Australian Assassins, Part II: A review of the new assassin spider genus Zephyrarchaea (Araneae, Archaeidae) from southern Australia

Abstract The Assassin Spiders of the family Archaeidae from southern Australia are revised, with a new genus (Zephyrarchaea gen. n.) and nine new species described from temperate, mesic habitats in southern Victoria, South Australia and south-western Western Australia: Zephyrarchaea austini sp. n., Zephyrarchaea barrettae sp. n., Zephyrarchaea grayi sp. n., Zephyrarchaea janineae sp. n., Zephyrarchaea marae sp. n., Zephyrarchaea marki sp. n., Zephyrarchaea melindae sp. n., Zephyrarchaea porchi sp. n. and Zephyrarchaea vichickmani sp. n. Specimens of the type species, Zephyrarchaea mainae (Platnick, 1991), comb. n., are redescribed from the Albany region of Western Australia, along with the holotype female of Zephyrarchaea robinsi (Harvey, 2002) comb. n. from the Stirling Range National Park. The previously described species Archaea hickmani Butler, 1929 from Victoria is here recognised as a nomen dubium. A key to species and multi-locus molecular phylogeny complement the species-level taxonomy, with maps, habitat photos, natural history information and conservation assessments provided for all species.


Introduction
Once considered among the most enigmatic and poorly known of spider families, recent research into the assassin spiders of the family Archaeidae ( Fig. 1) has revealed three diverse, highly endemic faunas from southern Africa, Madagascar and Australia, each of considerable evolutionary and conservation significance, and all the focus of modern revisionary systematic studies that have transformed our understanding of archaeid evolution and biogeography (see Platnick 1991a, 1991b, Lotz 1996, 2003, 2006, Harvey 2002a, Wood et al. 2007, Wood 2008, 2012. Although widely known from Mesozoic and Tertiary fossil deposits on multiple continents (see Dunlop et al. 2012), the Recent archaeid fauna consists of 54 described species in three genera (Platnick 2012), with numerous new species still to be described from African, Malagasy and Australian regions (H. Wood, pers. comm., M. Rix pers. obs.). These species, all remarkable for their araneophagic ecology and highly autapomorphic morphology, have highlighted the contrasting patterns of speciation and endemism that occur in Australian versus Old World taxa, and the importance of the unique archaeid carapace morphology in the evolution of the group (see Wood et al. 2007, Rix andHarvey 2012). The distinctive Australian fauna, while long neglected taxonomically and once presumed to be comparatively species-poor, has recently been shown to be far more widespread and species-rich than previously expected, thanks to dedicated field surveys and significant advances in our understanding of archaeid biology, ecology and biogeography (see Harvey 2011, 2012).
The history of the discovery and documentation of Archaeidae in Australia is one of significant recent, almost exponential progress, with over 85% of currently recognised species described in the past five years, and only four valid taxa described in the eight decades since the first species, Austrarchaea hickmani (Butler, 1929) was first recorded from Victoria. The second true archaeid to be recorded from Australia, A. nodosa (Forster, 1956), was described from the Lamington Plateau, southeastern Queensland, and nearly 30 years later Forster and Platnick (1984) described a third species, A. daviesae Forster & Platnick, 1984, from the Wet Tropics of north-eastern Queensland, further erecting the new genus Austrarchaea to include all of the Australian taxa. Archaeidae were unknown from Western Australia until A. mainae Platnick, 1991b (Figs 1E-F) was described from the Torndirrup Peninsula, near Albany (first reported by Dyer and Lyon 1983, Main 1987a, 1987b, and Main (1995) reported the discovery in the 1970s of six unidentified juvenile specimens from near Pemberton. The first archaeid specimen to be recorded from the Stirling Range National Park in southern Western Australia was collected in 1996 and this species, A. robinsi Harvey, 2002a, remained the only assassin spider to have been collected in Western Australia for a further 10 years. In 2006, living adult specimens of Zephyrarchaea janineae sp. n. (Fig. 1C) were discovered in wet forest near Pemberton, catalysing the first of many such discoveries in Western Aus-tralia, and highlighting the importance of temperate coastal heathlands as a habitat for assassin spiders. Indeed, since the pioneering revisionary work of Forster and Platnick (1984), when only eight adult archaeids had been recorded for the whole of Australia, over 500 specimens of Archaeidae have since been found throughout Queensland, New South Wales, Victoria, South Australia and Western Australia, revealing a diverse Australian fauna, characterised by numerous new species and mostly allopatric, relictual, short-range endemic taxa (Harvey 2002b, 2012. The current paper -the second in a series revising the Archaeidae of Australia -presents a taxonomic revision of the assassin spiders from temperate 'southern Australia', including those species from Victoria, South Australia and south-western Western Australia (Fig. 2). A new genus, Zephyrarchaea (Figs 1, 4), is described to include the type species, Z. mainae, along with Z. robinsi and nine new species. These taxa were found to form a monophyletic and highly divergent clade in a recent molecular phylogenetic analysis (Rix and Harvey 2012;Fig. 3), sister to all other species of Archaeidae from eastern Australia. This revision takes the total number of described Australian Archaeidae to 30 species, with the remaining, unrevised archaeids from north-eastern Queensland to be described in the third and final monograph of this series.

