Further revision of the genus Megalopsalis (Opiliones, Neopilionidae), with the description of seven new species

Abstract The Australian harvestmen genus Megalopsalis (Neopilionidae: Enantiobuninae) is recognised as a senior synonym of the genera Spinicrus and Hypomegalopsalis, and seven new species are described in Megalopsalis: Megalopsalis suffugiens, Megalopsalis walpolensis, Megalopsalis caeruleomontium, Megalopsalis atrocidiana, Megalopsalis coronata, Megalopsalis puerilis and Megalopsalis sublucens. A morphological phylogenetic analysis of the Enantiobuninae is also conducted including the new species. Monophyly of Neopilionidae and Enantiobuninae including ‘Monoscutidae’ is corroborated, with the Australasian taxa as a possible sister clade to the South American Thrasychirus.


introduction
The Enantiobuninae (previously Monoscutidae-see Taylor 2011) are the most speciose group of long-legged harvestmen (Eupnoi) in Australasia, with over 45 described species (Taylor 2004(Taylor , 2011. The taxonomy of Australasian Enantiobuninae was long based on a small number of characters of doubtful significance, such as the presence or absence of a distal apophysis on the pedipalpal patella, or a ventral tooth-row on the tarsal claw (Forster 1949). Recent reviews of the genera Pantopsalis Simon, 1879(Taylor 2004 and Megalopsalis Roewer, 1923(Taylor 2011 have refined the definitions of those genera and improved our understanding of enantiobunine systematics; however, the genus Spinicrus Forster, 1949 remained unrevised. The current paper completes the revision of genera of Australasian Enantiobuninae. Forster (1949) established Spinicrus with the type species Pantopsalis tasmanica Hogg, 1910 from Tasmania, distinguishing it from the New Zealand genus Pantopsalis (Hogg, 1910) on the basis of the presence of a ventral tooth-comb on the pedipalpal claw (a character that had been accorded high significance in the artificial classification of Eupnoi used by Roewer 1923). Forster (1949) assigned two new species S. camelus Forster, 1949 andS. stewarti Forster, 1949 to the genus, and suggested that Pantopsalis continentalis Roewer, 1923 of Queensland might also belong to Spinicrus. Hickman (1957) later described two new species from Tasmania, S. nigricans Hickman, 1957 andS. thrypticum Hickman, 1957. Around the same time, Kauri (1954) assigned two species from Western Australia to the genus, S. minimum Kauri, 1954 andS. porongorupense Kauri, 1954.
Despite the small number of included species, Spinicrus was a morphologically heterogeneous assemblage right from its initial establishment (Forster 1949). The absence of a pedipalp apophysis (distinguishing it from Megalopsalis) and the presence of a toothed pedipalp claw remained the only defining characters of the genus; neither of these characters was unique to Spinicrus and both are likely to be plesiomorphic for Enantiobuninae as a whole (Taylor 2011). Taylor and Hunt (2009) separated S. camelus and S. continentale from Spinicrus as part of the morphologically distinct genus Neopantopsalis Taylor & Hunt, 2009 but did not consider the status of the remaining species. The possibility that some of the species, particularly S. nigricans and the Western Australian species, might also need to be transferred to new genera had previously been raised by Hunt (1990) after examination of their distinct spiracle morphologies. In the phylogenetic analysis of Enantiobuninae by Taylor (2011), Spinicrus was not identified as monophyletic. The current paper expands Taylor's (2011) analysis with the addition of seven new species that would have previously been assigned to Spinicrus. Spinicrus is identified as paraphyletic with regard to the genera Megalopsalis and Hypomegalopsalis Taylor, 2011, and all three genera are combined into a single genus Megalopsalis. usually with patella shorter than tibia (slightly longer in M. nigricans); apophysis present or absent on patella; claw with ventral tooth-row. Glans relatively short, broad, more or less subtriangular in ventral view, proximal section usually somewhat inflated dorsally (except in M. nigricans); distal section more or less dorsoventrally flattened. Spiracle with reticulate or partially reticulate covering spines (reduced or absent in M. minima-species group); lace tubercles present or absent.
Comments. The genus Spinicrus as previously defined (Forster 1949) is likely to be non-monophyletic with regard to both Megalopsalis and Hypomegalopsalis, and lacks clear synapomorphies (see phylogenetic analysis below). In contrast, the clade uniting these three genera is characterised by a distinct penis morphology, and they are hence united into a single genus Megalopsalis. The species groups listed above are clades that were consistently recovered in the phylogenetic analysis under varying analytical conditions; those species not placed in groups did not form consistent subgeneric clades across all analyses. Members of the M. serritarsus-and M. leptekes-groups were described by Taylor (2011).
Megalopsalis tasmanica and M. thryptica were described in detail by Hickman (1957), and so are not redescribed here. Both sexes of M. tasmanica can be distin- guished from other Neopilionidae by their distinctive elongate opisthosoma as illustrated by Hickman (1957: fig. 29); this distinction is even more pronounced in the female. See below under M. stewarti for discussion of the distinction between this species and M. thryptica. Ozopore openings very small, circular, not raised on lateral lobes; pedipalpal femur relatively long, more than 1.5× length of prosoma ...Megalopsalis nigricans -Ozopore openings oblong, raised on lateral lobes; pedipalpal femur shorter than or subequal to prosoma .

Diagnosis.
Megalopsalis minima can be distinguished from other members of the M. minima-species group by the heavier denticulation on the dorsal prosomal plate (Fig. 3a); the major males can also be distinguished from other species by the proportionately much longer chelicerae (Fig. 3b). It can be distinguished from M. porongorupensis by the absence of spines on the pedipalpal femur and patella, and from M. suffugiens by the heavily denticulate leg femora (Fig. 3b).
Pedipalps Comments. Unfortunately, the type specimen(s) of M. minima were not available for the present study. This species has been identified based on its original description by Kauri (1954).
