses and with the salticoid heliophanines with
parsimony. Hispo is lacking data from 28S, which was the single most informative gene
region in the analyses of Maddison & Hedin (2003) as judged by its concordance with the
All Genes analysis. Tomocyrba fell with the corinnid by parsimony. Goleba fell with the
corinnid in all large taxon sample analyses. Given the strong support from various
partitions for salticid monophyly in the small taxon sample analyses, and their unique and
unreversed morphological synapomorphies, these placements would appear to be in error.
The dominance of salticoids within the larger matrix may have degraded either the
alignment in the basal regions of the tree, or the appropriateness of the estimated models of
evolution for the basal regions of the tree (for likelihood and Bayesian methods).
Otherwise, the basal relationships are as with the small taxon analyses, with lapsiines
sister to spartaeines (including Holcolaetis), and Lyssomanes sister to the two.
Discussion
Our molecular data clearly support a placement of the lapsiines (Galianora and
Thrandina) outside of the Salticoida. The All Genes analyses (Figs. 2–3) and various
partitions support this (28S, 18S, 16S-ND1, CO1; Figs. 4–8), as does the large taxon
sample analysis (Fig. 10). Some morphological support for the exclusion of lapsiines from
the Salticoida has already been mentioned—the presence of the tarsal claw in females and
median apophysis in males are both plesiomorphic and indicate lapsiines fall outside of the
Salticoida. Also supporting the basal position of lapsiines is the more or less equal
numbers of teeth on the anterior and posterior tarsal claws on the second pair of legs
© 2006 Magnolia Press 45
SALTICIDAE
1255
ZOOTAXA
FIGURE 1. Summary of phylogenetic analyses. Filled ovals indicate support from different data
partitions and analytical methods. Bayesian analyses: black indicates posterior probability (p.p.) =
0.75; checkered indicates p.p. 0.51–0.75. Other analyses: black indicates the clade appears in the
most parsimonious or highest likelihood trees; striped indicates the clade was supported but with
one taxon excluded, as explained in the following notes. Notes concerning numbered clades: (1)
28S+18S p.p. = 0.74; 18S p.p. = 0.58. (2) CO1 p.p. = 0.56; clade by CO1 likelihood and parsimony
excludes Orthrus. (3) 18S p.p. = 0.71. (4) H3+16S-ND1+CO1 p.p. = 0.69; CO1 p.p. = 0.67; clade
by 16S-ND1 Bayesian and parsimony excludes Tomocyrba; clade by CO1 parsimony excludes
Thrandina; clade by All Genes parsimony includes Hispo. (5) Clade by 28S parsimony excludes
Thrandina. (6) 18S p.p. = 0.57. (7) 16S-ND1 sequence not obtained for Galianora sacha. (8) 18S
p.p. = 0.56. (9) 18S sequences obtained only for Portia s. in this clade. (10) 28S+18S p.p. = 0.63;
CO1 sequences not obtained for any hisponine. (11) 28S p.p. = 0.53.
MADDISON & NEEDHAM
46 © 2006 Magnolia Press
1255
ZOOTAXA (Maddison, 2006; unequal in salticoids, Maddison, 1988, 1996). Two other plesimorphies
also indicating exclusion from the Salticoida are the relatively long fovea of the carapace
and, in Thrandina at least, the large posterior median eyes (Maddison, 2006). Live
specimens of Thrandina were observed to have a peculiar walk similar to that of
spartaeines (Maddison, 2006); it might be either a synapomorphy uniting lapsiines with
spartaeines or a plesiomorphy excluding both from the salticoids. Because so few lapsiine
specimens are available, we did not examine several relevant characters requiring
disarticulation (cheliceral mound, gnathocoxal glands, intercheliceral sclerite; Maddison
& Hedin, 2003).
FIGURES 2–3. Summary of analyses using All Genes (28S + 18S + H3 + 16S-ND1 + CO1).
Circled numbers mark clades from Figure 1. 2, Majority rule consensus tree of 9900 trees sampled
from Bayesian analysis; shown are estimated posterior probabilities (from last 9900000 of 10
million generations). 3, Strict consensus of three most parsimonious trees (treelength 6871 steps);
shown are bootstrap values, 1000 replicates.
