cla 12029 120. - Prashant P. SharmaWillemart et al. (2007, 2010) similarly reported the disposition of tegumental gland openings in the legs of numerous Opiliones, with emphasis on

Walk it off: predictive power of appendicular characters towardinference of higher-level relationships in Laniatores (Arachnida:

Opiliones)

Guilherme Gainetta, Prashant P. Sharmab,*, Ricardo Pinto-da-Rochaa, Gonzalo Giribetc

and Rodrigo H. Willemartd

aDepartamento de Zoologia, Instituto de Biociencias, Universidade de Sao Paulo, Caixa Postal 11461, 05422-970, Sao Paulo, SP, Brazil; bDivision

of Invertebrate Zoology, American Museum of Natural History, 200 Central Park West, New York, NY 10024, USA; cMuseum of Comparative

Zoology and Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA 02138, USA; dEscola de Artes Ciencias

e Humanidades, Universidade de S~ao Paulo, Rua Arlindo B�ettio, 1000—Ermelino Matarazzo, CEP: 03828–000, S~ao Paulo, SP, Brazil

Accepted 27 March 2013

Abstract

Morphological characters are essential for establishing phylogenetic relationships, delimiting higher-level taxa, and testing phy-logenetic relationships inferred from molecular sequence data. In cases where relationships between large clades remain unre-solved, it becomes imperative to establish which character systems are sound predictors of phylogenetic signal. In the case ofLaniatores, the largest suborder of Opiliones, some superfamilial relationships remain unresolved or unsupported, and tradition-ally employed phenotypic characters are typically of utility only at the family level. Here we investigated a promising set of mor-phological characters that can be discretized and scored in all Opiliones: cuticular structures of the distal podomeres (metatarsiand tarsi). We intensively sampled members of all known families of Laniatores, and define here three new, discrete appendicularcharacters toward refinement of Laniatores superfamilial systematics: metatarsal paired slits (MPS; occurring in all Laniatoresexcept Sandokanidae), proximal tarsomeric gland (PTG; in Icaleptidae, Fissiphalliidae, and Zalmoxidae), and tarsal aggregatepores (TAP; found in Gonyleptoidea, Epedanoidea, and Pyramidopidae). We conducted statistical tests on each character tocharacterize the strength of phylogenetic signal and assess character independence, based on alternative tree topologies of Lani-atores. All three characters had high retention indices and bore significantly strong phylogenetic signal. Excepting one pairwisecomparison, morphological characters did not evolve in a correlated manner, indicating that appendicular morphology does notconstitute a single character system. Our results demonstrate the predictive power and utility of appendicular characters in Opili-ones phylogeny, and proffer a promising source of diagnostic synapomorphies for delimiting superfamilies.© The Willi Hennig Society 2013.

Introduction

The diversity of the harvestman suborder Laniatoresconstitutes a compelling case of morphological diversi-fication. With over 4100 described species, Laniatoresencompass much of the breadth of morphological vari-ation observed in the arachnid order Opiliones, includ-ing numerous lineages with legs of variable length,embellished coloration, an array of armature types

and body shapes, and adaptations to various ecologi-cal niches (Kury, 1993; Rambla, 1993; Pinto-da-Ro-cha, 2002; Willemart et al., 2006; Derkarabetian et al.,2010). Recent efforts to infer higher-level Laniatoresphylogeny have been based principally on molecularsequence data (Giribet et al., 2010; Sharma and Giri-bet, 2011) and have contributed to testing the compo-sition and monophyly of the Laniatores superfamilies.An outstanding area of inquiry in Laniatores systemat-ics is the relationships among superfamilies of the in-fraorder Grassatores, a division that includes ca. 3600species.

*Corresponding author:E-mail address: [email protected]

CladisticsCladistics 30 (2014) 120–138

10.1111/cla.12029

© The Willi Hennig Society 2013

Some relationships among Grassatores superfamiliesbased on molecular data remain either unsupported orsensitive to algorithmic approaches (Fig. 1). The appli-cation of morphological data, either exclusively or inconcert with molecular data, has generally been directedtoward subordinal relationships within Opiliones (Mar-tens, 1976, 1980, 1986; Shultz, 1998; Giribet et al., 1999,2002; Garwood et al., 2011). Due to limitations in sam-pling Laniatores and the suitability of the charactersystems implemented, these previous efforts have sup-ported the monophyly of Laniatores and Grassatores,but have not been able to provide a stable set of relation-ships between constituent superfamilies. This is due inpart to the autapomorphic nature of genitalic charactersat higher taxonomic ranks. Although genitalic charac-ters are of great taxonomic utility and predominantlycomprise the synapomorphies of most Laniatores fami-

lies (with some defined exclusively based on genitalicstructures), the homology of various elements of themale copulatory apparatus becomes ambiguous withincreasing phylogenetic distance, namely at the superfa-milial level. By contrast, the female ovipositor has a lar-gely conserved architecture among Laniatores (Martens,1976; Kury, 1993; Sharma et al., 2011).A fundamental dimension that morphological char-

acters are anticipated to provide in higher-level Laniat-ores phylogeny is the identification of characters thatcan validate or refute phylogenetic hypotheses inferredon the basis of molecular sequence data alone. Particu-larly in cases where relationships inferred from molec-ular data are tenuous and/or challenge traditionalsystematics, both re-examination of the previouslyregarded morphological synapomorphies and surveysof novel putative synapomorphies are requisite, espe-cially if these are available from independent sourcesof phylogenetic data (reciprocal illumination, sensuHennig, 1966). In the case of Laniatores, the sistergroup of the most diverse superfamily, Gonyleptoidea(including Stygnopsidae), has long been contentiousand remained a recalcitrant problem for a 10-genedataset (Sharma and Giribet, 2011). Similarly, theplacement of the autapomorphic family Sandokanidaeremains uncertain even at the level of superfamilialplacement (Sharma and Giribet, 2009, 2011; Giribetet al., 2010). Finally, the identification of any morpho-logical characters uniting an assemblage of SoutheastAsian families consistently recovered by moleculardata (Epedanoidea, sensu Sharma et al., 2011) remainsof great potential value to Laniatores systematics.A promising set of characters for inferring relation-

ships between Laniatores superfamilies may be cuticu-lar structures such as glands, secretory pores, andsensory structures, as suggested for the suborder Cyp-hophthalmi (Willemart and Giribet, 2010). Gnaspiniand Rodrigues (2011) observed a correspondencebetween the morphology of the repugnatorial glandopening area of several Grassatores and the phylogenyof this group (tree topology based on Giribet et al.,2010). Willemart et al. (2007, 2010) similarly reportedthe disposition of tegumental gland openings in thelegs of numerous Opiliones, with emphasis on Laniat-ores. The correspondence between these glands and arevised Laniatores phylogeny was subsequently notedas a potential source of diagnostic characters (Sharmaand Giribet, 2011; Sharma et al., 2011), but the degreeof evolutionary lability of appendicular gland openingsas a viable character system for Opiliones phylogenet-ics remains untested.We therefore endeavored to sample a suite of mor-

phological characters in all known families of Laniat-ores (sensu Sharma and Giribet, 2011; but seeclassification of Kury, 2011). We focused on the mor-phology of two distal leg podomeres, metatarsi and

Fig. 1. Classification of Laniatores indicating Grassatores superfam-ilies as presently defined (sensu Sharma and Giribet, 2011). Questionmark in edge corresponding to Sandokanidae indicates uncertaintyin the phylogenetic placement of this family. Superfamilies of Insidi-atores are not used, as Triaenonychoidea (Triaenonychidae + Syn-theonychiidae) is not monophyletic (Giribet et al., 2010; Sharma andGiribet, 2011) and families of Travunioidea sampled herein havebeen synonymized as Travuniidae by Kury (2011).

G. Gainett et al. / Cladistics 30 (2014) 120–138 121

tarsi; identified and scored multiple discrete charactersfrom these segments; and conducted statistical tests tocharacterize inherent phylogenetic signal and degree ofcharacter independence.

Materials and methods

Species sampling

Specimens for study were accessed from the frozentissue collection at the Museum of Comparative Zool-ogy. We sampled 35 species of Grassatores and fivespecies of Insidiatores, with a large proportion of thesedrawn from material used for previous molecular phy-logenetic study (Sharma and Giribet, 2011). Our sam-pling of outgroups consisted of three Eupnoi, twoDyspnoi, and one Cyphophthalmi, and our ingroupincluded 63 individuals (27 males, 23 females, 13 unde-termined). The list of specimens, including vouchernumbers, is found in Appendix 1; collecting details areavailable in the MCZ collections database (mcz-base.mcz.harvard.edu).