Females have been found in association with males of Megalopsalis minima, M. porongorupensis and M. walpolensis (unpublished observations, specimens in WAM). However, as no distinct morphotypes have been distinguished among the likely females, while the ranges of these species overlap, it has not been possible as yet to determine which females are assignable to which species. (Kauri, 1954) Spinicrus porongorupensis Kauri, 1954:8-9, fig. 4c (incorrect original spelling). Spinicrus porongorupense Kauri -Taylor 2004: 76 (spelling  Diagnosis. Megalopsalis porongorupensis is distinguishable from other members of the M. minima-species group by the presence of denticulation on the pedipalp (Kauri 1954 Spiracle (Fig. 2a). Spines almost entirely absent, residual reticulate bases only towards lateral corner; dense field of lace tubercles at lateral corner. Comments. Unfortunately, the type specimen(s) of M. porongorupensis were not available for the present study. This species has been identified based on its original description by Kauri (1954), who clearly figured the denticulate pedipalp.

Megalopsalis porongorupensis
This species shows a relatively large degree of difference in cheliceral size between the largest and smallest individuals, but there is no clear clustering into a larger and a smaller morph. Diagnosis. Megalopsalis suffugiens is readily distinguished from other species of the M. minima-species group by its pale coloration without dark transverse bands on the opisthosoma ( Fig. 5a-b). The spines on the legs (if present) are also proportionately longer and more slender than those found in other species. It can also be distinguished from M. minima by the lack of denticles on the ocularium and median propeltidial area (Fig. 5a) and from M. porongorupensis by the lack of denticles on the pedipalps. Description. MALE (N = 5). Prosoma length 0.75 (0.68-0.83), width 1.69 (1.56-1.86); total body length 2.21 (2.08-2.30). Dorsal prosomal plate unarmed (Balgair Station specimens) or with few denticles on anterior propeltidial area (Hampton Tableland specimens); patched tan and iridescent white with scattered darker mottling. Mesopeltidium with distinct transverse row of black setae. Metapeltidium and anterior part of opisthosoma mottled tan and silver. Posterior part of opisthosoma silver with transverse bands of dark brown mottling.
Penis (Figs 5c-d). Shaft broad, tendon relatively short; bristle groups well-developed. Glans short, broad, triangular in dorsal view; in line with shaft; dorsal side evenly convex; not significantly flattened distally. Pores shallowly recessed.
Spiracle (Figs   Variation. Males show a noticeable variation in the size of the chelicerae that correlates with the development of armature on the legs; however, the variation is not as large as that found in Megalopsalis minima, and it is uncertain at present whether variation is continuous or a distinction occurs between major and minor males. Further specimens are also required to establish whether the difference in dorsal armature recorded above between Balgair Station and Hampton Tableland specimens indicate separate populations in these localities.
Etymology. From the Latin suffugio, to take shelter, to reflect the finding of specimens of this species within caves in the arid Nullarbor.
Comments. All specimens of Megalopsalis suffugiens recorded to date were collected in caves; however, M. suffugiens does not show any obvious troglobitic adaptations. The eyes remain well-developed and the legs are proportionately only slightly longer than in other Megalopsalis species. It seems more likely that M. suffugiens only uses the caves as damp refugia during the day, emerging at night to feed. This suggestion is supported by the collection of at least one specimen (WAM T72298) from a cave entrance at nightfall. Anterior propeltidial area cream, remainder of propeltidium golden-brown with mottled black patches on anterior corners of dorsal prosomal plate and lateral shelves. Prosoma mostly unarmed, except few small scattered denticles on lateral edge of dorsal prosomal plate near odoriferous glands. Odoriferous glands visible as black patches through cuticle. Ocularium dark golden-brown, with row of small low denticles around each eye. Mesopeltidium, metapeltidium and opisthosoma with transverse band of mottled black across goldenbrown background of each segment, broken by tan or iridescent white spots. Coxae cream with mottled purple patches at distal ends; venter of opisthosoma cream dusted with purple, condensing to more solid patches laterally.
Penis (Figs 8d-e). Left anterior bristle group slightly reduced, remaining bristle groups well developed. Glans of medium length, edges converging in ventral view.
Spiracle (Figs 9a-b). Dense curtain of robust reticulate spines extending partway across spiracle; terminations of spines multifurcate but not palmate; lace tubercles in lateral corner, with small number of reticulate spines at lateral end of posterior margin. FEMALE ( Fig. 8c; N = 1). Prosoma length 1.5, width 2.3; total body length 3.48. Anterior propeltidial area tan, remainder of propeltidium mottled medium brown. Ocularium with row of denticles on each side. Mesopeltidium medially medium brown, laterally tan with black mottling; small denticles medially. Metapeltidium and first three segments of opisthosoma medially medium brown with black patches on edge of medial area, laterally tan mottled with black. Metapeltidium and first four segments of opisthosoma with transverse rows of small denticles. Posterior part of opisthosoma tan mottled with black. Coxae patched tan and dark brown; venter of opisthosoma grey with longitudinal rows of dark brown patches.
Chelicerae. Segment I 0.77, segment II 1.65. Segment I tan with dark brown lateral patches proximodorsally; segment II golden brown with tan fingers. Unarmed.
Etymology. From the Latin atrox, cruel, and the goddess Diana. The transverse rows of mounds on the opisthosoma are reminiscent of the figure known as Diana of Ephesus, while the epithet 'cruel' refers to the addition of a spine on each of the mounds. Diagnosis. Megalopsalis caeruleomontium differs from other species of Megalopsalis in the presence of setae on the mobile finger of the chelicerae (Fig. 10b). Most males (except the smallest) can also be distinguished by the inflated second segment of the chelicerae (Fig. 10b). Megalopsalis caeruleomontium has a relatively flattened penis compared to other Megalopsalis species except M. nigricans; the glans is rather short, with the sides becoming subparallel distally (Figs 10d-e).
