Phascogale pirata Thomas, 1904

Have, Ten, 2015, Taxonomic revision of Phascogale tapoatafa (Meyer, 1793) (Dasyuridae; Marsupialia), including descriptions of two new subspecies and confirmation of P. pirata Thomas, 1904 as a ‘ Top End’ endemic, Zootaxa 4055 (1), pp. 1-73 : 43-48

Phascogale pirata Thomas, 1904

Holotype: BMNH 4.1.3.100. South Alligator River. Collected by J. T. Tunney, 24th May 1903. Puppet skin and skull .

Revised diagnosis. A relatively small-bodied Phascogale that differs from all races of P. tapoatafa in the following ways: externally, in having a distinctly brownish wash to fur of the dorsum and flanks, pure white fur on the upper surface of the pes, a more elongate outer metatarsal pad, a similarly elongate but narrower pinna, a pure white cheek patch, and a lack of differentiation in fur colour between the upper and lower surface of the tail ‘stem’; dentally, in having a less enlarged and procumbent I1, a less pronounced size gradient along I2-4, a more compact upper premolar series with overlap between each tooth, a relatively more reduced P1, a more elongate preparacrista on M1, smaller protocones and metacones on M4, smaller entoconids and less prominent buccal cingulids on M1-3, a more anteriorly positioned paraconid on M1, a more reduced taloned on M4, a less enlarged I1 and a less reduced I3; and cranially, in being generally narrower across the braincase and zygomatic arches and in having an anteriorly less extensive lacrimal, a less prominent lacrimal preorbital crest, typically paired lacrimal foramina, a less prominent maxillonasolabialis fossa on the zygomatic arch, less forwardly rotated orbits, a shorter but deeper alisphenoid auditory bulla, and a posteriorly less inflated rostral tympanic process of the petrosal. Other proportional differences include more pronounced post-orbital constriction, a narrower foramen magnum and relatively smaller molars.

It further differs from all southern and eastern races of P. tapoatafa in having shorter, crisper dorsal fur and extensive all-white ventral fur.

It further differs from P. t. kimberleyensis in having the mid-dorsal fur darker than the flanks, at least some white-tipped guard hairs, whiter ventral fur, a more elongate and less heavily furred pinna, grey fur on the vibrissal pads, a more heavily pigmented scrotum, less elongate claws on the manus and a relatively longer tail ( Table 4 ).

It differs from P. cal ura in numerous features including much larger adult body size, essentially grey rather than brown dorsal colouration, a more elongate tail with more prominent brush, pure white fur on the upper surface of the pes, and absence of a contrasting postauricular patch.

Genetically, P. p i r a t a differs from the various races of P. tapoatafa at 7.3–16.3 % sequence divergence for cyt b (10.8–16.3 % for P. t. tapoatafa ; 7.3–10.5 % for P. t. wambenger ; 11.7–14.4 % for P. t. kimberleyensis ); and it differs from P. c al ur a at 10.3–11.4 % for the same gene.

Description. External measurements taken from museum specimen records are summarised in Table 4 . The dorsal fur is relatively short and crisp, the contour hairs measuring 6–7 mm in the centre of the back. The overall colour of the back and flanks is dull grey but with a slight brownish tinge, overlain by a fine peppering of white and black. This effect is produced by a combination of cream tipping on grey-based underfur, white tipping on grey based contour hairs, and an abundance of weakly emergent guard hairs that vary in colour from dark russet on the upper back and flanks to black on the rump. A small proportion of the guard hairs end in short white tips. The middorsal zone is distinctly darker than the flanks, especially on the upper back and extending forward onto the crown.

The ventral fur is pure pale cream to the bases ranges along a central zone extending from the chin to the cloaca. On the central part of the body this merges laterally into a zone of fur that is pale grey-based but with progressively shorter cream tipping, giving an overall off-white appearance. The inner surface of the fore-limbs is clothed in short cream fur, while that on the hind-limbs is lusher and pale grey-based with cream tipping. The scrotum of adult males (e.g., CM7225, CM7729) is heavily pigmented and densely clothed in cream fur. The pouch of adult females (e.g., CM8808, CM11724) is thinly furred with pale russet hairs and contains six elongate teats. A pigmented gular gland is present in both sexes but is much larger and more conspicuous on adult male specimens where it measures 15–20 mm in length and 7–8 mm in width. The location of this gland is not marked by any obvious external discolouration of the fur in either sex.