Material and methods
All taxa were described and illustrated from specimens stored in 75% or 95% ethanol. Digital images were taken using a Leica MZ16A binocular microscope and a Leica DM2500 compound microscope, with auto-montage images captured using Leica DFC500 mounted cameras with Leica Application Suite Version 3.6.0 software. Male left pedipalps were dissected prior to imaging and bulbs were aligned for standardised comparison in the retrolateral and prolateral positions illustrated; expanded pedipalps were illustrated in a retro-ventral position. Female genitalia were dissected and cleared in a 10% lactic acid plus 90% glycerol solution, prior to mounting on temporary glass slides. Illustrations were made on Utoplex tracing paper, using printed template auto-montage images. Maps were generated using Arc-Map version 9.3.1 (ESRI Inc.) with Virtual Earth (Microsoft Corp.).
Measurements are in millimetres (rounded to the nearest hundredth of a millimetre) and were taken using an ocular graticule on a Leica M80 binocular microscope. Left legs were removed from specimens prior to taking measurements and imaging lateral body profiles. Lateral profile images were standardised for inter-specific comparison by vertically aligning the centre of each left anterior median eye with the lower anterior margin of the carapace (above the labrum) (Fig. 8A). Carapace height was measured in lateral view, from the margin of the pars thoracica above coxa II to the highest point of the pars cephalica (Fig. 7). Carapace length was measured from the lower anterior margin of the carapace (above the labrum) to the posterior margin of the pars thoracica (above the pedicel) (Fig. 7). 'Neck' width was measured in lateral view, at the narrowest point of the carapace, with total length, carapace width, abdomen length and abdomen width all measured in dorsal view. To quantify inter-specific variation in the shape of the cephalothorax and 'head', three morphometric ratios (the carapace height to carapace length [CH/CL] ratio; the post-ocular ratio [P.O. ratio]; and the ratio of highest point of pars cephalica [HPC] to post-ocular length ratio) were derived from lateral profile images (Figs 7-9) as defined and discussed by Rix and Harvey (2011) (see also Fig. 8A).