Description. MALE (N = 10). Prosoma length 1.17 (0.88-1.46), width 2.34 (1.98-2.50); total body length 3.25 (2.81-3.88). Propeltidium light orange-brown spotted with white and dark brown patches. Anterior propeltidial area pinkish-brown, with diverging dark-brown lines from ocularium to anterior margin, and dark-brown area around short supracheliceral groove on sharply downturned face. Prosoma unarmed. Ocularium bright white with light orange-brown base and behind eyes; postocularium bright white. Mesopeltidium, metapeltidium and first four segments of opisthosoma grey-brown with slightly lighter median band and distinctive transverse row of black setae in lighter spots across each segment. Posterior part of opisthosoma yellow-brown dusted with dark-brown; anal operculum silver. Coxae pinkish-brown with median white areas proximally; venter of opisthosoma grey-brown.
Chelicerae. Segment I 2.26 (0.57-3.26), segment II 3.36 (1.24-4.55). Segment I medium-brown dorsally and on proximal two-thirds laterally, peach-coloured ventrally and distolaterally, white patch at distolateralmost end with ventrolateral medium-brown patch directly underneath it; denticulate dorsally and on ventrolateral and ventromedial edges. Segment II strongly inflated in larger specimens to not inflated in smallest specimens, proximally mottled medium-brown and pink dorsally, mediumbrown laterally, peach ventrally; distally pink-cream, fingers yellow-cream; denticulate dorsally and ventrolaterally. Cheliceral fingers bowed apart proximally in larger specimens, less or not bowed in smaller specimens.
Penis (Fig. 10d-e). Tendon long; bristle groups well-developed. Glans in line with shaft; dorsoventrally flattened for entire length with bases of bristle groups (especially left) consequently more ventral than lateral; glans short, sides converging in ventral view. Deep pores.
Etymology. From the Latin words caeruleus, blue, and mons, mountain: "of the Blue Mountains", in reference to the distribution of this species.
Comments. Males of this species vary significantly between the largest and smallest individuals in the development of the chelicerae, from inflated with bowed fingers in the largest specimens to slender with unbowed fingers in the smallest. However, variation appears to be more or less continuous (albeit with large individuals distinctly more numerous than small individuals) without a clear division between morphs. Paratypes. 1 male, ditto (QM S40679); 1 male, ditto, Conondale Range, 900 m, 2 March-12 April 1992, D. J. Cook, rainforest pitfall (QM S74237).
Diagnosis. Megalopsalis coronata is distinguished from all other Megalopsalis species except M. tanisphyros, M. puerilis and M. sublucens by its small, unarmed chelicerae. It is distinguished from M. tanisphyros by the absence of a pedipalpal patellar apophysis, from M. puerilis by the presence of denticles on the ocularium, and from M. sublucens by the absence of ventral brush-like bristles on distitarsi III and IV. The glans of M. coronata is relatively long compared to other Megalopsalis species, with a lower angle of convergence between the sides in ventral view (Fig. 12c).
Penis (Figs 12c-d). Tendon relatively short; bristle groups well-developed. Glans of medium length; sides in ventral view subparallel, converging only slightly; dorsal side angled only slightly dorsad from shaft. Pores with surrounding rim, not raised.  Spiracle (Figs 13a-b). Sparse curtain of reticulate spines extending partway across spiracle; spines basally much broader than terminally; terminations of spines simple, pointed, some larger central spines multifurcate; dense patch of lace tubercles at lateral corner.
Variation. The paratype specimens differ in coloration from the holotype, but this may be due to preservation. QM S74237 has the prosoma golden-brown mottled with orange-brown patches, while both QM S74237 and the paratype QM S40679 have a brown transverse band, cream in the former and iridescent white in the latter, across the anterior part of the opisthosoma.
Etymology. From the Latin coronatus, crowned, referring to the denticulate ocularium. Comments. Leg I was only preserved in the holotype.
Megalopsalis nigricans (Hickman 1957 (1957), except minor male (Fig. 14a): As for major male, except chelicerae not enlarged relative to female. Armature of chelicerae reduced: segment I mostly unarmed with ventral spur, segment II with dorsal longitudinal row of denticles only, with long black seta on each denticle.
Comments. This species has already been described in detail by Hickman (1957). As noted above in the discussion of the phylogenetic analysis, this is the most divergent species here assigned to Megalopsalis. Its genital morphology is unique: the shaft of the penis is distinctly short and broad (Fig. 14b), in contrast to the elongate shaft of other Megalopsalis species (see figures in Taylor 2011). The glans is strongly flattened, not proximally inflated as in other species, and the overall shape is less distinctly subtriangular than in other Megalopsalis species, being rather more oblong over the greater part. The ozopores of M. nigricans are small and round, and not raised on lateral lobes like the larger ozopores of other Enantiobuninae.
Despite Hickman (1957) describing only the major male of this species, collections of this species in AMS and QM indicate that minor males are more abundant and majors relatively uncommon. Diagnosis. Megalopsalis puerilis is distinguished from all other Megalopsalis species except M. tanisphyros, M. coronata and M. sublucens by its small, unarmed chelicerae. It is distinguished from M. tanisphyros by the absence of a pedipalpal patellar apophysis, from M. coronata by the absence of denticles on the ocularium, and from M. sublucens by the absence of ventral brush-like bristles on distitarsi III and IV. The glans of M. puerilis is less triangular in overall shape than most other Megalopsalis species, with the sides distally subparallel in ventral view (Fig. 15c). (1.56-1.88); total body length 2.37 (2.35-2.38). Dorsum entirely unarmed. Anterior propeltidial area with central stripe of light orange-brown between ocularium and anterior margin of prosoma, flanked by two yellow stripes; remainder of anterior and median propeltidial areas mottled black and dark orange-brown with broad iridescent dark silver patches between ocularium and ozopores. Ocularium iridescent white with dark grey stripe down central groove. Mesopeltidium, metapeltidium and first four segments of opisthosoma orange-yellow medially and along segment boundaries with blackish brown patches laterally; fifth opisthosomal segment with transverse iridescent white stripe bordered by mottled black; remaining segments mottled black with yellow segment boundaries. Mouthparts brown-cream. Coxae dull orange proximally, mottled black distally; venter of opisthosoma iridescent white.
Pedipalps. on orange, remaining segments orange-yellow, with widely-spaced mottled black transverse stripes. Femora with scattered denticles, mostly dorsal; other segments unarmed.