The head is boldly patterned. The coronal patch is dark grey and continuous with an indistinct to moderately well-defined mid-rostral stripe. This is flanked on each side by a band of pale grey fur flecked with black that extends onto the vibrissal pads. A distinct, dark grey loreal stripe runs from behind the mystacial vibrissae to the anterior corner of the eye. The eye itself is narrowly rimmed with lines of intense black hairs that merge posteriorly into a well-marked black to russet-brown postorbital spot. Distinct, all white orbital crescents are also present. The pale cream fur of the chin and throat extends around the corner of the mouth to merge with a supralabial bar and associated cheek patch rising to the level of the genal vibrissae. A narrow band of off-white fur, flecked with russet, separates the pale cream cheek patch from the white infraorbital crescent. The vibrissae are moderately thick and black, occasionally with short gingery tipping, and the vibrissal papillae are pigmented and visible through the fur. The pinna is elongate but moderately narrow, with a weakly developed posterior lobe. The external surface is very thinly furred, while the inner surface supports patches of white hairs anteriorly and posteriorly. The anterior margin of the pinna is emarginated by a narrow line of short black hairs.

The basal portion of the tail is thickly furred above and on the sides, continuing the colour and texture of the rump, while the ventral surface is more sparsely clothed in long, adpressed pale cream hairs ( Fig. 6 ). The tail narrows rapidly to form an elongate ‘stem’ (to 85 mm) that is clothed above and below by elongate black-based but cream-tipped hairs, giving an overall grizzled appearance. The tail ‘brush’ is relatively unexpanded and measures up to 170 mm; it consists of jet black hairs except for a short section of dark russet hairs ventrally, at the base of the brush.

The upper surface of the pes is densely clothed in white hairs that extend around the heel and forward onto the digits. The plantar surface of the pes is granular and supports a total of five conspicuously striated pads. The inner metatarsal and first interdigital pads are fused with no obvious line of demarcation; their combined length is 11.7 mm. The outer metatarsal pad measures 7.5 mm and is separated by a gap of no more than 2 mm from the fourth interdigital pad. The second to fourth interdigital pads are elongate; the second is approximately 25 % longer than the third or fourth.

The cranium of the holotype (BM4.1.3.100, an adult male) is illustrated in Figure 24 alongside an adult male from the vicinity of el Sherana, South Alligator River area (CM7225). Significant features were noted in an earlier section. Cranio-dental measurements for each sex are provided in Table 5.

The small sample of P. p i r a t a shows clear male-biased sexual dimorphism in cranio-dental measurements ( Tables 5 & 6). For all dimensions bar one (FMW), mean values are higher in males than females and for seven dimensions (SL, OBW, C-M4, ARW, PRW, LCL, LCW) the contrast is statistically significant. Several other contrasts approach statistical significance, including key measures of overall cranial size (e.g. ZW). Sexual dimorphism in canine size is particularly striking in this species.

Individual and geographic variation. The venter in some individuals (e.g., CM7031, CM8808) is entirely pale cream without the flanking grey-based zone.

An adult male from Yirrkali (AM7269), preserved as a flat skin, has an unusually narrow and short tail brush. We suspect that the animal was in poor health or perhaps recovering from being singed, an interpretation prompted by the observation that some parts of the tail support longer hairs that project through the brush, and the fact that fur on the lower back is very sparse. The skull of this specimen is shattered.

Two specimens from Melville Island (CM9983, CM9989; both lacking skulls) differ from the mainland sample in having a browner tinge to the dorsal fur, considerably shorter tails with less densely furred brushes, and shorter pinnae. The ventral fur in the adult male (CM9989) is shorter and less dense than in any mainland individual; the juvenile (CM9983) has more luxuriant ventral fur but with the grey-based zone extending almost to the midline on the abdomen. This population may warrant subspecific distinction from typical pirata on the adjacent mainland; however, this action is deferred pending more detailed assessment of this population.