Conventions
Specimens sequenced for the molecular analysis of Rix and Harvey (2012) are denoted by superscript codes, which correspond to specimen codes as shown in Rix and Harvey (2012, table 1, fig. 4). For species diagnoses, molecular autapomorphies for mitochondrial cytochrome oxidase genes (COI-COII; see Harvey et al. 2008, Cook et al. 2010 are coded according their nucleotide number (1-1609), as defined in Rix and Harvey (2011, table 3).
Abbreviations used in the text are as follows:
As noted by Wood (2008) and Rix and Harvey (2011), the homology of the tegular sclerites among archaeid genera remains unclear. Rix and Harvey (2011) used a numbering system for comparing the moveable tegular sclerites among species of Austrarchaea from mid-eastern Australia, identifying four separate sclerites (TS 1, 2, 2a and 3) according to their relative position within the unexpanded tegular cavity (see Fig. 4C). These four sclerites can be broadly homologised with the tegular sclerites of species of Zephyrarchaea, although at least one sclerite appears to be absent or otherwise highly modified in the latter, with no evidence for an interlocking, differentiated TS 2-2a complex (as in Austrarchaea; Fig. 4C cf. Fig. 4D). Tegular sclerite 1 (TS 1) is a prominent and strongly developed process in all species of Zephyrarchaea (Figs 4D, 10F, 16F), originating proventrally adjacent to the base of conductor sclerite 1. Two additional tegular sclerites (here labelled collectively TS 2-3) are closely contiguous and not easily distinguished in the unexpanded state, usually embedded prodistally adjacent to the retro-distal rim of the tegulum. The larger of these two sclerites, presumably homologous to tegular sclerite 3 (TS 3) in species of Austrarchaea, has a shorter and broader, more plate-like morphology relative to TS 1, and is usually (but not always) visible as a pointed projection beyond the retro-distal rim of the tegulum (Figs 10E, 16E).
Distribution. Species of Zephyrarchaea occur in mesic habitats throughout southern Western Australia, South Australia and Victoria (Fig. 2), usually in coastal (Figs 20C, 22C), sub-coastal or montane (Figs 23C, 24C) temperate heathlands, but also in wet eucalypt forests (Figs 21C, 29C) and temperate rainforests (Figs 26C,27C). In Victoria they occur along the Great Dividing Range, from Grampians National Park and the Otway Range in south-western Victoria east to the Yarra and Strzelecki Ranges east of Melbourne . In South Australia they occur on Kangaroo Island, at a single known locality north of Flinders Chase (Fig. 30). In south-western Western Australia they occur in the southern high rainfall and south-eastern coastal provinces (see Hopper and Gioia 2004;Figs 20-25), from the Wellington and Leeuwin-Naturaliste National Parks (near Bunbury) east to Cape Le Grand National Park, with outlying populations in the Porongurup and Stirling Range National Parks.
Remarks. The genus Zephyrarchaea forms a monophyletic and highly divergent clade sister to all other Archaeidae from mid-eastern and north-eastern Australia (see Harvey 2011, 2012;Fig. 3). Three main lineages have been recognised within the genus, for species from south-eastern Australia (South Australia and Victoria), from the Stirling Range National Park and from elsewhere in south-western Western Australia (see Rix and Harvey 2012;Fig. 3). The genus is not known to occur north or east of the Australian Alps, which may be a vicariant biogeographic barrier between populations of Zephyrarchaea and Austrarchaea.
Distribution and habitat. Zephyrarchaea mainae is known from the greater Albany region of southern Western Australia, from Walpole-Nornalup National Park (near Walpole) east to Bremer Bay and north to the Porongurup National Park, with a range centred on the Torndirrup Peninsula south of Albany (Fig. 20). Specimens have been collected by beating and sifting sedges (Lepidosperma sp.), curly grass (Empodisma gracillimum) and low shrubs in dense coastal or near-coastal groves of Peppermint (Agonis sp.), with several outlying populations also known from wet Karri (Eucalyptus diversicolor) forest.
Conservation status. This species is listed as threatened under the Western Australian Wildlife Conservation Act 1950. It is a short-range endemic taxon (Harvey 2002b), with known populations threatened by fire, dieback disease (affecting coastal heathland vegetation), land-clearing and climate change. Etymology. The specific epithet is a patronym in honour of Dr Janine Wojcieszek, for helping to discover the first live specimens of this species in 2006, and therefore catalysing the Western Australian Museum's 'archaeid project' in the half decade since 2007.
Distribution and habitat. Zephyrarchaea janineae is known from the high rainfall province (see Hopper and Gioia 2004) of southern Western Australia, from the Leeuwin-Naturaliste and Wellington National Parks (near Bunbury) east to Pemberton (Fig.  21). It is the dominant assassin spider of the south-western Karri (Eucalyptus diversicolor) forest and surrounding areas, and has been collected by beating and sifting elevated leaf litter in wet forested habitats and in coastal groves of Peppermint (Agonis sp.). Six juvenile specimens first collected by M. Gray in 1979 in the Treen Brook and Dombakup State Forests near Pemberton (see Main 1995) almost certainly belong to this species.
Conservation status. This species has a relatively widespread distribution in several National Parks and State Forests, and is not considered to be of conservation concern. Etymology. The specific epithet is a patronym in honour of Mark Wojcieszek, for helping to discover the first specimens of this species at Cape Le Grand National Park in 2009.

