Mammal craniodental remains in owl pellets provide new tooth row measurements for the threatened Julia Creek dunnart (Sminthopsis douglasi)

Introduction

When Archer conducted his landmark morphological revision of the dasyurid genus Sminthopsis in 1981, he was able to include only three specimens of the Julia Creek dunnart (Sminthopsis douglasi), which had been described just 2 years earlier by the same author (Archer 1979, 1981). Thereafter, the species was presumed extinct until new records were found in Queensland in the early 1990s (Woolley 1992). Today, Sminthopsis douglasi is listed as Vulnerable under the Environment Protection and Biodiversity Conservation Act (EPBC Act 1999). The species has proven elusive, but it has been sporadically detected using live (Elliott) trapping (Mifsud 1999; Bakker et al. 2024a), camera traps (Bakker et al. 2024b) and owl pellets (Woolley 2010; Lockhart 2024; Charley et al. 2025).

Woolley (2010) examined eastern barn owl (Tyto javanica delicatula) pellets collected in north-western Queensland between 1994 and 2001, and S. douglasi was identified from 22 (of 28) locations, usually with fewer than 10 individuals per site. More recently, Charley et al. (2025) assessed owl pellets collected from 2023 to 2024 at a single location where S. douglasi occurs, namely, Toorak station in north-western Queensland. Sminthopsis douglasi comprised up to 31% of the vertebrate material found in the pellets (including >220 individuals). However, neither study published measurements of S. douglasi tooth rows. This means that the main reference source to aid S. douglasi identification is still based on the comparative measurements of only three individuals from Archer’s earlier work (Archer 1981). Such small sample sizes could over- or under-estimate the mean, variance and range of species diagnostics because of the higher weighting attributed to each sample and the increased likelihood that (by chance) one or more samples may not be representative of a ‘typical’ member of the population.

Here, we aimed to fill this knowledge gap by making a series of (upper and lower tooth row) measurements for S. douglasi craniodental remains sourced from owl pellets used by Charley et al. (2025), other unpublished material, and registered specimens held at the Queensland Museum. We then compared our measurements with those previously taken of S. douglasi (and co-occurring and geographically proximate congeners, including S. macroura, S. crassicaudata, S. murina and S. virginiae) by Archer (1981) and synthesised the data to provide a more comprehensive reference source for identifying Sminthopsis from Queensland.

Materials and methods

Sminthopsis douglasi specimens

Owl pellets were collected from four sites in north-western Queensland as part of parallel monitoring (Charley et al. 2025) and distribution (Lockhart 2024) studies of S. douglasi (Table 1, Fig. 1). The craniodental material of S. douglasi retrieved from these pellets was retained and forms the bulk of specimens used in the present study. For full details of owl pellet dissection and processing, see Charley et al. (2025).

Table 1.S. douglasi specimens and locations.