Scanning electron microscopy

Legs of specimens were cleaned using a Branson 200sonicator, in a 1 : 10 dilution of detergent, and subse-quently in deionized water only. The sonication stepwas also conducted to break the trichomes concealingtegumental structures. The duration of sonication wasdetermined empirically. Cleaned legs were dried in100% acetone and mounted on 12 mm SEM stubs(Electron Microscopy Sciences, Hatfield, PA, USA)using Ultra-Smooth carbon adhesive tabs (ElectronMicroscopy Sciences, Hatfield, PA, USA). Sampleswere coated with a Pt-Pd target using a Cressington208HR sputter coater and photographed on a ZeissUltra-Plus FESEM or on a Zeiss Supra FESEM (fieldemission scanning electron microscope) at the Centerfor Nanoscale Systems, Harvard University. Imageswere taken in mesal, ectal, and/or dorsal view byrotating the stub, remounting the appendage, ormounting multiple appendages in alternate placements.Presence of a given character was established in one ormore leg pairs, and absence was scored if a givenstructure was not observed in any of the legs.

Ancestral state reconstruction

Ancestral state reconstructions were conducted inMesquite ver. 2.75 (Maddison and Maddison, 2011)under equal weights parsimony. The bases for thereconstruction were two tree topologies from a previousstudy (Sharma and Giribet, 2011). The first was theultrametric dated tree topology generated using simul-

taneous estimation of topology and divergence times(Figure 17 of Sharma and Giribet, 2011). This topol-ogy, which favours a sister relationship of Sandokani-dae and Podoctidae (see also Sharma and Giribet,2009; Giribet et al., 2010), is consistent with a series ofmorphological characters supporting a Sandokani-dae + Podoctidae clade, such as the occurrence oftubercular bridges between the prosoma and opisthoso-ma in many species, reduced tarsalia, and first postem-bryonic stages with undivided tarsi on legs I–II. Thesecond topology used to infer character evolution wasthe maximum likelihood topology of Sharma and Giri-bet (2011), which placed Sandokanidae in a grade withPhalangodidae at the base of Grassatores. The use ofalternative topologies was prompted by the autapomor-phic nature of many aspects of sandokanid morphol-ogy, which renders the exact placement of this familywithin Laniatores contentious. In cases of ambiguity,both accelerated and delayed transformations of char-acter state (ACCTRAN and DELTRAN, respectively)were chosen to resolve equally parsimonious recon-structions, following Agnarsson and Miller (2008).

Analysis of phylogenetic signal

Phylogenetic signal of each morphological characterwas inferred using two metrics computed on the ultra-metric and ML topologies (above). The first was thetraditionally used character retention index (ri). Thesecond was the D statistic of Fritz and Purvis, wherebywe tested for phylogenetic signal of each discrete char-acter (Fritz and Purvis, 2010). D is defined as

D ¼ ½Rdobs �meanðRdBÞ�=½ðmeanðRdRÞ �meanðRdBÞ�where Σdobs is the number of changes required for theobserved distribution of character states on the inputphylogeny; dR is a distribution of d values generatedby randomly shuffling values of the character on thetips of the phylogeny; and dB is a distribution of d val-ues generated under the expectations of a Brownianmotion model of evolution (i.e. a random walk model)(Fritz and Purvis, 2010).D is close to or > 1 if character evolution cannot be

distinguished from random phylogenetic structure, 0 ifcharacter evolution corresponds to a Brownian phylo-genetic structure, and < 0 if a binary trait is more con-served than the Brownian expectation (Fritz andPurvis, 2010). We implemented 1000 permutations togenerate the null distribution of each character, usingthe caper package in R (Orme et al., 2012).

Statistical analysis of trait correlation

To test for character independence, we conductedtests for correlation of discrete binary characters forall pairwise comparisons of the four appendicular

122 G. Gainett et al. / Cladistics 30 (2014) 120–138

(a)

(e) (i)

(j)

(k)

(l)

(f)

(g)

(h)

(b)

(c)

(d)

Fig.2.Metatarsal-tarsaljoints

ofrepresentativespeciesshowingpaired

slitorgans(M

PS).(a)Astrobunusgranulator(Eupnoi,Sclerosomatidae),lateralview

ofrightlegIV

.Scale

bar:

20lm

.(b)Synthetonychia

sp.(Synthetonychiidae),dorsalview

ofrightlegIV

.Scale

bar:

10lm.(c)Erebomaster

flavencens(C

ladonychiidae),lateralview

ofrightlegIV

.Scale

bar:

10lm

.(d)Larifugacf.capensis(Triaenonychidae),lateralview

ofleft

legIV

.Scale

bar:50lm

.(e)Scotolemonlespesi(Phalangodidae),lateralview

ofrightlegIV

.Scale

bar:20lm.(f)

Tithaeussp.(Tithaeidae),dorsalview

ofleftlegIV

.Note

varyingdiametersofthefivesensillabases(fourbroken).Scale

bar:10lm.(g)Cobania

picea

(Gonyleptidae),dorsolateralview

ofrightlegIV

.Scale

bar:

100lm.(h)Montalenia

sp.(A

ssamiidae),dorsolateralview

ofleft

legIV

.Scale

bar:

10lm

.(i)Metabiantessp.(Biantidae),dorsolateralview

ofrightlegIV

.Scale

bar:

20lm

.(j)Icaleptessp.(Icaleptidae),lateralview

ofrightlegIV

.Note

enlarged

tuberclebearingapair

ofsensillabasiconica.Scale

bar:

25lm

.(k)Sandokantruncatus(San-

dokanidae),lateralview

ofleft

legIV

.Bracket

markseightsensillabasiconica.Note

absence

ofslitorgansin

thedistalmetatarsusjoint.Scale

bar:

100lm

.(l)SameindividualofS.

truncatusasin

(k),dorsalview

ofthedistaltibialjointofleft

legIV

.Paired

slitsin

thetibia

(arrowheads)

andasingle

slitin

theproxim

almetatarsus(dotted

arrowhead).Scale

bar:

100lm

.In

allfigures,

arrowsindicate

sensillabasiconica,whitearrowheadsindicate

proxim

al(short)slit,andblack

arrowheadsindicate

distal(w

ide)

slit.Mt:metatarsus;

Ta:tarsus;

Ti:tibia.


characters, using BayesTraits ver. 1.0 and the methodof Pagel and Meade (2006). Both competing treetopologies were used for this analysis. We tested twomodels of character evolution, a discrete independentmodel and a discrete dependent model, and analysedtrait evolution under maximum likelihood. For eachcombination of character pair and model, 500 maxi-mum likelihood optimizations were conducted.

Results

Identification of morphological characters

We identified and scored three morphological char-acters in the distal-most podomeres (metatarsi andtarsi) of Laniatores. To provide a basis for analyticalcomparison of phylogenetic signal, we included in the

(a) (b)

(c)

Fig. 3. Proximal tarsomeric gland (PTG) of Zalmoxoidea. (a) Zalmoxis sp. (Zalmoxidae) from Vanuatu, lateral view of left leg. Scale bar: 100 lm.(b) Fissiphallius sp. (Fissiphalliidae), lateral view of left leg. Scale bar: 50 lm. (c) Icaleptes sp. (Icaleptidae), lateral view of right leg. Arrowheadsshow pores in loose aggregation. Scale bar: 20 lm. Brackets in (a) and (b) indicate disposition of gland. Mt: metatarsus; Ta1: tarsomere 1.


analysis the double claws of tarsi III and IV, asynapomorphy of Grassatores. The double clawconstitutes a discrete binary character that undergoesa single change in the phylogeny.The three other characters are discussed in turn.

Metatarsal paired slits

The metatarsal paired slits (MPS) comprise a smallretrolateral slit followed by a longer distal slit in the

(a)

(b) (c)

(d)

Fig. 4. Proximal tarsomeric gland (PTG) of Zalmoxoidea. (a) Baculigerus sp. (Escadabiidae), lateral view of left leg. Marked area indicates spanof pore openings. Scale bar: 100 lm. (b) Same individual, detail. Scale bar: 10 lm. (c) Kimula goodnightorum (Kimulidae), lateral view of leftleg. Note absence of pore aggregations. Scale bar: 100 lm. (d) Guasinia sp. (Guasiniidae), lateral view of right leg. Arrows indicate sparsely dis-tributed pores. Scale bar: 20 lm. Mt: metatarsus; Ta1: tarsomere 1; Ta2: tarsomere 2.


