Taxonomic review of the genus Dasycercus (Dasyuromorphia: Dasyuridae) using modern and subfossil material; and the description of three new species

Abstract

The genus Dasycercus (Dasyuridae: Dasyuromorphia) has a long and ongoing history of taxonomic uncertainty. Currently two species are recognized: Dasycercus cristicauda and Dasycercus blythi, with the previously named Dasycercus hillieri considered a junior synonym of D. cristicauda. This investigation integrates modern, historical and subfossil material from across Australia to provide the most comprehensive review of the genus to date. Cranial and dental morphological characters that enable phylogenetic and morphometric analysis of taxa are established. Linear discriminant analysis (LDA) and principal component analysis (PCA) of craniodental measurements, including from the type specimens of named taxa (D. cristicauda, D. hillieri and D. blythi), are used to assess the validity of each taxon. The results confirm the validity of D. cristicauda and D. blythi as species, but also reconfirm the validity of D. hillieri. As a result, much of the modern ‘D. cristicauda’ material is reassigned to D. hillieri. Three new taxa are proposed: Dasycercus woolleyae sp. nov., Dasycercus archeri sp. nov. and Dasycercus marlowi sp. nov. These six Dasycercus species are distributed across Australia’s arid zone and beyond. Based on prior investigations and the results of this taxonomic review, it is likely that only D. hillieri and D. blythi are extant. The identification of four likely extinct taxa marks the first recorded instance of modern extinction within the family Dasyuridae.

Jake Newman-Martin [[email protected]], School of Earth and Planetary Sciences, Curtin University Kent St, Bentley, WA, 6102 Australia;

Kenny J. Travouillon [[email protected]], Western Australian Museum, Collections and Research, Locked Bag 49, Welshpool DC, WA, 6986 Australia;

Natalie Warburton [[email protected]], Harry Butler Institute, Murdoch University, 90 South Street, Murdoch Western Australia 6150 Australia;

Milo Barham [[email protected]], School of Earth and Planetary Sciences, Curtin University Kent St, Bentley, WA, 6102 Australia;

Alison J. Blyth [[email protected]], School of Earth and Planetary Sciences, Curtin University Kent St, Bentley, WA, 6102 Australia.

Keywords:

• Taxonomy
• Mulgara
• mammalogy
• palaeontology

MULGARAS (genus Dasycercus Peters, 1875) are small carnivorous marsupials (Dasyuridae: Dasyuromorphia) found throughout Australia’s arid and semi-arid bioregions (Woolley 2005, 2006, Woolley et al. 2013). Mulgaras prey on a large variety of small fauna, including small mammals, reptiles, and invertebrates (Fisher & Dickman 1993, Chen et al. 1998, Contos & Letnic 2020). Throughout their distribution, mulgaras aid in ecosystem management, including via bioturbation (soil turnover) during burrowing and foraging for insect prey, with individuals using multiple burrows over their home range (Thompson & Thompson 2007). Australia lacks many large carnivorous species, owing to the extinction of megafauna at the end of the Pleistocene (Johnson 2014, Hocknull et al., 2020). As such, the natural ecological roles of carnivores in many regions are filled by smaller taxa such as mulgaras (Pavey et al. 2018).

Since European arrival, mulgaras have experienced declines in their geographic distribution (Woolley 2005, Dickman & Glen 2014), most likely as a result of domestic cat (Felis catus; Linnaeus, 1758) predation (Dickman & Glen 2014). The decline of mulgaras has apparently been less dramatic than that observed in larger dasyurids, such as quolls (Dasyurus spp.; Geoffroy, 1796) (Masters & Dickman 2012). Currently there are two recognized Dasycercus species (Woolley 2005, 2006): Dasycercus cristicauda (Krefft, 1867) and Dasycercus blythi (Waite, 1904). Dasycercus blythi is known to occur over much of the arid zone, including sympatrically in some areas with D. cristicauda (Pavey et al. 2011, Woolley et al. 2013). Specimens collected on the Canning Stock Route indicate that D. cristicauda occurs in sand dune environments, while D. blythi occupies spinifex grasslands within the same geographic range (Woolley 2005, Woolley et al. 2013).

The species taxonomy of Dasycercus is debated (Woolley 2005). Dasycercus cristicauda was the first species to be described (Krefft 1867). The taxon was originally named Chaetocercus cristicauda (Krefft, 1867) before Peters (1875) recognized that this was a homonym of a genus of hummingbirds and renamed the genus Dasycercus. In the early twentieth century, the species D. blythi and Dasycercus hillieri (Thomas, 1905) were named, though both species were originally considered part of the genus Phascogale (Temminck, 1824). These two species were later synonymized within D. cristicauda by Mahoney & Ride (1988).

An unpublished genetic report by Adams et al. (2000) proposed that there are two genetically distinct extant taxa within the genus, but identified these as D. hillieri and D. cristicauda. The taxonomic decision was based on a combination of molecular work, using mitochondrial (mtDNA) genes cytb and CR, and morphological examination of specimens. Of particular note was the difference between the two taxa in the morphology of the hair on the plantar surface of the pes (Adams et al. 2000).

Woolley (2005) reported several new morphological differences between specimens collected from southwestern Queensland and the Northern Territory, including the number of nipples, number of premolars, and posterodorsal tail hair morphology. Additionally, behavioural differences associated with breeding were observed in captive specimens (Woolley 2005). Based on observations of specimens collected and those existing in museum collections, Woolley (2005) concluded that Adams et al. (2000) had incorrectly classified the species in Dasycercus. Woolley (2005) proposed that the genetic population of D. hillieri noted by Adams et al. (2000) was in fact the species D. cristicauda, while the genetic population that Adams et al. (2000) identified as D. cristicauda was D. blythi. This decision was based primarily on morphological comparisons of new specimens with existing museum material. Woolley (2005) also noted that the holotype specimen of D. cristicauda could not be genetically sequenced. The taxonomy was again revised by Woolley et al. (2013) who determined with a larger sample size that the number of premolars was not a diagnostic character. Woolley et al. (2013) also attempted to determine the identity of subfossil material; however, cave specimens were concluded to display too great a variation in dental measurements to ascribe a species with confidence.

Unidentified subfossil specimens of Dasycercus form a major gap in our understanding of this genus, resulting in many geographical areas being poorly represented in taxonomic assessments. Additionally, the subfossil record offers the best approach to surveying taxa present prior to, and immediately after, European colonization. Our knowledge of how and when these taxa declined will be limited until species can be assigned to this material with confidence. To assign taxa to subfossil material, and undertake a full taxonomic review of the genus, the identification of diagnostic dental and cranial characters, which are more readily preserved than soft-tissue features, is of critical importance. Some currently used characters for Dasycercus have been noted as polymorphic within the currently recognized species (Kealy & Beck 2017, Beck et al. 2022), and these could be indicative of unrecognized species-level diversity; thus, additional species-diagnostic characters are required.

Here, we build on the taxonomic work of Woolley (2005) and Woolley et al. (2013) to identify a consistent and useable suite of craniodental characteristics that distinguish species-level taxa within Dasycercus, and provide a comprehensive review of the genus, including subfossil material. Fossil and historical material is also compared with the named type specimens D. cristicauda, D. blythi and D. hillieri to address prior taxonomic disagreements. Specimens of the sister taxon Dasyuroides byrnei (Spencer & Horn, 1896), were also included as this taxon had previously been allied with Dasycercus.

Materials and methods

A total of 111 modern and subfossil specimens are reported in this study, of which 84 specimens were used in the morphometric and statistical analyses (Table 1); excluded material was in too poor a condition to take accurate measurements, or represented a juvenile individual (specimens with incomplete erupted molars). Specimens examined included subfossil skulls, modern skulls, skins, taxidermy mounts, and wet-preserved specimens.

Table 1. List of subfossil and modern Dasycercus specimens examined in this investigation, including study skin specimens examined and their plantar pes morphology.

Institution	Department	Specimen #	Taxa	Adams et al., 2000	Pes	Tail	Canine over-eruption	Type status
AMS	Mammalogy	M11342	D. cristicauda	D. cristicauda	–	Crest	Present	Holotype
WAM	Palaeontology	66.6.33	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	2021.12.27	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	72.1.1058	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	72.1.1061	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	67.10.74	D. cristicauda	–	–	–	Absent	–
WAM	Palaeontology	69.8.48/69.8.51	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	69.8.47/69.8.50	D. cristicauda	–	–	–	–	–
WAM	Palaeontology	67.10.290	D. cristicauda	–	–	–	–	–
AMS	Mammalogy	M2987	D. archeri	–	A + B	Crest	Present	Holotype
WAM	Palaeontology	68.2.285	D. archeri	–	–	–	–	–
WAM	Mammalogy	M65280	D. archeri	–	–	–	–	–
WAM	Mammalogy	M65281	D. archeri	–	–	–	–	Paratype
WAM	Mammalogy	M64827	D. archeri	–	–	–	–	Paratype
WAM	Mammalogy	M64833	D. archeri	–	–	–	–	Paratype
WAM	Mammalogy	M65340	D. archeri	–	–	–	–	Paratype
WAM	Palaeontology	72.1.1057	D. archeri	–	–	–	–	Paratype
WAM	Palaeontology	78.1.55	D. archeri	–	–	–	–	–
WAM	Palaeontology	78.1.55	D. archeri	–	–	–	–	–
WAM	Palaeontology	2022.10.3	D. archeri	–	–	–	–	–
WAM	Palaeontology	2022.10.1	D. archeri	–	–	–	–	–
WAM	Palaeontology	72.1.1060	D. archeri	–	–	–	–	–
WAM	Palaeontology	72.1.1062	D. archeri	–	–	–	–	–
WAM	Palaeontology	2022.10.2	D. archeri	–	–	–	–	–
WAM	Palaeontology	68.3.86	D. archeri	–	–	–	Absent	–
WAM	Palaeontology	67.4.136	D. archeri	–	–	–	Present	–
WAM	Palaeontology	67.5.29	D. archeri	–	–	–	–	–
SAMA	Mammalogy	M17499	D. archeri	–	–	–	–	Paratype
AMS	Mammalogy	M2988	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4356	D. archeri	–	A	Crest	–	Paratype
AMS	Mammalogy	M3025	D. archeri	–	–	–	–	Paratype
AMS	Mammalogy	M4355	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4862	D. archeri	–	B	Crest	–	Paratype
AMS	Mammalogy	M4864	D. archeri	–	B	Crest	–	Paratype
AMS	Palaeontology	F152782	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4923	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4924	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4925	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4926	D. archeri	–	–	–	–	–
AMS	Mammalogy	M4865	D. archeri	–	A	Crest	–	–
AMS	Mammalogy	M8466	D. archeri	–	A	Crest	–	–
AMS	Mammalogy	M4863	D. archeri	–	A + B	Crest	–	Paratype
NHMUK	Mammalogy	1905.3.28.1	D. hillieri	D. hillieri	A	Crest	–	Holotype
WAM	Mammalogy	M9671	D. hillieri	D. hillieri	–	–	Absent	–
WAM	Mammalogy	M9670	D. hillieri	–	–	–	Present	–
WAM	Mammalogy	M9673	D. hillieri	–	–	–	–	–
WAM	Mammalogy	M8207	D. hillieri	D. hillieri	A	Crest	–	–
WAM	Mammalogy	M13716	D. hillieri	–	–	–	–	–
WAM	Mammalogy	M1538	D. hillieri	–	B	Crest	Absent	–
SAMA	Mammalogy	M3072a	D. hillieri	–	–	–	–	–
SAMA	Mammalogy	M3072b	D. hillieri	D. hillieri	–	–	–	–
AMS	Mammalogy	M982	D. hillieri	–	–	–	–	–
NMV	Mammalogy	C480	D. hillieri	–	–	–	–	–
NMV	Mammalogy	C256	D. hillieri	–	–	–	–	–
WAM	Mammalogy	M41477	D. blythi	D. cristicauda	–	Brush	–	Syntype
WAM	Mammalogy	M41476	D. blythi	D. cristicauda	–	Brush	–	Syntype
WAM	Mammalogy	M1512	D. blythi	–	A	Brush	Present	–
WAM	Mammalogy	M44970	D. blythi	–	–	–	Absent	–
AMS	Mammalogy	M39636	D. blythi	–	–	–	–	–
AMS	Mammalogy	M9013	D. blythi	–	–	–	–	–
NTM	Mammalogy	U2233	D. blythi	–	–	–	–	–
WAM	Mammalogy	M51466	D. blythi	–	–	–	Present	–
WAM	Mammalogy	69.8.49	D. blythi	–	–	–	–	–
WAM	Mammalogy	M23366	D. blythi	–	–	–	–	–
WAM	Mammalogy	M7743	D. blythi	–	–	–	–	–
WAM	Mammalogy	M60778	D. blythi	–	–	–	–	–
WAM	Mammalogy	M24134	D. blythi	D. cristicauda	–	–	Present	–
WAM	Mammalogy	M348	D. blythi	–	B	Brush	–	–
WAM	Mammalogy	M349	D. blythi	–	B	Brush	Absent	–
WAM	Mammalogy	M1496	D. blythi	–	A	Brush	Absent	–
WAM	Mammalogy	M1497	D. blythi	–	A	Brush	Present	–
WAM	Mammalogy	M1498	D. blythi	–	A	Brush	Absent	–
WAM	Mammalogy	M1517	D. blythi	–	A	Brush	Absent	–
WAM	Mammalogy	M1521	D. blythi	–	A	Brush	Present	–
WAM	Mammalogy	M1534	D. blythi	–	A	Brush	Present	–
WAM	Mammalogy	M1530	D. blythi	–	A	Brush	Present	–
NMV	Mammalogy	C7923	D. blythi	–	–	–	–	–
AMS	Mammalogy	M8641	D. marlowi	–	A	Brush	Absent	Holotype
WAM	Mammalogy	M1513	D. woolleyae	–	A	Crest	Present	Holotype
WAM	Palaeontology	73.1.195	D. woolleyae	–	–	–	–	Paratype
WAM	Mammalogy	M41891	D. woolleyae	–	–	–	–	Paratype
WAM	Mammalogy	M1520	D. woolleyae	–	A	Crest	–	Paratype
WAM	Mammalogy	M1511	D. woolleyae	–	A	Crest	–	Paratype
WAM	Mammalogy	M1495	D. woolleyae	–	A	Brush	Present	–
WAM	Mammalogy	M1526	D. woolleyae	–	A	Crest	Present	–
WAM	Mammalogy	M1576	D. woolleyae	–	B	Brush	Absent	–
WAM	Mammalogy	M1525	D. woolleyae	–	B	Brush	Absent	–
WAM	Mammalogy	M1529	D. woolleyae	–	A	Brush	Present	–
WAM	Mammalogy	M1535	D. woolleyae	–	B	Brush	–	–
WAM	Mammalogy	M1522	D. woolleyae	–	A	Crest	Present	–
SAMA	Mammalogy	M3141	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3150	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3106	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3116	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3143	D. woolleyae	D. hillieri	–	–	–	–
SAMA	Mammalogy	M3144	D. woolleyae	–	–	–	–	Paratype
SAMA	Mammalogy	M3105	D. woolleyae	–	–	–	–	Paratype
SAMA	Mammalogy	M3110	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3107	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3140	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3111	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3706	D. woolleyae	–	–	–	–	–
SAMA	Mammalogy	M3154	D. woolleyae	D. hillieri	–	–	–	–
SAMA	Mammalogy	M6064	D. woolleyae	D. cristicauda	–	–	–	–
ANWC	Mammalogy	M00529	D. woolleyae	–	–	–	Absent	Paratype
ANWC	Mammalogy	M12783	D. woolleyae	–	–	–	–	–
AMS	Mammalogy	M8638	D. woolleyae	–	–	–	–	Paratype
AMS	Mammalogy	M1144	D. woolleyae	–	–	–	–	Paratype
AMS	Mammalogy	M8639	D. woolleyae	–	–	–	–	Paratype
NMV	Mammalogy	C25594	D. woolleyae	–	–	–	–	–

Pes morphology A is the “D. blythi” pes hair morphology outlined in Adams et al. (2000), and morphology B is the “D. hillieri” morphology.

Specimens were assessed from museums across Australia, ensuring that taxa and morphologies throughout the continent were sampled (Fig. 1). Specimens used were from: the Western Australian Museum (WAM), South Australian Museum (SAMA), Australian Museum (AMS), Museum and Art Gallery of the Northern Territory (MAGNT), Commonwealth Scientific and Industrial Research Organisation (ANWC), Museum of Victoria (NMV), Queensland Museum (QM), and the Natural History Museum of London (NHMUK). They included the type specimens of Dasycercus cristicauda (AMS M11342), Dasycercus hillieri (NHMUK 1905.3.28.1), and of the syntypes of Dasycercus blythi (WAM M41476 and M41477).

Figure 1. Map of the localities of modern and subfossil Dasycercus specimens used in this investigation. Specimens are labelled using their current museum identification.

All specimens measured were adults with fully erupted molar rows (M1–4), to ensure that there was minimal effect of ontogenetic variation on morphological measurements. All specimens examined exhibit little to no tooth wear, a common feature in small dasyurids due to their mostly insectivorous diet (KJT pers. obs.). Dasycercus specimens with dental wear were no larger in size than specimens with no wear, suggesting that the skull stopped growing once adulthood was reached. This is contra to what has been reported in some other dasyurids (Cockburn & Johnson 1988).

Morphometric analyses

Specimens were measured with callipers using traditional craniodental measurements (Supplementary Data 1) adapted from Umbrello (2018), with the addition of supraoccipital height (soh). Dental (Fig. 2) and cranial terminology follows Beck et al. (2022). The data were log-transformed to normalize and ensure all measurements had equal weight in the analysis.

