New species of Austropurcellia, cryptic short-range endemic mite harvestmen (Arachnida, Opiliones, Cyphophthalmi) from Australia’s Wet Tropics biodiversity hotspot

Abstract The genus Austropurcellia is a lineage of tiny leaf-litter arachnids that inhabit tropical rainforests throughout the eastern coast of Queensland, Australia. The majority of their diversity is found within the Wet Tropics rainforests of northeast Queensland, an area known for its exceptionally high levels of biodiversity and endemism. Studying the biogeographic history of limited-dispersal invertebrates in the Wet Tropics can provide insight into the role of climatic changes such as rainforest contraction in shaping rainforest biodiversity patterns. Here we describe six new species of mite harvestmen from the Wet Tropics rainforests, identified using morphological data, and discuss the biogeography of Austropurcellia with distributions of all known species. With this taxonomic contribution, the majority of the known diversity of the genus has been documented.


Introduction
The mite harvestmen (order Opiliones, suborder Cyphophthalmi) are a globally distributed suborder of tiny (1.5-5 mm), cryptic arachnids that are extremely dispersal-limited, making them ideal for fine-scale historical biogeographic studies. Nearly all species are known from pristine leaf-litter habitats in tropical, subtropical, and temperate forests, with a few others from caves (Juberthie 1971, Giribet et al. 2012 . The mite harvestmen that are endemic to the Wet Tropics World Heritage Area (WT) of Queensland, in northeast Australia, are members of the genus Austropurcellia Juberthie, 1988, with a range spanning the WT in the north to the Queensland-New South Wales border in the south. The highest diversity of species is found in the rainforests of the WT (Figs 1-4). Austropurcellia is a member of the family Pettalidae Shear, 1980, a lineage with a classical temperate Gondwanan distribution that includes species from Chile, South Africa, Madagascar, Sri Lanka, Western Australia, and New Zealand . Phylogenetic analyses of this group have demonstrated monophyly of all Queensland mite harvestmen , Giribet et al. 2012, Boyer et al. 2015. Therefore, all Queensland species were transferred to Austropurcellia by , including species originally described as members of the genera Neopurcellia Forster, 1948 andRakaia Hirst, 1925, whose type species occur in New Zealand.
Prior to 2012, only five Austropurcellia species were known (four from the WT and one from Central Queensland). Thus, little was known about the evolutionary history of the genus and its diversity in the region. Subsequently, Boyer and Reuter (2012) described four new species of mite harvestmen from the WT, and Popkin-Hall and Boyer (2014) described three new species from southeast Queensland. Boyer et al. (2015) presented six new species in a phylogenetic study of Austropurcellia, providing further insight into the historical biogeography of the genus. Together, these new species expanded Austropurcellia's known range to cover most of Queensland's coast. After intensive examination of museum collections and a series of collecting campaigns by the authors and collaborators, there are currently 19 described species within Austropurcellia, including 15 species from the WT alone.
Austropurcellia is an ancient lineage, and its evolutionary history has no doubt been shaped by the turbulent geological and climatic history of the Australian continent. Molecular dating suggests that Austropurcellia underwent initial diversification in the late Cretaceous (Giribet et al. 2012, Giribet et al. in press). Since then, the genus has persisted despite significant climatic changes in the region. Following the separation of the Australian continent from East Antarctica and the establishment of the Antarctic Circumpolar Current (ACC) in the Oligocene, global cooling occurred and latitudinal temperature gradients steepened (Crisp et al. 2004, Byrne et al. 2008. Australia drifted north into warmer latitudes, partially offsetting the cooling effects of the ACC, leading to a drier and more seasonal climate by the onset of the Miocene (~23 Ma). Rainforest habitats suitable for Austropurcellia were widespread throughout the Australian continent during the early Miocene, before they were largely replaced by sclerophyllous vegetation during a late Tertiary phase of long-term climate change and aridification (Adam 1992, Truswell 1993, Schneider et al. 1998, Crisp et al. 2004, Byrne et al. 2008. Miocene climatic changes have been invoked as a putative driver of speciation processes in other ancient Australian lineages. For example, phylogenetic and biogeographic analyses of Australian Archaeidae (assassin spiders), another limited-dispersal temperate Gondwanan arachnid group found in Queensland, point to evolutionary divergence as a result of Miocene aridification events (Rix and Harvey 2012). Austropurcellia provides a relevant point of comparison with this group, and work in preparation by the authors will examine tempo and age of speciation events within the genus.