To date no morphological synapomorphies of lapsiines have been found, although they
are easily recognized among neotropical salticids by their plesiomorphic features
(Maddison, 2006). The molecular data (node 6, Fig. 1), however, do provide support for
the monophyly of the three species sampled. This demonstration of monophyly does not
yet justify recognition of a formal taxon including these three species and Lapsias, because
the latter has not yet been studied phylogenetically. Nonetheless we provisionally refer to
our three sampled species as lapsiines (see Maddison, 2006). Our analyses (All Genes,
28S, 18S, 16S-ND1, CO1) give strong support for the lapsiines being in a clade including
the spartaeines and possibly Lyssomanes (nodes 4 and 5, Fig. 1). Within the lapsiines, all
© 2006 Magnolia Press 47
SALTICIDAE
1255
ZOOTAXA
analyses agree on the two species of Galianora being sister taxa (node 7, Fig. 1). This is
consistent with their sharing a round tegulum with peripheral embolus and reduced
posterior median eyes (Maddison, 2006), both features derived with respect to the
conditions in Thrandina.
FIGURES 4–9. Results from analyses on separate data partitions. Circled numbers mark clades
from Figure 1. Figures 4–8 show majority rule consensus trees from sampled trees from Bayesian
analysis, with posterior probabilities at nodes. Spots on nodes indicate clade was recovered also in
parsimony analysis. Figure 9 shows strict consensus of nine most parsimonious trees (treelength
602) for ND1-CO1 amino acid matrix, with bootstrap values (1000 replicates).
Hisponines (Massagris, Tomocyrba, Hispo) lie outside the Salticoida according to the
All Genes analysis (all methods), 18S (all methods), and 28S (parsimony). They also lie
outside the salticoids when the remaining gene regions (H3 + 16S-ND1 + CO1 ) are
combined into a single analysis (Fig. 1). The precise relationships of the hisponines are
unclear, however. Their monophyly is not supported universally in our analyses, but we
MADDISON & NEEDHAM
48 © 2006 Magnolia Press
1255
ZOOTAXA suspect this is a result of the sparse data we have for them, especially Hispo. There is little
evidence that hisponines are near the spartaeines and lapsiines, but rather most analyses
suggest they branch independently from the lineage leading to the Salticoida, possibly as
the sister group to the salticoids (Figs. 2, 10). As noted above, their exclusion from the
salticoids is supported by their retention of a median apophysis. In addition, Wanless
(1981) notes that Hispo has a simple tracheal system as in lyssomanines and Portia, in
contrast to the more complex salticoid tracheae.
FIGURE 10. Results from Bayesian analysis of large taxon sample, all genes combined. Shown is
the majority rules consensus tree of 9900 sampled trees. Among the salticoids, only taxa also used
in the small taxon sample are named; the unnamed taxa are those in Maddison & Hedin (2003).
Circled numbers mark clades from Figure 1. Estimated posterior probabilities (last 9900000 of 10
million generations) marked for clades outside of the Salticoida. Maximum likelihood tree identical
outside the Salticoida except it placed the corinnid/thomisid/Goleba group more basally. Black
spots on nodes indicate clade was recovered also in parsimony analysis; hatched spots for nodes 2
and 3 indicate Hispo was placed within the salticoids by parsimony (as a heliophanine).
© 2006 Magnolia Press 49
SALTICIDAE
1255
ZOOTAXA
FIGURES 11–12. Ventral view of epigyna of hisponine females used. 11, Hispo cf. frenata; 12,
Massagris cf. honesta.
Opinion on the relationships of the salticoids, the Cocalodes group, and the
spartaeines has varied. A relationship of Holcolaetis to the spartaeines has been proposed
based on sharing of apparent secretory organs on the abdomen (Wanless, 1985,
Wijesinghe, 1992). Rodrigo & Jackson's (1992) formal morphological analysis placed the
Cocalodes group (including Holcolaetis) with the spartaeines. However, Wijesinghe's
(1997) phylogenetic reconstruction considered the Salticoida and Spartaeinae as sister
groups, in part on the basis of their shared loss of the median apophysis. Salticids with
median apophyses (e.g., lyssomanines, Holcolaetis) are a series of lineages branching at
the base of Wijesinghe's tree. Our present results suggest instead that Holcolaetis is indeed
related to the spartaeines, and thus that the median apophysis has been lost at least four
times, once as a synapomorphy of the Salticoida, once within the spartaeines, once within
the hisponines, and once within the lyssomanines (Chinoscopus Simon, Galiano 1998).