(a)

(b)

(c)

(d) (h)

(g)

(f)

(e)

Fig. 5. Tarsal aggregate pores (TAP) of Gonyleptoidea. Lateral (a) and distal aspect (b–h) of the terminal tarsomere IV. (a) Pseudopucrolia muti-ca (Gonyleptidae). Scale bar: 50 lm. Inset: detail of six pores and associated trichomes. Scale bar of inset: 2 lm. (b) Hoplobunus sp. (Stygnopsi-dae), left leg. Scale bar: 10 lm. (c) Torreana spinata (Agoristenidae), left leg. Scale bar: 10 lm. Black arrowheads indicate intertwined setae. (d)Stygnus multispinosus (Stygnidae), left leg. Scale bar: 10 lm. (e) Rhopalocranaus albilineatus (Manaosbiidae), left leg. Scale bar: 20 lm. (f) Gnidiaholmbergi (Cosmetidae), right leg. Scale bar: 10 lm. (g) Phareicranaus serratotibialis (Cranaidae), left leg. Scale bar: 20 lm. (h) Discocyrtoidesnigricans (Gonyleptidae), left leg. Scale bar: 10 lm. White arrowheads indicate pore aggregations.


(a) (e)

(f)

(g)

(h)

(b)

(c)

(d)

Fig. 6. Tarsal aggregate pores (TAP) of non-gonyleptoid Grassatores. Distal aspect of the terminal tarsomere of legs III (c–e, g) or IV (a–b, f).(a) Alloepedanus sp. (Epedanidae), left leg. Scale bar: 10 lm. (b) Tithaeus sp. (Tithaeidae), left leg. Scale bar: 10 lm. (c) Santobius sp. (Podocti-dae), left leg. Scale bar: 10 lm. (d) Conomma oedipus (Pyramidopidae), right leg. Scale bar: 10 lm. (e) Metabiantes sp. (Biantidae), right leg.Scale bar: 25 lm. (f) Badessa sp. (Samoidae), left leg. Scale bar: 10 lm. (g) Stygnomma bispinatum (Stygnommatidae), right leg. Scale bar:10 lm. (h) Male Larifuga cf. capensis (Triaenonychidae), left leg. Scale bar: 10 lm. White arrowheads indicate pore aggregations white arrow-heads in 6h indicate sparser distribution of pores.


(a) (b)

(c) (d) (e)

(f) (g) (h)

(i) (j)

Fig. 7. Detail of tarsal aggregate pores (TAP). (a) Phareicranaus serratotibialis (Cranaidae). (b) Cobania picea (Gonyleptidae). (c) Discocyrtoidesnigricans (Gonyleptidae). (d) Heteromitobates discolor (Gonyleptidae) (e) Pseudopucrolia mutica (Gonyleptidae). (f) Alloepedanus sp. (Epedani-dae). (g) Tithaeus sp. (Tithaeidae). (h) Santobius sp. (Podoctidae). (i) Conomma oedipus (Pyramidopidae). (j) Stygnomma bispinatum (Stygnom-matidae). Scale bars: 5 lm.


vicinity of some sensilla basiconica1 (Willemart et al.,2007) in the most distal part of the calcaneus of allfour pairs of legs (arrowheads in Fig. 2b–j). The basicstructure of the MPS, as described by Willemart et al.(2007), is very conserved. MPS occur in all Laniatores,except in Sandokanidae (Fig. 2k). In cases where thesensilla were broken, we assumed that they were sen-silla basiconica based on positional homology, asinferred from other legs in the same species and/or clo-sely related species. We found interfamilial variation inthe number of sensilla basiconica and in the shape ofthe integument underneath. For instance, the numberof sensilla ranged from two in most cases (black

arrows in Fig. 2b,i) to five in Tithaeidae and Assami-dae species (black arrows in Fig. 2f,h). The cuticularsurface underneath the setae varied from flat tosmooth (Assamiidae; Fig. 2h) to a very pronouncedtubercle (Icaleptidae; bracket in Fig. 2j). No intra-indi-vidual variation was detected within legs III and IV.We corroborate the presence of the MPS in all Lani-

atores families except Sandokanidae (Fig. 2k), andabsence in the outgroup suborders, Eupnoi, Dyspnoi,and Cyphophthalmi (e.g. Fig. 2a). Intriguingly, wefound a very similar pair of dorsal slits in the distalpart of the tibia in all leg pairs of Sandokan truncatus.The distal slit is wider (black arrowhead in Fig. 2l)and the proximal is shorter (white arrowhead inFig. 2l), as with the MPS. Sandokan truncatus alsobore a group of sensilla (8–14) in the most distal partof the metatarsus of all four legs (bracket in Fig. 2k).Another sandokanid, Gnomulus sp., also lacked thesensilla in the metatarsus.

Fig. 8. Ancestral state reconstruction of morphological characters on dated (left) and maximum likelihood (right) tree topologies of Laniatores(topologies simplified to indicate families as terminals). Characters are (i) double claws on legs III and IV [red], (ii) metatarsal paired slits [yel-low], (iii) proximal tarsomeric gland [green], and (iv) tarsal aggregate pores [blue]. Coloured squares to the right of either topology indicate pres-ence of character in terminals. Squares on branches indicate gains (coloured) or losses (blank) of characters, as optimized under DELTRAN.

1“Sensillum basiconicum” is used sensu Willemart et al. (2007).

Though this term is usually not used for long-shafted sensilla like

those described in Figs 2c,j, we use this name to avoid confusion

because these are in the same position as the typical sensilla basico-

nica described herein.


Proximal tarsomeric gland

A previously undescribed aggregation of poresoccurring on the mesal and ectal surfaces of the mostproximal tarsomere of leg IV was inferred to constitutethe proximal tarsomeric gland (Figs. 3 and 4). Theproximal tarsomeric gland (PTG) occurs in all Icalepti-dae, Fissiphalliidae, and Zalmoxidae specimens analy-sed (Fig. 3). This structure has an elongate ovularshape, with variation in the number of pores (12–30pores). Dorsoventral placement of the PTG variesamong families (Fig. 3a–c). The three Zalmoxidae spe-cies we investigated showed the gland in leg IV of bothsexes (bracket in Fig. 3a). We also observed the struc-ture in one female Fissiphalliidae (bracket in Fig. 3b)and one female Icaleptidae (white arrowheads inFig. 3c), but males of these two families were notavailable for study.A different, but also undescribed, glandular struc-

ture was observed in the most proximal tarsomere oflegs III and IV of a female Escadabiidae (Fig. 4a,b).This aggregation is very elongated along the proximo-

distal axis of the tarsomere (bracket in Fig. 4a), whichin turn is much longer than its counterparts in Zalm-oxidae, Fissiphalliidae, and Icaleptidae. The homologyof the escabadiid gland and the PTG is not definitive;we alternately scored the PTG as absent or present forEscadabiidae in two separate analyses (see below).We did not observe the PTG in Kimulidae, which

also lack an elongate proximal-most tarsomere(Fig. 4c). In Guasiniidae, pores resembling those ofEscadabiidae occur sparsely throughout the elongateproximal tarsomere (which is not as elongate as inEscadabiidae; white arrowheads in Fig. 4d). However,no aggregation of pores is formed in Guasiniidae.

Tarsal aggregate pores

A distinct aggregation of pores occurs in the dorso-lateral regions on the mesal and ectal surfaces of thedistal-most tarsomere of legs III and IV, named thetarsal aggregate pores (TAP). As previously describedby Willemart et al. (2007), this aggregation is associ-ated with the base of a group of trichomes. To be

Fig. 9. Ancestral state reconstruction of morphological characters (3) and (4), broadly interpreted. Tree topologies, colour notations, and opti-mization as indicated in Fig. 8. Cladograms based on Sharma and Giribet (2011).


scored as a presence, the structure had to meet threecriteria: (i) possess more than two pores, in a discreteaggregation, (ii) be associated with an aggregation oftrichomes, and (iii) be sagittally symmetrical in the dis-tal-most tarsomere (as in Fig. 5a,b). We observed nei-ther sexual dimorphism nor differences between theTAP of legs III and IV within species. In many caseswhen the TAP trichomes were not broken, weobserved an intertwined disposition of these setae,which was never observed in the rest of the tarsomere(black arrowheads in Fig. 5c).Glandular openings with these three characteristics

were observed in all families of Gonyleptoidea(Agoristenidae, Cosmetidae, Cranaidae, Gonyleptidae,Manaosbiidae, Stygnidae, and Stygnopsidae), threefamilies of Epedanoidea (Epedanidae, Tithaeidae, andPodoctidae), and Pyramidopidae. Glandular openingsmeeting at least two of the defining criteria were alsoobserved in Samooidea (Biantidae, Samoidae, andStygnommatidae; all three bear pore aggregationsthat are sagittally symmetrical, but an associationwith trichomes is ambiguous due to the incidence ofventral scopulae in the vicinity of the pores; discussedbelow). Pores in similar position were observed inGuasiniidae and one of the two Zalmoxidae specieswe examined, but were neither associated with tric-homes nor sagittally symmetrical in the zalmoxoids(data not shown).