Zephyrarchaea marki
Diagnosis. Zephyrarchaea marki can be distinguished from Z. janineae and Z. mainae by the absence of dorsal hump-like tubercles on the abdomen (Fig. 12A); from Z. marae sp. n. and Z. vichickmani sp. n. by the presence of a proximal tuft of accessory setae on the male chelicerae (Fig. 12B); from Z. barrettae sp. n. and Z. melindae sp. n. by the shape of tegular sclerites 2-3, which project well beyond the retro-distal rim of the tegulum (Figs 12C-D); and from Z. porchi sp. n. by the larger, more protuberant proximal bulge on the male chelicerae (Fig. 12B).
Distribution and habitat. Zephyrarchaea marki is known only from Thistle Cove at Cape Le Grand National Park, on the far south-eastern coast of Western Australia (Fig. 22). Specimens have been collected by beating and sifting elevated leaf litter in a dense coastal thicket of Banksia speciosa.
Conservation status. This species appears to be a rare short-range endemic taxon (Harvey 2002b), with the single known population in the Cape Le Grand National Park potentially threatened by fire, dieback disease (affecting Banksia heathland vegetation) and climate change. (Harvey, 2002a)  Diagnosis. Females of Zephyrarchaea robinsi can be distinguished from Z. janineae and Z. mainae by the absence of dorsal hump-like tubercles on the abdomen (Fig. 13A); from Z. austini sp. n., Z. grayi sp. n., Z. marae sp. n. and Z. vichickmani sp. n. by the shallow post-ocular depression in lateral view (Fig. 9G); and from Z. barrettae sp. n. and Z. melindae sp. n. by the much shorter carapace (CH/CL ratio < 1.70) (Figs 7, 13A).

Male: Unknown.
Distribution and habitat. Zephyrarchaea robinsi is known only from Ellen Peak, Bluff Knoll and the south face of Pyungoorup Peak, on the eastern massif of the Stirling Range National Park of southern Western Australia (east of Chester Pass) (Fig.  23). Specimens have been collected by beating and sifting sedges (Lepidosperma sp.) in montane heathland habitats and along mesic, shaded creek lines.
Conservation status. This species is a short-range endemic taxon (Harvey 2002b), with a maximum total range of less than 10 km 2 , and all known populations in the eastern Stirling Range National Park potentially threatened by fire, dieback disease (affecting montane vegetation) and climate change. Etymology. The specific epithet is a patronym in honour of Dr Melinda Moir, in recognition of her contributions to biodiversity research, especially in the Stirling Range National Park of southern Western Australia.
Diagnosis. Zephyrarchaea melindae can be distinguished from Z. janineae and Z. mainae by the absence of dorsal hump-like tubercles on the abdomen (Figs 14A-B); from Z. marae sp. n., Z. marki, Z. porchi sp. n. and Z. vichickmani sp. n. by the shape of tegular sclerites 2-3, which do not project beyond the retro-distal rim of the tegulum (Figs 14D-E); and from Z. barrettae sp. n. by the shape of the anterior margin of the diastema adjacent to the 'neck', which is straight (in females) or only slightly concave in lateral view (in males) (Figs 14A-B cf. Figs 15A-B). Females further distinguished from other known congeners by the combination of a spherical abdomen (Fig. 14A), shallow post-ocular depression in lateral view (Fig. 9H), taller carapace (CH/CL ratio > 1.70) (Figs 7, 14A) and straight, almost vertical anterior margin of the diastema adjacent to the 'neck' (Fig. 14A).
Distribution and habitat. Zephyrarchaea melindae is known only from the summits of Toolbrunup Peak and nearby Mount Hassell, in the western Stirling Range National Park of southern Western Australia (west of Chester Pass) (Fig. 24). Specimens have been collected by beating and sifting sedges (Lepidosperma sp.) and low shrubs in montane heathland habitats.
Conservation status. This species is a short-range endemic taxon (Harvey 2002b), with a maximum total range of less than 10 km 2 , and all known populations in the western Stirling Range National Park potentially threatened by fire, dieback disease (affecting montane vegetation) and climate change. Etymology. The specific epithet is a patronym in honour of Sarah Barrett, for first discovering assassin spiders in the Stirling Range National Park in 1996.
Diagnosis. Zephyrarchaea barrettae can be distinguished from Z. janineae and Z. mainae by the absence of dorsal hump-like tubercles on the abdomen (Figs 15A-B); from Z. marae sp. n., Z. marki, Z. porchi sp. n. and Z. vichickmani sp. n. by the shape of tegular sclerites 2-3, which do not project beyond the retro-distal rim of the tegulum (Figs 15D-E); and from Z. melindae by the shape of the anterior margin of the diastema adjacent to the 'neck', which is slightly (in females) or strongly concave in lateral view (in males) (Figs 15A-B cf. Figs 14A-B). Females further distinguished from other known congeners by the combination of a spherical abdomen (Fig. 15A), shallow post-ocular depression in lateral view (Fig. 9I), taller carapace (CH/CL ratio > 1.70) (Figs 7, 15A) and slightly concave anterior margin of the diastema adjacent to the 'neck' (Fig. 15A).
Distribution and habitat. Zephyrarchaea barrettae is known only from the summit of Talyuberlup Peak, in the western Stirling Range National Park of southern Western Australia (west of Chester Pass) (Fig. 25). Specimens have been collected by beating and sifting sedges (Lepidosperma sp.) in montane heathland.
Conservation status. This species is a short-range endemic taxon (Harvey 2002b), with a maximum total range of less than 10 km 2 , and all known populations in the western Stirling Range National Park potentially threatened by fire, dieback disease (affecting montane vegetation) and climate change.