Taxonomic remarks. Thomas’ (1904) description, based on two individuals of each sex, drew attention to all of the key diagnostic differences between pirata and tapoatafa . Nevertheless, he concluded that pirata ‘is in general appearance exceedingly like the common Brush-tailed Phascogale of temperate Australia’.

Spencer et al. (2001) reported the genetic distinction of P. p i r a t a from eastern and southwestern populations of P. tapoatafa but gave overinflated values due to sequence misalignment. The previous morphometric comparisons of ‘ pirata’ by Rhind et al. (2001) were compromised by the a priori pooling of the Top End, Kimberley and Cape York Phascogale populations on the assumption that these represented a single taxon.

Recognition of P. pirata as a full species within the Phascogale tapoatafa group is justified by the numerous points of morphological distinction that distinguish P. p i r a t a from all other phascogales, including other northern populations formerly included within the composite subspecies ‘ P. t. pirata ’, and by the high level of mitochondrial sequence divergence between this taxon and all populations of P. tapoatafa (individual contrasts range 9.4–12.3 % sequence divergence for for cyt b). The differences in dental morphology between P. pi r a t a and P. tapoatafa are suggestive of significant ecological contrasts that remain to be elucidated by field observations. The patchwork geographic distribution of the two species also highlights their separate evolutionary histories since the range of P. pi r at a is effectively embedded within the more widely distributed and regionally differentiated populations of P. tapoatafa .

Geographic distribution. The historical mainland range of P. pirata probably extended across much of the ‘Top End’, with records extending from Litchfield in the Daly River catchment to Yirrkala on Gove Peninsula ( Fig. 22 ). Most records come from the western part of this range, probably reflecting the historical focus of exploration and collecting. The only confirmed island population of the species is from Melville Island. A record from 1988 of a Phascogale on West Island in the Sir Edward Pellew Group is unvouchered; while this was referred to P. pirata by Woinarski et al. (2011), as noted above, it might also represent an outlier of the north Queensland population of P. tapoatafa .

Phascogale pirata has declined across most of its former range and now appears to be everywhere scarce. The most recent sightings come from Kakadu and Litchfield National Parks and from near Batchelor. A recent survey of the Sir Edward Pellew Island Group failed to detect populations of phascogales on any island ( Woinarski et al. 2011).

The geographic relationship between populations of P. pirata and P. tapoatafa in the east is unclear owing to uncertainty over the specific identity of the populations on the Sir Edward Pellew islands and at Edward River on the western side of Cape York. However, some kind of range disjunction is more or less guaranteed by the presence of extensive tracts of essentially treeless habitats that occupy the country at the head of the Gulf of Carpenteria, the so-called ‘Carpenteria Gap’ ( MacDonald 1969; Ford 1987; Schodde & Mason 1999). On the western side, a narrower gap of around 300 km separates records of P. p i r a t a from the nearest known P. t. kimberleyensis . This gap corresponds to a stretch of essentially treeless black soil country in the Victoria and Keep River basins, which most likely represents a real barrier to dispersal by Brush-tailed Phascogales.

The occurrence of P. tapoatafa populations both to the east and to the west of P. p i r a t a raises the possibility that populations of P. tapoatafa may still occur in the Northern Territory. The specific identity of all populations in this region should be reassessed in view of the newly identified specific distinction between P. p i r a t a and P. tapoatafa .

The label attached to the holotype of P. p i r a t a contains the annotation “caught in hollow tree, Aboriginal name ‘Woomboo’”.

Figures

FIGURE 6. Ventral surface of the tail in representatives of four populations of the Phascogale tapoatafa group; a) nominotypical P. tapoatafa (CM 20990); b) P. t. pirata (CM 7225); c) P. taoatafa southwestern W. A. (WAM 5006, from Balingup Brook); d) P. tapoatafa Kimblerley, W. A. (WAM 16028, from Pago Mission).

FIGURE 22. Map of northern Australia, showing the distributions of P. tapoatafa kimberleyensis subsp. nov (circles) and P. pirata (triangles). The square symbol represents the location of a sight record that might pertain to either P. pirata or P. tapoatafa.

FIGURE 24. Crania and dentaries of Phascogale pirata: a) CM 7225, b) holotype, BMNH 4.1. 3.100, an adult male from South Alligator River, Northern Territory, and c) CM 7617.