Penis (Figs 15c-d). Bristle groups well-developed on both sides, left groups set slightly back and longer than right groups. Glans short, sides in ventral view subparallel, dorso-ventrally flattened distally, dorsal surface in plane with shaft. Deep pores.
Spiracle (Figs 13c-d). Curtain of robust reticulate spines extending only partway across spiracle; terminations of spines multifurcate but not palmate; lace tubercles on margin of lateral corner only.
Etymology. From the Latin puerilis, childish, referring to the lack of ornamentation or significant secondary sexual characteristics in the adult male. (Forster 1949 Description. MALE. Description as in Forster (1949), except following.

Megalopsalis stewarti
Legs. Leg I with anterior longitudinal row of enlarged denticles from proximal end of femur to distal end of tibia (males with smaller chelicerae with reduced denticle row on femur only), scattered denticles on basitarsus I; smaller row on femora and patellae of other legs.
Penis (Fig. 16). Shaft and tendon elongate; all four bristle-groups well-developed. Distinct lateral protrusion of glans above left anterior bristle group. Glans short, broad, triangular in dorsal view, dorsoventrally flattened at distal end. Pores shallowly recessed.
FEMALE. Description as in Forster (1949), except all legs with dorsal longitudinal row of small denticles along femora.
Comments. Forster (1949) stated that the holotype and female paratype had been deposited in NMV, while the remaining male paratypes (of unspecified number) had been deposited in the Canterbury Museum, Christchurch, New Zealand (CMNZ). However, as indicated above, NMV holds eight male specimens of this species labelled as paratypes. Because these specimens share the same collection data as the female paratype, it seems likely that these correspond to the specimens that Forster (1949) intended to place in CMNZ.

M. thryptica is very similar to M. stewarti, from which it can be distinguished by having distitarsus IV basally swollen. Hickman (1957) initially distinguished
Megalopsalis stewarti from M. thryptica by the presence of denticles on tibia I in males of M. thryptica, compared to their supposed absence in M. stewarti as described by Forster (1949). However, specimens of M. stewarti with longer chelicerae also have more extensive denticulation on leg I, extending as far as the basitarsus in some specimens. There is insufficient data as yet to determine whether this indicates a division between major and minor males or whether variation is continuous. Megalopsalis thryptica may still be distinguished from M. stewarti by the male of the former having distitarsus IV basally inflated (personal examination of holotype, AMS). Paratypes. 1 male, 1 female, as above (QM S2857). Diagnosis. Megalopsalis sublucens can be readily distinguished from other Megalopsalis species by the absence of a pedipalpal patellar apophysis or enlarged chelicerae, in combination with the presence of ventral brush-like bristles on distitarsi III and IV. It can also be distinguished from other species except M. stewarti by the presence of a lateral protrusion on the left side of the glans near the shaft-glans junction (Fig. 17c). (1.98-2.00); total body length 2.66 (2.59-2.72). Dorsum entirely unarmed. Dark tan stripes from ocularium to anterior margin, remainder of anterior iridescent white. Median propeltidial area white with dark brown patches. Posterior propeltidial area mottled dark brown. Lateral shelves dark brown anteriorly, iridescent white around and posterior to ozopores. Ocularium golden-brown with anterior face silver. Mesopeltidium, metapeltidium and first four segments of opisthosoma medially yellow-brown edged with dark brown with medial iridescent white spots; laterally iridescent white with dark purple mottling. Posterior of opisthosoma white with purple mottling. Venter of prosoma cream, coxae mottled black distally; opisthosoma purple-brown with transverse rows of iridescent white spots.

Megalopsalis sublucens
Chelicerae. Segment I 0.68, segment II 1.65 (1.57-1.73). Yellow-cream with tan mottling; not particularly enlarged compared to female. Segment I unarmed, segment II with proximodorsal denticles only. Fingers long, apotele closely opposed to finger of segment II.
Penis (Figs 17c-d). Left anterior bristle group reduced, but left posterior group well-developed. Glans short, triangular in dorsal view, dorsoventrally flattened, dorsal edge in plane with shaft.
Spiracle (Figs 18b-c Etymology. From the Latin sublucens, gleaming faintly, in reference to the iridescent patches covering part of the dorsum.
Comments. Those measurements for which a range is not given in the description of the male were only preserved in the holotype. While Megalopsalis sublucens possesses brush-like bristles on distitarsi III and IV as found in M. stewarti, M. tasmanica and species of the M. serritarsus-group, the number of bristles is reduced and they are proportionately more widely spaced and less regular. This may be related to M. sublucens' smaller size.