Zephyrarchaea sp. (unidentified juvenile specimens)
Note. In the absence of adult specimens or molecular data, the following juvenile specimens from Western Australia could not be confidently identified as a known species. Remarks. These specimens are the first Archaeidae to be collected from lowland habitats in the Stirling Range National Park, and the first members of the Western Australian High Rainfall Zone Clade (Fig. 3) to be discovered in the Stirling Range. All four juveniles possess paired dorsal tubercles on the abdomen, clearly aligning them with Z. mainae and Z. janineae. Zephyrarchaea mainae has a known distribution that extends north to the Porongurup National Park (see Fig. 20), and it is possible that the Talyuberlup Picnic Area may represent a northern extension of this range. Adult specimens or sequence data are required to confirm the identification of this population. Nomenclatural remarks. Archaea hickmani was first described by Butler (1929) from a juvenile specimen of unspecified providence, labelled and listed by Butler simply as "Victoria". Forster and Platnick (1984) examined this holotype, stating that it was in rather poor condition, and noting that a second adult female (labelled as a "homotype") accompanied the specimen, the latter apparently collected after the original description in 1929. The genitalia of this adult female specimen were illustrated in Forster and Platnick (1984, fig. 69), the specimen was briefly described, and the species was transferred to the genus Austrarchaea. Forster and Platnick (1984, figs 40-50) also presented scanning electron micrographs of a juvenile archaeid from near Sydney, erroneously regarded as being conspecific or very closely related to A. hickmani based on the absence of setose tubercles on the carapace. However, this specimen is clearly a juvenile of an unrelated species of Austrarchaea, as evidenced by the abdominal tubercles (see Forster and Platnick 1984, fig. 40) and New South Wales distribution. The Australian Museum collection also has an additional juvenile specimen of A. hickmani collected by C. Oke in 1936, similarly labelled as being from "Victoria".

The Victorian fauna
Based on the three known Victorian specimens identified by Butler as Archaea hickmani, the species is clearly congeneric and probably even conspecific with one of the four new Victorian species of Zephyrarchaea described in this paper. Unfortunately, given the unspecified collection locality of all three specimens, and thus the inability to unequivocally link the single adult female to the holotype or the type locality, this species must be regarded as a nomen dubium. Etymology. The specific epithet is a patronym in honour of the late Professor Victor Hickman, for his extraordinary contributions to arachnology and in honour of L. S. Butler's original (1929) patronym.