There was support for the placement of Lyssomanes with the spartaeines and lapsiines
from the All Genes analysis and from the partitions 28S, 16S-ND1, and CO1 (node 4, Fig.
1). With this placement, Lyssomanes would be separate from the Old World "lyssomanine"
Goleba, and thus the lyssomanines would be polyphyletic (or paraphyletic, depending on
one's interpretation of ancestors). Wanless (1980) was correct therefore in being skeptical
about the monophyly of the lyssomanines. Wanless suggested that one of the most
distinctive features uniting lyssomanines, the transparent green or yellow colouration and
long thin legs, may simply be a derived and convergent adaptation. This body form has
arisen several times independently in other salticid groups (e.g., Epeus Peckham &
Peckham and Orthrus). Goleba's placement is unclear, but it may branch deeply, perhaps
even as the sister group to all remaining salticids. Our results therefore indicate the
lyssomanines should be divided into two groups. However, we will postpone revising their
MADDISON & NEEDHAM
50 © 2006 Magnolia Press
1255
ZOOTAXA classification until other Old World lyssomanine genera are better studied. The results
also argue against placing the lyssomanines in a separate family, unless the spartaeines and
lapsiines were to go with them.
Although our analyses were not designed to examine relationships within the
salticoids, their deep relationships may be more easily resolvable now that we have a more
complete representation of their outgroups, i.e. the basal salticids. The amycoids are
resolved as sister group to the rest of the salticoids by 28S and 18S as well as by the All
Genes analyses, although this result did not hold for the other genes individually.
Future work should seek to resolve the position of the hisponines and the
decomposition of the lyssomanines. In addition, the finding that lapsiines are sister group
to the spartaeines motivates work on their natural history. Do lapsiines share with their
spartaeine relatives the behaviours, rare among salticids, of making webs and eating other
spiders (Jackson & Pollard, 1996)? With these phylogenetic results it is apparent that
there are new opportunities in the lapsiines and hisponines to study basal salticids and the
family's early evolution.
Acknowledgments
Research was funded by an NSERC Discovery grant to WPM. Research and collecting
permits in Ecuador were obtained through the Instituto Ecuatoriano de Areas Protegidas y
Vida Silvestre, with the assistance of Marco Altamirano of the Museo Ecuatoriano de
Ciencias Naturales. We thank Gita Bodner, Charles Griswold, Greta Binford, and David
Maddison for their efforts in collecting specimens for this study. Ingi Agnarsson, Nikolaj
Scharff, and Martín Ramírez provided helpful comments on earlier drafts of this paper.
Literature cited
Blest, A.D., O'Carroll, D.C. & M. Carter, M. (1990) Comparative ultrastructure of Layer I receptor
mosaics in principal eyes of jumping spiders: the evolution of regular arrays of light guides.
Cell and Tissue Research, 262, 445–460.
Colgan, D.J., McLauchlan, A., Wilson, G.D.F., Livingston, S.P., Edgecombe, G.D., Macaranas, J.,
Cassis, G., & Gray, M.R. (1998) Histone H3 and U2 sn RNA DNA sequences and arthropod
molecular evolution. Australian Journal of Zoology, 46, 419–437.
Ewing, B. & Green, P. (1998) Basecalling of automated sequencer traces using phred. II. Error
probabilities. Genome Research, 8, 186–194.
Ewing, B., Hillier, L., Wendl, M. & Green, P. (1998) Basecalling of automated sequencer traces
using phred. I. Accuracy assessment. Genome Research, 8, 175–185.
Felsenstein, J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolu-
tion, 39, 783–791.
Galiano, M.E. (1998) Revision of the genus Chinoscopus (Araneae, Salticidae, Lyssomaninae).
Bulletin of the British Arachnological Society, 11, 1–9.
Giribet, G., Carranza, S., Baguna, J., Riutort, M. & Ribera, C. (1996) First molecular evidence for
© 2006 Magnolia Press 51
SALTICIDAE
1255
ZOOTAXA
the existence of a Tardigrada + Arthropoda clade. Molecular Biology and Evolution, 13,
76–84.