The WT is considered to be a model system for studying biogeographic processes that shape rainforest diversity because it contains disproportionately large percentages of Australia's fauna (despite comprising only 0.12% of the continent by area), as well as unusually high rates of endemism (Nix 1991, Williams 2006, Rix and Harvey 2012. Palynological records from the WT suggest that significant range contractions and expansions of forest habitats have occurred as a result of more recent climate change. In particular, angiosperm rainforests were replaced by sclerophyllous or drier gymnosperm-dominated forests during Pleistocene glacial and interglacial cycles prior to the establishment of the current climate (Kershaw 1994, Graham et al. 2006, Bell et al. 2007. Rainforests have persisted in some areas of the WT much more consistently than others, leading to identification of potential species refugia by Webb and Tracey (1981), which take the form of small upland rainforest fragments scattered throughout areas of warmer and drier habitats (Schneider et al. 1998, Graham et al. 2006, Graham et al. 2010. Mite harvestmen only need small patches of suitable habitat to persist, and are thus able to survive these severe rainforest contraction events, making them an ideal group to study historical biogeography and speciation in the WT (e.g. Boyer et al. 2005, 2007b, Boyer and Giribet 2009, Giribet et al. 2012, Boyer et al. 2015. Boyer et al. (2016) modeled suitable climatic conditions for Austropurcellia and projected them onto paleoclimate data layers from time slices going back to the Last Glacial Maximum (LGM). They found that differences in LGM climatic suitability across the WT were a strong predictor of present-day diversity, outperforming current climatic suitability. This suggests that the LGM climatic refugia acted as museums of biodiversity, preserving lineages during a restrictive climatic regime and shaping the distribution of biodiversity across the WT that is seen today.
Mite harvestmen are known to have very low dispersal rates, with species even in well-surveyed areas generally found in only a few localities within a 50-km radius (Boyer and Giribet 2009, Boyer et al. 2015, Clouse et al. 2016 (Figs 1-4). Previous phylogenetic and biogeographic work has indicated that different closely related groups of species within Austropurcellia occupy distinct geographic areas of the WT. As examples, species north of the Black Mountain Corridor (BMC), an area that experienced loss of rainforest habitat during the Last Glacial Maximum, such as A. articosa and A. giribeti, form a distinct clade. Austropurcellia from the north-central WT and central WT uplands regions comprise another species group (Boyer et al. 2015) Figure 1. Distribution of all Austropurcellia species found throughout Queensland, Australia. Largest white box outlines the Wet Tropics World Heritage Area, shown in closer detail to the right. Smaller boxes within inset map represent groups of closely related species found within the Wet Tropics by Boyer et al. (2015), shown in larger detail in Figs 2, 3, & 4. Each circle denotes a locality and each colored circle denotes a different species found in one of the three relevant groups. Colored squares indicate the two species found in the southern Wet Tropics, which was excluded for the purposes of this study because it does not contain any of the new species presented.   (Figs 1-4). Unique morphological features also tend to be shared between closely related species. Therefore, combining morphological and geographic data can provide reciprocally corroborative insights into the evolutionary history of the genus.
Here we present six new species of mite harvestmen from the WT that are morphologically distinct from other members of Austropurcellia. We identify several diagnostic characters that vary between groups of species whose ranges are geographically proximate, and use this information to form hypotheses about the new species' phylogenetic relationships.

Methods
Specimens were hand-collected by the authors and colleagues in the WT of Queensland, Australia by sifting leaf litter during 2011-2015 and preserved in 95% ethanol. Additional specimens were provided by collections from Harvard's Museum of Comparative Zoology (MCZ), the Queensland Museum (QM), and the Australian National Insect Collection (ANIC). GPS data were recorded at each locality.
Collected specimens were examined for morphological differences using light microscopy and sorted into putative morphospecies. Due to their small size and highly conserved morphology, species-level differences are often only visible using an SEM. Therefore, males from different localities were examined on a scanning electron microscope (SEM). Only males possess characters that are diagnostic at the level of species.
Holotype specimens were photographed using an Olympus SZX10 light microscope driven by Leica Acquire software (Leica Microsystems) at multiple focal planes. Image series were integrated using Helicon Focus (Helicon Soft Limited). Specimens were placed in hand sanitizer for lateral images.