Figure 2. Traditional cranial morphology measurements used in this investigation, adapted from Umbrello (2018). A full list of measurements can be found in Supplementary Data 2. Additionally, illustration of the upper and lower molars of Dasycercus, in occlusal view. Diagram indicates the morphology of the molars, and terminology used in this investigation.

Principal component analysis (PCA) and linear discriminant analysis (LDA) were used to visualize morphometric patterns within the data collected, and to group specimens based on the component measurement that most distinguished their morphology. LDA was used to identify specific morphological discrimination among groups, as an LDA analysis will fail to discriminate groups on intraspecific traits. Similar methods have previously been used in Travouillon & Phillips (2018); this analysis has also been referred to as a canonical variates analysis (CVA).

As is often the case when working with subfossil or fragmentary material, some measurements could not be taken on all specimens. This mainly affected specimens from the Nullarbor, and the two species-level taxa from this locality (Dasycercus cristicauda and Dasycercus archeri sp. nov.). To account for the missing values in the PCA analyses, iterative imputation was applied in the statistical programme PAST v. 4.03 (Hammer et al. 2001), where the missing values are first replaced by the column average, before an initial PCA run is used to compute regression values for the missing data.

PCA and LDA were initially conducted on data uncorrected for size. However, measurements were also transformed in PAST to correct for allometry, and a second set of PCA and LDA were then run on the size-corrected data. Allometry was removed following the methodology of Rowley et al. (2021), using the ‘allometric vs. standard’ transformation. In this transformation, the mean of a reference measurement is used to standardize the size of each taxon, and allometric coefficients are estimated based on the reference measurement. In this investigation, the zygomatic width (zw) was used as the reference measurement for cranial measurements, and the length of the M1 (M1L) was used for dental measurements. The allometric vs. standard transformation was performed on each taxon separately, to ensure that the mean used is correct in each case. The single specimen of Dasycercus marlowi sp. nov. was grouped with D. cristicauda or Dasycercus blythi, with D. marlowi being grouped with the latter when specimens were separated by sex. This was done because the mean of a group is required to perform an allometric vs. standard transformation, so a single specimen cannot be used; instead, the mean of the combined taxa is used. It should be noted that this means that the single specimen of D. marlowi sp. nov. is more likely to group with D. blythi or D. cristicauda in the results, and appear more closely related. However, the authors feel that not to include an allometry-corrected analysis that includes D. marlowi would lead to an inaccurate representation of Dasycercus. Similarly, in analyses of D. cristicauda where sex was factored, there is only one specimen with an identified sex (the holotype), and this was grouped with those identified as D. archeri, as again allometry vs. standard requires more than a single individual. Similarly to the situation for D. marlowi, this means that the results are more likely to group D. cristicauda and D. archeri as related taxa; however, we feel that the alternative of including unsexed D. cristicauda would introduce more issues. As such, we have decided to accept that the previously mentioned taxa will appear more closely related rather than introduce other variables that could have unpredictable results on the analysis. Additional statistical analyses were run on female specimens only (n = 24), and male specimens only (n = 31), ensuring that sexual dimorphism is separated from interspecific variation.

A priori groups were identified by comparing morphology to the type specimens. Any specimens that did not match the morphology of the type specimens were assigned to a new group (sexual dimorphism was taken into account by comparing with a male and a female from the same locality). In total, three new morphological groups were identified. These morphological groups were tested using the previously mentioned morphometric analyses.

One-way PERMANOVAs were undertaken in PAST to test for significant differences between groups. The first two PERMANOVAs were performed on the cranial measurements, with allometry removed using an allometric vs. standard in the second analysis. As for PCAs, data were log transformed to normalize and give all measurements equal weight in the analysis. P values below 0.05 were treated as statistically significant. The second two PERMANOVAs were conducted on the dental measurements, using the same methods. To distinguish variance from sexual dimorphism, the PERMANOVAs were also conducted on male and female cranial measurements separately, with and without allometry removed. Males and females were not separated for dental PERMANOVAs, as no sexual dimorphism in size or shape was observed on the premolars or molars during the a priori comparison. In total, eight PERMANOVAs were conducted.

Phylogenetic analyses

The phylogenetic matrix of Kealy & Beck (2017) was used to conduct a total evidence analysis. This analysis contains morphological characters as well as DNA sequence data for a wide range of taxa within Dasyuromorphia. The total evidence analysis was adapted for use in this investigation, with all three Dasycercus species described to date (Dasycercus cristicauda, Dasycercus blythi and Dasycercus hillieri) and the additional species-level taxa identified here (Dasycercus archeri, Dasycercus woolleyae sp. nov. and Dasycercus marlowi) evaluated and added (Supplementary Data 3). Existing molecular characters from Kealy & Beck (2017) were left within the matrix, as a combination of both molecular and morphological characters serves to give a more robust understanding of dasyurid phylogeny. However, as specimens were being reassigned to taxa based on the morphology of types, this also needed to be addressed. For the purposes of this investigation, the D. cristicauda DNA sequence data from Kealy & Beck (2017) were reassigned to D. hillieri. This was done because the extant population in central Australia which was used to sequence the species matches the morphology of D. hillieri. The Kealy & Beck (2017) morphological dataset for D. cristicauda was collected on specimens that are assigned to D. archeri. Note that some of the DNA sequence data used by Kealy & Beck (2017) for D. cristicauda could be a mix of D. cristicauda and D. blythi, as some of the sequences predate the taxonomic revision by Woolley (2005; as mentioned in Westerman et al. 2016). Additionally, the postcranial material examined by Kealy & Beck (2017) was reassigned after comparisons with the type material. As such, postcranial material from Kealy & Beck (2017) were assigned to D. archeri (defined herein).

The analysis was performed in MrBayes 3.2 (Huelsenbeck & Ronquist 2001, Ronquist et al. 2012). The Bayesian analysis followed Kealy & Beck (2017), comprising four runs of four chains each (three heated and one cold), and sampling trees every 5000 generations. The analysis was run on four independent runs of four MCMC chains (one cold, three heated, with temp = 0.1), for 15 million generations, sampling every 5000 generations. The MrBayes output was examined in Tracer v1.7.1 (Rambaut et al. 2018), which indicated that stationarity and convergence between runs was achieved within the first 7.5 million generations (= 50%), which was excluded as burnin; post-burnin trees were summarized using 50% majority rule consensus. Bayesian posterior probabilities (BPP) are given as support values.

Results

PCA & LDA results—cranial morphology (full dataset)

The PCA results for cranial measurements on the full dataset (n = 84), using non-allometrically corrected data (Fig. 3A; Supplementary Data 2), shows that PC 1 accounts for 45% of the variance, and separates taxa based on nps, jh, lpr and bol. PC 2 accounts for 15% of the variance and separated taxa based on biw, apw, lpr, bl and bw. However, when corrected for allometry (Fig. 3B), PC 1 accounts for 77% of the variance, and instead separates taxa based on onl, bcl, jl, mw, zw and bow. The influence of PC 2 is much reduced, accounting for only 3% of variance, and separated taxa based on fs, zw, oP3, onl, biw and pnw.

Figure 3. Cranial morphometric analysis of all individuals. A, PCA; B, allometry-corrected PCA; C, LDA; D, allometry-corrected LDA. Males are indicated by filled squares, females by hollow circles, and individuals of indeterminate sex by dashes.

Males and females separate out for most groups where sex identification is possible, but there is a large amount of overlap between most taxa (Fig. 3A). However, when the data are corrected for allometry (Fig. 3B), Dasycercus hillieri and Dasycercus cristicauda separate out more readily, with only Dasycercus archeri, Dasycercus blythi and Dasycercus woolleyae overlapping.

In the LDA for the cranial measurements on the full dataset (Fig. 3C), Axis 1 accounts for 58.36% of variance and separates taxa by onl, pw, ctl, ppw, jl, lmr and bl. Axis 2 accounts for 27.31% variance and separates taxa by bcl, mw, lmr, bw, onl, soh and oP3. When corrected for allometry (Fig. 3D), Axis 1 accounts for 63.47% variance and separates taxa by onl, ppl, rwc, cw, lmr and bl. Axis 2 accounts for 21.73% of variance, and separates taxa by pw, ail, apl, oP3, jl and bl. In the non-allometrically corrected data (Fig. 3C), there is a minor overlap present between D. archeri and D. woolleyae, with all other taxa being separated. When corrected for allometry (Fig. 3D), this overlap is removed and all taxa are separated.

PCA & LDA results—cranial morphology (females)

In the PCA of the female specimens (n = 24) using non-allometrically corrected data (Fig. 4A), PC 1 accounts for 34.23% of the variance, and separates taxa based on nps, anw, bol, apw and jh. PC 2 accounts for 16.65% of the variance, and separates taxa based on nps, jh, ppl, apl and uMR. When the data are corrected for allometry (Fig. 4B), PC 1 accounts for a much increased 98.26% of the variance, and separates taxa based on bsl, soh, bcl, onl and jl. PC 2’s contribution is correspondingly minor, accounting for only 1.01% of variance, and separates taxa based on onl, soh, bsl, bcl and nps. All taxa overlap, except for Dasycercus hillieri which only overlaps slightly with Dasycercus woolleyae (Fig. 4A). However, when allometry is corrected for (Fig. 4B), only Dasycercus archeri and D. woolleyae overlap, with Dasycercus blythi slightly overlapping D. woolleyae, D. hillieri is separated from the other taxa.

Figure 4. Cranial analysis with females and males separate. A, PCA with females only; B, allometry-corrected PCA with females only; C, PCA with males only; D, allometry-corrected PCA with males only. Not that due to presence of indeterminate material, fewer specimens were available for these analyses.

PCA & LDA results—cranial morphology (males)

The PCA of male specimens (n = 31) using non-allometrically corrected data (Fig. 4C) shows that PC1 accounts for 47.8% of the variance and separates taxa based on biw, nps, jh and ppl. PC 2 accounts for 16.14% of the variance and separates taxa based on biw, pnw, jh and bl. When corrected for allometry (Fig. 4D), the influence of PC1 again increases, accounting for 79.63% of the variance, and separating taxa based on bcl, onl, jl, zw and mw. PC2 accounts for only 3.74% of variance and separates taxa based on mw, pnw and biw. In the non-allometrically corrected data (Fig. 4C) there is an overlap present between Dasycercus blythi and Dasycercus woolleyae, as well as between D. woolleyae and Dasycercus archeri. Both Dasycercus marlowi and Dasycercus hillieri are separate from other taxa, and each other. When allometry is corrected for (Fig. 4D), D. blythi no longer overlaps with D. woolleyae; however, D. marlowi plots within the D. blythi group.

PCA & LDA results—dental morphology (full dataset)

For the measurements of dental morphology on the full dataset (n = 84), the PCA of non-allometrically corrected data (Fig. 5A) shows PC1 accounting for 33.53% of variance, and separating taxa based on M4L, P1L and P2L. PC 2 accounts for 19.91% of the variance, and separates taxa based on P2W, P1W and M1PW. When allometrically corrected data are used (Fig. 5B), PC 1 accounts for 29.65% of variance, and separates taxa based on M3PW, M2PW, M1PW and M4AW, whereas PC 2 (21.76% of variance) separates taxa based on M2L, M3L, M4L and P2L.

Figure 5. Dental morphometric analysis. A, PCA; B, allometry-corrected PCA; C, LDA; D, allometry-corrected LDA. Males are indicated by filled squares, females by hollow circles, and individuals of indeterminate sex by dashes.

For the LDA on the full dataset of non-allometrically corrected dental measurements (Fig. 5C), Axis 1 accounts for 57.3% of variance and separates taxa based on M3AW, M3PW, P2L and M4L. Axis 2 (accounting for 29.92% of variance) separates taxa based on M3L, P2W, M2PW and M3AW. When allometry is corrected for (Fig. 5D), axis 1 (58.4% of variance) instead separated taxa based on P2W, P2L and M4L, while Axis 2 (32.35% of variance) separated taxa based on P2W, M2PW, P2L and M3L. In both analyses, there is minimal overlap present, with each group being distinguishable from each other. This overlap decreases when allometry is removed.

PCA & LDA results—dental morphology (males)

The results of the PCA of dental measurements performed on identified male specimens (n = 31), and using non-allometrically corrected data (Fig. 6A) are as follows. PC 1 accounts for 30.43% of the variance and separates taxa based on M4L, P2L and P1L. PC 2 accounts for 20.67% of the variance, and separates taxa based on P2W, M4AW and P1W. When the data are corrected for allometry (Fig. 6B), PC 1 (accounting for 32.33% of the variance) instead separates taxa based on M3PW, M2PW, M1PW and M4AW. PC 2 (20.97% of variance) separates taxa based on M4AW, M3L, M4L and M2PW. Some overlap is present between the taxa (Fig. 6A); however, six distinct groups exist across component 1. When allometry is corrected for, the overlap is decreased (Fig. 6B).

Figure 6. Dental analysis with females and males separate. A, PCA with males only; B, allometry-corrected PCA with males only; C, PCA with females only; D, allometry-corrected PCA with females only. Not that due to presence of indeterminate material, fewer specimens were available for these analyses.

PCA & LDA results—dental morphology (females)

The PCA of dental measurements on identified female specimens (n = 24) using non-allometrically corrected data (Fig. 6C) shows PC 1 accounting for 32.64% of the variance and separating taxa based on M4L, P2L and M1L. PC 2 accounts for 20.66% of the variance and separates taxa based on P1L, P2L and M3L. When allometry is corrected for (Fig. 6D), PC 1 (accounting for 28.89% of the variance) instead separates taxa based on M3L, P2L, M2L and M4L, while PC 2 (25.15% of variance) separates taxa based on M2PW, M3PW and M1PW.

PERMANOVA results—cranial characteristics

The one-way PERMANOVAs for the non-allometry corrected and allometry corrected cranial measurements on the full specimen dataset, and for the corrected and non-corrected measurements on the male specimens, all demonstrated significant differences between the taxa, with all pairs found to be significantly different from one another. For the full dataset of non-allometry corrected cranial measurements, p = 0.0001, F = 8.703 (pairwise results in Table 2). For the allometry corrected cranial measurements, p = 0.0001, F = 958.3 (pairwise results in Table 3). In the male specimens (n = 31), p = 0.0001, F = 7.06 for the non-corrected data (pairwise results in Table 4). In the allometry corrected male data, p = 0.0001, F = 25.37 (pairwise results in Table 5).

Table 2. P-values of the pairwise one-way PERMANOVA performed on cranial measurements.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0001	0.0001	0.0001	0.0024
D. cristicauda	0.0001	–	0.0001	0.0001	0.0001
D. hillieri	0.0001	0.0001	–	0.0001	0.0001
D. blythi	0.0001	0.0001	0.0001	–	0.0147
D. woolleyae	0.0024	0.0001	0.0001	0.0147	–

Values <0.05 are considered significant.

Table 3. P-values of the pairwise one-way PERMANOVA performed on allometry corrected cranial measurements.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0001	0.0001	0.0001	0.0044
D. cristicauda	0.0001	–	0.0001	0.0001	0.0001
D. hillieri	0.0001	0.0001	–	0.0001	0.0001
D. blythi	0.0001	0.0001	0.0001	–	0.0047
D. woolleyae	0.0044	0.0001	0.0001	0.0047	–

Values <0.05 are considered significant.

Table 4. P-values of the pairwise one-way PERMANOVA performed on cranial measurements of male specimens.

Taxa	D. archeri	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0002	0.001	0.0028
D. hillieri	0.0002	–	0.0014	0.0001
D. blythi	0.001	0.0014	–	0.0073
D. woolleyae	0.0028	0.0001	0.0073	–

Values <0.05 are considered significant.

Table 5. P-values of the pairwise one-way PERMANOVA performed on allometry corrected cranial measurements of male specimens.

Taxa	D. archeri	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0002	0.0066	0.0079
D. hillieri	0.0002	–	0.0011	0.0002
D. blythi	0.0066	0.0011	–	0.0005
D. woolleyae	0.0079	0.0002	0.0005	–

Values <0.05 are considered significant.

For the female specimens (n = 24), the one-way PERMANOVAs on the cranial data showed a significant difference between the taxa, with p = 0.0012, F = 2.085 for the non-allometrically corrected measurements (pairwise results in Table 6), and p = 0.0001, F = 8.103 (pairwise results in Table 7) for the allometrically corrected data. In both analyses, pairs were found to be significantly different from one another, except for: Dasycercus cristicauda and Dasycercus archeri, D. cristicauda and Dasycercus hillieri, D. cristicauda and Dasycercus blythi, D. cristicauda and Dasycercus woolleyae, D. blythi and D. woolleyae, and D. woolleyae and Dasycercus archeri.

Table 6. P-values of the pairwise one-way PERMANOVA performed on cranial measurements of female specimens.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.5053	0.0062	0.0115	0.9675
D. cristicauda	0.5053	–	0.179	0.3291	0.4294
D. hillieri	0.0062	0.179	–	0.0168	0.0168
D. blythi	0.0115	0.3291		–	0.5129
D. woolleyae	0.9675	0.4294	0.0168	0.5129	–

Values <0.05 are considered significant.

Table 7. P-values of the pairwise one-way PERMANOVA performed on performed on allometry corrected cranial measurements of female specimens.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.3779	0.0012	0.004	0.1081
D. cristicauda	0.3779	–		0.3323	0.2904
D. hillieri	0.0012	0.1757	–	0.0091	0.0018
D. blythi	0.004	0.3323	0.0091	–	0.1459
D. woolleyae	0.1081	0.2904	0.0018	0.1459	–

Values <0.05 are considered significant.

PERMANOVA results—dental characteristics

For the dental measurements, the one-way PERMANOVA on the non-allometry corrected data demonstrated significant differences between taxa (p = 0.0001, F = 6.649, pairwise results in Table 8). All pairs were found to be significantly different from one another except Dasycercus cristicauda and Dasycercus blythi (p = 0.0501). For the allometry corrected dental data, the test demonstrated significant differences for the taxa (p = 0.0001, F = 6.824, pairwise results in Table 9). All pairs were found to be significantly different from one another.