Location (site symbol; N individuals)	Latitude	Longitude	Collector	Date collected	Source
Toorak Woolshed, 40 km south of Julia Creek. Collection 1 (T; 48)	21°01′58″S	141°47′59″E	Gray, Mifsud, Baker	13 Apr. 2023	Owl pellets QUT
Toorak Woolshed, 40 km south of Julia Creek. Collection 2 (T; 8)	21°01′58″S	141°47′59″E	Gray	18 July 2023	Owl pellets QUT
Toorak Woolshed, 40 km south of Julia Creek. Collection 3 (T; 7)	21°01′58″S	141°47′59″E	Gray	2 Oct. 2023	Owl pellets QUT
Toorak Woolshed, 40 km south of Julia Creek. Collection 4 (T; 9)	21°01′58″S	141°47′59″E	Gray	21 Apr. 2024	Owl pellets QUT
Stamford Racecourse, Stamford Collection 1 (S; 2)	21°14′56″S	143°48′53″E	Gray, Mifsud, Baker	12 Apr. 2023	Owl pellets QUT
Stamford Racecourse, Stamford Collection 2 (S; 1)	21°14′56″S	143°48′53″E	Gray	6 June 2023	Owl pellets QUT
Woodsberry, via Stamford (W; 1)	21°34′38″S	143°53′03″E	Gray	3 Oct. 2023	Owl pellets QUT
Nelia, 48 km E of Julia Creek (N; 1)	20°39′16″S	142°12′18″E	Gray, Mifsud, Baker	14 Apr. 2023	Owl pellets QUT
Garomna, Julia Creek (G; 1)	20°42′S	141°53′E	M Browne	24 May 1933	J5459 Male Paratype QM
Eureka, via Julia Creek (E1; 1)	20°50′S	141°48′E	P Woolley	24 Apr. 1992	JM9805 Male QM
Bladensburg National Park, via Winton (B; 1)	22°31′58″S	143°02′46″E	R Hobson	2 Oct. 2001	JM14810 Male QM
Ashton Station (A; 1)	21°09′59″S	144°44′59″E	A Kutt	25 Sep. 2000	JM14336 QM
Edith Downs Station, south-west of Julia Creek (E3; 1)	20°56′S	142°12′E	P Woolley	17 Apr. 1992	JM18108 Male QM
Lyrian Station, via Cloncurry (L1; 3)	19°21′S	141°11′E	P Woolley	Apr. 1991	JM18139 Owl Pellet QM
Lyrian Station, via Cloncurry (L2; 1)	19°28′S	141°12′E	P Woolley	27 Apr. 1992	JM18112 Male QM
Lyrian Station, via Cloncurry (L2; 1)	19°28′S	141°12′E	P Woolley	28 Apr. 1992	JM18114 Male QM
Near Edith Downs, via Julia Creek (E2; 1)	20°53′S	142°12′E	C Tritton	Oct. 1991	JM18122 Male QM
Proa, via Nelia (P; 1)	20°54′S	142°08′E	G Mifsud	3 June 1998	JM18601 Male QM
Proa, via Nelia (P; 1)	20°54′S	142°08′E	D Fysh	Jan. 1999	JM18602 QM
Proa, via Nelia (P; 1)	20°54′S	142°08′E	G Mifsud	15 Nov. 1997	JM18603 Male QM
North-east of Cloncurry (C; 1)	20°26′54″S	140°46′42″E	N Godwin	28 Aug. 2012	JM19866 QM

Four collections were made from Toorak and two collections from Stamford Racecourse. The majority of specimens were taken from Toorak Collection 1, which represented a broader age range of individuals. Representative samples were taken from each of the subsequent three Toorak collections, each of which represented a maximum age of 6 months. For the three other owl pellet sites, there was a limited amount of material available, yielding representative samples at each location. Site symbols in column 1 match those in Fig. 1. QM indicates Queensland Museum (registration number is provided). Note: J5173, the holotype, could not be used in the analysis because it is a juvenile female.

Fig. 1.

Map of Queensland, Australia (on right), and the Mitchell Grass Downs Bioregion (on left), with Julia Creek dunnart (Sminthopsis douglasi) collection locations highlighted in yellow. Figure on the left is labelled as follows: L1, Lyrian Station; L2, Lyrian Station; C, north-east of Cloncurry; G, Goromna; N, Nelia; E1, Eureka; P, Proa; E2, near Edith Downs Station; T, Toorak; E3, Edith Downs Station; B, Bladensburg National Park; S, Stamford Racecourse; W, Woodsberry; and A, Ashton Station. Site symbols match those in column 1 of Table 1. Red and orange regions display the Government habitat distribution model for S. douglasi (Department of Climate Change, Energy, the Environment and Water 2024).

Comparative measurements from all craniodental material of S. douglasi registered in the Queensland Museum were also included in this study, to improve the geographic coverage of specimens (Table 1, Fig. 1).

Measurements

Tooth row measurements were taken from both the left and right side of the craniodental material. Multiple t-tests comparing left and right tooth row measurements for S. douglasi and corrected using the Holm–Bonferroni method (Holm 1979) indicated that there was no significant difference between the sides for any measurement (see Appendix 1). However, only one side of each specimen was used in the analysis, as per Archer (1981); the left side was nominally chosen, because this side had more complete samples.

Five craniodental measurements (three skull and two mandible) were made to the nearest 0.1 mm with Mititoyo digimatic calipers.

Superscripts designate tooth number in the skull (e.g. M³), subscripts define the tooth number in the dentary (e.g. M₃), and neither subscript nor superscript indicates a reference to both upper and lower teeth (e.g. M3), as per Archer (1981).