Green, P. (1999) Phrap, version 0.990329, distributed by the author. See http://www.phrap.org
(accessed 15 March 2006)
Green, P. & Ewing, B. (2002) Phred, version 0.020425.c, distributed by the authors. See http://
www.phrap.org (accessed 15 March 2006)
Hedin, M.C. & Maddison, W.P. (2001) A combined molecular approach to phylogeny of the jump-
ing spider subfamily Dendryphantinae (Araneae, Salticidae). Molecular Phylogenetics and
Evolution, 18, 386–403.
Huelsenbeck, J. P. & Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogeny. Bioinfor-
matics, 17, 754–755.
Jackson, R. R. & Pollard, S.D. (1996) Predatory behavior of jumping spiders. Annual Review of
Entomology, 41, 287–308.
Li, D.Q. (2000) Prey preferences of Phaeacius malayensis, a spartaeine jumping spider (Araneae:
Salticidae). Canadian Journal of Zoology, 78, 2218–2226.
Logunov D.V. (2004) On the taxonomic position of "Lyssomanes" karnatakaensis and other Indian
species formerly assigned to Lyssomanes (Araneae, Salticidae). Bulletin of the British Arach-
nological Society, 13, 73–75.
Maddison, D.R. (1991) The discovery and importance of multiple islands of most-parsimonious
trees. Systematic Zoology, 40, 315–328.
Maddison, D.R. & Maddison, W.P. (2005) MacClade version 4: Analysis of phylogeny and charac-
ter evolution. Version 4.08. Sinauer Associates, Sunderland, Massachusetts.
Maddison, W.P. (1988) A revision of jumping spider species groups formerly placed in the genus
Metaphidippus, with a discussion of salticid phylogeny (Araneae). Unpublished PhD. thesis,
Harvard University.
Maddison, W.P. (1996). Pelegrina Franganillo and other jumping spiders formely placed in the
genus Metaphidippus (Araneae: Salticidae). Bulletin of the Museum of Comparative Zoology,
154, 215–368.
Maddison, W.P. (2006) New lapsiine jumping spiders from Ecuador (Araneae: Salticidae). Zootaxa,
1255, 17–28.
Maddison, W.P. & Hedin, M.C. (2003) Jumping spider phylogeny (Araneae: Salticidae). Inverte-
brate Systematics, 17, 529–549.
Maddison, W.P. & Maddison, D.R. (2005) Mesquite: A modular system for evolutionary analysis.
Version 1.06. Available from http://mesquiteproject.org (accessed 15 March 2006)
Maddison, W.P. & Zhang, J.X. (2006) New hisponine and lyssomanine jumping spiders from
Africa (Araneae: Salticidae). Zootaxa, 1255, 29–35.
Platnick, N.I. (2005) World spider catalog version, 6.5. Available from http://research.amnh.org/
entomology/spiders/catalog/SALTICIDAE.html. (accessed 15 March 2006)
Prószy ski, J. & abka, M. (1983) Genus Tomocyrba (Aranei, Salticidae)—hypothetic survivor of
the Amber fauna. Systematic study with description of four new species. Acta Zoologica Cra-
coviensia, 26, 563–578.
Rodrigo, A.G. & Jackson, R.R. (1992). Four jumping spider genera of the Cocalodes-group are
monophyletic with genera of the Spartaeinae (Araneae: Salticidae). New Zealand Natural Sci-
ences, 19, 61–67.
Ronquist, F. & Huelsenbeck, J.P. (2003) MRBAYES 3: Bayesian phylogenetic inference under
mixed models. Bioinformatics, 19, 1572–1574.
Simon, E. (1901) Histoire Naturelle des Araignées. Deuxième édition. Tome 2, fasc. 3, pp.
381–668. Paris.
Stamatakis, A. (2005) An efficient program for phylogenetic inference using simulated annealing.
Proceedings of the 19th International Parallel and Distributed Processing Symposium, Den-
MADDISON & NEEDHAM
52 © 2006 Magnolia Press
1255
ZOOTAXA ver, Colorado, April 2005.
Stamatakis, A., Ludwig,T. & Meier, H. (2005) RAxML–III: A fast program for maximum likeli-
hood-based inferrence of large phylogenetic trees. Bioinformatics, 21, 456–463.
Swofford, D.L. (2002) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods).
Version 4.0b10. Sinauer Associates: Sunderland, Massachusetts.
Thompson, J.D., Gibson,T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. (1997) The ClustalX
windows interface: flexible strategies for multiple sequence alignment aided by quality analy-
sis tools. Nucleic Acids Research, 25, 4876–4882.