Diagnosis.
Zephyrarchaea vichickmani can be distinguished from other known congeners except Z. marae sp. n. by the absence of a proximal tuft or brush on the male chelicerae (Fig. 16C); and from Z. marae sp. n. by the less sinuous shape of conductor sclerite 2 (C2) (Figs 16D-E) and the more concave anterior margin of the male diastema adjacent to the 'neck' (Fig. 16B). Females further distinguished from other known congeners except Z. marae sp. n. by the combination of a spherical abdomen (Fig. 16A), strongly concave post-ocular depression in lateral view (Fig. 9A), and moderately elevated 'head' dorsally (post-ocular ratio ≥ 0.25) (Fig. 9A).

Distribution and habitat.
Zephyrarchaea vichickmani is known only from temperate Nothofagus rainforest habitats in the Victorian Central Highlands, north-east of Melbourne (Fig. 26).
Conservation status. This species has an imperfectly known distribution, and although potentially restricted, the abundance of protected forested habitats near the type locality would suggest that the species is unlikely to be of conservation concern. Etymology. The specific epithet is a patronym in honour of Dr Māra Blosfelds, in recognition of her love for small spiders and the Australian forests.

Zephyrarchaea marae
Diagnosis. Zephyrarchaea marae can be distinguished from other known congeners except Z. vichickmani by the absence of a proximal tuft or brush on the male chelicerae (Fig. 17C); and from Z. vichickmani by the more sinuous, slender, S-shaped conductor sclerite 2 (C2) (Figs 17D-E) and the less concave, slightly convex anterior margin of the male diastema adjacent to the 'neck' (Fig. 17B). Females further distinguished from other known congeners except Z. vichickmani by the combination of a spherical abdomen (Fig. 17A), strongly concave post-ocular depression in lateral view (Fig. 9B), and moderately elevated 'head' dorsally (post-ocular ratio ≥ 0.25) (Fig. 9B).
Distribution and habitat. Zephyrarchaea marae is known only from temperate rainforest and mesic closed forest habitats in the Dandenong and Strzelecki Ranges of West Gippsland, east and south-east of Melbourne, Victoria (Fig. 27).
Conservation status. This species has a relatively widespread distribution in several National Parks and State Forests, and is not considered to be of conservation concern.
Zephyrarchaea porchi Rix  Etymology. The specific epithet is a patronym in honour of Dr Nicholas Porch, for first discovering this species in the Otway Range.
Diagnosis. Zephyrarchaea porchi can be distinguished from Z. janineae and Z. mainae by the absence of dorsal hump-like tubercles on the abdomen (Fig. 18A); from Z. marae and Z. vichickmani by the presence of a proximal tuft of accessory setae on the male chelicerae (Fig. 18B); from Z. barrettae and Z. melindae by the shape of tegular sclerites 2-3, which project well beyond the retro-distal rim of the tegulum (Figs 18C-D); and from Z. marki by the smaller, less protuberant proximal bulge on the male chelicerae (Fig. 18B).
Female: Unknown. Distribution and habitat. Zephyrarchaea porchi is known only from north of Cape Otway, in the Otway Range of southern Victoria (Fig. 28). The single known specimen was collected in a dry vertebrate pitfall trap in eucalypt forest with a dense bracken fern understorey.
Conservation status. This species has an imperfectly known distribution, and although potentially restricted, the abundance of protected forested habitats near the type locality would suggest that the species is unlikely to be of conservation concern.
Zephyrarchaea grayi Rix & Harvey, sp. n. Etymology. The specific epithet is a patronym in honour of Professor Andy Austin, in recognition of his contributions to biodiversity research.
Male: Unknown. Distribution and habitat. Zephyrarchaea austini is known only from eucalypt woodland and associated heathland habitats near 'Billy Goat Falls', in the Western River Wilderness Protection Area of Kangaroo Island, South Australia (Fig. 30).