In Gonyleptoidea (Fig. 5) and Epedanoidea(Fig. 6a–c) with a TAP, the character had very similarposition and general morphology (white arrowheads inFigs 5 and 6). The number of pores in each aggrega-tion varied from four to five in Stygnopsidae and Cos-metidae (Fig. 5b,f, respectively) up to 40 in Cranaidae(Figs 5g and 7a). Within the family Gonyleptidae,there was also variation in pore counts between sub-families. A cobanine and a goniosomatine (Cobaniapicea [Fig. 7b] and Heteromitobates discolor [Fig. 7d])bore a maximum of 21 pores per aggregation, whereasthe heteropachyline Pseudopucrolia mutica bore four tonine pores per aggregation (Fig. 7e).Epedanoidea (Fig. 7f–h) generally bore the same

number of pores per aggregation as basally branchingGonyleptoidea. Tithaeids bore four to five pores(Fig. 7g), whereas epedanids bore two to three (Pseud-oepedanus dolensis; not shown) or up to seven pores(Alloepedanus sp.; Fig. 7f). Similarity in position andclear association with the base of trichomes make thepresence of the TAP unambiguous in theseEpedanoidea. In some cases, epedanoids also boreintertwined setae (e.g. Fig. 7f), comparable with thoseof gonyleptoids.Similarly, Pyramidopidae bore a reduced number of

pores (two to four), associated with very characteristiccurved trichomes in a symmetrical disposition (Figs 6dand 7i). The position is comparable with that seen inthe Gonyleptoidea families.The aggregate pores of the three Samooidea families

(Figs 6e–g and 7j) were different from those ofGonyleptoidea (Fig. 5) and Epedanoidea (Fig. 6a–c).The position was more lateral than in Gonyleptoideaand fewer pores were present (two to three). Associa-

Table 1Tests for phylogenetic signal for each morphological character, indi-cating retention index and Fritz and Purvis D test

CharacterRetentionindex (ri) D

P (H0)—randomphylogeneticstructure

P (HB)—Brownianor moreconservedphylogeneticstructure

Ultrametric topologyDC 1.000 �2.019 0 1.000MPS 0.909 �2.002 0 1.000PTG 1.000 �2.581 0 1.000PTG* 0.889 �2.230 0 1.000TAP 0.897 �1.143 0 0.968TAP* 0.905 �1.377 0 1.000ML topologyDC 1.000 �1.903 0 1.000MPS 0.909 �1.950 0 1.000PTG 1.000 �2.192 0 1.000PTG* 0.889 �1.898 0 1.000TAP 0.897 �1.547 0 1.000TAP* 0.929 �1.462 0 1.000

Probabilities of random and Brownian phylogenetic structure asshown (H0, null hypothesis of random phylogenetic structure; HB,alternative hypothesis of Brownian phylogenetic structure). DC, dou-ble claws on legs III and IV; MPS, metatarsal paired slits; PTG,proximal tarsomeric gland; TAP, tarsal aggregate pores. PTG andTAP with asterisks indicate broadly interpreted character states, asdepicted in Fig. 9.

Table 2Pairwise tests for correlated evolution using BayesTraits ver. 1.0

Character pair �ln (LI) �ln (LD)Likelihoodratio P

Ultrametric topologyDC + MPS �16.940 �15.847 2.187 0.7014DC + PTG �12.702 �11.552 2.301 0.6805DC + TAP �27.000 �23.897 6.206 0.1842MPS + PTG �17.261 �16.310 1.902 0.7537MPS + TAP �31.560 �25.775 11.570 0.0208*PTG + TAP �27.321 �26.897 0.850 0.9316

ML topologyDC + MPS �15.799 �14.739 2.119 0.7139DC + PTG �12.598 �11.491 2.214 0.6964DC + TAP �26.345 �22.624 7.441 0.1143MPS + PTG �16.241 �15.305 1.873 0.7590MPS + TAP �29.988 �27.595 4.787 0.3098PTG + TAP �26.787 �26.279 1.017 0.9071

Likelihoods of character evolution under dependent and indepen-dent evolution, and associated likelihood ratios, are indicated. Aster-isk indicates statistically significant trait correlation. Abbreviationsas in Table 1.


tion with trichomes was ambiguous, because the poreswere very close to or at the base of trichomes of theventral scopulae (characteristic of Samooidea), whichwere all broken during the SEM preparations(Fig. 6g).In the triaenonychid Larifuga cf. capensis Lawrence,

1931, pores were present but more sparsely distributedand did not form an aggregation (Fig. 6h). We observedpores in the same region of legs I and II, also in smallnumbers and not positioned in sagittal symmetry.

Ancestral state reconstruction

Reconstruction of morphological character evolutionwas based on two alternative topologies of Laniatoresrelationships (Sharma and Giribet, 2011). For eithertopology, MPS was gained at the root of Laniatoresand secondarily lost in Sandokanidae (cost = 2)(Fig. 8).Under conservative coding (presence in Icaleptidae,

Fissiphalliidae, and Zalmoxidae), PTG was gainedonce in the most recent common ancestor of thesethree Zalmoxoidea families (cost = 1) (Fig. 8). Abroader definition of this character (i.e. additionalpresence in Escadabiidae) resulted in two independentgains of this character (cost = 2; Fig. 9); an equallyparsimonious reconstruction (under ACCTRAN; Fig.S1) is a gain at the base of Zalmoxoidea, with subse-quent loss in the clade Kimulidae + Guasiniidae.Conservative scoring of TAP resulted in alternative

reconstructions contingent upon tree topology. Underthe dated tree topology, TAP is independently gainedat the base of Gonyleptoidea + Epedanoidea and inPyramidopidae, and subsequently lost in the epeda-noid families Sandokanidae and Petrobunidae(cost = 4). Under the maximum likelihood topology,four independent acquisitions of TAP are recon-structed (cost = 4) (Fig. 8). However, an alternativemost parsimonious reconstruction recovers a gain ofTAP at the node corresponding to the non-phalango-did and non-sandokanid Grassatores, followed bythree independent losses in Petrobunidae, Assamiidae,and the clade Samooidea + Zalmoxoidea (ACC-TRAN; Fig. S2).A broader interpretation of TAP scored a presence

of this character for all Samooidea (Fig. 9). As withthe conservative scoring, the reconstruction is sensitiveto the tree topology. Based on the dated tree, threeindependent acquisitions of TAP are reconstructed—in Pyramidopidae, Samooidea, and the clade Gonylep-toidea + Epedanoidea. The character is secondarilylost in Sandokanidae and Petrobunidae (cost = 5).Under the maximum likelihood topology, TAP isreconstructed as under ACCTRAN optimization ofthe conservative coding on the same tree (cost = 4)(Fig. 9).

Phylogenetic signal and character independence

Inference of phylogenetic signal was based on theretention index (ri) and the Fritz and Purvis D test fordiscrete characters. For comparison with a non-homo-plastic character, we also scored the presence of dou-ble claws on legs III and IV, an unambiguoussynapomorphy of Grassatores (ri = 1). Parsimony-based reconstruction of all three morphological char-acters reported here (MPS, PTG, and TAP) on bothalternative tree topologies resulted in low homoplasyand demonstrable phylogenetic signal (ri � 0.85)(Table 1). Broader interpretation and scoring of PTGand TAP had different effects on the ri, with a lowervalue for PTG (ri = 0.889 for either topology) but ahigher value for TAP (ri = 0.917 for ultrametric topol-ogy; ri = 0.929 for ML topology).Consistent with the high values of ri, negative values

of D were obtained for all characters, indicative ofnon-random phylogenetic signal and high evolutionaryconservation. Tests for phylogenetic signal for all char-acters and character codings rejected the null hypothe-sis of random phylogenetic signal (P << 0.01),favouring instead a model of strong phylogeneticstructure (P � 0.95) (Table 1).Pairwise tests for trait correlation revealed no signifi-

cant associations (at a = 0.05) between characters, savefor one. The null hypothesis of independent evolutionin MPS and TAP was rejected (P = 0.0208) in favourof a model of correlated evolution, but only for theultrametric topology (Table 2).

