comparative postnatal ontogeny of the skull in the ... · comparative postnatal ontogeny of the...

Comparative Postnatal Ontogeny of the Skull in theAustralidelphian Metatherian Dasyurus albopunctatus(Marsupialia: Dasyuromorpha: Dasyuridae)David A. Flores,1,2* Norberto Giannini,1,2 and Fernando Abdala3

1Division of Vertebrate Zoology, Department of Mammalogy, American Museum of Natural History,New York, New York 100242Programa de Investigaciones de Biodiversidad Argentina, Facultad de Ciencias Naturales e Instituto Miguel Lillo,CP 4000, Tucuman, Argentina3Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, Johannesburg, South Africa

ABSTRACT We describe the cranial ontogeny of an aus-tralidelphian marsupial, Dasyurus albopunctatus, using acombination of qualitative and quantitative approaches.We examined in detail qualitative morphological changesof just-weaned individuals as compared to old adults; spe-cifically, changes in 31 morphological structures (e.g., pro-cesses, foramina) and 38 changes in cranial joints. We alsointerpreted growth-invariant structures in terms of theirfunctional relevance. We performed a multivariate allom-etry analysis based on 14 cranial measurements takenfrom 31 specimens encompassing the entire postweaningperiod. Three variables (height of occipital plate, breadthof braincase, and height of mandible) showed the sameallometric trends in D. albopunctatus and the three mar-supial species studied previously in the same framework(Didelphis albiventris, Lutreolina crassicaudata, andDromiciops gliroides). In addition, D. albopunctatusshared allometric trends in two variables (length of theupper postcanine row and length of the orbit) with themicrobiotheriid D. gliroides. Most of the growth trendsobserved are interpreted as linked to the predominantlycarnivorous dietary habit of adult D. albopunctatus. Be-cause dasyuromorphians are most likely basal to the ma-jor Australasian radiation of marsupials, knowledge ofontogenetic changes in D. albopunctatus may shed lighton the evolution of ontogeny in the highly diverse Aus-tralasian marsupial fauna. J. Morphol. 267:426–440,2006. © 2006 Wiley-Liss, Inc.

KEY WORDS: Dasyurus; marsupials; skull; ontogeny; al-lometry

The postweaning cranial ontogeny of two closelyrelated didelphid species, Didelphis albiventris andLutreolina crassicaudata, and the microbiotheriid,Dromiciops gliroides, has recently been studied indetail (Abdala et al., 2001; Flores et al., 2003; Gian-nini et al., 2004). Based on comparisons of the twodidelphid species, we hypothesized the existence ofgeneral cranial growth trends in the family (Floreset al., 2003). These trends indicated that neurocra-nial and sphlanchnocranial components interact in acomplex way, producing morphological arrange-ments with important relationships to trophic func-

tions, particularly the transition from suckling tomastication of solid food (Flores et al., 2003). Ingeneral, neurocranial components grew at a slowerpace than the skull as a whole, while splanchnocra-nial components exhibited a range of growth ratesfrom strongly negative allometry to strongly positiveallometry. A subsequent study of the South Ameri-can australidelphian Dromiciops confirmed most ofthe allometric trends seen in didelphids, but alsoshowed some surprising differences, such as theisometry of the orbit, a typically negatively allomet-ric structure (Giannini et al., 2004). Six variables(out of 14 analyzed) that define much of the skullproportions showed the same allometric trends inthe three species studied, regardless of the adult sizedifferences involved. Therefore, a common pattern ofgrowth in marsupials seemed to emerge. However,the existence of this putatively plesiomorphic pat-tern has not been tested in Australasian forms,which in fact constitute the largest fraction of extantmarsupial diversity.

Previous ontogenetic studies of Australasian mar-supials have focused on developmental descriptionand age estimation (e.g., Fleay, 1935; Hill and Hill,1955; Shield and Woolley, 1961; Sharman et al.,1964; Nelson and Smith, 1971; Inns, 1982; Merchantet al., 1984). To our knowledge, only one study

Contract grant sponsors: American Museum of Natural History(Kalbfleisch Postdoctoral Research Fellowship to D.A.F.) (ColemanPostdoctoral Research Fellowship); LIMBO grants, Argentina (toN.P.G.); National Research Foundation, South Africa (PostdoctoralResearch Fellowships to F.A.); PAST (Palaeontological ScientificTrust, Johannesburg) (to F.A.).

*Correspondence to: David Flores, Division of Vertebrate Zoology,Department of Mammalogy, American Museum of Natural History,Central Park West at 79th St., New York, NY 10024.E-mail: [email protected]

Published online 18 January 2006 inWiley InterScience (www.interscience.wiley.com)DOI: 10.1002/jmor.10420

JOURNAL OF MORPHOLOGY 267:426–440 (2006)

© 2006 WILEY-LISS, INC.