Table 8. P-values of the pairwise one-way PERMANOVA performed on dental measurements.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0014	0.0001	0.0171	0.0001
D. cristicauda	0.0014	–	0.0001	0.0501	0.0001
D. hillieri	0.0001	0.0001	–	0.0001	0.004
D. blythi	0.0171	0.0501	0.0001	–	0.0006
D. woolleyae	0.0001	0.0001	0.004	0.0006	–

Values <0.05 are considered significant.

Table 9. P-values of the pairwise one-way PERMANOVA performed on allometry corrected dental measurements.

Taxa	D. archeri	D. cristicauda	D. hillieri	D. blythi	D. woolleyae
D. archeri	–	0.0072	0.0001	0.025	0.0001
D. cristicauda	0.0072	–	0.0001	0.0187	0.0001
D. hillieri	0.0001	0.0001	–	0.0001	0.0003
D. blythi	0.025	0.0187	0.0001	–	0.0007
D. woolleyae	0.0001	0.0001	0.0003	0.0007	–

Values <0.05 are considered significant.

Phylogenetic placement

Our analysis (Supplementary Data 4) grouped five of the six Dasycercus taxa (Fig. 7) forming a soft polytomy (BPP = 0.65), and separated Dasycercus hillieri (BPP = 0.87). The most closely related taxon to Dasycercus is Dasyuroides byrnei, which is strongly supported (BPP = 1). It should be noted that the fossil species Dasyuroides achilpatna (Archer, 1982), did not group with D. byrnei, but rather within Sacrophilus (Cuvier, 1837); this is strongly supported (BPP = 0.96).

Figure 7. Phylogenetic relationships of taxa within the genus Dasycercus, and the phylogenetic relationships of Dasycercus within Dasyuromorphia. Bayesian posterior probability support is represented at nodes, as follows: black = BPP ≥ 0.95; dark grey = BPP 0.75–0.94; light grey = BPP 0.5–0.74. Adapted from Kealy & Beck (2017).

Systematics

DASYUROMORPHIA (Gill, 1872)

DASYURIDAE (Goldfuss, 1820)

DASYCERCUS (Peters, 1875)

Dasycercus cristicauda (Krefft, 1867)

Diagnosis

Dasycercus cristicauda differs from Dasycercus hillieri in the following ways: markedly smaller (approximately 25% smaller mean occipitopremaxillary length); no masseteric process present; no additional foramen present on the palatine anteroventral to the orbitosphenoid; postorbital constriction lesser; postorbital process smaller; frontal crests connect to the nuchal crest; accessory palatal fenestrae absent; no sagittal crest present; the supraoccipital is concave; interparietal present; internal jugal canal posteriorly positioned; minute anterior and posterior cusp on C1; presence of P3 variable; P3, small and rounded where present; M1 anterior cingulum does not connect to the talon via the preprotocrista; stylar cusp B absent; M1 postparacrista does not meet the premetacrista to form a centrocrista; postprotocrista ends on the lingual flank of the metacone; paracristid of m1 is notched; cristid obliqua does not connect to the protoconid. Dasycercus cristicauda differs from Dasycercus blythi in the following ways: minute foramen present dorsal to the foramen rotundum; accessory palatal fenestrae always absent; interparietal present; minute anterior and posterior cusp present on C1; M2–3 protoconule present; no posterior cusp present on c1; crested tail of black dorsal hairs; no proximal fattening of the tail. Dasycercus cristicauda differs from Dasycercus woolleyae in the following ways: overall smaller (by approximately 11%, based on the mean onl measurement); lateral nasal expansions present; no slight lacrimal process present between the sphenopalatine foramen and the maxillary foramen; minute foramen present dorsal to the foramen rotundum (this foramen seems absent in all other dasyurids, and is unnamed); frontal crests only present in males; the inflation of the squamosal located lateral to the subsquamosal foramen is kidney shaped; interparietal present; postparacrista of M1 less posteriorly directed; anterior cingulum is not connected to the talon via the preprotocrista; postprotocrista ends at the lingual flank of the metacone; paracristid of the M1 is notched; posthypocristid is straight. Dasycercus cristicauda differs from Dasycercus archeri in the following ways: lateral nasal expansion present; accessory palatal fenestrae always absent; interparietal unfused to the occipital; inflation caused by the posterior epitympanic sinus is kidney-shaped; P3 located directly anterior to M1; anterior cingulum of the M1 does not connect to the talon via the preprotocrista; postprotocrista ends on the lingual flank of the metacone; i1 bilobed; diastema present between p2 and m1. Dasycercus cristicauda differs from Dasycercus marlowi in the following ways: Lateral nasal expansions present; minute foramen present dorsal to the foramen rotundum; postorbital crests present; frontal crests present; interparietal present; anterior and posterior cusp of C1 are taller than the crown of P1; anterior cingulum present on M1; no stylar cusp B present on M1; anterior cingulum present in M2–3; notch present on the paracristid of the m1; tail crested, and no proximal fattening present.

Holotype

AM M11342, mounted adult, suspected female (Parnaby et al. 2017), with associated skull and dentaries. Referred material is listed in Table 1.

Type locality

North Spencer Gulf around Port Augusta region, South Australia (Woolley 2005, Parnaby et al. 2017)

Description

This description is based on the type specimen (AM M.11342), as well as the subfossil skull specimen of indeterminate sex (WAM 67.10.74), and subfossil dentary specimen of indeterminate sex (WAM 2020.1.487). The craniodental morphology of Dasycercus cristicauda is depicted in Figs 8–10.

Skin

The original description on the specimen was conducted by Krefft (1867). No other skins exist in museum collections. Known external characters cannot be assessed on the holotype specimen, as per Woolley (2005). The specimen is known to be made of multiple individuals, potentially even multiple species (Parnaby et al. 2017). Due to the quality of the specimen and the fact that is a mount, the pes hair and nipples are unidentifiable (Woolley 2005). When re-examining the type in this investigation, it was confirmed that these characters could not be assessed.

Facial bones

In dorsal view, the nasals are narrow anteriorly, widening posteriorly before narrowing towards the frontals. There is also a lateral expansion present on the posterior of the nasal. The frontonasal suture is undulated. The posterior margin of the nasals terminates level with the anterior margin of the orbits. The premaxillary–nasal suture is longer than the maxillary–nasal suture. In anterior view, the nares are as tall as they are wide, and are ovate. Incisors are not preserved in either cranial specimen. In lateral view, four alveoli are present for four unirooted incisors, all contained within the premaxilla. Between I4 and the C1, the paracanine fossa is present, with the premaxillary–maxilla suture running through it. The C1 to M4 are located within the maxilla. There is no diastema present between C1 and M1. In some other specimens (e.g., WAM 66.6.33), the P3 is absent. The infraorbital foramen is located centrally above M1, with the maxillary foramen located below the orbit, in line with M2. The lacrimal extends slightly past the orbit, and is in line with the posterior margin of M1. It is bordered by the maxilla anteriorly, frontal dorsally, and palatine ventrally. The lacrimal has a single lacrimal foramen; it is large and located anterior to the jugal–maxillary suture. A lacrimal crest is present on the rim of the orbit and continues into the anterior of the frontal, flaring laterally as a postorbital process.

Neurocranium

In lateral view, the maxillary foramen is wide, being twice as wide as it is tall, and triangular. This foramen is bordered lateroventrally by the maxilla, medially by the palatine, and dorsally by the lacrimal. The sphenopalatine foramen sits posterior to the maxillary foramen, within the palatine. The sphenorbital fissure is large, connecting to the sphenorbital fissure on the opposite side of the skull. The sphenorbital fissure is bordered by the orbitosphenoid anteriorly, the squamosal dorsally as well as anteriorly, and by the presphenoid and basisphenoid ventrally. The foramen rotundum is minute, circular, positioned posteriorly to the sphenorbital fissure, and opens anteriorly; it is entirely within the alisphenoid. The alisphenoid and parietal contact each other and prevent the frontal and squamosal from contacting. The pterygoids are poorly preserved.

In dorsal view, the frontals narrow posteriorly to the postorbital process. Beginning at the postorbital processes, two frontal crests run posteromedially along the cranium, continuing through the parietals. This feature is present in WAM 67.10.74, which is a subfossil specimen of indeterminate sex, but absent in the holotype (AM M.11342), which is female. This may be a sexually dimorphic feature. No sagittal crest is present. There is a swelling on the external surface of the squamous part of the squamosal, posterior to the subsquamosal foramen, caused by the presence of a very large posterior epitympanic sinus of the squamosal (see Archer, 1976, p. 260; Beck et al., 2022, character 84). This swelling is kidney shaped. There is an interparietal present on the posterior of the cranium between the parietal, which is unfused to the supraoccipital.

Palate

In ventral view, the incisive foramina are long and narrow, extending from the posterior alveolus of I2 to the anterior of C1. The palate is mostly flat between the incisive foramina and the maxillopalatine fenestrae. The maxillopalatine fenestrae are ovoid in shape, being longer than they are wide. The maxillopalatine fenestrae begin in line with the protocone of M2, and end level with the posterior end of the metacrista of M3. The maxillopalatine fenestrae are separated by a very thin central septum, which is often broken or not preserved. The central septum is thickest anteriorly and posteriorly, where it is triangular in shape. The maxillopalatine fenestrae are bordered on their posteromedial margins by the palatines, and otherwise enclosed by the maxillae. Posterior to the maxillopalatine fenestrae are a distinct pair of large palatine foramina. Located behind M4 is the posterolateral palatal foramen, represented by an incomplete notch.

Cranial floor

In ventral view, the presphenoid is located within the nasopharyngeal passage and is visible only as a small triangular bone. It is obscured by the vomer, which overlaps the presphenoid anteriorly. The vomer and presphenoid are bordered laterally by the palatine. On the posterior half of the basisphenoid, a weak lateral crest is present; this crest is in line with the crest of the pterygoid and continues as far as the basioccipital. The transverse canal foramen is present laterally to the basisphenoid and is enclosed by the alisphenoid; the canal opens posteriorly. The alisphenoid is inflated, forming a large auditory bulla. The foramen ovale is present just anterior to the basisphenoid–basioccipital suture, between the auditory bullae. The carotid canal is present just anterior to the foramen ovale. The foramen for the greater petrosal nerve is present posterior to the foramen ovale, and rests medially against the auditory bulla. The paroccipital process of the exoccipital and the rostral and caudal tympanic processes of the petrosal are inflated and affixed to the auditory bulla. This character is also present in Dasyuroides. The inner chambers of the auditory bulla, paroccipital process of the exoccipital, and rostral and caudal tympanic processes of the petrosal are connected. The foramen for the inferior petrosal sinus is small and located on the basioccipital–petrosal suture. The jugular foramen is half the size of the foramen for the inferior petrosal sinus, and is positioned posterior to it. The postglenoid process is present and large. There are two hypoglossal foramina present on the anterolateral edge of the occipital condyle.

Posterior cranial wall

In posterior view, the posterior cranial wall is triangular with rounded edges. The supraoccipital is convex and rhomboid in shape. The supraoccipital forms a significant part of the foramen magnum, with a small contribution from the exoccipital. The foramen magnum is wide and oval in shape, being wider than it is tall. The mastoid is large and flat. The posterior part of the paroccipital process is inflated. When viewing inside the cranium, the internal jugal canal is located posterior to the basispheno–occipital suture on the occipital.

Mandible

In lateral view, there are alveoli for three single rooted incisors (i1–3), a canine (c1), two premolars (p1–2), and four molars (m1–4) on the mandible. The mandibular symphysis extends from between the alveoli of the incisors to level with the anterior of the m1. There is a minute diastema present between p2 and m1. The ventral edge of mandibular body is slightly curved, increasing in height posteriorly until the base of the ascending ramus. There are two mental foramina visible on the mandibular ramus: below the posterior root of p1, and below the posterior root of m1. The ascending ramus and the mandibular ramus form an angle of ca 130°. The coronoid process is broad at the apex, rounded, posteriorly directed, and has a strong crest that connects to the mandibular ramus. The coronoid process is otherwise flat lateromedially. The mandibular condyle is ovoid in shape, being wider than it is long. The angular process is long and wide, emerging level with the mandibular body, and is posteromedially directed. The mandibular foramen is small and circular in shape, opening posteriorly. The posterior shelf of the masseteric fossa is wide and connects to a weakly developed fossa, which is V-shaped.

Upper dentition

I1 is not preserved in any known specimen. Based on its alveolus, I1 is large, slightly anteriorly directed, and no diastema is present between the left and right I1. I2–4 have a posterior and an anterior cusp present. The crowns are broader than the roots, giving the crown a broad triangular shape. The teeth are unirooted. I3 is the largest tooth by crown size, followed by I2, and then I4. The incisors form a U shape within the premaxilla. C1 is large and caniniform. There is a minute cusp present on the posterior side, and an even smaller cusp on the anterior side of the tooth. In occlusal view, the crown recurves gently lingually.

P1 is tricuspid, having a tall central cusp and a small anterior and posterior cusp. The tooth is ovoid in shape in occlusal view. In lateral view, it is triangular. The tooth is double rooted, and rests against both C1 and P2, with no diastema present anteriorly or posteriorly. P2 is wider than P1, and the central cusp is also taller. In occlusal view, the tooth is angled lingually. In specimens where P3 is absent (e.g., WAM 72.1.1057), there is a short diastema between P2 and M1. P3, when present, is minute, rounded, single rooted, and unicuspid.

M1, in occlusal view, is triangular, triple rooted and longer than it is wide. The metacone is the tallest cusp on the crown, followed in order by stylar cusp D, stylar cusp E, paracone, stylar cusp A, and the protocone. There is a minute anterior cingulum connected to stylar cusp A on the anterior side of the molar. The anterior cingulum is not connected to the talon. The parastylar shelf is reduced when compared to the metastylar shelf, being about one third of the size. The paracone is anterobuccal to the metacone, and anterolingual to stylar cusp D. The preparacrista and postparacrista are parallel to one another and posterobuccally directed. They end just anterior to the anterior flank of stylar cusp D, with no evidence of stylar cusp B or C being present. A minute stylar cusp A is present at the anterior-most point of the tooth, with a posterior blade ascending the anterior flank of the parastylar shelf. Stylar cusp E is small and connected to the metacone by the postmetacrista, and to a stylar crest that ascends the posterior flank of stylar cusp D without connecting to this cusp. The talon is small, with only the protocone, preprotocrista and postprotocrista present. The preprotocrista departs from the protocone anterobuccally, and ends just anterior to the anterior flank of the paracone. The postprotocrista departs from the protocone posterobuccally, and then descends the posterior end of the talon, level with the metacone.

M2 differs from M1 in the following ways. The M2 is longer and wider than M1. All crests are longer, and all cusps are taller. The anterior cingulum is greatly reduced. Stylar cusp A is obscured by the metastyle of M1. Stylar cusp B is present connected to the preparacrista. A stylar crest departs from stylar cusp B and runs posteriorly, ending anterior to the anterior flank of stylar cusp D. The parastylar shelf is larger in size. The postparacrista is connected to the premetacrista as a centrocrista. On the talon, a minute protoconule is present at the anterobuccal end of the preprotocrista.

M3 is similar to M2 in size, but differs in the following ways. The preparacrista is longer, increasing the width of the parastylar shelf. Stylar cusp B is taller, and stylar cusp D is smaller but rounder. The stylar crest posterior to stylar cusp B is posterolingually directed and the stylar crest anterior to stylar cusp E is anterolingually directed. There are two kinks in the buccal side of the tooth on either side of stylar cusp D. The talon is reduced in size.

M4 is heavily reduced, without a metastylar shelf. The preparacrista is longer and straighter. There is no anterior cingulum present. There is no stylar cusp present other than stylar cusp A. The postparacrista ends at the most posterior end of the tooth. The talonid is heavily reduced, with a minute protocone present, and the preprotocrista and postprotocrista reduced in length. The protoconule is absent.

Lower dentition

The i1–4 teeth are not preserved in any specimen known. The c1 is similar to C1; however, it lacks anterior and posterior cuspids, and is also smaller in size. The p1–2 are similar to P1–2 in morphology, but more elongated. The central cuspid is bladed, with the crest running posteriorly to the posterior cuspid.

The m1 is longer than it is wide. The tallest cuspid on the crown is the protoconid, followed by the hypoconid, hypoconulid, metaconid and the paraconid. The paraconid is heavily reduced, with a short paracristid ascending the anterior flank of the protoconid without reaching it. The paracristid has a small notch near the paraconid, with a distinct posterolingual groove. The metacristid links the protoconid to the very small metaconid posterolingually. The crista obliqua descends the hypoconid anterolingually, and then ascends and connects to the protoconid. The posthypocristid is oblique and connects the hypoconid to the hypoconulid. There is no anterior cingulid present, but there is a small posterior cingulid present on the posterior buccal side of the tooth. There is no entoconid present.

The m2 is similar to m1 but it is longer and wider. It differs in morphology in the following ways. The protoconid is taller and more buccally positioned. The paraconid is enlarged. The paracristid connects the paraconid to the protoconid with a distinct carnassial notch at its midpoint. The metaconid is also enlarged, and connected via the metacristid to the protoconid. The crista obliqua does not ascend, nor connects to, the protoconid. The posthypocristid is longer and the hypoconulid taller. The entoconid is present anterolingual to the hypoconulid. There is a small anterior cingulid present on the tooth that begins at the base of the protoconid and terminates anterior to the paraconid, with a hypoconulid notch present. The posterior cingulid is larger than on m1.

The m3 is slightly larger overall than m2. The trigonid is wider with paracristid and metacristid more elongated but the protoconid, paraconid and metaconid are smaller in size. The distance between the paraconid and the metaconid is shorter. The anterior cingulid is larger.