Phylogenetic analysis
A maximum parsimony analysis was conducted using the programme TNT (Goloboff et al. 2008); apomorphies were mapped onto the resulting trees using the programme Winclada (Nixon 1999). An initial analysis was conducted with all characters given equal weight, followed by successive implied weighting analyses with k varying in values of 1.0 from 1.0 to 6.0. Heuristic ("traditional") searches were conducted using a Wagner-tree random seed, 10 replicates holding 10 trees per replication and constructing trees using a tree bisection-reconnection (TBR) swapping algorithm. Changes from an unambiguous character state to a polymorphic character are counted by TNT as a step, even if the ancestral character state is included in the polymorphism. Jack-knife resampling analysis was conducted with a 36% removal probability over 1000 replications. The character matrix was based on that used in Taylor (2011); see therein for descriptions and discussion of characters not elaborated on below, and for specimen details of outgroup taxa (note that the name Pantopsalis luna [Forster, 1944] has been replaced by Pantopsalis listeri [White, 1849] as per Taylor 2013). As the monophyly of Eupnoi is well supported (Shultz 1998;Giribet et al. 2002Giribet et al. , 2009), representatives of Dyspnoi were excluded in order to reduce the effects of possible homoplasy. The recently described Mangatangi parvum Taylor, 2013 was also added to the matrix; see Taylor (2013) for specimen details for this taxon. Americovibone lanfrancoae Hunt & Cokendolpher, 1991 was coded based on the description given by Hunt and Cokendolpher (1991). Characters with more than two states have been treated as additive in the order given in initial analyses unless otherwise specified; a further set of analyses was conducted with all characters unordered. The characters used in the analysis are as follows: (1) present. These last two characters were treated as separate states of a single character in Taylor (2011;character 25 therein). This was probably inappropriate: there is no a priori reason to believe that a single species could not possess both pore morphologies, and they are here treated separately. Taylor (2011) also differentiated character states for shallowly recessed or shallowly raised genital pores; upon re-examination of the electron micrographs used in coding these states, I am not convinced that they are not artefactual (possibly related to the preservation of the specimen used). I have therefore not included these states in the current analysis. 22: Glans length: (0) short; (1) long. 23: Glans depth: (0) shallow in lateral view; (1) distinctly deep. In Pantopsalis albipalpis, the glans is surmounted by a relatively high, narrow dorsal keel, but the main body of the glans is still recognisable a narrow triangular shape in lateral view, contrasting with the more uniformly deep glans in Tercentenarium linnaei (Taylor 2008b) and Australiscutum species (Taylor 2009). Pantopsalis albipalpis is therefore coded as lacking this character. 24: Glans shape in ventral view: (0) subtriangular; (1) parabolic; (2) narrow or constricted. This character replaces character 19 in Taylor (2011), in which the alternate character states were poorly distinguished. A subtriangular glans characterises the species discussed in the current paper (as well as those attributed to Megalopsalis in Taylor 2011), in which the sides converge along the greater length of the relatively broad glans. Species of genera such as Pantopsalis (Taylor 2004(Taylor , 2013 and Australiscutum (Taylor 2009) have a parabolic glans shape, in which the glans is relatively longer and the rate of convergence of the sides less rapid. Species of outgroup taxa such as Ballarrinae (Hunt and Cokendolpher 1991), as well as Monoscutum titirangiense (Taylor 2008a (1) present. In species exhibiting this character, the glans is essentially subtriangular in overall shape, but the distal end of the glans has become extended, with the lateral margins becoming more subparallel distally after converging in the proximal part of the glans (Figs 10d, 15c). 29: Triangular dorsolateral keel on glans: (0) absent; (1) present. 30: Sharp dorsal papillae on glans: (0) absent; (1) present. 31: Longitudinal membrane on stylus: (0) absent; (1) present; (2) enlarged. Species of Neopilionidae have a longitudinal membranous banner running along the stylus, and often wrapped around it. This membrane has become particularly large in Australiscutum. Ballarra longipalpus is coded as lacking this character, though its presence in Americovibone lanfrancoae suggests its absence may be secondary for the former species; small membranous flanges at the base of the stylus in Ballarra species (Hunt and Cokendolpher 1991)  Thrasychirus-form spines or lace tubercles. This character has not been treated as additive. See Taylor (2011) for a discussion of the potential homology or otherwise between the spiracular spines present in Enantiobuninae and those present in Dyspnoi and Caddo. Ballarra longipalpus was coded by Taylor (2011) as exhibiting a fourth state of this character; however, as illustrated by Taylor (2011: fig. 15), the spiracular spines of B. longipalpus are derived from hypertrophy of the micro-ornamentation covering the broader venter, and are almost certainly not directly homologous with the spines present in Enantiobuninae, which are restricted to the spiracle and differ from the surrounding micro-ornamentation. The character state 'Ballarra-form spines' has therefore been removed from the current analysis, and B. longipalpus has been coded as lacking this character. Neopilio australis, previously coded as unknown for this character, has been re-coded as possessing Thrasychirus-form spines on the basis of the figure provided by Hunt and Cokendolpher (1991). (1) bows away from immobile finger proximally. 46: Setae on mobile finger of chelicera: (0) absent; (1) present. 47: Medial side of pedipalpal coxae: (0) unarmed; (1) with covering of blunt denticles. 48: Plumose setae on pedipalp: (0) absent; (1) present. 49: Length of pedipalpal femur: (0) shorter than or subequal to prosoma length; (1) more than 1.5 × as long as prosoma. 50: Pedipalpal patella vs tibia: (0) patella shorter than tibia; (1) patella longer than tibia. 51: Medial side of pedipalpal patella: (0) sparsely setose; (1) hypersetose. 52: Apophysis on pedipalpal patella: (0) absent; (1) poorly developed (less than one-third of patella length); (2) well-developed (about one-half of patella length). In Taylor (2011), the presence of a pedipalpal patellar apophysis was coded separately for males and females; this risked inappropriately weighting the significance of the patellar apophysis as the two characters are closely correlated. Instead, the presence of a pedipalpal patellar apophysis has here been coded as a single character, with a second character (character 53 below) coded only for species possessing the apophysis to account for the reduction of the apophysis in the male of Pantopsalis species relative to the females. 53: Pedipalpal patellar apophysis in male: (0) reduced relative to female apophysis; (1) fully developed as in female. 54: Shape of apophysis on pedipalpal patella: (0) rounded; (1) triangular. As with character 52, this character has been coded for both males and females rather than the two being coded separately as in Taylor (2011). The only species in which males and females are known to differ in apophysis shape, Forsteropsalis grimmetti, has been coded as polymorphic for this character. 55: Pedipalp with reflexed tibia relative to patella: (0) absent; (1) present. A dorsally reflexed pedipalpal tibia is characteristic of the Ballarrinae (Hunt and Cokendolpher 1991). 56: Shape of pedipalpal tibia: (0) straight; (1) bent mediad from patella. 57: Distribution of microtrichia on pedipalp: (0) absent; (1) distal half to third of tarsus only; (2) full length of tarsus; (3) tibia and tarsus. 58: Pedipalpal claw: (0) reduced or absent; (1) present. Neopilio australis has been coded as possessing a reduced pedipalpal claw, as per Hunt and Cokendolpher (1991). 59: Teeth on pedipalpal claw: (0) absent or only one or two teeth; (1) tooth-comb. 60: Armature of coxa I: (0) absent; (1) present. 61: Armature of trochanter I: (0) absent; (1) present in the form of prolateral denticles only; (2) present in the form of prolateral and retrolateral denticles; (3) present in the form of retrolateral denticles only. This character has not been treated as additive. 62: Leg I length and shape: (0) long and slender; (1) short and sturdy. 63: Leg I armature in male: (0) absent; (1) present on femur; (2) present on femur to patella; (3) present on femur to tibia; (4) present on femur to basitarsus; (5) present on femur to distitarsus.