Wanless, F.R. (1980). A revision of the spider genera Asemonea and Pandisus (Araneae: Salti-
cidae). Bulletin of the British Museum of Natural History (Zoology), 39, 213–257.
Wanless, F.R. (1981). A revision of the spider genus Hispo (Araneae, Salticidae). Bulletin of the
British Museum of Natural History (Zoology), 41, 179–198.
Wanless, F.R. (1982). A revision of the spider genus Cocalodes with a description of a new related
genus (Araneae: Salticidae). Bulletin of the British Museum of Natural History (Zoology), 42,
263–298.
Wanless, F.R. (1984). A review of the spider subfamily Spartaeinae nom. n. (Araneae: Salticidae)
with descriptions of six new genera. Bulletin of the British Museum of Natural History (Zool-
ogy), 46, 135–205.
Wanless F.R. (1985) A revision of the spider genera Holcolaetis and Sonoita (Araneae: Salticidae).
Bulletin of the British Museum of Natural History (Zoology), 48, 249–278.
Weso owska, W. (1993) A revision of the spider genus Massagris Simon, 1900 (Araneae, Salti-
cidae). Genus (Wroclaw), 4, 133–141.
Wijesinghe, D.P. (1992) A new genus of jumping spider from Borneo with notes on the spartaeine
palp (Araneae: Salticidae). Raffles Bulletin of Zoology, 40, 9–19.
Wijesinghe, D.P. (1997) Relationships of Spartaeinae and other primitive jumping spiders, with
reviews of Cyrba and Cocalus and description of a new genus (Arachnida: Araneae: Salti-
cidae). Unpublished PhD. thesis, City University of New York.
abka M. & Kovac D. (1996) Paracyrba wanlessi—a new genus and species of Spartaeinae from
Peninsular Malaysia, with notes on its biology (Arachnida: Araneae: Salticidae). Senckenbergi-
ana biologica, 76, 154–156.
© 2006 Magnolia Press 53
SALTICIDAE
1255
ZOOTAXA
APPENDIX 1. Specimens sequenced for molecular phylogenetic analysis. All specimens depos-
ited in Spencer Entomological Museum (UBC-SEM) except those marked CASENT, which are
deposited in the California Academy of Sciences. Specimens (m) male, (f) female, and (j) juvenile
as marked by species name. Columns marked 28S, 18S, H3, 16S-ND1, and CO1 show sequence
length in parentheses and GenBank accession numbers. * Specimen used by Maddison & Hedin,
2003.
Species # Locality 28S 18S H3 16S-ND1 CO1
Lapsiines
Galianora sacha
Maddison (j) d116 ECUADOR:
Napo: S 1.067 W
77.617
(772)
DQ665766 (1313)
DQ665734 (331)
DQ665716 (975)
DQ665754
Galianora bryi-
cola Maddison (m) d124 ECUADOR:
Napo: S 1.067 W
77.617
(756)
DQ665771 (1313)
DQ665741 (325)
DQ665717 (692)
DQ665727 (972)
DQ665758
Thrandina paroc-
ula Maddison (m) d123 ECUADOR:
Morona Santi-
ago: S 2.9227 W
78.4079
(773)
DQ665779 (1320)
DQ665751 (311)
DQ665718 (793)
DQ665726 (970)
DQ665761
Hisponinae
Massagris schisma
Maddison & Zhang
(m)
d081 SOUTH
AFRICA: North-
ern Cape, Oor-
logskloof Nature
Reserve
(778)
DQ665762 (1661)
DQ665731 (898)
DQ665728
Massagris cf. hon-
esta Weso owska
(f)
d082 SOUTH
AFRICA: Kwa-
zulu-Natal: Lake
St. Lucia S
28.1021 E
32.4279
(597)
DQ665772 (1677)
DQ665743 (331)
DQ665705 (661)
DQ665722
Hispo cf. frenata
Simon (f) d126 MADAGAS-
CAR: Prov. Toa-
masina S 18.944
E 48.418 CASE
NT 9005643
(1319)
DQ665739 (736)
DQ665724
Tomocyrba anda-
sibe Maddison &
Zhang (m)
d127 MADAGAS-
CAR: Prov. Toa-
masina S 18.944
E 48.418 CASE
NT 9005649
(513)
DQ665780 (1317)
DQ665752 (229)
DQ665706 (630)
DQ665725
Lyssomaninae
Goleba lyra Madd-
ison & Zhang (m) d051 MADAGAS-
CAR: Fianarant-
soa: S 22.592 E
45.128 CASENT
9005863
(771)
DQ665768 (1304)
DQ665737 (323)
DQ665707 (964)
DQ665755
................to be continued
MADDISON & NEEDHAM
54 © 2006 Magnolia Press
1255
ZOOTAXA APPENDIX 1 (continued)
Species # Locality 28S 18S H3 16S-ND1 CO1
Lyssomanes viri-
dis Walckenaer (f) d129 U.S.A.: Missis-
sippi, Wall Doxey
State Park
(1694)
DQ665742 (327)
DQ665715
Spartaeinae
Holcolaetis sp. (H.