The m4 differs from m3 in having a very reduced talonid, shortening the length and width of the tooth. The trigonid is narrower, with the paracristid and metacristid being shorter. The hypoconid and hypoconulid are extremely reduced. The crista obliqua is very short. The posthypocristid is very short and runs posterolingually from the hypoconid to the hypoconulid. The entoconid is absent. The hypoflexid basin begins at the posterobuccal side of the protoconid and terminates posterior to the hypoconulid.

Remarks

The species can retain its common name of ‘Crested-tailed Mulgara’, although it should be noted that this is not an adequate way to identify the taxa, as the crested tail is a shared trait amongst D. cristicauda, D. archeri, D. hillieri and D. woolleyae.

Dasycercus hillieri (Thomas 1905)

Diagnosis

Dasycercus hillieri differs from Dasycercus cristicauda as follows: overall larger (approximately 25% larger based on mean occipitopremaxillary length); masseteric process present; additional foramen present on the palatine, located anteroventrally to the orbitosphenoid; accessory palatal fenestrae present; postorbital constriction much greater; postorbital process more developed; frontal crests connect to the sagittal crest; sagittal crest present; interparietal absent; foramen magnum oval in shape; supraoccipital is convex; internal jugular canal anteriorly positioned; males lack additional cusps on C1, while females have a posterior cusp present; P3 is always present; P3 enlarged and tricuspid in males; anterior cingulum of M1 connects to the talon via the preprotocrista; stylar cusp B present; M1 is sexually dimorphic, and varies as follows: in males the crest departing from stylar cusp A connects to the parastylar shelf, and the preparacrista ascends the anterior flank of stylar cusp D and meets this cusp, whereas in females the preparacrista ends at the base of the anterior flank of stylar cusp D, and stylar cup A does not connect to the parastylar shelf; postparacrista of the M1 meets the premetacrista to form a complete centrocrista; postprotocrista ends on the posterior flank of the metacone; paracristid of m1 is not notched; crista obliqua does not connect to the protoconid. Dasycercus hillieri differs from Dasycercus blythi in the following ways: overall larger (approximately 13% larger, based on mean onl measurement); masseteric process present; foramen present dorsal to the foramen rotundum; accessory palatal fenestrae always present; sagittal crest present; internal jugular canal anteriorly positioned; posterior cusp present on I1; no minute anterior or posterior cusp present on C1 in males, and a single posterior cusp in females; P3 always present; P3 tricuspid in males; anterior cingulum of the M1 connects to the talon; postparacrista of M1 meets the premetacrista to form a centrocrista; postprotocrista ends at the posterior flank of the metacone; anterior cingulum and protoconule present on M2–3; c1 lacks additional cusps; crested tail with black dorsal hairs and no proximal fattening. Dasycercus hillieri differs from Dasycercus woolleyae in the following ways: overall larger (approximately 12% larger based on mean onl measurement); lateral nasal expansion present; foramen present dorsal to the foramen rotundum; accessory palatal fenestrae present; frontal crests connect to the sagittal crest in males; the inflation of the squamosal located lateral to the subsquamosal foramen (caused by the epitympanic sinus) is kidney-shaped; internal jugular canal anteriorly positioned; C1 has no additional cusps in males; P3 is always present; P3 tricuspid in males; postparacrista of M1 meets the preparacrista to form a centrocrista; M2–4 minute anterior cingulum present; paracristid of m1 not notched; posthypocristid is straight; m2 protoconid, metaconid and paraconid more rounded. Dasycercus hillieri differs from Dasycercus archeri in the following ways: overall larger (approximately 14% larger based on mean onl measurement); lateral nasal expansion present; accessory palatal fenestrae always present; sagittal crest present; internal jugular canal anteriorly placed; no additional cusps present on C1 of males; only posterior cusp present on C1 of females; P3 always present; P3 located directly posterior to P2; postparacrista of M1 meets the premetacrista to form a centrocrista; anterior cingulum present on M4; i3 bilobed; diastema present between p2 and m1; paracristid of the m1 unnotched. Dasycercus hillieri differs from Dasycercus marlowi in the following ways: overall markedly larger (approximately 32% larger, based on the mean onl measurement of D. hillieri); lateral nasal expansion present; foramen present dorsal to the foramen rotundum; accessory palatal fenestrae present; postorbital crests present; sagittal crest present; frontal crests present; internal jugular canal anteriorly positioned; no additional cusps present on C1 of males; only posterior cusp present on C1 of females; anterior cingulum present on M1; no stylar cusp B present; postparacrista meets the premetacrista to form a complete centrocrista; postprotocrista ends on the posterior flank of the metacone; anterior cingulum present on M2–3; no notch present on the paracristid of m1; crested tail with black hairs dorsally, and fattening absent.

Holotype

NHMUK 1905.3.28.1, skin, adult male, no skull is associated. Referred material is listed in Table 1.

Type locality

Killalpanima, East of Lake Eyre, South Australia (28°00′ S, 138°30′ E).

Description

No skull is associated with the holotype specimen of Dasycercus hillieri, and as such, a cranial and dental description has never been completed. For this reason, a brief description of D. hillieri is presented here. The specimens used are an adult male (WAM M9670) and adult female (WAM M9671) collected by P. A. Woolley from Sandringham homestead, Queensland. In order to avoid repetition, the following description focuses on morphological differences with Dasycercus cristicauda. The craniodental morphology of D. hillieri is depicted in Figs 8–10.

Skin

The skin and tail of Dasycercus hillieri is represented in Figs 11A, 12. The skin is not described here, as the skin has been described previously by Thomas (1905). The morphology of the plantar pes hair was observed in both states described by Adams et al. (2000). It is not known if this is a polymorphic character or if this was a result of preservation, as not enough intact skins of D. hillieri could be examined (with most being mounted). A full list of specimens and their plantar pes hair morphology is given in Table 1. The tail is proximally reddish-brown, with the distal two thirds being black. The tail has long hairs arranged in a linear crest dorsally; these hairs are orientated dorsoposteriorly, and are only located on the black portion of the tail. No suitable specimens could be found to count the number of nipples to compare with the results of Woolley (2005).

Figure 8. Upper dentition of each taxon of Dasycercus identified in this investigation, specimens are presented in occlusal view. A, D. cristicauda (WAM 67.10.74); B, D. woolleyae (WAM M1513, holotype); C, D. blythi (WAM M1512); D, D. archeri (AMS M2987, holotype); E, D. hillieri (WAM M9670); F, D. marlowi (AMS M8641, holotype). All specimens shown are male.

Figure 9. Lower molar rows of each taxon of Dasycercus identified in this investigation (specimens are shown in occlusal view). A, D. hillieri (WAM M9670); B, D. woolleyae (WAM M1513, holotype); C, D. blythi (WAM M1512); D, D. archeri (AMS M2987, holotype); E, D. cristicauda (WAM 67.10.74); F, D. marlowi (AMS M8641, holotype). All specimens shown are male.

Facial bones

In anterior view, the nares are larger and wider than those of Dasycercus cristicauda. There is a very small masseteric process present on the maxilla, in line with the posterior of the M3 and anterior of the M4. The paracanine fossa is larger and deeper into both the premaxilla and the maxilla than in D. cristicauda. There is a diastema present between right and left I1. The rostrum is also sexually dimorphic, with WAM M9670 (male) having an overall longer and wider rostrum than WAM M9671 (female). This is also observed in all Dasycercus hillieri specimens examined in this study.

Neurocranium

In lateral view, the maxillary foramen is markedly wider and rectangular in shape, with the lacrimal contributing further posteriorly to the process lining the dorsal wall of the foramen. The foramen rotundum is larger. The minute foramen present dorsal to the foramen rotundum is larger than in Dasycercus cristicauda and more clearly visible. There is an additional foramen present on the palatine, located anteroventrally to the orbitosphenoid.

In dorsal view, the postorbital constriction is much greater than in D. cristicauda, with the postorbital process more developed and robust; this is more strongly developed in M9670 (and other male specimens) than in M9671 (and other female specimens). As in D. cristicauda, there is a crest on the cranium that begins at the frontal crest and that runs posteromedially from the postorbital process continues through the frontal and parietal, before connecting with the sagittal crest. In males, the sagittal crest is more pronounced and defined, while in females it is reduced. The nuchal crest in males flares posteriorly. The inflation of the squama of the squamosal due to the squamosal epitympanic sinus is more prominent and encompasses more of the subsquamosal foramen than in D. cristicauda. There is no interparietal present on the neurocranium, unlike in D. cristicauda.

Palate

In ventral view, the paired incisive foramina begin at the level of I4 and terminate at the midpoint of C1. The maxillopalatine fenestrae extend from level with the protocone of M2 to level with the protocone of M4, and are separated by a thick central septum that is hourglass in shape. The palatine foramina are smaller than those in Dasycercus cristicauda. An additional pair of minute accessory palatal fenestrae are present in Dasycercus hillieri.

Cranial floor

In ventral view, the auditory bulla is larger than that in specimens of Dasycercus cristicauda. The petrosal is smaller, and the paraoccipital process contributes more to the posterior inflation of the bulla in Dasycercushillieri than in D. cristicauda. The pterygoid is preserved in D. hillieri (not preserved in D. cristicauda): it is long, and hook shaped, facing posteriorly and ending anterior to the transverse sinus canal. In lateral view, the postglenoid process is much larger and more elongated than in D. cristicauda.

Posterior cranial wall

In posterior view, the foramen magnum is rectangular in shape and wider than it is tall. The supraoccipital is concave, compared to Dasycercus cristicauda in which it is convex. The nuchal crest is more robust than those present in D. cristicauda and merges dorsally with the sagittal crest. In lateral view, the occipital condyle is larger and posteriorly directed. Inside the cranium, the internal jugular canal is anteriorly positioned compared to D. cristicauda, and is positioned just posterior to the basioccipital suture.

Mandible

In lateral view, the mental foramina are located below the posterior root of p1 and anterior root of m1. There is a diastema present between p2 and m1. In dorsal view, the mandibular condyle is large and more rounded. The posterior shelf of the masseteric fossa is large and angular.

Upper dentition

I1 is preserved in Dasycercus hillieri (not preserved in Dasycercus cristicauda). The tooth is slightly procumbent, departing from the premaxilla at approximately 30°. There is a minute rounded posterior cusp on the crown. The I1 is longer than the other incisors and peglike, with crowns only slightly wider than the roots. I2–4 are larger in D. hillieri than those in D. cristicauda. Additionally, they increase in size posteriorly, with I2 being the smallest and I4 being the largest.

C1 is larger than in D. cristicauda and has no additional cusps present in the adult male specimen WAM M9670. In the adult female specimen, WAM M9671, the canines are smaller than the male and have a minute anterior cusp.

There is no observable difference between the P1–P2 of D. hillieri and those of D. cristicauda aside from size. P3 is present in all known specimens of D. hillieri, unlike in D. cristicauda where presence of this tooth is variable. The P3 is also better developed and larger. In males of D. hillieri, P3 is broader and triangular with additional cusps, whilst female specimens retain a P3 that is smaller and lacks additional cusps.

M1 is longer and wider in D. hillieri than in D. cristicauda. The anterior cingulum is much larger and connects to the talon via the preprotocrista. There is an enlarged stylar cusp A present on the anterior of the tooth. In the male, the crest departing from stylar cusp A connects to the parastylar shelf, but in the female, this crest does not. A hint of stylar cusp B is present on the preparacrista where the preparacrista changes direction and runs posteriorly. In the male, the preparacrista ascends the anterior flank of stylar cusp D and meets this cusp, but in the female the preparacrista ends at the base of the anterior flank of stylar cusp D. The postparacrista is more posteriorly directed and meets the premetacrista to form a complete centrocrista. The talon is larger. The postprotocrista ends on the posterior flank of the metacone. The metastylar shelf is larger, with the distance between stylar cusp D and the metacone being greater, as is the distance between stylar cusp D and E.

M2 is longer and wider in D. hillieri than in D. cristicauda. The parastylar shelf is larger than in D. cristicauda, with the preparacrista being longer and straighter. Stylar cusp A has an anterobuccally directed crest that rests against the posterior surface of the metastyle of M1. The anterior cingulum is reduced compared to the M1. The kinks on the buccal side of the tooth between stylar cusp B and D, and D and E, are wider than in D. cristicauda but are of equal depth. The talon is enlarged. The protoconule is enlarged and more visible than in D. cristicauda.

M3 is longer and wider than in D. cristicauda, with deeper and wider kinks on the buccal surface. The anterior cingulum is not as wide as in the M2, but is slightly longer. The protoconule is enlarged.

M4 is larger than in D. cristicauda. The tooth has a larger talon with a wider and taller protocone. A minute anterior cingulum is present that rests against the lingual side of the metastyle of the M3.

Lower dentition

The i1 is slightly procumbent, but i2–3 are not. There is a diastema present between the left and right i1. The i3 is bilobed and is the smallest incisor. In dorsal view, the incisors decrease in size slightly, with i1 being the largest. All incisors have a crown that is not larger than the roots.

The c1 shares a similar morphology to C1, but is larger in Dasycercus hillieri when compared to Dasycercus cristicauda. The c1 is sexually dimorphic in size, with males having longer and wider canines than females.

The p1–2 are not morphologically different from those of D. cristicauda other than being larger.

The m1 in D. hillieri differs from that of D. cristicauda in the following ways: the tooth is longer and slightly wider; the paracristid is unnotched; the metaconid is larger and more pronounced; the protoconid is more massive and rounder, giving the trigonid shelf a bulbous look; the cristid obliqua does not connect to the protoconid and instead ends on its the posterior flank; the posterior cingulid extends further buccally around the hypoconid. The m2–4 are similar to those of D. cristicauda but larger overall.

Remarks

This taxon was considered a junior synonym of D. cristicauda by Woolley (2005). However, based on the morphological findings of this study, the taxon is valid. ‘Ampurta’ was the original common name for D. hillieri, and is a name of the Arunta Aboriginal people for the species (Woolley 2005, Finlayson 1961). It is suggested that this name be retained.

Dasycercus blythi (Waite, 1904)

Diagnosis

Dasycercus blythi differs from Dasycercus cristicauda as follows: no minute foramen present dorsal to the foramen rotundum; accessory palatal fenestrae presence variable; interparietal absent; anterior cusp is present on C1; P3 varies in morphology; protoconule of M2–3 is absent; posterior cusp present on c1 in females; brushed tail with proximal fattening present. Dasycercus blythi differs from Dasycercus hillieri as follows: overall smaller (approximately 13% smaller based on mean onl measurements); accessory palatal fenestrae not always present; lacks a masseteric process; no foramen present dorsal to the foramen rotundum; lacks a sagittal crest; internal jugular canal posteriorly positioned; no posterior cusp present on I1; minute anterior cusp is present on the C1; P3 varies in morphology; anterior cingulum of M1 does not connect to the talon; postprotocrista of M1 ends at the lingual flank of the metacone; anterior cingulum and protoconule absent on M2–3; posterior cusp present on c1 of females; brushed tail with proximal fattening. Dasycercus blythi differs from Dasycercus woolleyae as follows: overall similar size, with males of D. woolleyae slightly larger and males of D. blythi similar in size to female D. woolleyae; lateral nasal expansion present; accessory palatal fenestrae presence variable; frontal crests only observed in males; the inflation of the squama of the squamosal due to the squamosal epitympanic sinus is kidney-shaped; minute posterior cusp absent on I1; C1 has a minute anterior cusp only; P3 varies in morphology; M1 postparacrista less posteriorly directed; anterior cingulum of M1 is not connected to the talon via the preprotocrista; the postprotocrista of the M1 ends at the lingual flank of the metacone; diastema between p2 and m1 present in both sexes; m1 posthypocristid is straight; brushed tail with proximal fattening. Dasycercus blythi differs from Dasycercus archeri as follows: lateral nasal expansion present; lacks a interparietal; the inflation of the squama of the squamosal due to the squamosal epitympanic sinus is kidney-shaped; C1 has only an anterior cusp is present; where present, P3 located directly posteriorly to P2; P3 varies in morphology; protoconule absent on M2; anterior cingulum present on M4; i3 is bilobed; females have a posterior cusp present on c1; diastema is present between p2 and m1; no notch present on the paracristid of the m1; brushed tail with proximal fattening. Dasycercus blythi differs from Dasycercus marlowi as follows: overall larger (approximately 16% based on mean onl measurements of D. blythi); lateral nasal expansions present; accessory palatal fenestrae present; postorbital crests present; frontal crests present; central septum is hourglass in shape; the foramen magnum is proportionally smaller compared to the posterior cranial wall; interparietal present; mandibular condyle is ovoid; only the anterior cusp is present on C1; anterior cingulum, protoconule, and metaconule present on M1; M1 lacks a stylar cusp B; anterior cingulum present on M2–4; there is a diastema is present between p2 and m1.

Holotype

The syntype specimens are a male (WAM M41476) and female (WAM M41477) mounted specimens, with a single associated skull belonging M41477. Referred material is listed in Table 1.

Type locality

Pilbara, Western Australia.

Description

In order to avoid repetition the following description will focus on morphological differences with Dasycercus cristicauda and Dasycercus hillieri. The following description is based on specimens WAM M44970, WAM M348 (Fig. 13), WAM M1512, WAM M1530, and the syntype WAM M41477. The craniodental morphology of Dasycercus blythi is depicted in Figs 8–10.

Figure 10. Cranium of each taxon of Dasycercus; D. hillieri (WAM M9670; A, G, M), D. woolleyae (WAM M1513, holotype; B, H, N), D. blythi (WAM M1512; C, I, O), D. archeri (AMS M2987, holotype; D, J, P), D. cristicauda (WAM 67.10.74; E, K, Q), and D. marlowi (AMS M8641, holotype; F, L, R). Specimens are shown in dorsal (A–F), ventral (G–L), and lateral (M–R) views. All specimens shown are male.