Results and discussion
The analysis with all characters given equal weight, and multistate characters treated as ordered, produced two equally supported trees of 321 steps (CI = 0.290, RI = 0.626), the consensus of which is shown in Fig. 19. Bremer support for most clades was low, and very few can be regarded as numerically well supported. Nevertheless, monophyly was recovered for Neopilionidae (including Ballarrinae) and Enantiobuninae sensu Taylor (2011). Enantiobuninae was placed as sister to a clade of Neopilio australis and Ballarrinae. All implied weights analyses returned a single best tree, agreeing with the equal weight analysis on the monophyly of Neopilionidae and Enantiobuninae. Figure  20 shows the majority-rule consensus for all trees recovered under various weighting schemes with ordered characters; Table 1 shows support for selected clades under individual weighting schemes. A second round of analyses was also conducted with multistate characters treated as unordered. When all characters were given equal weight, this analysis produced two equally supported trees of 309 steps (CI = 0.301, RI = 0.633), the consensus of which is shown in Fig. 21. Though Neopilionidae was recovered as monophyletic in both trees, Enantiobuninae was only recovered in one tree, in which Neopilio australis and Ballarrinae were sister taxa as in the analyses with ordered characters. In the other best tree, Neopilio australis alone was sister to the remaining Neopilionidae, while Tercentenarium linnaei was placed outside a clade of Ballarrinae and the remaining Enantiobuninae. All implied weights analyses with unordered characters recovered a single best tree in which both Neopilionidae and Enantiobuninae were monophyletic. Figure  22 shows the majority-rule consensus for all analyses with unordered characters, with Figure 19. Consensus from equal weights analysis with multistate characters ordered. Numbers above branches indicate Bremer support values; numbers below branches indicate jack-knife support. Those branches without lower numbers received less than 50% jack-knife support.  Table 2. Figure 23 shows the strict consensus of clades recovered under all analytical conditions, with a total majority-rule consensus shown in Fig. 24. Figure 25 shows the distribution of apomorphies when mapped onto the consensus tree in Fig. 24; some of these apomorphies are discussed below.
Neopilionidae is potentially supported by the absence of setae on the glans itself (char. 19), and by the presence of a longitudinal banner on the stylus (char. 31: lost within Ballarrinae, unknown for Thrasychirus). Reduction of the number of seminal receptacles to two (char. 32) is also mapped as a potential synapomorphy, but this character is reversed in Australiscutum and Pantopsalis + Forsteropsalis. Within Neopilionidae, most analyses supported a basal split between Enantiobuninae on one side and Neopilio + Ballarrinae on the other. The relationship between the latter two taxa is supported by one unique synapomorphy, the reduction in the pedipalpal claw (char. 58; Hunt and Cokendolpher 1991). None of the potential synapomorphies of Enantiobuninae are unique to that clade; they include the presence of large oblong ozopores on raised lateral lobes (chars 2, 3; reversed in Megalopsalis nigricans) and the presence of lateral processes behind the shaft-glans junction on the penis (char. 15; possibly homologous with the barbed process found in Australian Ballarrinae). Other potential synapomorphies, the presence of a distinct space between the ocularium and mesopeltidium (char. 7) and the presence of a pedipalpal patellar apophysis (char. 52) are both commonly reversed within Enantiobuninae.
The current analysis is, admittedly, limited in its ability to test neopilionid monophyly by the low number of outgroup taxa included. Members of all other families of Phalangioidea possess setae on the glans (Cokendolpher et al. 2007), as do males of Caddo agilis (Gruber, 1974). Males of Caddidae have been recorded on only rare occasions; other than Caddo agilis, males of Acropsopilioninae have genitalia that are highly modified relative to other Eupnoi (Shear 1975), making homologisation difficult. The lateral processes on the penis of Enantiobuninae (bristles or bristle groups) differ from the penile setae of other Palpatores in lacking a basal socket (Taylor 2011). Hedin et al. (2012) recovered a relationship between Thrasychirus and 'Megalopsalis' (probably Figure 21. Consensus from equal weights analysis with multistate characters unordered. Numbers above branches indicate Bremer support values; numbers below branches indicate jack-knife support. Those branches without lower numbers received less than 50% jack-knife support. Forsteropsalis, as the original specimen was collected in New Zealand) under most analytical conditions in a molecular phylogeny focused on Sclerosomatidae, but did not include Neopilio or Ballarrinae.
Many of the relationships within the Enantiobuninae are sensitive to weighting regimes, but the results are more stable than those found by Taylor (2011). Major clades supported by all analyses include Australiscutum, a clade of Templar + Monoscutum (henceforth referred to as the 'Monoscutum clade'), Pantopsalis, Neopantopsalis, and a clade containing all species assigned herein to Megalopsalis except M. nigricans. Implied weights analyses consistently placed Thrasychirus and Australiscutum outside a clade of the remaining Enantiobuninae; Australiscutum was sister taxon to the remaining Australasian Enantiobuninae when multistate characters were ordered, but was placed as sister to Thrasychirus in some analyses with multistate characters unordered. The equal weights analyses, in contrast, did not support monophyly of the Australasian Enantiobuninae. When multistate characters were ordered, a clade of Tercentenarium + the Monoscutum clade was positioned basalmost within the Enantiobuninae. When multistate characters were unordered, the equal weights analysis placed Thrasychirus    and Australiscutum as a clade in a more nested position close to Pantopsalis, similar to the results found by Taylor (2011). However, as noted by Taylor (2011), Thrasychirus and Australiscutum differ from the remaining Enantiobuninae in their retention of a plesiomorphic mobile junction between the basitarsus and distitarsus (char. 73), and such a nested position would require that this character be reversed in these two taxa. Separation of Thrasychirus from the Australasian Enantiobuninae also aligns with the presence of bristle groups rather than isolated bristles at the shaft-glans junction in the latter (char. 14). Within Australiscutum, all analyses agree with Taylor (2011) in placing A. triplodaemon and A. graciliforceps closer to each other than A. hunti.