zuluensis
Lawrence?) (j)
d036 SOUTH
AFRICA: Kwa-
zulu-Natal: Lake
St. Lucia S
28.2369 E
32.4100
(659)
DQ665770 (1305)
DQ665740 (213)
DQ665721 (971)
DQ665757
Phaeacius cf. fim-
briatus Simon (m) d111 MYANMAR:
Yangon Division:
N 17.045 E
96.095 CASENT
9019139
(789)
DQ665775 (973)
DQ665759
Portia cf. schultzi
Karsch (f) d131 MADAGAS-
CAR: Fianarant-
soa: S 22.592 E
45.128 CASENT
9005865
(642)
DQ665776 (859)
DQ665747 (326)
DQ665708
Salticoida
Dendryphantes
hastatus Clerck (f) d043 POLAND:
Siedlce N 52.127
E 22.271
(661)
DQ665763 (1323)
DQ665732
Euophrys monad-
nock Emerton (m) d029 CANADA: Nova
Scotia: N 44.5279
W 64.6405
(596)
DQ665764 (1471)
DQ665733 (326)
DQ665714
Evarcha proszyn-
skii Marusik &
Logunov (m)
d096 CANADA: Brit-
ish Columbia:
Richmond
(781)
DQ665765 (939)
DQ665723
Ghelna castanea
Hentz (f) d005 U.S.A.: North
Carolina (744)
DQ665767 (1686)
DQ665735 (325)
DQ665709
Heliophanus
cupreus Walcke-
naer (m)
d044 POLAND: Mie-
lik, N 52.331 E
23.042
(534)
DQ665769 (1573)
DQ665738 (327)
DQ665710 (975)
DQ665756
Orthrus bicolor
Simon * S192 PHILIPPINES:
Luzon (773)
DQ665773 (1441)
DQ665745 (305)
DQ665719
Pellenes peninsu-
laris Emerton (m) d057 CANADA: Nova
Scotia: N 45.5862
W 62.2271
(702)
DQ665774 (1579)
DQ665746 (325)
DQ665712
Salticus scenicus
Clerck (j) d003 CANADA: Brit-
ish Columbia:
Mission
(699)
DQ665777 (1488)
DQ665748 (323)
DQ665713
................to be continued
© 2006 Magnolia Press 55
SALTICIDAE
1255
ZOOTAXA
APPENDIX 1 (continued)
Species # Locality 28S 18S H3 16S-ND1 CO1
Sitticus palustris
Peckham & Peck-
ham (m)
d030 CANADA: Nova
Scotia: N 44.4318
W 64.6075
(721)
DQ665778 (1336)
DQ665749 (770)
DQ665729 (971)
DQ665760
Zenodorus cf.
orbiculatus Key-
serling (m)
d008 AUSTRALIA:
Queensland: S
16.2 E 145.4
(743)
DQ665781 (1286)
DQ665753 (294)
DQ665711
Outgroups
Thomisidae: Xysti-
cus sp. * S316 U.S.A.: Colorado (1680)
DQ665750 (325)
DQ665704
Anyphaenidae:
Hibana sp. * S318 MEXICO:
Sonora (1534)
DQ665730
Gnaphosidae:
Cesonia sp. * S319 MEXICO:
Sonora (579)
DQ665736 (320)
DQ665720
Miturgidae: Chei-
racanthium sp. * S321 MEXICO:
Sonora (822)
DQ665744
56 © 2006 Magnolia Press
1255
ZOOTAXA