Discussion

The pursuit of morphological characters to eluci-date Laniatores superfamilial relationships offersmany character systems for study, but few withdemonstrable utility at the superfamilial level. Herewe examined cuticular structures of the distal walkingleg podomeres, defining three new morphologicalcharacters for study. We review morphological evolu-tion and systematic contributions of each character inturn.

Metatarsal paired slits reinforce the autapomorphicnature of Sandokanidae

Among Opiliones, metatarsal paired slits (MPS)occur exclusively in Laniatores and constitute a con-served character with respect to the number of slitsand their position on the calcaneus of the podomere(Willemart et al., 2007, 2009a) (Fig. 2). Interfamilialdifferences exist in the underlying integument, but thismay constitute a separate character associated with thecalcaneus, and was not investigated here. The MPS


almost constitute an unambiguous synapomorphy ofLaniatores, save for the secondary loss of this charac-ter in Sandokanidae (Fig. 2k).Sandokanidae is the only family of Laniatores with

phylogenetic affinities highly sensitive to algorithmictreatment. It constitutes a curious lineage, once heldto be the sister group of the remaining Grassatores(e.g. Martens, 1976; Martens et al., 1981), whose phy-logenetic position has remained ambiguous in morerecent years (Schwendinger, 2007; Sharma and Giri-bet, 2009, 2011). Morphologically, the family har-bours numerous autapomorphies, such as a scutumcompletum (which occurs in the distantly related Cyp-hophthalmi), highly reduced tarsalia, and a laterallycompressed ovipositor, among others (Schwendinger,2007; Sharma and Giribet, 2009). The secondary lossof MPS reconstructed on both tree topologies rein-forces the atypical morphology and uncertain place-ment of this family among Grassatores. The absenceof MPS in Sandokan truncatus is all the more intrigu-ing due to the presence of paired slits on the tibiaeof all walking legs of this species (arrowheads inFig. 2l). These tibial slits may represent a transloca-tion of the MPS. Mechanistically, translocation canbe achieved by modification of a developmental pro-gram along the proximo-distal axis, displacing slitorgans from the metatarsi to the tibiae. Alternatively,independent evolution of slit organs on a more proxi-mal podomere, with subsequent loss of the metatarsalcounterparts, may result in the state observed inextant sandokanids. The developmental genetic basisof proximo-distal axial patterning in harvestmen isknown to be highly comparable with that of spidersat the level of leg gap gene activity (Sharma et al.,2012). The subsequent mechanism by which segmen-tal identities are conferred upon individual podomeresis unknown in arachnids (Pechmann et al., 2010), butmay be analogous to its counterpart in insects (Koj-ima, 2004; Angelini and Kaufman, 2005). If thismechanism were deduced for arachnids, and a reliablegenetic marker for the MPS were discovered, it maybe possible to test by gene expression whether thepaired slits of Sandokanidae and other Laniatores arehomologous and represent translocation, or whetherthey are patterned by independently evolved mecha-nisms. The recent advent of gene-silencing techniquesin harvestmen also proffers the possibility of testingfunctionally the association between cuticular struc-tures and the podomeres that harbour them (Sharmaet al., 2013).As few behavioural data have been collected for

most Laniatores, much less sandokanids, the functionsof tibial and metatarsal slits are not well understood.For the present, we treated the tibial slits of Sandok-anidae as a different character from the MPS despitetheir superficial similarity.

Proximal tarsomeric glands support relationships withinZalmoxoidea

On the basis of molecular sequence data and a fewmorphological characters (including appendicular glandstructures; Willemart et al., 2010), two erstwhile con-stituents of Samooidea—Escadabiidae and Kimulidae—were transferred to Zalmoxoidea (Sharma and Giri-bet, 2011; Sharma et al., 2011). The occurrence of agland on the proximal-most tarsomere of leg IV (PTG)supports a clade comprised of Icaleptidae, Fissiphallii-dae, and Zalmoxidae, commonly termed “core Zal-moxoidea”. This relationship was previously supportedby other morphological characters pertaining to themale copulatory apparatus (Kury and P�erez-Gonz�alez,2002; Kury and Perez-Gonzalez, 2007; Pinto-da-Rochaand Kury, 2003), as well as a molecular phylogeny ofZalmoxoidea (Sharma and Giribet, 2012). Absence ofthe PTG in Guasiniidae also corroborates the exclusionof this cryptic family from the core Zalmoxoidea(Fig. 4d); previous description of the male copulatoryapparatus had similarly suggested placement of guasi-niids among basal groups (Pinto-da-Rocha, 2007).A gland comparable with the PTG occurs on the

elongate proximal tarsomere of legs III and IV in Es-cadabiidae (Fig. 4a)—a character we found difficult tointerpret unambiguously. We therefore scored the PTGalternately as present and absent for Escadabiidae intwo separate reconstructions. Two equally parsimoni-ous reconstructions indicate that the PTG is indepen-dently acquired in Escadabiidae and the coreZalmoxoidea (under DELTRAN; Fig. 9), or acquiredat the base of Zalmoxoidea and lost in Guasinii-dae + Kimulidae (under accelerated transformation[ACCTRAN]; Fig. S1). We note that the former recon-struction is more parsimonious under the alternativetopology of Zalmoxoidea proposed by Sharma andGiribet (2012), wherein Guasiniidae is nested withinEscadabiidae. These results disfavour the homology ofthe core Zalmoxoidea PTG and the escadabiid struc-ture. However, we note that these results are highlycontingent upon taxonomic sampling, and numerouszalmoxoids of uncertain familial placement (e.g. Co-stabrimma, Phalangodinella; Kury, 2003; Giribet et al.,2010; Sharma and Giribet, 2012) should be scored forthis character and placed definitively in the phylogenyof Zalmoxoidea in order to elucidate the nature of theescadabiid proximal tarsomeric character state.Moreover, the comparably elongate proximal-most

tarsomere of Escadabiidae and Guasiniidae, both ofwhich bear pores along the proximo-distal axis, accordwith the alternative topology of Zalmoxoidea (Sharmaand Giribet, 2012). The denser sampling of Escadabii-dae from the Guyana Shield in that study, togetherwith the similar morphology of the guasiniid and esca-dabiid tarsomeric pores, potentially threatens the sys-


tematic validity of Guasiniidae. Whether guasiniidsconstitute troglomorphic escadabiids is an intriguingtopic for future investigation.

Tarsal aggregate pores suggest phylogenetic proximityof Epedanoidea and Gonyleptoidea

The identity of the sister group of Gonyleptoidea,the most speciose superfamily of Laniatores, remainsan outstanding and unresolved issue in Opiliones phy-logeny. Gonyleptoidea (including Stygnopsidae) is pos-tulated to be sister to Assamioidea on the basis ofsimilarities in copulatory structure (Kury, 1993; Giri-bet and Kury, 2007), whereas molecular sequence dataindicate either sister relationship to, or placement in agrade with, Epedanoidea (Sharma and Giribet, 2011)or some combination of epedanoids and assamioids(Giribet et al., 2010).The discovery and sampling of the TAP (Willemart

et al., 2007), strictly defined, represents a morphologicalcharacter potentially revealing of basal Grassatores rela-tionships. TAP occur in one of two Assamioidea families(Pyramidopidae), all Gonyleptoidea, and three of fiveEpedanoidea families (Epedanidae, Podoctidae, and Tit-haeidae). On the ultrametric tree topology, the most par-simonious reconstruction indicates two acquisitions ofTAP, in Pyramidopidae and at the base of Gonyleptoi-dea + Epedanoidea; two independent losses are recon-structed in Petrobunidae and Sandokanidae. On themaximum likelihood tree topology, two most parsimoni-ous reconstructions are possible. Under DELTRAN,TAP is acquired four times independently (Fig. 8), butunder ACCTRAN, TAP is gained once at the base of thenon-phalangodid and non-sandokanid Grassatores, andsubsequently lost in three lineages (Fig. S2).The occurrence of this character in the majority of

Epedanoidea and all Gonyleptoidea is suggestive ofphylogenetic proximity between these superfamilies,consistent with the topologies based on 10 genes (Shar-ma and Giribet, 2011). The incidence of this characterin Pyramidopidae may represent either retention of theplesiomorphic state or an independent acquisition.However, given the diversity of Assamiidae, it is possi-ble that unrepresented basal assamiids also have thischaracter. While we endeavoured to sample multipleassamiids, phylogenetic relationships within this fam-ily, and ipso facto the identities of the basal lineages,are effectively unknown (Kury, 2007).Nevertheless, to determine whether stringent coding

of TAP engendered predisposition to being scored aspresent in Gonyleptoidea and Epedanoidea, weemployed an alternative coding, broadening the defini-tion of TAP to mean a character that met any two ofthe three criteria defined above. The results were gen-erally similar to those under a strict definition of TAP,under either topology (Fig. 9).