Skin

The skin and tail of Dasycercus blythi (Figs 11B, 13). The skin is not described here as it has been previously described by Waite (1904), and further by Woolley (2005). Similar to Dasycercus hillieri, the pes hair morphology of specimens observed matches both the described morphologies by Adams et al. (2000). This indicates that this character is polymorphic. A full list of specimens and their plantar pes hair morphology is in Table 1. The tail is a lighter colour than D. hillieri, with 50% of the tail covered in black hairs. The morphology is brushed, as described by Woolley (2005). Nipples cannot be assessed in the female syntype specimen (Woolley 2005). Unfortunately, none of the specimens examined preserved the conditions required (lactating when preserved, nipples had been sucked, etc.) so that they could be counted.

Figure 11. Lateral view of the tails of different Dasycercus taxa, demonstrating the morphology of each. A, D. hillieri (WAM M9670); B, D. blythi (WAM M348); C, D. woolleyae (WAM M1513); D, D. marlowi (AMS M8641); E, D. archeri (AMS M2987). D. cristicauda is not represented due to the poor condition of the holotype specimen. Figures are not to scale.

Facial bones

In lateral view, the sulcus between I4 and C1 is shallow, like that of Dasycercus cristicauda. This is unlike that of Dasycercus hillieri, where the sulcus is enlarged and deep. Dasycercus blythi also lacks a masseteric process on the maxilla, similar to D. cristicauda. There is very little difference between males and females in the shape of the rostrum, with males having a slightly wider rostrum.

Neurocranium

In lateral view, the maxillary foramen is equal in size to that of Dasycercus cristicauda, but taller and more circular in shape. The canal is similar to D. cristicauda with no additional slight lacrimal process present. There is no minute foramen present dorsal to the foramen rotundum. In dorsal view, the frontals narrow posteriorly to the postorbital processes, which are markedly reduced. Beginning at the postorbital processes, two slight frontal crests run posteriorly along the cranium, forming a ‘V’ shape before merging anterior to the parietofrontal suture and continuing posteriorly along the mid-parietal suture. This character is only observed in male specimens examined, such as WAM M1512, whilst all female specimens examined lacked this character, such as WAM M44970. This indicates that it is likely sexually dimorphic. As in D. cristicauda and Dasycercus hillieri, the inflation of the squama of the squamosal due to the squamosal epitympanic sinus is smaller than that of D. hillieri, but larger than those of D. cristicauda.

Palate

In ventral view, the incisive foramina extend from I3 to mid-point of the C1. The presence of paired accessory palatal fenestrae is variable (e.g., present in WAM M44970). No other morphological differences have been observed.

Cranial floor

In ventral view, the auditory bulla is smaller than Dasycercus hillieri, but larger than that of Dasycercus cristicauda. The hypoglossal foramina are paired each side, as in other species, except in specimen M1530 which has a pair of hypoglossal foramina on the right side, but a single hypoglossal foramen on the left. The cranial floor is otherwise the same as D. cristicauda.

Posterior cranial wall

There are no morphological differences in this region between Dasycercus blythi and Dasycercus cristicauda.

Mandible

In lateral view, there is a slight diastema present between p2 and m1; this diastema is shorter than the one present in Dasycercus hillieri. On the lateral side of the mandibular ramus, there are two mental foramina present, one below the anterior root of P2, and the other below the centre of the m1, similar to that of D. hillieri. The posterior shelf of the masseteric fossa is reduced as in Dasycercus cristicauda.

Upper dentition

I1 is less procumbent, facing straight down, and there is no diastema present between them, unlike Dasycercus hillieri, which has a diastema present between I1 and more procumbent teeth. There is no posterior cusp present on the I1, unlike D. hillieri, which has a very small cusp present. There is a short diastema present between I1 and I2. I2–4 are similar to Dasycercus cristicauda. C1 is larger than that of D. cristicauda, but smaller than that of D. hillieri. An anterior cusp is present on C1.

P1–2 have no observable differences with those of D. cristicauda, besides being relatively larger in size (Appendix). P3 is variably present, being observed only in some individuals (e.g., WAM M1512). Where present, the P3 varies in morphology, even on the same individual, having either an anterior and posterior cusp present, or having a posterobuccal shelf.

M1 is transversely wider relative to its length than that of D. cristicauda, but smaller overall than that of D. hillieri (Appendix 3). The talon is proportionally smaller than in D. hillieri. Stylar cusp A is larger than in D. cristicauda, but smaller than in D. hillieri. The anterior cingulum is equal in size to D. cristicauda and does not connect to the talon. The postmetacrista is longer than in D. cristicauda. The postprotocrista ends at the lingual flank of the metacone, as in D. cristicauda.

M2 is longer and wider than in D. cristicauda, but smaller overall than in D. hillieri. In Dasycercus blythi, there is no anterior cingulum present, similar to D. cristicauda but different from D. hillieri. Stylar cusp B is larger and less rounded than in D. hillieri and D. cristicauda. The preparacrista is longer than in D. cristicauda, but similar in length to that of D. hillieri. The protoconule is absent, unlike in D. cristicauda and D. hillieri.

M3 is similar to M2 except as follows: stylar cusp A is larger than in D. cristicauda or D. hillieri. The buccal kinks are not as deep. Overall morphology of the tooth is similar to that of D. cristicauda, but lacking a protoconule. M4, besides being relatively larger in size to that of D. cristicauda, has no noticeable morphological differences.

Lower dentition

The i1–3 have no morphological differences compared to those of Dasycercus hillieri, besides being relatively smaller in size (Appendix). The c1 is similar in morphology to C1 but is smaller and lacks the posterior cusp in males.

The p1–2 show no morphological differences compared to the same teeth in D. hillieri and Dasycercus cristicauda. There is a very short diastema presents between p2 and m1.

The m1 shares a similar morphology to that of D. cristicauda. The tooth is the same length but narrower than in D. cristicauda, but overall smaller than in D. hillieri. The paracristid is not notched. The protoconid is less rounded than in D. hillieri and is similar in appearance to that of D. cristicauda. The metaconid is more pronounced than in either D. hillieri or D. cristicauda.

The m2 is similar in morphology and size to that of D. cristicauda, but smaller than that of D. hillieri with a less rounded trigonid shelf. The anterior and posterior cingulid are smaller than in D. cristicauda and D. hillieri.

The m3 is similar in morphology and size to that of D. cristicauda, but smaller in size than that of D. hillieri. The anterior and posterior cingulid are smaller than in D. cristicauda, but larger than in D. hillieri.

The m4 is similar in morphology to that of D. cristicauda, being smaller than that of D. hillieri. The anterior cingulid is smaller than in D. cristicauda but larger than in D. hillieri. The talonid is larger than in D. cristicauda but smaller than in D. hillieri.

Remarks

The common name of ‘Brush-tailed Mulgara’ was assigned by Woolley (2005) for ease of differentiation between the recognized taxa of the time. This common name can remain, but the brush-tail is no longer unique to this species. For this reason, ‘Spinifex Mulgara’ has been suggested as an additional common name, based on the preferred habitat of the taxa.

Dasycercus woolleyae sp. nov. Newman-Martin & Travouillon

Zoobank identifier

urn:lsid:zoobank.org:act:3E172626-4752-4CF3-A11C-62305CA7AB31

Diagnosis

Dasycercus woolleyae differs from Dasycercus cristicauda in the following ways: overall larger (approximately 11% larger based on mean occipitopremaxillary length); lacks lateral nasal expansion; slight process present between the sphenopalatine foramen and the maxillary foramen; no minute foramen present dorsal to the foramen rotundum; frontal crests present in both males and females, with the ‘V’ ending more posteriorly in females; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; interparietal absent; the diastema between p2 and m1 is present only in males of D. woolleyae; the M1 postparacrista is more posteriorly directed; the anterior cingulum of the M1 is connected to the talon via the preprotocrista in D. woolleyae; the postprotocrista of M1 ends at the posterior flank of the metacone; the paracristid of the m1 is not notched; the posthypocristid of m1 is curved. Dasycercus woolleyae differs from Dasycercus hillieri in the following ways: lateral nasal expansion absent; there is no minute foramen present dorsal to the foramen rotundum; the frontal crests can be observed in both males and females, but with the ‘V’ ending more posteriorly in females; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; lacks accessory palatal fenestrae; the internal jugular canal is posteriorly located; C1 has minute cusps anteriorly and posteriorly; the morphology of the C1 is not sexually dimorphic; P3 presence is variable; P3, if present, is rounded, circular, and weakly developed; M1 postparacrista does not meet the preparacrista to for a centrocrista; M2–4 have a minute anterior cingulum present; posthypocristid of m2–4 is curved; m2 protoconid, metaconid and paraconid are less rounded. Dasycercus woolleyae differs from Dasycercus blythi in the following ways: overall similar size, with males of D. woolleyae slightly larger and males of D. blythi similar in size to female D. woolleyae; lacks lateral nasal expansions; frontal crests present in both males and females, with the ‘V’ ending more posteriorly on the parietal in females; accessory palatal fenestrae are always absent; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; minute posterior cusp present on I1; C1 has a minute anterior and posterior cusp; P3 presence variable, but is minute and rounded where present; M1 postparacrista is more posteriorly directed; anterior cingulum of M1 is connected to the talon via the preprotocrista; postprotocrista ends at posterior flank of metacone; diastema between p2 and m1 is present only in males examined; posthypocristid of the m1 is curved; crested tail with black dorsal hairs, and proximal fattening absent. Dasycercus woolleyae differs from Dasycercus archeri in the following ways: similar in size with males of D. woolleyae larger and males of D. archeri subequal in size to female D. woolleyae; lacks interparietal; P3, where present, is located directly anterior to M1; M1 postparacrista of M1 meets premetacrista; anterior cingulum present on M4; i3 is bilobed; the diastema between p2 and m1 is present, but only in males examined; the paracristid of the m1 is not notched. Dasycercus woolleyae differs from Dasycercus marlowi in the following ways: overall larger (approximately 18% larger based on mean onl measurements); frontal crests present; postorbital process present and large; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; the central septum between the maxillopalatine fenestrae is hourglass in shape; diastema present between p2 and m1, but only in males examined; mandibular condyle is ovoid in shape; C1 has a minute anterior and posterior cusp present, the posterior cusp is taller than the crown of P1; M1 postparacrista is more posteriorly directed; anterior cingulum of M1 is connected to the talon via the preprotocrista; stylar cusp B is absent on M1–4; postprotocrista of M1 ends at the posterior flank of the metacone; protoconule and metaconule are present on M1; anterior cingulum present on M2–3; m1 posthypocristid is curved; the tail is crested with black hairs dorsally, with proximal fattening absent.

Etymology

The taxon is named in recognition of mammologist Dr Patricia Woolley for her taxonomic investigations of the genus. Prior to this work, the mulgaras were poorly studied. Many of the museum study skin specimens examined in investigation were collected by Dr Woolley, who greatly expanded the representation of Dasycercus within museum collections. Dr Woolley has also contributed greatly to the understanding of mammalogy, and specifically dasyurids, over a long and distinguished career.

Holotype

The type specimen for this species (WAM M1513) is an adult male cranium, dentaries, and skin collected by Otto Lipfert in 1931 on the Canning Stock Route expeditions. The label attached to the skin indicates that, at the time of collection, the specimen had the following measurements: head–body: 15.7 cm, tail length: 10.5 cm, hind foot: 3 cm and ear: 2.6 cm.

Paratypes

SAM M3144, adult female cranium, dentaries and skin; SAM M3105, adult male skull, dentaries and skin; WAM M41891, adult male skull and dentaries; WAM M1520, adult male skull, dentary and skin; WAM 73.1.195, indeterminate adult subfossil skull and dentary; WAM M1511, adult male skull, dentary and skin; AMS M8639, adult female skull, dentaries and skin; AMS M1144, adult male skull, dentaries and skin; AMS M8638, adult male skull, dentaries and skin; ANWC M00529, adult female skull, dentaries and skin. Referred material is listed in Table 1.

Type locality

Point Massie (Well 46, Canning Stock Route), Western Australia (20°40′00ʺ S, 126°26′00ʺ E)

Description

The following description relates to the holotype specimen, WAM M1513, unless otherwise stated. To avoid repetition, the description will focus of morphological differences compared to Dasycercus cristicauda, Dasycercus hillieri and Dasycercus blythi. The craniodental morphology of Dasycercus woolleyae is depicted in Figs 8–10.

Skin

The skin and tail of D. woolleyae is represented in Figs 11c, 14. The pelage is an orange-brown colour dorsally; the dorsal colour is redder and lighter than that of Dasycercus blythi while also much less brown than Dasycercus cristicauda, but darker than Dasycercus hillieri. This colour lightens laterally, similarly to other Dasycercus taxa. Laterally, the colour is a brownish creamy colour that continues to a lighter colour at the feet. The lateral colour is less brown than that of D. blythi, and lighter than D. cristicauda, but similar to that observed in D. hillieri. Ventrally, the specimen is an uninterrupted light creamy colour until the base of the tail, as in other Dasycercus taxa. The plantar pes hair morphology of Dasycercus woolleyae appears to be polymorphic, with the left and right pes even displaying different morphologies in WAM M1513. It is believed that this could be a result of preservation, as degradation and alteration of pelage is likely to happen when the specimen is being skinned or preserved (JN-M and KJT pers. obs.). However, it is not known for certain if the polymorphism is due to preservation. The morphology of the left pes is as follows: the lateral plantar pes hairs ‘fold’ over and towards the medial surface; the medial pes hairs on the lower half of the plantar surface ‘fold’ over laterally; the hair on the plantar pes covers roughly 75% of the surface. The morphology of the right pes differs in that on the lateral surface no hairs fold over, resulting in roughly 50% hair coverage. Multiple individuals were observed with both the D. hillieri plantar pes hair morphology and the D. blythi morphology. A full list of specimens and their plantar pes hair morphology is in Table 1.

The tail is the orange-brownish colour of the dorsal pelage. The tail is thinner and browner than in D. blythi, and has a black dorsoventral crested tip. This crest is not as substantial as is observed in D. hillieri, with the dorsal hairs being much shorter. A few individuals have been observed with tail fattening at the base, similar to D. blythi, indicating a polymorphic character. No specimens were identified with visible nipples that could be counted.

Facial bones

In lateral view, the premaxillary–nasal suture is longer than the maxillary–nasal suture. Dasycercus woolleyae lacks a posterior nasal lateral expansion, unlike Dasycercus cristicauda. Similarly to Dasycercus hillieri, the rostrum is sexually dimorphic, with males having longer and wider rostra than females (e.g., specimen ANWC CM529); however, the degree of dimorphism between the sexes is not as great as is seen in D. hillieri. The facial region of the skull is largely similar to that of D. hillieri.

Neurocranium

In lateral view, there is a slight lacrimal process present between the sphenopalatine foramen and the maxillary foramen, similar to the one seen in Dasycercus hillieri. However, the process is smaller and does not extend as far posteriorly as in D. hillieri. As in Dasycercus blythi, there is no minute foramen present dorsal to the foramen rotundum.

In dorsal view, two slight frontal crests run posteriorly along the cranium and form a ‘V’ shape, similar to that seen in Dasycercus blythi. However, unlike other in Dasycercus taxa, this character can be observed faintly in both males and females, but with the ‘V’ ending more posteriorly in females. The postorbital process is markedly larger than in D. cristicauda and D. blythi, but smaller than in males of D. hillieri. Male Dasycercus woolleyae specimens have a postorbital process slightly larger than that of female D. hillieri. The inflation of the squamosal caused by the squamosal epitympanic sinus is kidney shaped, similar to D. cristicauda, and lateral to the subsquamosal foramen. Additionally, the squamosal epitympanic sinus is larger than in D. blythi and D. cristicauda, but smaller than in D. hillieri. No interparietal is present on the posterodorsal surface of the cranium, similar to D. hillieri and D. blythi. The neurocranium displays a degree of sexual dimorphism, with the zygomatic width and overall size of the cranium being larger in all males examined. D. woolleyae males and D. hillieri females can be difficult to differentiate due to size similarities. Additionally, female D. woolleyae (specimen ANWC CM529) and male D. blythi are similar in size, so can also be difficult to differentiate.

Palate

Dasycercus woolleyae lack accessory palatal fenestrae, similar to Dasycercus cristicauda. There are no morphological differences on the palate.

Cranial floor

In ventral view, the auditory bulla is larger than in Dasycercus cristicauda but smaller than in Dasycercus hillieri. All the examined males of Dasycercus blythi and Dasycercus woolleyae have similarly sized bullae, but all examined females of D. blythi have larger bullae than female D. woolleyae. Posterior to the auditory bulla, the paroccipital process of the exoccipital, and the rostral and caudal tympanic processes of the petrosal are equal in size to those of D. cristicauda, but smaller than that of D. hillieri and D. blythi. There is a single hypoglossal foramen on the left side of the holotype (WAM M1513), but the right side is obscured by tissue, and two foramina are present in other specimens (e.g., SAMA M3144).

Posterior cranial wall

Inside the skull, the internal jugular canal is posteriorly positioned, similar to Dasycercus cristicauda and Dasycercusblythi. There are no noticeable morphological differences in this region of the skull compared to D. cristicauda or D. blythi.

Mandible

In lateral view, the diastema between p2 and m1 is present only in males, and is markedly shorter than Dasycercus blythi and Dasycercus hillieri. In females, it is still absent as in Dasycercus cristicauda.

Upper dentition

I1 has a diastema present between the right and left I1, and this tooth is slightly procumbent. There is a minute posterior cusp present, larger than in Dasycercus hillieri. I2–4 increase in size posteriorly, with the I4 being the largest, similar to D. hillieri.

C1 is large and caniniform, with minute cusps anteriorly and posteriorly. The morphology of C1 is not sexually dimorphic, with no observable differences between sexes other than size.

P1–2 have no morphological differences to the other Dasycercus species. P3, if present, is rounded, circular, and weakly developed, as in Dasycercus cristicauda.