Paratype males chosen for SEM were dissected under the light microscope and mounted on stubs. One of each walking leg (I-III) as well as one palp and one chelicera were removed and mounted on a single stub. Both legs IV were mounted to provide a lateral and medial view of distinguishing features. One female leg IV was also mounted for comparison. Males were mounted ventrally on another stub to allow for close examination of the anal plate and scopula, and remounted for examination of dorsal ornamentation. Stubs were coated with gold-palladium alloy using a Denton Vacuum Desk III sputter coater and imaged using a JEOL JSM-6610LV SEM. Appendage measurements were made using the digital scalar tool included in the JEOL software package. New species were diagnosed based on several key character systems that are demonstrably informative in Austropurcellia taxonomy: male anal plate shape, scopula size and shape, tarsus IV segmentation and shape, and adenostyle shape (Fig. 5) Reuter 2012, Boyer et al. 2015).
SEM images for new species were edited to have a uniform black background using Adobe Photoshop CS6 Extended and compiled into plates using Adobe Illustrator CS6. ArcGIS 10.2.1 was used to create distribution maps for species.
Legs with all claws smooth, without ventral dentition or lateral pegs (Fig. 17). All tarsi smooth (Fig. 17). Distinct solea present on ventral surface of tarsus I (Fig. 17A). Metatarsi I and II heavily ornamented on proximal half, with smooth distal half (Fig. 17A, B). Remaining metatarsi with full ornamentation (Fig. 17C-F). Male tarsus IV completely divided into two tarsomeres (Fig. 17D, E). Adenostyle with relatively robust, blunt claw, wide base, and small pore at apex on lateral (external) side (Fig. 17D). Long seta on lateral surface of adenostyle from below pore to above apex (Fig. 17D, E); very short seta rising from adenostyle base below pore ( Fig. 17D)      Diagnosis. Distinguished from congeners by an usually wide and long scopula emerging from anterior quarter of male anal plate and easily visible in lateral view. Anal plate is very flat compared to the more rounded anal plates of geographically proximate species such as A. tholei and A. despectata. Distinctive areas lacking granulation cause ventral sutures to appear fused. Male tarsus IV is fully bisegmented rather than partially bisegmented as in A. tholei and A. despectata.
Chelicerae (Fig. 22A) short and relatively robust. Proximal article of chelicerae with dorsal crest, without ventral process. Median article with prominent apodeme. Chela with two types of dentition typical in pettalids (Fig. 22A) Legs with all claws smooth, without ventral dentition or lateral pegs (Fig. 23). All tarsi smooth (Fig. 23). Distinct solea present on ventral surface of tarsus I (Fig. 23A). Metatarsi I and II heavily ornamented on proximal half, with distal half smooth (Fig.  23A, B). Remaining metatarsi with full ornamentation (Fig. 23C-F). Male tarsus IV fully divided into two tarsomeres (Fig. 23D, E). Adenostyle with relatively robust claw, wide base, and small pore at apex on lateral (external) side (Fig. 23D). Long seta rising from medial (internal) face of adenostyle from below pore to above apex (Fig. 23D, E); very short seta rising from adenostyle base below pore on lateral (external) face ( Measurements from male paratype of leg articles from proximal to distal (in mm): leg I 0.14, 0. Diagnosis. Distinguished from congeners by an unusually wide scopula emerging from anterior margin or anterior quarter of male anal plate and covering entire width of anal plate. Closely resembles A. megatanka sp. n., due to full scopula covering most of anal plate, but distinguished from A. megatanka by differences in scopula shape and ubiquity of ornamentation on opisthosomal sternites.
Ozophores tall and conical, of type III sensu Juberthie (1970) (Figs 25A, 27B). Coxae of legs I and II mobile, coxae of remaining legs fixed. Male coxae II-IV meeting in the midline (Fig. 25B). Male gonostome small, subtriangular, wider than long (Fig.  25B). Spiracles circular and C-shaped with slightly recurved edges (Fig. 27A), as found in "open circle" type of Giribet and Boyer (2002). Anal region of "pettalid type" (Giribet and Boyer 2002). Anal plate convex and sparsely granulated near anterior margin, with granulation density increasing laterally (Fig. 26B). Very wide scopula emerging from anterior quarter of anal plate or from anterior margin and continuing past posterior margin of anal plate (Fig. 26B). Two anal pores visible, one at suture between anal plate and tergite IX and one between lobes of tergite VIII (Fig. 26B).