The TAP and the PTG are structures of unknownfunction awaiting both histological and behaviouralinvestigation. Previously, Willemart et al. (2007)reported that both males and females of two gonylep-tid species rub legs III and IV on the substrate whilewalking, and males of a gonyleptine (Neosadocus sp.)also engage in this behaviour during male–male com-bat (Willemart et al., 2009b). The execution of thisbehaviour, specifically the dragging of the tarsus, withtwisting of the leg to allow contact between the lateralpodomere and the substrate, is suggestive of a role forthese structures in chemical marking of the substrate,at least in Gonyleptidae.

Appendicular characters bear strong phylogenetic signal

Often the character systems utilized in Laniatores sys-tematics are specific to derived lineages of interest andare inapplicable for broad assessment of basal relation-ships. Such is the case for a suite of characters pertain-ing to coloration, armature, or elements of thecopulatory apparatus that are of great utility, but suitedfor the systematics of Gonyleptoidea (e.g. Pinto-da-Ro-cha, 2002; da Silva and Gnaspini, 2009; Yamaguti andPinto-da-Rocha, 2009), insofar as most other Laniat-ores lack coloration and differ from the gonyleptoidbauplan. The advantage of the appendicular charactersreported here is that they can be scored for all Laniat-ores, in spite of alternative interpretations of characterstates (Figs. 8 and 9). But applicability to all Laniatoresis hardly any guarantee of phylogenetic utility, and wetherefore characterized phylogenetic signal of each char-acter in order to test their predictive power. We usedthe double claws (DC), a synapomorphy of Grassat-ores, as a benchmark for comparison (ri = 1, D < 0).High retention indices (ri � 0.85) and large negative

values of the Fritz and Purvis D for all characters indi-cate strong inherent phylogenetic signal across the La-niatores tree (Table 1). Even when broadly interpreted(PTG and TAP), trait evolution was more strongly con-served (D � –1) than the Brownian expectation (i.e.random walk model; D � 0), with no support for a nullmodel of random phylogenetic structure (i.e. D � 1).These results disfavour the lability of either characterat the level of Laniatores superfamilial relationships.The comparability of D values for PTG and TAP withthat of the double claw—a traditional and bona fidesynapomorphy of Grassatores—supports the use ofsuch appendicular characters for defining superfamily-level synapomorphies. Furthermore, tests for correlatedtrait evolution favour independence of all pairwisecharacter combinations, save for the MPS and TAP(Table 2). A significant correlation between the twomay be driven by secondary losses of both charactersin Sandokanidae, but this correlation was obtainedonly on the ultrametric tree topology.


Our results are consistent with evolutionary conser-vation of appendicular cuticular characters and theirphylogenetic utility for higher-level laniatorid system-atics. Ongoing efforts toward identifying morphologi-cal synapomorphies of Laniatores superfamilies andconstituent clades are anticipated to sample these andother appendicular characters more broadly (Willemartet al., 2009a).

Conclusion

The three new morphological structures investigatedhere constitute independent characters with strongphylogenetic signal that shed light on Laniatores sys-tematics. MPS constitute a synapomorphy for Laniat-ores, with a secondary loss in the autapomorphicSandokanidae. PTG is a synapomorphy of the coreZalmoxoidea. TAP is suggestive of a close relationshipbetween Gonyleptoidea and Epedanoidea.

Acknowledgments

We are indebted to Adam Graham and Dave Lange,Center for Nanoscale Systems, and to Enio Mattos andPhilip Lenktaitis, IB-USP, for support with the scan-ning electron microscopy. Training and assistance toG. Gainett were provided by Gisele Y. Kawauchi andErin McIntyre. Comments from Lorenzo Prendini andtwo anonymous reviewers improved an earlier draft ofthe manuscript. G. Gainett was supported by FAPESP(Fundac�~ao do Amparo �a Pesquisa de S~ao Paulo) 2011/11527-4 and FAPESP grant 2010/00915-0 to R.H.W.P.P. Sharma was supported by an NSF PostdoctoralResearch Fellowship in Biology DBI #1202751. R.Pinto-da-Rocha was supported by FAPESP grant2008/06604-7. This work was supported by NSF DEB#1144417 (Collaborative Research: ARTS: Taxonomyand systematics of selected Neotropical clades of arach-nids) to G. Giribet.

References

Agnarsson, I., Miller, J.A., 2008. Is ACCTRAN better thanDELTRAN? Cladistics 24, 1032–1038.

Angelini, D.R., Kaufman, T.C., 2005. Insect appendages andcomparative ontogenetics. Dev. Biol. 286, 57–77.

Derkarabetian, S., Steinmann, D.B., Hedin, M., 2010. Repeated andtime-correlated morphological convergence in cave-dwellingharvestmen (Opiliones, Laniatores) from montane western NorthAmerica. PLoS ONE 5, e10388.

Fritz, S.A., Purvis, A., 2010. Selectivity in mammalian extinctionrisk and threat types: a new measure of phylogenetic signalstrength in binary traits. Conserv. Biol. 24, 1042–1051.

Garwood, R.J., Dunlop, J.A., Giribet, G., Sutton, M.D., 2011.Anatomically modern Carboniferous harvestmen demonstrateearly cladogenesis and stasis in Opiliones. Nat. Commun. 2, 444.

Giribet, G., Rambla, M., Carranza, S., Bagu~n�a, J., Riutort, M.,Ribera, C., 1999. Phylogeny of the arachnid order Opiliones(Arthropoda) inferred from a combined approach of complete18S and partial 28S ribosomal DNA sequences and morphology.Mol. Phylogenet. Evol. 11, 296–307.

Giribet, G., Edgecombe, G.D., Wheeler, W.C., Babbitt, C., 2002.Phylogeny and systematic position of Opiliones: a combinedanalysis of chelicerate relationships using morphological andmolecular data. Cladistics 18, 5–70.

Giribet, G., Kury, A.B., 2007. Phylogeny and Biogeography. In:Pinto-da-Rocha, R., Machado, G., Giribet, G. (Eds.),Harvestmen: The Biology of Opiliones. Harvard University Press,Cambridge, pp. 62–87.

Giribet, G., Vogt, L., P�erez-Gonz�alez, A., Sharma, P., Kury, A.B.,2010. A multilocus approach to harvestman (Arachnida:Opiliones) phylogeny with emphasis on biogeography and thesystematics of Laniatores. Cladistics 26, 408–437.

Gnaspini, P., Rodrigues, G.C.S., 2011. Comparative study of themorphology of the gland opening area among Grassatoresharvestmen (Arachnida, Opiliones, Laniatores). J. Zool. Syst.Evol. Res. 49, 273–284.

Hennig, W., 1966. Phylogenetic Systematics. University of IllinoisPress, Urbana, IL, USA.

Kojima, T., 2004. The mechanism of Drosophila leg developmentalong the proximodistal axis. Dev. Growth Differ. 46, 115–129.

Kury, A.B., 1993. An�alise filogen�etica de Gonyleptoidea (Arachnida,Opiliones, Laniatores). PhD thesis, Universidade de S~ao Paulo,S~ao Paulo.

Kury, A.B., 2003. Annotated catalogue of the Laniatores of the NewWorld (Arachnida, Opiliones). Revta ib�er. Aracnol., vol. especialmonografico. 1: 1–337.

Kury, A.B., 2007. Assamiidae Sørensen, 1886. In: Pinto-da-Rocha,R., Machado, G., Giribet, G. (Eds.), Harvestmen: The Biologyof Opiliones. Harvard University Press, Cambridge, pp. 173–176.

Kury, A.B., 2011. Order Opiliones. In: Zhang, Z.-Q. (Ed.), AnimalBiodiversity. Zootaxa 3148, 112–114.

Kury, A.B., P�erez-Gonz�alez, A., 2002. A new family of Laniatoresfrom northwestern South America (Arachnida, Opiliones). Revtaib�er. Aracnol., 6: 3–11.