M1 is similar in width and length to that of D. hillieri, but larger than that of D. cristicauda and Dasycercus blythi. M1 is similar in overall morphology to that D. hillieri, except that the postparacrista does not meet the preparacrista to form a centrocrista. The postparacrista is more posteriorly directed than in D. cristicauda and D. blythi. The anterior cingulum is smaller than in D. hillieri, but larger than in D. blythi and D. cristicauda. The anterior cingulum is connected to the talon via the preprotocrista. The postprotocrista ends at the posterior flank of the metacone.

M2 is similar in morphology to the M1. The tooth is shorter in width than D. hillieri, but wider and longer than in D. blythi and D. cristicauda. The anterior cingulum is smaller than seen in D. hillieri. The protoconule is smaller than in D. hillieri and D. cristicauda. Overall tooth morphology is closer to that of D. hillieri than that of D. cristicauda and D. blythi.

M3 is similar in size to that of D. hillieri, with this tooth being slightly larger in D. hillieri. There is a minute anterior cingulum. A minute stylar cusp C is present just anterior to stylar cusp D on the left M3 of the holotype, but this is absent in other specimens. The talon is smaller than in D. hillieri, but larger than in D. blythi and D. cristicauda.

M4 is similar to D. hillieri, but is narrower. The tooth is wider than in D. blythi and D. cristicauda. A minute anterior cingulum is present. The talonid is small and proportionally similar to that in D. cristicauda and D. blythi.

Lower dentition

The i1–3 i1 is procumbent to the jaw, with i3 bilobed. The c1 is similar morphologically to C1. The p1–2 are morphological similar to these teeth in other taxa.

The m1 is similar in size to m1 in Dasycercus hillieri, but larger than this tooth in Dasycercus cristicauda and in Dasycercus blythi. The paracristid is unnotched. The posthypocristid is not straight as in other taxa, but is instead curved. The posterior cingulid is smaller than in D. hillieri, but larger than in D. blythi and D. cristicauda. The m2 has the protoconid, metaconid and paraconid less rounded than in D. hillieri. The trigonid valley is deeper than in D. hillieri. The tooth is longer and wider than in D. cristicauda and D. blythi, but similar in size to that of D. hillieri. The anterior cingulum is larger than in D. blythi and D. cristicauda, and the posthypocristid is longer posteriorly than in D. blythi and D. cristicauda. The m3 is similar in morphology and size to that of D. blythi. The m4 is similar to that seen in D. hillieri, D. cristicauda and D. blythi, but differs in that there is a large anterior cingulid present. The talonid is similar in size to that of D. blythi.

Remarks

The common name of ‘Northern Mulgara’ is suggested by the authors. This is based on the taxon’s northern distribution (particularly in the case of the holotype, which was collected on the Canning stock route), while also keeping the common name simple. It should be noted that future efforts should be made to determine the name used by the Aboriginal peoples where the species occurred.

Dasycercus archeri sp. nov. Newman-Martin & Travouillon

Zoobank identifier

urn:lsid:zoobank.org:act:DBCF4171-78A7-4F55-AE9A-02CD085A30F2

Diagnosis

Dasycercus archeri differs from Dasycercus cristicauda in the following ways: nasal lateral expansions are absent; accessory palatal fenestrae variable in presence; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; the interparietal is fused to the supraoccipital; P3, where present, is located anterolingual to M1; the anterior cingulum of M1 connects to the talon via the preprotocrista; the postprotocrista of the M1 ends on the posterior flank of the metacone; i3 is wide and triangular, with a crown that is wider than the root; i3 is also not bilobed; there is no diastema between p2 and m1. Dasycercus archeri differs from Dasycercus hillieri in the following ways: lateral nasal expansion absent; accessory palatal fenestrae presence variable; a sagittal crest is absent; interparietal present; internal jugular canal posteriorly positioned; C1 has an anterior and posterior cusp present; P3 is not always present; where present, P3 is minute and rounded and located anterolingual to the M1; M1 postparacrista is posterobuccally directed but does not meet the premetacrista; M4 lacks the anterior cingulum; i3 is wide and triangular, with a crown that is wider than the root; i3 is not bilobed; no diastema is present between p2 and m1; m1 paracristid is notched. Dasycercus archeri differs from Dasycercus blythi in the following ways: lateral nasal expansion absent; the inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid in shape; interparietal present; C1 has an anterior and posterior cusp; P3, where present, is minute, rounded, and located anterolingual to the M1; M2 has a large protoconule; M4 lacks an anterior cingulum; i3 is wide and triangular, with a crown that is wider than the root; i3 is not bilobed; c1 lacks additional cusps; no diastema is present between p2 and m1; m1 paracristid is slightly notched; tail is crested with black dorsal hairs, and with proximal fattening absent. Dasycercus archeri differs from Dasycercus woolleyae in the following ways: similar in size with males of D. woolleyae larger and males of D. archeri subequal in size to female D. woolleyae; interparietal present; where present, P3 is located anterolingual to the M1; M1 postparacrista does not meet premetacrista; M4 lacks an anterior cingulum; i3 is wide and triangular, with a crown that is wider than the root; i3 is not bilobed; no diastema present between p2 and m1; m1 paracristid is notched. Dasycercus archeri differs from Dasycercus marlowi in the following ways: postorbital crests present; frontal crests present; interparietal present; an anterior and posterior cusp are present on C1, the posterior cusp is taller than the crown of the P1; anterior cingulum of the M1 connects to talon via preprotocrista; postprotocrista of the M1 ends on posterior flank of metacone; M2 has a large protoconule, and the anterior cingulum is present; anterior cingulum present on M3; i3 is wide, triangular and not bilobed; m1 paracristid notched; tail crested with black dorsal hairs, and without proximal fattening.

Etymology

The taxon was named in recognition of mammologist and palaeontologist Professor Michael Archer, who has contributed greatly to our understanding of Australian mammals and their history across the continent. Professor Archer was the director of the Australian Museum and has been instrumental in uncovering the secrets of the Riversleigh World Heritage fossil site. Many of the subfossil specimens used in this investigation were also collected by Professor Archer, with the first specimens identified of this species coming from Horseshoe Cave, which was excavated by him (Archer 1974). Professor Archer has also worked extensively with Dasyuridae, having named multiple species over his career.

Holotype

The type specimen is an adult male skull and dentaries with associated skin (AMS M2987).

Paratypes

AMS M4356, adult female skull, dentaries and skin; AMS M3025, adult male skull, dentaries and skin; AMS M4862, adult male skull, dentaries and skin; AMS M4864, adult female skull, dentaries and skin; AMS M4863, adult female skull, dentaries and skin; SAM M17499, adult skull and dentaries; WAM M65281, adult of indeterminate sex, subfossil skull; WAM M64827, adult of indeterminate sex, subfossil skull; WAM M64833, adult of indeterminate sex, subfossil skull; WAM M65340, adult of indeterminate sex, subfossil skull; WAM 72.1.1057, adult of indeterminate sex, subfossil skull, jaw and postcranials. Referred material is listed in Table 1.

Type locality

Ooldea, Trans-Australian railway line, South Australia.

Description

The following description relates to the holotype specimen, AMS M2987, and listed paratypes only unless otherwise stated. As previously, the description will focus on morphological differences between Dasycercus archeri and Dasycercus cristicauda, Dasycercus hillieri, Dasycercus blythi and Dasycercus woolleyae. The craniodental morphology of Dasycercus archeri is depicted in Figs 8–10.

Skin

The skin and tail of Dasycercus archeri is represented in Figs 11E, 15. The pelage of the holotype specimen is reddish-brown dorsally; the dorsal coloration is browner than in Dasycercus woolleyae, redder than in Dasycercus blythi, darker than in Dasycercus hillieri, and lighter than in Dasycercus cristicauda. Laterally, the colour becomes a light-brown colour, similar to other Dasycercus species. Ventrally and around the feet, the fur is a tan colour. The hair of the left pes and right pes of the holotype specimen (AMS M2987) display different morphologies. It is not known if this is the result of preservation or polymorphic pes hair. The upper plantar surface of the right pes is covered by long hairs that extend and ‘fold’ in over from medial to lateral side of the plantar pes. The lateral side of the pes is covered by hairs that direct distally. Hair on the right pes covers roughly 33% of the plantar surface. The left pes differs in that the hair on the upper plantar surface covers less of the plantar surface and is not as long. The hair on the left pes covers roughly 25% of the plantar surface. A full list of specimens and their plantar pes hair morphology is in Table 1.

Figure 12. The skin of specimen WAM M8207. This specimen was used to represent D. hillieri. Views are as follows; A, dorsal; B, ventral; C, lateral.

The tail is reddish-brown proximally and black distally, with crested hairs dorsally. The tail morphology is similar to that of D. hillieri; however, the dorsal hairs are shorter and orientated more distally, which is similar to the condition in Dasycercus woolleyae. The tail exhibits no proximal fattening, as opposed to the condition in D. blythi. The number of nipples was observed to be variable, with two specimens (AMS M4355 and AMS M4864) having eight nipples, and another (AMS M4863) having six. It is not known if this is due to poor preservation or if the character is polymorphic.

Facial bones

Dasycercus archeri lacks a posterior lateral expansion on the nasal, similar to Dasycercus woolleyae. No differences in this region have been observed other than being larger than Dasycercus cristicauda.

Neurocranium

In lateral view, the sphenopalatine foramen and maxillary foramen are similar in size and appearance to Dasycercus cristicauda, with no additional process present. In dorsal view, the two frontal crests that run posteriorly along the cranium form a ‘V’ shape in the same way that can be seen in that of Dasycercus blythi and Dasycercus woolleyae. The inflation of the squamosal caused by the squamosal epitympanic sinus is reduced and ovoid in shape, equal in size to that seen in D. cristicauda. There is an interparietal present on the posterior of the cranium, similar to D. cristicauda. However, the interparietal is fused to the supraoccipital in Dasycercus archeri.

Palate

The accessory palatal fenestrae are variable in presence, similar to the condition in Dasycercus blythi. No differences in this region compared to that of Dasycercus cristicauda are observed, other than being larger.

Cranial floor

In ventral view, the auditory bulla is larger than that of Dasycercus blythi and Dasycercus cristicauda, but smaller than that of Dasycercus woolleyae and Dasycercus hillieri. Posterior to the auditory bulla, the paroccipital process of the exoccipital, and the rostral and caudal tympanic processes of the petrosal, are cylindrical and elongated, and as a result they are equal in size to that found in D. cristicauda.

Posterior cranial wall

The internal jugular canal is posteriorly placed, similar to Dasycercus cristicauda. No observable morphological difference in this region with D. cristicauda other than being larger.

Mandible

No morphological differences to Dasycercus cristicauda other than being larger.

Upper dentition

I1 is less procumbent, similar to that of Dasycercus blythi. There is a small diastema present between the left and right I1, smaller than in Dasycercus woolleyae. I2–4 are rounded and similar in morphology to those of D. woolleyae. The incisors increase in size, with I2 being the smallest and I4 being the largest. I4 is similar in morphology to that seen in D. woolleyae, with a wider crown than the root.

C1 has an anterior and posterior cusp present. This is similar to D. woolleyae and Dasycercus cristicauda, as well as Dasycercus hillieri females.

P1–2 have no difference in morphology to other Dasycercus species; however, the diastema between P2 and M1 is smaller than is seen in other Dasycercus species. P3 is not always present, as in other Dasycercus taxa besides D. hillieri. Where present, it is minute and rounded, as in D. cristicauda and D. woolleyae, but it is located anterolingual to the M1, instead of directly anterior.

M1 is similar to that of D. woolleyae, but not as wide. The postparacrista is posterobuccally directed, but does not meet the premetacrista as it does in D. woolleyae. The anterior cingulum connects to the talon via the preprotocrista, as in D. woolleyae and D. hillieri. The postprotocrista ends on the posterior flank of the metacone, as in D. woolleyae and D. hillieri. M2 is similar to that of D. woolleyae, but smaller in length and width. It has a larger protoconule than observed in D. cristicauda and D. woolleyae, but smaller than in D. hillieri. This cusp is absent in D. blythi. M3 is longer and wider than in D. cristicauda and D. blythi, but is smaller than in D. woolleyae and D. hillieri. The anterior cingulum is minute and almost absent. The protoconule is larger than in D. cristicauda. M4 is similar in morphology to that of D. cristicauda as it lacks the anterior cingulum, which is present in D. hillieri, D. woolleyae and D. blythi. The M4, however, is larger than that of D. cristicauda, but smaller than that seen in D. hillieri, D. woolleyae and D. blythi.

Lower dentition

The i1–3 are similar to those of Dasycercus hillieri except that the i1 is more procumbent to the mandible. In the holotype specimen, the right i1 is missing, but based on the dental alveolus there appears to be a diastema present between both the i1s. The i3 is wide and triangular, with a crown that is wider than the root. Additionally, the i3 is not bilobed, as is seen in other Dasycercus species. The c1 is similar to C1, but smaller and lacking additional cusps. The p1–2 show no morphological difference compared to these teeth in other Dasycercus species. No diastema is present between p2 and m1.

The m1 is similar to that of D. hillieri. This tooth is smaller than in Dasycercus woolleyae and D. hillieri, but larger than in Dasycercus blythi and Dasycercus cristicauda. The protoconid is bulbous, and the metaconid is minute, as in D. cristicauda. The paracristid is only slightly notched, as in D. cristicauda, with a slight groove running posterolingually. The m2 is similar in morphology to that of D. hillieri, except the trigonid shelf is less rounded, with the paraconid, protoconid and the metaconid being more defined. The anterior cingulid is smaller than in D. hillieri, being closer in size and morphology to that of D. blythi. The m3 has a larger anterior cingulid than in D. hillieri. The entoconid is smaller than in D. hillieri, as in all other species of Dasycercus. The m4 is identical morphologically and in size to D. woolleyae.

Remarks

The suggested common name for this taxon is the ‘Southern Mulgara’, owing to the species’ southern distribution, especially in the case of the type specimen and paratypes, which were collected around Ooldea. As previously, it is suggested that future efforts be made to work with Aboriginal communities in the areas where specimens were collected to determine more appropriate common names.

Dasycercus marlowi sp. nov. Newman-Martin & Travouillon

Zoobank identifier

urn:lsid:zoobank.org:act:D957B0D9-225C-4911-8273-F0B9E8750EB3

Diagnosis

Dasycercus marlowi differs from Dasycercus cristicauda in the following ways: nasal lateral expansion absent; there is no minute foramen present dorsal to the foramen rotundum; postorbital crests absent; frontal crests absent; interparietal absent; the anterior and posterior cusps present on C1 are both shorter than the crown of P1, possibly as the canine is less erupted; M1 lacks an anterior cingulum; stylar cusp B is present on the M1; M2–3 lack an anterior cingulum; no notch present on m1 paracristid; tail brushed and fattens proximally. Dasycercus marlowi differs from Dasycercus hillieri in the following ways: no nasal lateral expansion present; overall markedly smaller (approximately 32% smaller based on mean onl measurements); no minute foramen present dorsal to the foramen rotundum; postorbital crests absent; sagittal crest absent; accessory palatal fenestrae absent; frontal crests absent; internal jugular canal is posteriorly positioned; anterior and posterior cusps present on C1; M1 lacks an anterior cingulum; stylar cusp B is present; postparacrista of the M1 does not meet premetacrista to form a complete centrocrista; postprotocrista ends on lingual flank of metacone; M2–3 lack an anterior cingulum; tail is brushed and fattens proximally. Dasycercus marlowi differs from Dasycercus blythi in the following ways: overall smaller; lacks lateral nasal expansion; lacks postorbital crests; frontal crests absent; central septum between the maxillopalatine fenestrae is rectangular; accessory palatal fenestrae absent; mandibular condyle circular; C1 has a minute anterior and posterior cusp present located below the crown of P1; M1 lacks an anterior cingulum; stylar cusp B of M1 is present; protoconule and metaconule absent on M1; M2–4 lack an anterior cingulum; no diastema is present between p2 and m1. Dasycercus marlowi differs from Dasycercus woolleyae in the following ways: overall smaller; process between the sphenopalatine foramen and the maxillary foramen absent; frontal crests absent; postorbital process absent; the inflation of the squamosal caused by the squamosal epitympanic sinus is kidney shaped; the central septum of the maxillopalatine fenestrae is rectangular in shape; diastema absent between p2 and m1; mandibular condyle is circular; C1 has a minute anterior and posterior cusp, both cusps are shorter than the crown of P1; M1 postparacrista is less posteriorly directed; anterior cingulum absent on M1; stylar cusp B present on M1; postprotocrista ends at lingual flank of metacone; protoconule and metaconule absent; M2–3 lack an anterior cingulum; m1 posthypocristid straight; tail brushed, with proximal fattening. Dasycercus marlowi differs from Dasycercus archeri in the following ways: overall smaller (approximately 16% based on the mean onl measurements of D. archeri); postorbital crests absent; frontal crests absent; accessory palatal fenestrae consistently absent; interparietal absent; minute anterior and posterior cusps on C1, both cusps are shorter than the crown of P1; M1 anterior cingulum does not connect to talon via preprotocrista; postprotocrista of the M1 ends on lingual flank of metacone; anterior cingulum absent on M2–3; there is no diastema present between left and right i1 on the dentary, contra to D. archeri; i3 is bilobed, contra to D. archeri; notch absent on m1 paracristid; tail brushed, and fattened proximally.

Etymology

The taxon is named after mammologist Basil J. Marlow, who collected the only known specimen. Marlow was previously the Curator of Mammals at the Australian Museum, and contributed greatly to the collection and updating the techniques used during his time at the facility. Marlow also researched the behaviour of Australian mammals and helped to further our understanding of them. Through his books, such as Marsupials of Australia (Marlow 1962), this knowledge and much more was made more accessible to the public.