Chelicerae (Fig. 28A) short and relatively robust. Proximal article of chelicerae with dorsal crest, without ventral process. Median article with apodeme. Chela with two types of dentition typical in pettalids (Fig. 28A). Measurements from male paratype of cheliceral articles from proximal to distal (in mm): 0.61, 0.83. Palp (Fig. 28B)    Legs with all claws smooth, without ventral dentition or lateral pegs (Fig. 29). All tarsi smooth (Fig. 29). Distinct solea present on ventral surface of tarsus I (Fig. 29A). Metatarsi I and II heavily ornamented on proximal half, with distal half smooth (Fig.  29A, B). Remaining metatarsi with full ornamentation (Fig. 29C-F). Male tarsus IV fully divided into two tarsomeres (Fig. 29D, E). Adenostyle with relatively robust claw, wide base, and small pore at apex on lateral (external) side (Fig. 29D). Long seta rising from medial (internal) face of adenostyle from below pore to above apex (Fig. 29D, E); very short seta rising from adenostyle base below pore on lateral (external) face (    Etymology. The specific epithet is a tribute to the legendary Queensland field biologist Geoff Monteith for his invaluable knowledge of Wet Tropics entomology, which guided much of our fieldwork. The authors also wish to recognize his outsize generosity and hospitality to visiting researchers. In addition, he collected many of the specimens used in this study, including the holotype for A. monteithi sp. n. Diagnosis. Distinguished from congeners by lack of scopula on the male anal plate, a trait shared only with A. absens. Anal plate is flat and entirely ungranulated; A. absens anal plate is convex, bilobed, and mostly granulated.

Discussion
To develop Austropurcellia further as a system for studying WT biodiversity and biogeography, it is critical to describe and map the diversity of these cryptic, dispersallimited arachnids. Significant progress has been made toward this aim; with the six new Austropurcellia species described here in addition to other recent work (Boyer and Reuter 2012, Popkin-Hall and Boyer 2014, Boyer et al. 2015, the majority of mite harvestman diversity throughout Queensland is presently thought to be documented. These new species bring the total number of described species within Austropurcellia to 25, 21 of which are found within the WT biodiversity hotspot (Fig. 1).
With a greater knowledge of the diversity of Austropurcellia comes greater challenge in diagnosing the genus, as previously discussed by Popkin-Hall and Boyer (2014). In their 2007 phylogenetic analysis of the family Pettalidae, Boyer and Giribet re-diagnosed Austropurcellia to include the presence of a scopula in the anal plate. However, we now know of two species of Austropurcellia that lack a scopula in the anal plate, A. absens (Fig. 44E) and A. nuda sp. n. That said, the most recent phylogenetic study of the genus indicates that this loss is secondary (i.e. not the ancestral state). The re-diagnosis identified the robustness of the adenostyle, with height no more than twice base length, as an important feature uniting Austropurcellia. Although this is present in many species (e.g. A. daviesae, Fig. 45C), there are also species within the genus that have as thin, bladelike adenostyles (e.g. A. acuta, Fig. 45F), and in this case the diagnostic character is likely ancestral with respect to Austropurcellia. Other diagnostic characters of  are still valid, including prominent ventral process on trochanter of the palp, lack of robust ventral process on the chelicera, solea in tarsus I, and male tarsus IV bisegmented dorsally to fully bisegmented (Fig. 45). However, all of these characters are shared with another lineage of pettalids, the New Zealand genus Rakaia. Several phylogenetic analyses of the family Pettalidae have indicated that these two genera are reciprocally monophyletic , Boyer and Giribet 2009, Giribet et al. 2012, Boyer et al. 2015. However, those same analyses have remained equivocal about the relationship of Austropurcellia and Rakaia to each other. While there is some suggestion that they may be sister taxa (e.g. Boyer et al. 2015), support for that hypothesis is low and studies with extensive taxon sampling across the genus have suggested other possible relationships, also with low support (e.g. Giribet et al. 2012). Regardless of the relationship of Austropurcellia to other pettalids, it is clearly reciprocally monophyletic with all other genera. Therefore, we are left with a genus that is valid based on molecular characters and the phylogenetic criterion of monophyly, but currently lacks a robust diagnosis grounded in anatomy. Fortunately, Austropurcellia does not co-occur geographically with any other pettalid genus.