Skull: C¹–M⁴; M¹–M⁴; M¹–M³ (see Fig. 2). These matched Archer (1981): e, f, g respectively.

• C¹–M⁴: length of maxillary tooth row: from anterior edge of C¹ root to posterior edge of M⁴ root.

• M¹–M⁴: length of upper molar row: from anterior side of M¹ paracone root to posterior side of M⁴ posterior root.

• M¹–M³: length of first three upper molars: from anterior side of M¹ paracone root to posterior side of M³ metacone root.

Fig. 2.

Photograph of a S. douglasi skull from a Toorak owl pellet, showing tooth numbers and the position that the C¹–M⁴; M¹–M⁴; M¹–M³ measurements (described in text) should be taken.

Mandible: M₁–M₄; M₁–M₃ (see Fig. 3). These matched Archer (1981): m, n respectively.

• M₁–M₄: length of lower molar row: from anterior face of the anterior root of M₁ to the posterior edge of M₄ talonid root.

• M₁–M₃: length of the first three lower molars: from anterior face of the anterior root of M₁ to the posterior edge of the M₃ posterior root.

Fig. 3.

Photograph of a S. douglasi mandible from a Toorak owl pellet, showing tooth numbers and the position that the M₁–M₄; M₁–M₃ measurements (described in text) should be taken.

Results and discussion

We aimed to assess variation in craniodental (tooth-row) measures of a threatened dasyurid, the Julia Creek dunnart. We measured up to 85 S. douglasi individuals across 14 locations (Table 2, Fig. 1). As expected, the increased number of individual S. douglasi assessed meant that compared with Archer (1981), both minimum and maximum ranges of all tooth-row measures were extended (Table 2).

However, despite the increased sample size, S. douglasi was still largest in size, and there was no overlap in ranges in any of the comparative measurements made from three co-occurring or geographically proximate congeners (S. macroura, S. crassicaudata and S. murina) (Fig. 4). Of measured congeners, in terms of size, S. virginiae most closely approached S. douglasi, with the former being only slightly smaller in several measures and marginally overlapping in one (C¹–M⁴) (Fig. 4).

Fig. 4.

The absolute measurement range (mm) for each measured craniodental feature split across assessed species. Coloured bars represent the range of each species’ measurement, and light grey bars represent the mean ± standard deviation for the respective species. Black dots represent the mean of each measurement.

In short, a dunnart specimen from Queensland is likely to be an S. douglasi if in the skull: M¹–M³ ≥ 5.6 mm and C¹–M⁴ is ≥ 11.8 mm; and in the dentary: M₁–M₄ ≥ 7.3 mm. If the specimen is a juvenile, the M4 may not be fully emergent and is typically inflected in relation to the rest of the dental row. In this case, M1–M3 measures should be used, because they are least likely to vary between age classes, and even in juveniles, the molars of S. douglasi are distinctively large.

Thus, the clear craniodental size differentiation of S. douglasi will aid confident identification of the species if present in owl pellet material collected in the grasslands of central and north-western Queensland, even if the skull and dentary are damaged, incomplete or the teeth are loose. See Appendix 2 for a more detailed description of the condition of the craniodental material sourced from owl pellets and some suggested preservation methods.

Variation in S. virginiae

Although we were able to measure only two further S. virginiae skulls, one of them (JM18722) fell outside the reported range for the species in Archer (1981) for one measurement, M¹–M⁴ – 6.4 mm for JM18722, versus the maximum in upper range of 6.0 mm reported in Archer (1981). Despite the increase in maximum range, this was still less than the lower range for the comparative measurement for S. douglasi of 6.6 mm (Table 2).

Implications

This research has added significantly to the limited craniodental measurements available for S. douglasi. It should provide a useful resource for future studies assessing craniodental material to help inform the distributional range of this threatened species. The easiest and most accurate measurements proved to be C¹–M⁴ and M¹–M³/M⁴, and at least one of these (usually the latter) is likely to be possible for dunnarts found in eastern barn owl pellets. Researchers should refer to Archer (1981) when further details of comparative craniodental characters among dunnart species are required.

Supplementary material

Supplementary material is available online.