Holotype

The holotype (AMS M8641) is an adult male skull and skin, collected in 1965 by Basil J. Marlow. The attached label indicates that at the time of capture the specimen had the following measurements: total length: 195 mm; tail–vent: 84 mm; hind foot: 25 mm; ear: 21 mm; weight: 40 g.

Type locality

East-northeast of Numery Station, Northern Territory (23°52′ S, 125°17′ E).

Description

The following description for Dasycercus marlowi is for the holotype specimen (AMS M8641). Unfortunately, no other specimens are currently known. To avoid repetition the description will focus on morphological differences with Dasycercus cristicauda, Dasycercus hillieri, Dasycercus blythi, Dasycercus woolleyae and Dasycercus archeri. The craniodental morphology of Dasycercus marlowi is depicted in Figs 8–10.

Skin

The skin and tail of Dasycercus marlowi is represented in Figs 11D, 16. The pelage is near identical to specimens of Dasycercus blythi from similar localities; however, the animal is overall smaller. Coloration is dark brown dorsally, darker than that of Dasycercus woolleyae, Dasycercus hillieri, Dasycercus archeri and Dasycercus cristicauda. The fur lightens laterally to a creamy colour, similar to other Dasycercus species. Ventrally and around the feet, the fur colour becomes an off white; this is lighter than in D. archeri, D. hillieri, D. woolleyae and D. cristicauda. The pes hair morphology is similar to that described by Adams et al. (2000) for D. blythi. The tail is fattened proximally, similar to D. blythi, and also brushed.

Figure 13. The skin of specimen WAM M348. This specimen was used in this investigation to represent D. blythi. Views are as follows; A, ventral; B, dorsal; C, lateral.

Facial bones

In lateral view, the rostrum is proportionally shorter than other Dasycercus taxa, but it is taller than in Dasycercus cristicauda, as tall as Dasycercus blythi. The maxillary–nasal suture is almost as long as the premaxillary–nasal suture, shorter than in other Dasycercus species. There is no posterior nasal lateral expansion present, similar to Dasycercus woolleyae and Dasycercus archeri. There is no masseteric process present, as in all other Dasycercus except Dasycercus hillieri.

Neurocranium

In lateral view, the sphenopalatine foramen and maxillary foramen are similar in size and appearance to those of Dasycercus cristicauda, with no additional process present. There is no minute foramen present dorsal to the foramen rotundum. In dorsal view, the squamosal epitympanic sinus is larger than in D. cristicauda, with the subsquamosal foramen being smaller than that of D. cristicauda. The cranium lacks postorbital crests over the orbits, contra all other Dasycercus taxa. The skull lacks any major crests on the frontals or parietals, and is smooth dorsally. No interparietal is present. The nuchal crest is reduced and not very prominent.

Palate

The palate is similar in morphology to that of Dasycercus cristicauda, but differs in the length of the maxillopalatine fenestrae, which begin at the midpoint of the M1 and terminate at the anterior of the M3. The central septum is wider than in D. cristicauda and rectangular, unlike in other Dasycercus, where it is hourglass. There are no accessory palatal fenestrae present, similar to the condition in D. cristicauda.

Cranial floor

The foramen ovale is small and hidden below the auditory bulla, located inside a canal that runs anteriorly and connects to the transverse sinus canal. The auditory bulla, although broken, is similar in size to Dasycercus cristicauda but smaller than in all other Dasycercus taxa. The paroccipital process of the exoccipital is similar in morphology and size, but taller than in Dasycercus woolleyae, being elongated as well as inflated.

Posterior cranial wall

In posterior view, the foramen magnum is proportionally larger in Dasycercus marlowi than in other Dasycercus taxa, with the opening being equal in size to Dasycercus woolleyae, despite the posterior cranial wall being markedly smaller. The supraoccipital is convex, similar to Dasycercus cristicauda. The internal jugular canal is posteriorly positioned.

Mandible

Similar in morphology to Dasycercus cristicauda; however, the mandibular condyle is circular as opposed to ovoid, similar to Dasycercus hillieri but markedly smaller.

Upper dentition

I1 is less procumbent than other Dasycercus taxa, similar to Dasycercus blythi with a very slight diastema present between both teeth. I2–4 increase in size, with I4 being the widest and I2 being the thinnest. I4 is wide and spadelike, with the crown being wider than the root. I2 is peglike, with the crown being similar in size to the root. The crown of I3 is slightly wider than the root and I3, is intermediate in width between I2 and I4.

C1 has a minute anterior and posterior cusp present; however, these cusps are located below the crown of P1, whereas they are above the crown of P1 in other Dasycercus taxa. However, this could be due to ‘canine over eruption’ (Jones 2023), where the canines of the Dasycercus marlowi holotype specimen (AMS M8641) may not be fully erupted. The canines are similar sized to Dasycercus cristicauda, but smaller than all other Dasycercus taxa.

P1–2 have no morphological differences, besides being markedly smaller than those of other Dasycercus species. P2 is narrower than in D. cristicauda. P3 is absent, with a small diastema present between P2 and M1.

M1 is similar in size to that of D. blythi. The M1 is most similar in morphology to that of D. cristicauda. D. marlowi lacks an anterior cingulum, and stylar cusp A is markedly reduced. A small stylar cusp B is present on the parastylar shelf, directly buccal to the paracone, on the preparacrista. The postparacrista does not meet with the premetacrista to form a centrocrista. The cusps of the metastylar shelf are more defined than on D. cristicauda and less rounded. The talon is larger than in D. cristicauda, similar in size to D. blythi, with the postprotocrista ending at the lingual flank of the metacone. The protoconule and metaconule are absent. M2–3 are morphologically similar to those of D. blythi. They lack an anterior cingulum, and stylar cusp A is markedly reduced. The protoconule is minute on the talon. M4 has a minute talon, the smallest of all the Dasycercus taxa discussed here.

Lower dentition

The i1–3 are similar to those of Dasycercus blythi. i1 is procumbent, at an angle of ca 70^° to the mandibular ramus of the jaw. No diastema present between the i1s. The i3 is bilobed; however, the posterior lobe is minute, as in D. blythi. The c1 is similar to C1, but smaller and sloping further posteriorly. The p1–2 have no morphological differences compared to the P1–2. No diastema is present between p2 and m1, similar to Dasycercus archeri.

The m1 is similar in morphology to D. blythi, and is similar in size. The protoconid is taller than that of D. blythi but rounder. The paraconid is small, and no notch is present on the paracristid. The metaconid is much smaller and rounded than in D. blythi. The m2 is overall similar in morphology to D. blythi and similar in size, but the anterior cingulid is reduced. The m3 is closest in morphology to D. blythi and similar in size, but the anterior cingulid is reduced in comparison. The hypoconid is less rounded than in D. blythi. The m4 has a similar in morphology to D. blythi, but is narrower and shorter in length. The anterior cingulid is also reduced, being similar in size to that of Dasycercus hillieri. The talonid is reduced, similar in size to that of Dasycercus cristicauda.

Remarks

The common name ‘Little Mulgara’ has been suggested for this taxon, in reference to its small size compared with other mulgaras.

Discussion

Taxonomic diversity

The results of this investigation indicate that both Dasycercus cristicauda and Dasycercus blythi represent distinct taxa, with craniodental characters that can be used to identify subfossil specimens. This confirms the original assessment of Woolley (2005), and demonstrates new information for identification of specimens in subfossil assemblages. However, it was also determined that, as proposed by Adams et al. (2000), D. hillieri has enough distinct morphological characters to be recognized as a valid taxon. This is contra Woolley (2005), where Dasycercus hillieri was synonymized within D. cristicauda. Additionally, three new taxa are also identified in the taxonomic evaluation, those being Dasycercus woolleyae, Dasycercus archeri and Dasycercus marlowi.

The results of our PCA using craniodental measurements indicate a total of six distinct taxa within Dasycercus, with separation maintained and increased when allometry was removed (Figs 3–6). The LDA (Fig. 3C) discriminated between each grouping. The results of the cranial one-way PERMANOVA demonstrated that the groups were significantly different (p = 0.0001, F = 8.703; Table 2). This was likewise the case when the data are corrected for allometry (p = 0.0001, F = 858.3; Table 3). The dental PERMANOVAs concurred with the results of the cranial measurements (p = 0.0001, F = 6.649; Table 8), including when corrected for allometry (p = 0.0001, F = 6.824; Table 9). Some overlap was observed between taxa in the PCA due to possible sexual dimorphism (Figs 3, 5). This is seen in many marsupial species, such as bandicoots (Travouillon & Phillips 2018). However, PERMANOVAs performed on the males and females (Tables 4–7) were found to be significantly different between species (p = <0.05).

The total evidence analysis (Fig. 7) was unable to separate five of the six species beyond the genus level, forming a soft polytomy (BPP = 65%). This strongly indicates that all six species belong in the genus Dasycercus, but unfortunately these could not be separated further. The species D. hillieri was able to be separated from other Dasycercus (BPP = 87%), strongly indicating that it is a separate taxon from other Dasycercus. While the outcome of the total evidence analysis is unresolved for D. cristicauda, D. blythi, D. woolleyae sp. nov., D. archeri sp. nov. and D. marlowi sp. nov., this does indicate that further work is needed on the genus with a greater set of characters that can be used for morphological and phylogenetic work.

A number of characters were identified for each taxon, with a full list of different characters found in the diagnosis section of the systematics. Dasycercus archeri is characteristically different from D. cristicauda, despite their being sympatric taxa. Bilobed i3 are present in all Dasycercus, except for D. archeri; instead, the i3 is wide and triangular. Dasycercus archeri is the only species of Dasycercus to have an interparietal fused to the supraoccipital (unfused in D. cristicauda, and absent in other Dasycercus). Additionally, D. archeri lacks nasal lateral expansions. Specimens of D. woolleyae have superficial similarities with smaller individuals of D. hillieri. However, D. woolleyae demonstrates key differences in morphology; for example, there is no minute foramen present dorsal to the foramen rotundum. This minute foramen is present in both D. hillieri and D. cristicauda. It should be noted that in individuals of the didelphid genus Monodelphis (Burnett, 1830), the number and size of cranial foramina on the lateral surface of the cranium were shown to be variable (Wible 2003). However, multiple individuals were examined in this investigation, and it was not found to be polymorphic in any taxa. The inflation of the squamosal caused by the squamosal epitympanic sinus is ovoid as opposed to kidney-shaped. Externally, both crested tail and brush tailed morphologies have been observed in D. wolleyae resulting in some specimens previously being identified as D. blythi. However, the crested tail morphology is more common in D. woolleyae.

Dasycercus marlowi is the most morphologically distinct of the Dasycercus species. Most notably, the taxon lacks any cranial crests, including the postorbital process, which is present in all other Dasycercus taxa. D. marlowi also has larger auditory bulla than D. cristicauda, despite being smaller in size. The dentition of D. marlowi also display unique characters. The minute anterior and posterior cusps on C1 of D. marlowi are located below the crown of the P1, in contrast to all other Dasycercus taxa. However, it is possible that the specimen is a young adult with C1 not yet fully erupted (Jones 2023). The anterior cingulum of M1 does not connect to the talon via the preprotocrista. The postprotocrista ends on the lingual flank of the metacone. There is no anterior cingulum present on M2–3. Externally, the skin is identical to D. blythi, although the specimen is smaller in size. In this sense, D. marlowi could be mistaken for a juvenile; however, the dentition confirms that it is a full-grown adult.

Unfortunately, no craniodental morphological comparison could be conducted with the holotype of D. hillieri (NHMUK 1905.3.28.1) due to that specimen lacking a skull. However, it should be noted that, when compared to the holotype of D. cristicauda (AMS M11342), specimens identified in this investigation as D. hillieri differ in several aspects (Fig. 12). Several key characters were identified in this investigation that support the notion that D. hillieri is a separate taxon from D. cristicauda. For instance, the M1 of D. hillieri has a character unique to the taxon, with a postparacrista that meets the preparacrista to form a centrocrista. Other dental characters that separate D. hillieri from D. criticauda are the presence of stylar cusp B on the M1, which is absent in D. cristicauda, and the postprotocrista of the M1 ends on the posterior flank of the metacone.

Interestingly, all specimens examined of D. hillieri retained a P3, which was enlarged and triangular. When present in D. cristicauda (and other Dasycercus), this tooth is minute and rounded. This character was previously proposed by Woolley (2005), but later shown to be undiagnostic (Woolley et al. 2013); this could be due to the influence of multiple taxa in that sample. Additionally, D. hillieri features bullae than are larger and more inflated than D. cristicauda. There is a correlation in small mammals, such as rodents, between the size of the auditory bullae and their habitat; larger bullae are present in arid adapted species, and smaller bullae in more mesic ones (Yazdi, Colangelo, & Adriaens 2015). All specimens examined in this investigation are from arid regions, including the Lake Eyre basin, Simpson Desert, Strezlecki Desert, and Sturt Stony Desert. This xeric distribution and the larger bullae could suggest that D. hillieri is a more arid-adapted taxon than other Dasycercus.

Several of the craniodental characters for D. cristicauda in previous taxonomic matrices (Kealy & Beck 2017, Beck et al. 2022) demonstrate variability. Beck et al. (2022) noted that the i3 of D. cristicauda is bilobed in some individuals and not in others. Our results show that the variability of characters within D. cristicauda is due to the prior lumping of multiple taxa under that umbrella taxon. Inquiry into the origin of the genetic material used by Kealy & Beck (2017) to represent D. cristicauda suggests it is likely a specimen of D. hillieri. This was determined based on the specimen’s origin and the current lack of known modern D. cristicauda material, beyond the holotype specimen, which failed to sequence (Adams et al. 2000). Apart from the holotype specimen, all skins previously identified as D. cristicauda have been reassigned to D. hillieri based on the specimens’ craniodental morphology.

Initially it would be easy to dismiss the variation within the Dasycercus taxa as a cline, due to the gradual change in size present across the taxa. However, when geographic distribution is considered, it is evident that a cline is not present as the taxa do not experience a gradual change in size with distribution. For instance, D. hillieri is the largest Dasycercus and D. marlowi is the smallest, although both taxa are found in the Simpson Desert. Similarly, throughout the geographic range of the genus all taxa exhibit sympatry, with multiple species having been determined to occur in a region. For example, in the Great Victoria Desert the highest diversity of Dasycercus was recorded, with four taxa (D. blythi, D. woolleyae, D. archeri and D. cristicauda) occurring within the same assemblages. Morphological characters are, likewise, not related to size.

Discrepancy in size and some cranial characters are known to be affected by ontogenetic change in dasyurids. However, the variation observed in these analyses cannot only be due to ontogenetic change as no juvenile specimens were included in the analysis. All specimens examined were adults with fully erupted molar rows (M1–4). No significant difference in skull size was observed between individuals from the same locality. This implies that the skull does not grow continuously in Dasycercus. The degree of canine eruption (Jones 2023) was checked for specimens examined, to further investigate the potential impact of ontogeny on our results. Specimens of all taxa were found to have canine overeruption present (Table 1). While this does mean that some individuals that were assessed appeared to be older adults (even with relatively unworn teeth), it also shows that, because canine overeruption is present and absent in all individuals of all taxa, no one taxon is comprised of elderly individuals only. Note that D. marlowi was assessed as not having canine overeruption present; however, as it is only known from a single individual, the same conclusion cannot be made. Additionally, juvenile specimens of both D. archeri and D. hillieri were identified, with the former being well represented in the wet specimen collection of the Australian Museum (AMS M4923-M4926). A skin specimen of D. hillieri is also known from the WAM (WAM M1538). This suggests that the taxa do not represent mature forms of other Dasycercus. A PCA analysis was performed on the dental measurements of males (Fig. 6A, B) and females (Fig. 6C, D). These analyses indicated that, whilst there is some overlap between species, size is not the only variable separating species. This is supported by the fact that both cranial and dental measurements separated each taxon equally, and we found that all examined taxa have a unique combination of characters based on their dentitions.

Distal hair of the tail has previously been argued to be a key character when identifying the species of Dasycercus (Adams et al. 2000; Wooley, 2005). The dorsal tail hair of the holotype specimen of D. hillieri is crested, long, black, and faces dorsally (Fig. 11A), which matches those specimens identified as D. hillieri in this investigation. The holotype skin for D. cristicauda is in very poor condition (Woolley 2005, Parnaby et al. 2017), so the tail is difficult to assess. Morphology of the tail hair of D. blythi matches that described in Woolley (2005), being brushed, distally facing, and proximally fattened. The new taxon D. archeri has a tail hair morphology similar to that of D. hillieri, where the dorsal hairs are crested; however, the hair is short (Fig. 11E). This morphology is also seen in D. woolleyae, but this taxon is polymorphic and can also have a tail morphology similar to that of D. blythi (Fig. 11B). The only known specimen of D. marlowi has a tail hair morphology similar to that of D. blythi (Fig. 11D). As each taxon displays different tail morphology, this character can be used to differentiate between taxa. However, due to the character of ‘brushed’ and ‘crested’ tail hair being shared, it is possible to confuse multiple taxa and caution should be taken.

The morphology of the plantar pes is another character that has previously been used to differentiate taxa (Adams et al. 2000, Woolley 2005). In this investigation it was determined that the character is polymorphic within species, with specimens examined of all taxa displaying both described characters. As a result, the morphology of the plantar pes hair is an unreliable character to differentiate species, with some individuals even having polymorphic plantar pes hair from the left to right hind leg.

The number of nipples is a key character outlined by Woolley (2005) for differentiating between the species of Dasycercus, and was examined in this investigation and compared to previous observations. However, the number of nipples present is a hard character to assess, visible only on lactating female specimens with nipples that have been sucked (Woolley 2005). It was highlighted in Woolley (2005) that the type specimens of D. hillieri, D. cristicauda, and D. blythi could not be assessed for this character.