Kury, A.B., Perez-Gonzalez, A., 2007. Zalmoxidae Sørensen, 1886.In: Pinto-da-Rocha, R., Machado, G., Giribet, G. (Eds.),Harvestmen: The Biology of Opiliones. Harvard University Press,Cambridge, pp. 243–246.

Maddison, W.P., Maddison, D.R., 2011. Mesquite: a modularsystem for evolutionary analysis. Version 2.75 (http://mesquiteproject.org).

Martens, J., 1976. Genitalmorphologie, system und phylogenie derweberknechte (Arachnida, Opiliones). Entomologica Germanica,Zeitschrift f€ur das Gesamtgebiet der wissenschaftlichenEntomologie 3, 51–68.

Martens, J., 1980. Versuch eines Phylogenetischen Systems derOpiliones. In: Gruber, J. (Ed.), Proceedings 8th InternationalCongress of Arachnology, Vienna. Verhandlungen. Verlag H.Egerman, Vienna, Austria. pp. 355–360.

Martens, J., 1986. Die Grossgliederung der Opiliones und dieEvolution der Ordnung (Arachnida). Actas del X CongresoInternacional de Aracnolog�ıa, Jaca 1, 289–310.

Martens, J., Hoheisel, U., G€otze, M., 1981. Vergleichende Anatomieder Leger€ohren der Opiliones als Beitrag zur Phylogenie derOrdnung (Arachnida). Zoologische Jahrb€ucher, Abteilung f€urAnatomie und Ontogenie der Tiere 105, 13–76.

Orme, C.D.L., Freckleton, R.P., Thomas, G.H., Petzoldt, T., Fritz,S.A., Isaac, N.J.B., 2012. Caper: comparative analyses ofphylogenetics and evolution in R. R package version 0.5.

Pagel, M., Meade, A., 2006. Bayesian analysis of correlatedevolution of discrete characters by reversible-jump Markov chainMonte Carlo. Am. Nat. 167, 808–825.

Pechmann, M., Khadjeh, S., Sprenger, F., Prpic, N.-M., 2010.Patterning mechanisms and morphological diversity of spiderappendages and their importance for spider evolution. Arthr.Struct. Dev. 39, 453–467.


Pinto-da-Rocha, R., 2002. Systematic review and cladistic analysis ofthe Caelopyginae (Opiliones, Gonyleptidae). Arquivos deZoologia, S~ao Paulo 36, 357–464.

Pinto-da-Rocha, R., 2007. Guasiniidae Gonzalez-Sponga, 1997. In:Pinto-da-Rocha, R., Machado, G., Giribet, G. (Eds.),Harvestmen: The Biology of Opiliones. Harvard University Press,Cambridge, pp. 204–205.

Pinto-da-Rocha, R., Kury, A.B., 2003. Third species of Guasiniidae(Opiliones, Laniatores) with comments of the familialrelationships. J. Arachnol. 31, 394–399.

Rambla, M., 1993. Maiorerus randoi n. gen., n. sp., the firstlaniatorid from a Canary Island cave (Opiliones, Phalangodidae).M�em. Biosp�eol. 20, 177–182.

Schwendinger, P.J., 2007. Oncopodidae Thorell, 1876. In: Pinto-da-Rocha, R., Machado, G., Giribet, G. (Eds.), Harvestmen: TheBiology of Opiliones. Harvard University Press, Cambridge, pp.211–214.

Sharma, P., Giribet, G., 2009. Sandokanid phylogeny based on eightmolecular markers—the evolution of a southeast Asian endemicfamily of Laniatores (Arachnida, Opiliones). Mol. Phylogenet.Evol. 52, 432–447.

Sharma, P.P., Giribet, G., 2011. The evolutionary and biogeographichistory of the armoured harvestmen – Laniatores phylogenybased on ten molecular markers, with the description of two newfamilies of Opiliones (Arachnida). Invertebr. Syst. 25, 106–142.

Sharma, P.P., Giribet, G., 2012. Out of the Neotropics: lateCretaceous colonization of Australasia by American arthropods.Proc. Biol. Sci. 279, 3501–3509.

Sharma, P.P., Prieto, C.E., Giribet, G., 2011. A new family ofLaniatores (Arachnida: Opiliones) from the Afrotropics.Invertebr. Syst. 25, 143–154.

Sharma, P.P., Schwager, E.E., Extavour, C.G., Giribet, G., 2012.Evolution of the chelicera: a dachshund domain is retained in thedeutocerebral appendage of Opiliones (Arthropoda, Chelicerata).Evol. Dev. 14, 522–533.

Sharma, P.P., Schwager, E.E., Giribet, G., Jockusch, E.L.,Extavour, C.G., 2013. Distal-less and dachshund pattern bothplesiomorphic and apomorphic structures in chelicerates: RNAinterference in the harvestman Phalangium opilio (Opiliones).Evol. Dev. doi:10.1111/ede.12029

Shultz, J.W., 1998. Phylogeny of Opiliones (Arachnida): anassessment of the “Cyphopalpatores” concept. J. Arachnol. 26,257–272.

da Silva, M.B., Gnaspini, P., 2009. A systematic revision ofGoniosomatinae (Arachnida: Opiliones: Gonyleptidae), with acladistic analysis and biogeographical notes. Invertebr. Syst. 23,530–624.

Willemart, R.H., Giribet, G., 2010. A scanning electron microscopicsurvey of the cuticle in Cyphophthalmi (Arachnida, Opiliones)with the description of novel sensory and glandular structures.Zoomorphology 129, 175–183.

Willemart, R.H., Farine, J.P., Peretti, A.V., Gnaspini, P., 2006.Behavioral roles of the sexually dimorphic structures in the maleharvestman, Phalangium opilio (Opiliones, Phalangiidae). Can. J.Zool. 84, 1763–1774.

Willemart, R.H., Chelini, M.C., de Andrade, R., Gnaspini, P.,2007. An ethological approach to a SEM survey on sensorystructures and tegumental gland openings of two neotropicalharvestmen (Arachnida, Opiliones, Gonyleptidae). Ital. J. Zool.74, 39–54.

Willemart, R.H., Farine, J.-P., Gnaspini, P., 2009a. Sensory biologyof Phalangida harvestmen (Arachnida, Opiliones): a review, withnew morphological data on 18 species. Acta Zool. 90, 209–227.

Willemart, R.H., Osses, F., Chelini, M.C., Mac�ıas-Ord�o~nez, R.,Machado, G., 2009b. Sexually dimorphic legs in a neotropicalharvestman (Arachnida, Opiliones): ornament or weapon? Behav.Process. 80, 51–59.

Willemart, R.H., Perez-Gonzalez, A., Farine, J.-P., Gnaspini, P.,2010. Sexually dimorphic tegumental gland openings inLaniatores (Arachnida, Opiliones), with new data on 23 species.J. Morphol. 271, 641–653.

Yamaguti, H.Y., Pinto-da-Rocha, R., 2009. Taxonomic review ofBourguyiinae, cladistic analysis, and a new hypothesis ofbiogeographic relationships of the Brazilian Atlantic rainforest(Arachnida: Opiliones, Gonyleptidae). Zool. J. Linn. Soc. 156,319–362.

Supporting Information

Additional Supporting Information may be found inthe online version of this article:Fig. S1. Ancestral state reconstruction of the proxi-

mal tarsomeric gland under ACCTRAN optimizationon dated (left) and maximum likelihood (right) treetopologies of Laniatores.Fig. S2. Ancestral state reconstruction of the tarsal

aggregate pores under ACCTRAN optimization onmaximum likelihood tree topology of Laniatores.


Appendix

1

Listofspeciesexamined

inthisstudy,withaccessionnumbersandgender

whereavailable

Species

Taxonomy

Voucher

#Locality

Latitude

Longitude

Erebomaster

flavescensflavescens

Cope,

1872

Travunioidea:Travuniidae

MCZ

DNA101444

USA:PattonCave,

Deam

Wilderness

Area,Indiana

39.044586

�86.338429

Synthetonychia

glacialisForster,1954

Synthetonychiidae

MCZ

DNA101718

New

Zealand:Minnehahawalk,Fox

Glacier,South

Island

�43.46861

170.0181

Synthetonychia

sp.