Differences in morphological features within the genus can provide insight into evolutionary relationships, especially in the context of geographic distributions of character states. For example, Popkin-Hall and Boyer (2014) described the geographic variation of two morphological features: adenostyle shape and posterior lobe shape in tergite VIII, and found that all species from the WT tend to have a robust, blunt adenostyle morphology (e.g. Fig. 45A, D, E, C), while those south of the WT possess a thin, blade-like adenostyle shape (e.g. Fig. 5F). The six new species described herein are all distributed within the WT, and accord with the pattern of robust, blunt adenostyles being geographically concentrated in the northern end of Austropurcellia's range. Previous work also demonstrates that male body shape varies with geography; species within the WT tend to have a rounded posterior, in contrast to the more triangular posterior and pointed lobes of tergite VIII found in species south of the WT. The six new species we describe support the pattern, as they all have rounded posterior lobes like those found in previously described WT species (Figs 7,13,19,25,31,37,42,43,44). However, a handful of species arguably defy this trend-A. giribeti and A. articosa (Fig. 42D, F), both found in the northern WT, share the more triangular posterior shape typically found in species south of the WT, while A. superbensis from Southeast Queensland has a more rounded posterior shape.
Further defining characters that vary significantly between species include the size, position, and shape of both the anal plate and the scopula (Figs 42-44). The anal plate is flat in some species and convex to a variable degree in others. Scopula morphology varies in terms of its position on the anal plate and its size; in some species it emerges from the anterior portion of the anal plate while in others it emerges near the posterior margin of the anal plate. Species found in the northernmost WT tend to have a long, narrow scopula that emerges from the posterior end of the anal plate and is oriented parallel to the body (Fig. 42). Although they both possess a somewhat shorter scopula, both A. finniganensis sp. n. and A. riedeli sp. n. (blue and green points, respectively, on Fig. 2) have the same scopula orientation and scopula placement within the anal plate as other northernmost WT species (Fig. 42A, C). A. fragosa sp. n. (white points on Fig.  2) has a unique scopula and anal plate shape that distinguishes it from all other Austropurcellia species found in the WT (Fig. 42B). Within species found further south in the rest of the WT, there is much greater variation in scopula morphology (Fig. 43,  44). However, the scopula emerges from the center or near the anterior margin of the anal plate in all of these species, distinguishing them from the northernmost WT species. One exception is a A. nuda sp. n., (blue points on Fig. 3) which lacks a scopula on its anal plate (Fig. 32B), a trait that it shares only with A. absens.
While this pattern of geographic distributions of shared characters states suggests closer relationships of geographically proximate species, an alternative interpretation is that unrelated species in certain regions of the WT have experienced morphological convergence. However, Boyer et al. (2015Boyer et al. ( , 2016 found a distinct correlation between phylogenetic position within Austropurcellia and geographic proximity of species' ranges, as inferred from a four-gene phylogeny of the genus; species that formed monophyletic or paraphyletic groups in the phylogeny were also recovered in close      proximity to one another on distribution maps. Distinct geographic and phylogenetic groups emerged from the northernmost WT, the north-central WT, the central WT uplands, and the southern WT (Boyer et al. 2015) (Figs 1-4). Based upon these trends reported by Boyer et al. (2015), we postulate that geographic distance serves as a generally reliable proxy for phylogenetic affinity in Austropurcellia. Using distribution maps of the new species described here, we predict below the putative clades to which the new species described herein would belong, and used the distribution of shared morphological character states as a corroborative litmus test for inferred relationships.
Two of the new species described here have already been included in a recent molecular phylogenetic analysis . One of them, A. megatanka sp. n., is found at the top of Mt. Baldy as well as several localities within the Lamb Range including Mt. Tiptree, Mt. Haig, and a CSIRO trail in Danbulla State Forest (Fig. 4). These sites in the center of the Atherton Tablelands place A. megatanka sp. n. in the northern end of the central WT uplands region (Fig. 4), suggesting that its closest relatives should include A. daviesae, A. cadens, and A. tholei. Apropos, A. megatanka sp. n. is similar to these species with regard to overall body shape and degree of tarsus IV bisegmentation, with its unusually long and wide scopula being its main distinguishing feature (Figs 19, 20, 44). Boyer et al. (2016) found that A. megatanka (identified as "Austropurcellia sp. n. Baldy Mt.) is indeed a member of a clade that also includes A. daviesae, A. cadens, and A. tholei.