For the most part, the remaining relationships between genera in Enantiobuninae were not robust to analytical conditions. Analyses with multistate characters ordered supported a relationship between Tercentenarium and the Monoscutum clade, but this was not supported when multistate characters were unordered. Most analyses placed these taxa towards the base of the enantiobunines, which accords with their morphological distinctiveness. The New Zealand Templar incongruens and Monoscutum titirangiense (together with Acihasta salebrosa, not included in the current analysis but expected to be closely related to these two taxa) are short-legged, heavily sclerotised species previously classified as their own subfamily Monoscutinae (Forster 1948, Crawford 1992, Taylor 2008a. Tercentenarium linnaei possesses a number of features unique within the Neopilionidae, including a unilateral process on the left side of the penis at the shaft-glans junction, and a 'keyhole' emargination on the anterior margin of the female genital operculum (Taylor 2008b(Taylor , 2011.
Implied weights analyses supported a relationship between the New Zealand genera Pantopsalis and Forsteropsalis as found by Taylor (2011), though this clade was not recovered in equal weights analyses. The genus Forsteropsalis formed a paraphyletic series in the equal weights analyses, but was returned to monophyly in implied weights analyses at k values of 2.0 and above. Implied weights analyses at a k value of 1.0 supported a clade of the New Zealand genera Forsteropsalis, Pantopsalis and Mangatangi, but Pantopsalis and Mangatangi were both nested within Forsteropsalis. Forsteropsalis species share two distinctive characters within the Enantiobuninae, a small triangular pedipalpal patellar apophysis (except in the female only of F. grimmetti) and an array of blunt denticles on the inner margin of the pedipalpal coxa (Taylor 2011). Taylor (2013 suggested that the newly described Mangatangi parvum might be the sister taxon to Pantopsalis + Forsteropsalis, but its position was variable depending on analysis conditions. At higher k values with multistate characters ordered, it was placed as sister to the eastern Australian genus Neopantopsalis.
Neopantopsalis was supported as monophyletic in all analyses, though its internal topology was sensitive to analysis conditions. Varying analyses placed it closer to either Megalopsalis or Pantopsalis + Forsteropsalis.

Phylogeny of Megalopsalis
As found by Taylor (2011), the genus Spinicrus as hitherto recognised (Forster 1949, Hickman 1957, Kauri 1954 was not monophyletic, even with the segregation of Neopantopsalis (Taylor & Hunt, 2009). Instead, 'Spinicrus' species were placed as paraphyletic to Megalopsalis in sense of Taylor (2011). Members of Megalopsalis and Spinicrus are united by penis morphology, possessing a relatively short, broad, shallow glans that is more or less subtriangular in ventral view (char. 24). With the exception of 'Spinicrus' nigricans, the glans is also dorsally inflated near the shaft-glans junction, becoming distally narrower in lateral view, giving the glans a 'bellied' profile (char. 26;Figs 3e,4e). In members of other enantiobunine genera, the glans is relatively much longer and/or deeper (Taylor 2004, 2008a, 2008b, 2009, 2011, 2013, Taylor and Hunt 2009. Spinicrus and Megalopsalis were distinguished by the presence in the latter of a lateral apophysis on the patella of the pedipalp (Forster 1949). However, this feature is not unique to Megalopsalis, being also found in the Monoscutum clade, Forsteropsalis and females of Pantopsalis, and has probably appeared within the Enantiobuninae on multiple occasions. Even within the clade of Spinicrus and Megalopsalis, the presence of a pedipalpal apophysis may be homoplasious. While species with this character form a single clade in the equal weights analyses, the implied weights analyses all separate the M. leptekes-group from the M. serritarsus-group (see below for group definitions), placing the latter as sister to 'Spinicrus' stewarti. As the distinction between Megalopsalis and Spinicrus thus appears less reliable phylogenetically than the synapomorphies connecting them, the two genera are here synonymised. Hypomegalopsalis tanisphyros, described by Taylor (2011) in its own genus owing to its then phylogenetically uncertain position, shares these synapomorphies and is also placed within Megalopsalis. In this broadened sense, Megalopsalis admittedly encompasses a higher morphological diversity than most other genera in Enantiobuninae (with the possible exception of Forsteropsalis: Taylor 2011). Nevertheless, a broadened Megalopsalis represents a far more practical solution to the apparent paraphyly of 'Spinicrus' than division of the latter between a number of small genera that would have to be distinguished by relatively small-scale and difficult characters.
For the most part, this expanded Megalopsalis forms a distinct clade in all analyses. The only exception is M. nigricans, which is placed as the basalmost species of Megalopsalis in the implied weights analyses but does not form a clade with Megalopsalis in the equal weights analyses. The glans of Megalopsalis nigricans lacks the bellied profile of other Megalopsalis species, and is less regularly subtriangular than most other species. Megalopsalis nigricans also possesses a distinctly short and broad shaft to the penis compared to the relatively elongate and narrow shaft of other Megalopsalis species (Fig.  14b). Nevertheless, M. nigricans is more similar in genital morphology to Megalopsalis species than to other Enantiobuninae. Compared to other species of Enantiobuninae, M. nigricans is a particularly small taxon with reduced armature, and it is possible that its position is being distorted in the equal weights analyses by homoplasy with similarly small-bodied taxa such as Mangatangi parvum. Rather than creating a new monotypic genus for this species, M. nigricans is here provisionally assigned to Megalopsalis.
Within the clade of Megalopsalis species other than M. nigricans, a small number of clades are recovered in all analyses, treated here as the M. leptekes-, M. minima-and M. serritarsus-species groups. The M. leptekes-group includes the two Western Australian species M. leptekes and M. tanisphyros. Though a number of potential synapomorphies are identified for these two species in Fig. 25, none are unique to this clade; these include the presence of an elongate, hypersetose pedipalpal patellar apophysis, and an opisthosomal spiracle with relatively long, non-reticulate covering spines extending across the entire breadth of the spiracle rather than only partway as in other Megalopsalis species (Taylor 2011).