Synthetonychiidae

MCZ

64591

New

Zealand:RahuScenic

Reserve,

Nea

Springs

�42.33222

172.1711

FumontanadeprehendorShear,1977

Triaenonychoidea:

Triaenonychidae

MCZ

DNA100701

USA:YanceyCO

Hwy19W,EofSpivey

Gap,~3

miWofSioux

36.04833

�82.40714

Larifugacf.capensis

Triaenonychoidea:

Triaenonychidae

MCZ

DNA100727

South

Africa:New

landsForest.Table

Mountain,CapeProvince

�33.97917

18.45

NeopygoplussiamensisSuzuki,1985

Assamioidea:Assamidae

MCZ

DNA104860

Thailand:Kanchanaburi

14.63567

98.99733

Montalenia

sp.

Assamioidea:Assamidae

MCZ

DNA105667

Cameroon:CampoReserve,

Littoral

Province

�2.746111

53.51333

ConommaoedipusRoew

er,1949

Assamioidea:Pyramidopidae

MCZ

DNA101051

EquatorialGuinea:Musola,Bioko

13.43139

8.618889

cf.Pyramidopssp.

Assamioidea:Pyramidopidae

MCZ

DNA101432

EquatorialGuinea:MontanaChocolate,

NiefagoDistrict

1.756944

10.28417

Pseudoepedanusdolensis

Suzuki,1969

Epedanoidea:Epedanidae

MCZ

DNA101438

Thailand:HuaykhokMa,DoiSuthep,

ChiangMai

�11.26

122.48

Alloepedanussp.

Epedanoidea:Epedanidae

MCZ

DNA104862

Thailand:ChiangMaiDoi

18.55267

98.48

Petrobunusspinifer

Sharm

a&

Giribet,2011

Epedanoidea:Petrobunidae

MHNG

PAL

09/04

Philippines:BulalacaoWaterfall,ElNido,

Palawan

11.22806

119.4667

Petrobunusschwendingeri

Sharm

a&

Giribet,2011

Epedanoidea:Petrobunidae

MCZ

DNA103572

Philippines:Panay,Sibaliw

�11.46

122.48

Santobiussp.

Epedanoidea:Podoctidae

MCZ

DNA104930

Fiji:VituLevuSavura

Park

�18.07083

178.4444

Sandokantruncatus

Thorell,1891

Epedanoidea:Sandokanidae

MCZ

DNA101099

Singapore:BukitTim

ahNature

Reserve,

Jungle

FallValley

10.34814

103.7765

Gnomulussp.(rostratusgroup)

Epedanoidea:Sandokanidae

MCZ

DNA102592

Thailand:KoSiray,Phuket

Province

7.885278

98.43861

Tithaeussp.

Epedanoidea:Tithaeidae

MCZ

DNA104074

Malaysia:Terengganustate

5.897778

102.7339

Torreanaspinata

Avram,1977

Gonyleptoidea:Agoristenidae

MCZ

DNA105839

Cuba:SantiagodeCuba

20.01667

�75.63333

Gnidia

holm

bergi,Sørensen,1884

Gonyleptoidea:Cosm

etidae

MCZ

DNA100398

Argentina

n/a

n/a

Phareicranausserratotibialis

Roew

er,1932

Gonyleptoidea:Cranaidae

MCZ

DNA100426

Trinidad&

Tobago:Mt.St.Benedict’s,

Trinidad

10.66361

�61.39889

Cobania

picea

Bertkau,1880

Gonyleptoidea:Gonyleptidae:

Cobaniinae

MZSP21709

Brazil:Itamonte,Brejo

daLapa,

MinasGerais

�22.3625

�44.735

Heteromitobatesdiscolor

Sorensen,1884


Goniosomatinae

MZSP30051

Brazil:Rio

deJaneiro,Estrada

Parati-C

unha

�23.19916667

�44.83

Iporangaia

pustulosa

Mello-Leitao,1935


Pro-gonyleptoidellinae

MZSP16709

Brasil:SaoPaulo,Iporanga

�24.58333

�48.58333

Pseudopucroliamutica

(Perty,1833)


Heteropachylinae

MZSP26934

Brasil:Bahia,Salvador,Campus

UFBA

–Ondina

�15.61667

�39.7

Discocyrtoides

nigricans

Mello-Leitao,1922


Mitobatinae

IBSP7940

Brasil:S~ ao

Paulo,Jundia� ıReservaBiol� ogica

MunicipaldaSerra

doJapi

�23.2333333

�46.9666666


Appendix

1(C

ontinued)

Species

Taxonomy

Voucher

#Locality

Latitude

Longitude

Rhopalocranausalbilineatus

Roew

er,1932

Gonyleptoidea:Manaosbiidae

MCZ

DNA100331

Trinidad&

Tobago:Mt.St.Benedict’s,

Trinidad

10.66361

�61.39889

Stygnusmultispinosus

Piza,1938

Gonyleptoidea:Stygnidae

MZSP15166

�14.01667

�48.3

Hoplobunussp.

Gonyleptoidea:Stygnopsidae

MCZ

DNA101418

Mexico:25km

ofValle

NacionalonHighway

125Ooixaca-Tuxtepec,Ooaxaca

17.05

�96.05

Scotolemonlespesi

Lucas,1860

Phalangodoidea:Phalangodidae

MCZ

DNA100326

Spain:LaFagedad’enJorda,Girona

42.1575

�2.516667

Metabiantessp.

Samooidea:Biantidae

MCZ

DNA100704

Swaziland:Sarahcampsite,Mlabula

Nature

Reserve

�26.19556

31.99

Badessa

sp.

Samooidea:Samoidae

MCZ

DNA104600

Fiji:Viro,Ovalau

�17.05

178.7611

Neoscotolemons.sp.

Samooidea:Stygnommatidae

MCZ

DNA101427

Cuba:Fuentescavesystem

,Pinardel

R� ıo

22.46472

�83.99611

Stygnommabispinatum

Goodnight&

Goodnight,1953

Samooidea:Stygnommatidae

MCZ

DNA105636

Mexico:SierraMorena,Chiapas

16.15342

�93.60078

BaculigerusmilenaeKury,2012

Zalm

oxoidea:Escadabiidae

MCZ

DNA100640

Brazil:ParqueEcol� ogicodeCoc� o,Fortaleza

�3.716667

�38.5

Fissiphalliussp.

Zalm

oxoidea:

Fissiphalliidae

MCZ

DNA104057

Colombia:SantuariodeF

aunayFlora

Iguaque,

Departamento

deBoyac~ a

Brazil:Trailleading

5.711944

�73.46222

Guasinia

sp.

Zalm

oxoidea:Guasiniidae

MCZ

DNA107084

East

ofCaicub�ı,Rio

Jafari,Municipalidade

Caracara� ı,Roraim

aColombia:Reserva

�1.02897

�62.08722

Icaleptessp.

Zalm

oxoidea:Icaleptidae

MCZ

DNA101420

NaturalR� ıo

~ Namb�ı,Munic� ıpio

de

Barbacoas

1.285

�78.07361

Kim

ula

goodnightorum,

Silhavy,1969

Zalm

oxoidea:Kim

ulidae

MCZ

DNA105837

Cuba:SantiagodeCuba,GranPiedre

20.01667

�75.63333

Zalm

oxissp.

Zalm

oxoidea:Zalm

oxidae

MCZ

DNA106885

Vanuatu:EspirituSanto

�15.25

166.8333

Zalm

oxissp.

Zalm

oxoidea:Zalm

oxidae

MCZ

DNA102502

Indonesia:Bantimurung-Bulusaraung

N.P.,SulawesiSelatan

�5.042222

119.735278

Sabaconsp.

Sabaconidae

MCZ

DNA100715

Spain:Valledel

R� ıo

Barragan,Moscoso,

Pontevedra

42.31778

�8.487778

Astrobunusgranullator

Sclerosomatidae/Scle

rosomatinae

MCZ

DNA100707

Spain:Montseny,Barcelona

41.7625

2.364667

Protolophussingularis

Banks,1893

Protolophidae

MCZ

DNA101033

USA:East

base

ofGuatayMountain,

SanDiegoCounty,California

32.83327

�116.54624

CaddoagilisBanks,1892

Caddidae/Caddidae

n/a

Storrs,MansfieldCounty,Connecticut,

USA

41.791216

�72.252081

Heperonem

astomamodestum

(Banks,1894)

Ceratolasm

atidae

MCZ

DNA100312

USA:Oregon

n/a

n/a

Troglosiro

longifossa

Sharm

a&

Giribet,2005

Troglosironidae

MCZ

DNA100398

New

Caledonia:Yahoue,

Province

Sud

�22.1945

166.5017


Documents

cla 12029 120. - Prashant P. SharmaWillemart et al. (2007, 2010) similarly reported the disposition of tegumental gland openings in the legs of numerous Opiliones, with emphasis on