The other species whose phylogenetic relationships have already been investigated is A. monteithi sp. n., which is known from five localities throughout the Lamb Range (Davies Creek, Chujeba Peak summit, Mt. Edith summit, Mt. Williams summit, and the Kahlpahlim Rock trail), geographically placing this species in the north-central WT (Fig. 3). Boyer et al. (2016) found that this species (identified as "Austropurcellia sp. n. Lamb Range") is related, as we would expect based on geography and morphology, as a member of a clade that also includes A. culminis, A. scoparia, and A. vicina. A. nuda sp. n. is found from only two localities (Black Mtn. and Black Mtn. summit), both located in the center of the distribution of this same north-central WT group (Fig. 3). We expect that it, too, is a close relative of A. culminis, A. scoparia, and A. vicina. Both of these new species share morphological features such as body shape and adenostyle morphology with the north-central WT group, with the main exception again being differences in the scopula and anal plate. A. monteithi sp. n. has a very long and wide scopula and A. nuda sp. n. lacks a scopula entirely, both in contrast to the relatively narrow, short scopula of A. vicina and A. culminis and the very unusual scopula of A. scoparia, which originates from the anterior margin of the anal plate (Figs 25,26,31,32,43).
A. finniganensis sp. n., A. fragosa sp. n., and A. riedeli sp. n. all have distributions in the northernmost WT (Fig. 2). A. finniganensis sp. n. is found in two localities on Mt. Finnigan, A. fragosa sp. n. is found in three very proximate localities by Roaring Meg Creek as well as a locality in the McDowall Range, and A. riedeli sp. n. was collected from a single locality along the Rossville-Bloomfield Road (Fig. 2). This suggests that these three species may be closely related to the other northernmost WT species such as A. articosa, A. giribeti, A. forsteri Juberthie, 2000 and A. sharmai, which have been found to form a paraphyletic grade at the base of the WT endemic clade within Austropurcellia (Boyer et al. 2015). This prediction is partially supported by morphology; A. finniganensis sp. n., A. riedeli sp. n. and A. fragosa sp. n. all share morphological features such as body shape with A. sharmai and A. forsteri, but they lack the more pointed, elongated body shape shared by A. arcticosa and A. giribeti. A. finniganensis sp. n., A. fragosa sp. n., and A. riedeli sp. n. also share another trait that central WT species lack-a very defined, ungranulated dorsal medial sulcus (Figs 7,8,13,14,37,38,42).
The WT is an important system for studying patterns and causes of rainforest endemism in both vertebrates and invertebrates. Yeates et al. (2002) investigated patterns and levels of endemism in 274 flightless insects at the scale of the 23 subregions within the WT (as defined by Winter 1984 andWilliams andPearson 1997). They found that 50% of species were endemic not only to the WT, but also to a single subregion within the WT, compared to only 15% subregional endemism within WT vertebrates (Yeates et al. 2002). Four subregions were found to contain the highest levels of subregional endemism in flightless insects: Finnigan Uplands, Carbine Uplands, Bellenden Ker/Bartle Frere Uplands, and Atherton Uplands (Yeates et al. 2002). When distributional data for the new species described here are combined with unpublished data collected in the lab of author SLB, we find that Austropurcellia conforms to the patterns found in other small flightless arthropods, with 50% of species endemic to single subregions. Only two of the 23 subregions contain more than one subregional endemic mite harvestman: Atherton Uplands and Finnigan Uplands. Of the six new species presented here: A. finniganensis sp. n. and A. riedeli sp. n. are both from the Finnigan Uplands, and indeed all of the species except for A. megatanka are subregional endemics. Williams and Pearson (1997) articulated the hypothesis that the distribution of diversity and endemism across WT subregions could be explained by differential extinction during Pleistocene glacial cycling, when rainforest persisted in some subregions but was extirpated from others. Work in the lab of SLB modeling historical distribution of climatic conditions suitable for Austropurcellia from the Last Glacial Maximum to the present confirms that this pattern holds for our system .
These predictions, and the postulated covariance of geographic distance, morphology, and phylogenetic relatedness in Cyphophthalmi more broadly, should be tested in the future using multilocus molecular phylogenies including the new species of Austropurcellia described herein. Such an approach would enable quantification of phylogenetic signal inherent to male morphological characters such as scopula and adenostyle shape, toward integrative taxonomy of mite harvestmen. Two putative new species from the WT remain undescribed due to incomplete collections and a lack of sufficient data, both in terms of morphology and DNA; however, we are confident that the majority of Queensland's mite harvestmen diversity has now been documented. As we continue to approach complete species-level sampling of Austropurcellia's extant diversity, we anticipate this genus will serve as robust model system to test hypotheses of how climatic and geologic changes in the WT have affected the distribution of genetic diversity, and how such processes leave their signature in the evolutionary history of Queensland's paleoendemic fauna.