The species of the Megalopsalis minima-group are similarly Western Australian in distribution, comprising M. minima, M. porongorupensis, M. suffugiens and M. walpolensis. Members of this group are also united by spiracle morphology, with the covering spines being mostly lost (char. 37) though remnant lace tubercles are retained except in M. suffugiens (in which they have been reduced to a patch of reticulate ornamentation only). All analyses also agree in placing M. suffugiens, found in the Nullarbor in south-east Western Australia, as the sister taxon to the remaining species found in the south-west corner of Western Australia.
The Megalopsalis serritarsus-group corresponds to the 'core Megalopsalis clade' of Taylor (2011). Members of this group possess an elongate pedipalpal patellar apophysis, and males have specialised brush-like setae in a ventral double row on basally inflated distitarsi III and IV (Taylor 2011). With the exception of M. hoggi, males of the M. serritarsus-group also have modified second tarsi with ventrodistal swellings on the basal pseudosegments (Taylor 2011), and the majority of analyses agree with Taylor (2011) in placing M. hoggi as sister to the remaining species. The brush-like setae on distitarsi III and IV are shared with four further species: M. stewarti, M. sublucens, M. tasmanica and M. thryptica. For the most part, these species do not also have the distitarsi inflated as in the M. serritarsus-group, though M. thryptica has distitarsus IV only basally inflated (Hickman 1957; M. thryptica was omitted from the phylogenetic analysis owing to the unavailability of specimens). In the implied weights analyses, all species with brush-like setae are placed in a single clade; as this character is not found elsewhere in the Enantiobuninae, this result seems more credible than that of the equal weights analysis which unites the M. serritarsus-group with the M. leptekes-group on the basis of the probably homoplasious pedipalpal patellar apophysis.
Relationships between the remaining species of Megalopsalis are not consistently recovered by all analyses, and they are left unplaced in species groups. Clades recovered by implied weights analyses but not by the equal weights analyses include a sister relationship between the Tasmanian species M. tasmanica and M. sublucens, and a clade including M. atrocidiana, M. caeruleomontium and M. puerilis. The latter three species share a genital morphology with the sides of the glans closer to parallel in ventral view than in other Megalopsalis species (char. 28).

Biogeography and male variation
Though neither Australian nor New Zealand Enantiobuninae are identified as monophyletic, there is an overall separation between the fauna of the two land masses. New Zealand taxa may belong to as few as two separate clades (depending on the position of Mangatangi parvum). A relatively low level of interchange is also indicated between the western and eastern sides of the Australian continent, with the Western Australian species of Megalopsalis mostly assignable to the endemic clades of the M. minima-and M. leptekes-groups. Megalopsalis epizephyros is a member of the otherwise eastern Australian M. serritarsus-group and may represent a more recent immigration; it is notable in this regard that the only Enantiobuninae known to date from South Australia are representatives of the M. serritarsus-group (Forster 1949, Taylor 2011. In general, members of the M. leptekes-and M. serritarsus-groups tend to be found in more inland, and presumably drier, localities than other Megalopsalis species (Taylor 2011).
The feature of Australasian Enantiobuninae that has perhaps caused the most comment is the presence in males of most species of greatly enlarged chelicerae relative to the females. Though this has been cited as a diagnostic characteristic of the group (e.g. Dunlop and Mammitzsch 2010), many species also exhibit minor males with chelicerae that are less exaggerated than those of majors (Taylor and Hunt 2009). For a number of species of Megalopsalis (M. coronata, M. puerilis, M. sublucens, M. tanisphyros and M. walpolensis), the males appear to lack enlarged chelicerae (Taylor 2011 and below), though the possibility cannot be excluded that major males of these species remain to be discovered. The ratios of major to minor males in a population may vary between species: in Neopantopsalis species, for instance, the majority of specimens in collections are majors, while in M. minima and M. nigricans, majors are relatively rare and greatly outnumbered by minors (personal observations).
Variation in male cheliceral development has been identified for species of Megalopsalis, Neopantopsalis and Pantopsalis (Taylor 2004, 2013, Taylor and Hunt 2009; as noted by Taylor 2013, the 'effeminate' males described by Forster 1964 andTaylor 2004 for certain Pantopsalis species probably represent young specimens that have yet to complete cuticular hardening). Specimens of Forsteropsalis chiltoni vary in the degree of inflation of the second cheliceral segment, though they are not divisible between discrete morphs (Taylor 2011). In Pantopsalis, broad-and slender-chelicerate males differ in cheliceral length and inflation of the second segment, but males otherwise do not differ significantly in body size (Taylor 2004). In Neopantopsalis and Megalopsalis, major and minor males distinctly differ in body size as well as cheliceral size. It seems likely that discrete male dimorphism has developed independently in Pantopsalis versus Neopantopsalis + Megalopsalis, though it is possible that a more general variability as seen in Forsteropsalis chiltoni may be ancestral for the larger clade containing all three genera.
Within the Australian taxa, discrete major and minor morphs (i.e. without intermediate-sized individuals) are identifiable in Neopantopsalis species, Megalopsalis minima and M. nigricans. However, M. caeruleomontium, M. porongorupensis and M. stewarti exhibit more continual variation without clearly discrete morphs (variation is also present in M. suffugiens, but sample size is not large enough to identify whether it is discrete or continuous). Most other Megalopsalis species are known from too few specimens to discount the possibility of male variation, with the possible exception of M. epizephyros (Taylor 2011). It seems likely that at least some degree of male variation is ancestral for Megalopsalis, though we cannot say whether it was discrete or continuous.   Americovibone_lanfrancoae  0  0  0  0  0  0  1  0  0  0  1  0  0  ?  ?  ?  0  0  0  ?  ?  1  0  2  0  0  0  0  0  0  1  0  ?  ?  ?  ?  ?  ?  ?  ?  1  0  0  0  0  0  0