Specimens from this investigation (Table 1) and Woolley (2005; Tables 1, 2), were cross examined to compare specimens examined in both studies. Of these, specimens identified by Woolley (2005) from the Nullarbor, with eight nipples and a crest tail, were identified as D. archeri in this investigation. This is also the case with some D. woolleyae specimens, suggesting that the character is likely shared by D. archeri, D. woolleyae and D. hillieri. Interestingly, a crested tailed specimen (AMS M4863) that Woolley (2005) gave a nipple count of ‘-’ was, in this investigation, identified as having six nipples. This nipple count does not fit with other Dasycercus from the region, where other specimens were noted as having eight nipples (Woolley, 2005).

All taxa vary greatly in colour throughout their distribution, making it of little use as a taxonomic characteristic. For example, D. hillieri specimens collected by Woolley from Sandringham station are redder in colour than the holotype specimen, and near identical to the patterning and colour of D. woolleyae sp. nov. specimens collected on the Canning stock route by Lipfert. Additionally, the pelage colour can change with age, quality, and method of preservation in museum collections. The coloration of the pelage is also not a useful character when attempting to identify subfossil or osteological material.

The genetic report by Adams et al. (2000), indicated that based on DNA sequence data from existing populations, two taxa were present within living Dasycercus, determined to be D. hillieri and D. cristicauda. Although the present investigation reinstates D. hillieri as a valid taxa, at an individual level, there is conflict with the results of Adams et al. (2000). Specimens SAMA M3143 and SAMA M3154 were identified as D. hillieri by Adams et al. (2000), but as D. woolleyae in this investigation. Specimen SAMA M6064 was identified as D. woolleyae in this investigation, but as D. cristicauda by Adams et al. (2000). This indicates that the genetic work performed by Adams et al. (2000) does not support the monophyly of the species D. woolleyae.

However, we believe that future DNA work is required to allow genetic validation of these taxa. The molecular work in Adams et al. (2000) is based on mitochondrial DNA (mtDNA), specifically mitochondrial genes cytb and CR, which are frequently utilized in molecular taxonomic evaluations. However, mtDNA has been shown to be of limited utility when assessing closely related taxa (Rubinoff, Cameron & Will 2006). It has also been suggested that mtDNA should not be used exclusively when identifying species (Rubinoff & Holland 2005). As such, it is likely that the molecular work of Adams et al. (2000), though fundamental at a time when only one taxon was valid, did not completely capture the genetic diversity of Dasycercus.

We suggest that addition of nuclear DNA (nDNA) could bring clarity to the genus. This has been observed in the taxonomy of rock wallabies (Petrogale spp.; Gray, 1837), where mtDNA has been found to be inconsistent with nDNA, and as a result of the sequences’ individual inability to resolve the taxonomy, a combination of both was required to bring clarity to the diversity of the taxa (Potter et al. 2012). Similarly, single gene investigations have provided conflicting results to those that have combined methodologies in the taxonomy of bats (Solari, Sotero-Caio & Baker 2019), and mtDNA has had limitations when used to resolve issues with didelphid taxonomy (Voss 2022).

Of the new taxa, D. woolleyae is the best represented in terms of the quality of specimens, as multiple live-caught specimens were collected on the Canning Stock route by Otto Lipfert. These specimens were divided and shared between the WAM and SAMA, and both institutes have skins and skulls within their collections. Specimens of D. archeri are the most numerous, especially among subfossil material from the Nullarbor Plain. These specimens are subfossil in nature and fragmentary. Only a few live-caught specimens are known; thus, subfossil material was critical to the identification of this taxon. Of the proposed new taxa, D. marlowi is the most poorly understood, with only one specimen in existence. Due to the timing of this specimen’s collection, and the lack of any other identified specimens in Australian Museum collection, it is unlikely that additional extant specimens will be recovered.

Of the existing taxa, D. cristicauda is fairly numerous in the subfossil record across the Nullarbor Plain and Roe Plains. However, the only known skin is now the holotype specimen. D. blythi specimens remain largely unchanged as a result of the current revision, with only a few specimens reassigned to D. woolleyae and the type specimen of D. marlowi. D blythi is also fairly well represented in subfossil material from the Great Victoria Desert, with a single specimen present on the Northern Swan Coastal Plain. Specimens now assigned to D. hillieri are restricted to live caught animals. Subfossils from the areas where D. hillieri occur were identified as Dasyuroides or were too poorly preserved to identify.

Common names

Given the findings of this comprehensive taxonomic review, the common names associated with Dasycercus cristicauda and Dasycercus blythi, namely the Crest-tailed Mulgara and Brush-tailed Mulgara respectively, are no longer distinctive. Woolley (2005) assigned these names based on tail morphology. However, of the six taxa of Dasycercus, four have a crested tail: D. cristicauda, Dasycercus hillieri, Dasycercus woolleyae and Dasycercus archeri. The remaining two have a brushed tail, those being D. blythi and Dasycercus marlowi. Interestingly, it was also observed that in D. woolleyae the tail is usually crested, but very rarely exhibits proximal fattening and a brush tail similar to that observed in D. blythi. Both the ‘Crest-tailed Mulgara’ and ‘Brush-tailed Mulgara’ can be kept by their respective taxa, but it should be noted that this is no longer an appropriate way of differentiating the species.

Taxa within the genus are known by many names to Aboriginal peoples, including ‘Mudagoora’, ‘Muritcha’, ‘Ilyowra’, ‘Ampurta’, ‘Narloodi’, and ‘Tajinna’ (Finlayson 1961, Baker et al. 1997). Of these names, D. hillieri originally had the common name ‘Ampurta’. ‘Mulgara’, which has become vernacular term for the entire genus Dasycercus, was originally the name Aboriginal people of Ooldea used (Woolley, 2005). This means that D. archeri is most likely to be the taxon originally associated with this name. However, due to the repeated changes to the scientific taxonomy of the genus, and the cultural considerations around naming within Aboriginal communities (Baker et al. 1997), it is difficult to link previously recorded Indigenous names to the other four taxa.

We suggest the following provisional common names for the six taxa: D. hillieri to retain the existing name ‘Ampurta’ (Baker et al. 1997); D. blythi to be the ‘Spinifex Mulgara’ in reference to its preferred habitat, or retain the name ‘Brush-tailed Mulgara’; D. cristicauda to retain ‘Crest-tailed Mulgara’; D. woolleyae to be referred to as ‘Northern Mulgara’ or ‘Sand Mulgara’; D. archeri to be ‘Southern Mulgara’; D. marlowi to be the ‘Little Mulgara’. Future research in collaboration with Aboriginal communities is merited to identify the relationship between the newly refined taxa and their original names.

Distribution

Subfossil specimens of Dasycercus cristicauda have been identified on the Nullarbor Plain, Roe Plains, Northern Swan Coastal Plain, and the Great Victoria Desert (Fig. 1). However, subfossil material from these and other assemblages have not been previously identified beyond the genus level, due to a lack of characters (Woolley et al. 2013, Fusco et al. 2017). D. cristicauda is common throughout assemblages, especially on the Nullarbor and Roe plains. This indicates that the species occurred in open arid woodland and scrubland environments, but could also occur in coastal regions outside of the arid zone. This is further supported by the findings of Fusco et al. (2017), where subfossil material of Dasycercus was uncovered within 40 km of Lake Alexandrina, which is the disputed original type locality of D. cristicauda. However, it should be noted that the type locality of D. cristicauda was reassessed by Parnaby et al. (2017), and determined to instead be from near Port Augusta. This conclusion was based on the lack of collecting near Lake Alexandrina at the relevant time, and letters to Krefft indicating that collecting should be conducted. Fusco et al. (2017) suggested that their findings indicated that the original locality cannot be so easily dismissed. However, material from Fusco et al. (2017) could not be identified to species. We propose that it may be a combination of the two conclusions, with Dasycercus indeed occurring outside of the arid zone, but the type specimen having been collected near Port Augusta.

Modern and subfossil specimens of Dasycercus woolleyae have been identified from arid regions throughout central Australia. Notably, specimens collected by Otto Lipfert were documented in the collection notes as being collected in the dune crest systems of the Canning Stock Route, indicating that the species prefers sandy dune habitats (Woolley et al. 2013). Specimens of Dasycercus blythi also collected by Otto Lipfert on the Canning Stock Route expedition were noted as living on spinifex (Woolley et al. 2013). Live D. blythi specimens have been caught throughout central Australia from the Great Victoria Desert, Gibson Desert, Little Sandy Desert, Great Sandy Desert, Tanami Desert and Simpson Desert. Additional subfossil specimens have also been collected from the Northern Swan Coastal Plain and Great Victoria Desert.

Subfossil material of Dasycercus archeri is commonly found throughout the Nullarbor Plain, Roe Plains, Great Victoria Desert, Northern Swan Coastal Plain, and Dirk Hartog Island in Shark Bay. Most D. archeri specimens have been recovered from the Nullarbor Plain and Great Victoria Desert, indicating that it prefers arid habitats. Modern specimens of D. archeri were also collected at Ooldea on the Nullarbor Plain (Woolley et al. 2013). Interestingly, specimens from Madura cave on the Nullarbor, which were originally identified as Dasyuroides by Lundelius & Turnbull (1978), are assigned as Dasycercus archeri herein. As this is the only record of Dasyuroides from this region, it is possible that Dasyuroides never occurred on the Nullarbor, with the specimens previously misidentified. Live collected Dasycercus hillieri specimens are known from the Simpson Desert, Sturt Stony Desert, Strzelecki Desert and throughout the southern Lake Eyre Basin. D. marlowi is known only from a single modern specimen in the Simpson Desert, and without more specimens understanding of the full distribution cannot be attained. Based on the Dasycercus distribution results from this investigation (Fig. 17), and the past distribution of the taxa (fig. 2 in Woolley et al. 2013), it is likely, but not certain, that four species—D. archeri, D. woolleyae, D. marlowi, and the redefined D. cristicauda—are extinct.

Figure 14. The skin of the holotype specimen of D. woolleyae (WAM M1513), an adult male. Views are as follows: A, ventral; B, dorsal; C, lateral.

Figure 15. The skin of the holotype specimen of D. archeri (AMS M2987). Views are as follows: A, dorsal; B, ventral; C, lateral.

Figure 16. The skin of the holotype specimen of D. marlowi (AMS M8641). Views are as follows: A, dorsal; B, ventral; C, lateral.

Figure 17. Distribution of Dasycercus material examined in this investigation, specimens have been mapped using taxonomic named assigned in this investigation. Note: material localities have been approximated to protect sensitive sites.

The only modern specimens of D. archeri were live collected in 1920–1925 from Ooldea (Woolley et al. 2013), and are present in the Australian Museum’s collection. No Dasycercus specimens were recorded in Ooldea in 1987 by Boscacci et al. (1987), indicating that the species declined between the 1920s and the 1987 surveys (Boscacci et al. 1987, McKenzie, & Robinson 1987). This timeframe for extinction matches with other extinct taxa in the arid zone (Burbidge et al. 1988), likely being caused by feral cats and foxes. Of the examined material in this investigation, D. woolleyae was last live caught in 1962. D. marlowi is only known from a single specimen that was live caught in 1965. The status of this species is unknown until further specimens are identified. The only live caught specimen of D. cristicauda is the holotype, with the date of collection unclear. The current population of living Dasycercus in the Lake Eyre Basin that were previously recognized as D. cristicauda should now be reclassified as D. hillieri.

If it is the case that any or all of these taxa are extinct, it would mark the first recorded recent extinction in Dasyuridae. This suggests that the small carnivores have suffered a greater loss in diversity since European arrival than previously believed. More research is needed into other taxa within Dasyuridae to determine if such losses have occurred across the family.

Future work

A complete genetic investigation of the taxa presented here is planned and will improve taxonomic resolution within Dasycercus. DNA analysis of Dasycercus hillieri and Dasycercus blythi should be straightforward, as both taxa have published genetic data available (noting that the published genetic information assigned to Dasycercus cristicauda is that of the realigned D. hillieri). An analysis of Dasycercus archeri, Dasycercus woolleyae and Dasycercus marlowi will require the use of tissue samples from preserved skins in museum collections.

Extraction of DNA has previously been unsuccessfully attempted on the holotype specimen of D. cristicauda (AMS M11342), with the failure attributed to the mounted skin containing a ‘tar like substance’ (Adams et al. 2000). Improvements in methodology for DNA extraction could mean that the successful extraction and sequencing of DNA from the holotype of D. cristicauda is achievable. Alternatively, extraction of aDNA (ancient DNA) from subfossil material is also possible.

Field surveys to determine the status of the newly named taxa will also be of importance. Recently a Dasycercus specimen was seen on camera trap footage at the AWC (Australian Wildlife Conservatory) facility in Mt Gibson, Western Australia (Georgina Anderson pers. comm. to KJT, 2021). This is further west than previous observations of live Dasycercus. It is possible that there may be a remnant population of D. archeri or D. cristicauda present, as specimens of both taxa have been recorded as far west as the Northern Swan Coastal Plain. Of the known extant taxa, it is also possible that this is D. blythi, as this species did once occur in this region.

Conclusions

We reviewed the taxonomy of Dasycercus by assessing a suite of craniodental characteristics not previously used in this genus in 84 specimens, including modern skins and skulls and subfossil material. It is concluded that Dasycercus hillieri is a valid taxon distinct from Dasycercus cristicauda, and includes the extant modern population previously attributed to the latter. In addition to the three previously recognized taxa, Dasycercus blythi, D. hillieri and D. cristicauda, three new taxa are described. These are named Dasycercus woolleyae, Dasycercus archeri and Dasycercus marlowi. The inclusion of subfossil material greatly expanded the number of assessable specimens within the genus, and provided insights on the past distributions and histories of the taxa. Based on material examined and available surveys, these new taxa and D. cristicauda likely represent the first recorded recent dasyurid extinctions.

Zoobank UUID

http://zoobank.org/C91DD4B0-2C8D-4B89-8D6A-35F846F69AE8

Acknowledgements

We acknowledge the Traditional Owners of the Country where all specimens were collected, and particularly the Whadjuk people (Perth) and the Gadigal peoples (Sydney) as the Traditional Owners where this research was conducted. Specimens were contributed by the WAM, SAMA, AMS, MAGNT, NMV, QM, ANWC and NHMUK, and we acknowledge all museum curators, staff and volunteers. In particular, numerous specimens were collected by Michael Archer (UNSW) and Patricia Woolley (La Trobe University), which made this review possible. We thank Alexander Baynes (WAM) for all his collection work and access to material from the Great Victoria Desert. Thanks also go to Mark Adams (SAMA) for sharing the unpublished genetics report ‘Revision of Dasycercus Systematics’, Robin Beck (University of Salford) for assistance with the phylogenetic analysis, and the reviewers and Editor for their constructive feedback. J.N.-M. was supported by a RTP scholarship from Curtin University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03115518.2023.2262083.

Appendix

Anatomical abbreviations

Upper dentition
P1L	=	1st upper premolar length
P1W	=	1st upper premolar width
P2L	=	2nd upper premolar length
P2W	=	2nd upper premolar width
M1L	=	1st upper molar length
M1AW	=	1st upper molar anterior width
M1PW	=	1st upper molar posterior width
M2L	=	2nd upper molar length
M2AW	=	2nd upper molar anterior width
M2PW	=	2nd upper molar posterior width
M3L	=	3rd upper molar length
M3AW	=	3rd upper molar anterior width
M3PW	=	3rd upper molar posterior width
M4L	=	4th upper molar length
M4AW	=	4th upper molar anterior width

Lower dentition
p1L	=	1st lower premolar length
p1W	=	1st lower premolar width
p2L	=	2nd lower premolar length
p2W	=	2nd lower premolar width
m1L	=	1st lower molar length
m1AW	=	1st lower molar anterior width
m1PW	=	1st lower molar posterior width
m2L	=	2nd lower molar length
m2AW	=	2nd lower molar anterior width
m2PW	=	2nd lower molar posterior width
m3L	=	3rd lower molar length
m3AW	=	3rd lower molar anterior width
m3PW	=	3rd lower molar posterior width
m4L	=	4th lower molar length
m4AW	=	4th lower molar anterior width

Cranium
aIL	=	length of anterior upper incisor series (I1–4)
anw	=	combined anterior width of the nasal bones
apl	=	length of incisive (anterior palatal) foramen
apw	=	combined width across the paired anterior palatal foramina
bcl	=	condylobasal length
bh	=	bullae height
biw	=	bullae inner width
bl	=	bullae length
bol	=	length of basioccipital bone
bow	=	bullae outer width
bsl	=	length of basisphenoid bone
bw	=	bullae width
cctl	=	length of cheektooth series (canine + premolars + molars)
ctl	=	length of cheektooth series (premolars + molars)
cw	=	combined width of occipital condyles
fs	=	length of midline suture of frontal
iow	=	width of greatest constriction of orbitotemporal fossa
jh	=	height of jaw measured between m2 and m3
jl	=	length of jaw, from anterior-most point of dentary (excluding incisors) to condyle
lmr	=	combined length of lower molar series (m1–4), measured at crowns
lpr	=	combined length of lower premolar series (p1–2), measured at crowns
mw	=	maximum width across braincase
nl	=	maximum length of nasal bone
nps	=	length of sutural contact between nasal bone and premaxilla
onl	=	greatest skull length (occipitopremaxillary length)
oP3	=	palatal width, measured across P3
paw	=	palatal width across M3
pnw	=	combined posterior width of the nasal bones, measured at intersection with width across outside of paroccipital
ppl	=	length of maxillopalatine fenestra
ppw	=	maximum width across postorbital ridge
rwc	=	anterior rostral width, measured across outer surface of each canine
rwi	=	posterior rostral width, measured across infraorbital
soh	=	supraoccipital height
uML	=	combined length of M1–3, measured at crowns
uMR	=	combined length of M1–4, measured at crowns
zw	=	maximum width of cranium, measured across zygomatic arches