(Moeller, 1973) approached allometry in australidel-phians, but an unfortunate choice of measurementsmakes these results noncomparable with other stud-ies. With 61 currently recognized species (Groves,1993), Dasyuridae is a highly diverse clade thatrepresents the principal insectivore-carnivore radi-ation in the recent Australasian mammal fauna(Wroe, 1997). Dasyurids are relatively conservativein cranial shape (e.g., Wroe, 1997, and citationstherein), and are currently considered as eithersister-group to the remaining Australasian marsu-pials (Horovitz and Sanchez Villagra, 2003), or sis-ter to the highly diverse diprotodontian clade (Asheret al., 2004; Kavanagh et al., 2004). As part of anongoing effort to document comparative cranial on-togeny in marsupials, we analyzed the postweaningcranial development of Dasyurus albopunctatus inthe same methodological framework previously usedfor didelphids and microbiotheriids (Abdala et al.,2001; Flores et al., 2003; Giannini et al., 2004). Thepostweaning cranial ontogeny of a dasyurid mayprovide an important comparative context for theunderstanding of cranial ontogeny in derived aus-tralidelphians. Comparisons of the cranial develop-ment found in D. albopunctatus with that of didel-phids and microbiotheriids allow us to test andexpand previous hypotheses of cranial ontogeny inmarsupials, as well as to refine prior generalizationsin the light of new data.

MATERIALS AND METHODSStudy Specimens

For this study, we chose the New Guinean quollDasyurus albopunctatus on the basis of the avail-ability of specimens of widely different ages. Dasyu-rus albopunctatus is the smaller of the two NewGuinean species of quoll; its natural prey includessmall mammals, birds, and reptiles, as well as in-sects (Flannery, 1990; Menzies, 1991). There is littleinformation about the reproductive biology of NewGuinean quolls, although their Australian conge-ners have been well studied in this respect (e.g.,Fleay, 1940; Hill and Hill, 1955; Collins, 1973; Set-tle, 1978; Godsell, 1982, 1983; Edgar, 1983; Mer-chant et al., 1984; Bryant, 1988; Burnett, 2000).According to Menzies (1991), D. albopunctatus gavebirth to 1–6 neonates that spent 2 months in thepouch and another 2 months in the den before dis-persing; the lifespan is about 2 years (an indicationof high mortality).

We analyzed an ontogenetic series of 31 specimensof Dasyurus albopunctatus in the mammal collectionof the American Museum of Natural History(AMNH). The specimens examined were: AMNH104044-5, 109407, 109410, 109416, 109418-20,109424, 151971-6, 157081-4, 190924-9, 194704-7,194709, and 221650. In our sample, nine specimensdo not have a fully adult dentition, whereas theremaining individuals were adults of different sizes.

The smallest specimen (AMNH 190927; condylo-basal length 35.6 mm) has the lower (m3) and upperthird molar (M3) erupting, and the lower (dp3) andupper deciduous premolar (dP3) present. The larg-est specimen (AMNH 151976, male) has a condylo-basal length of 66.2 mm. Although no temporal on-togenetic data are available for D. albopunctatus,some inference about the age of the specimens ispossible based on data from D. viverrinus (Hill andHill, 1955; Merchant et al., 1984). In D. viverrinus,tooth eruption begins at �90 days and is completedat �177 days (Nelson and Smith, 1971; Collins,1973; Bryant, 1988). Our younger specimens, then,are clearly older than 90 days.

Study of Growth

Following previous studies (Abdala et al., 2001;Flores et al., 2003; Giannini et al., 2004), we tooktwo descriptive approaches. First, we comparedqualitative features of the skull of young specimenswith those of fully adult specimens and describedobserved differences using anatomical terminologydefined by Wible (2003). Second, we used 14 linearmeasurements representing important skull dimen-sions (Fig. 1) to estimate allometric growth of skullcomponents (Giannini et al., 2004). Adult males ofDasyurus albopunctatus tend to be larger than adultfemales, but the difference is not statistically de-monstrable at conventional � levels (t � 1.81, d.f. �19, P � 0.087). Therefore, we pooled males andfemales in a single sample. For the study of allom-etry, we took a multivariate approach based on thegeneralization of the allometry equation proposed byJolicoeur (1963a,b). In multivariate allometry, sizeis regarded as a latent variable affecting all mea-sured variables simultaneously. The first eigenvec-tor of a principal components analysis (PCA) ex-presses various allometric relationships of allvariables with the latent size, provided that thisvector is extracted from a variance-covariance ma-trix of log-transformed variables and scaled to unity(i.e., with all elements scaled so that the sum ofsquared elements equals 1; Jolicoeur, 1963a). Forthis purpose, we used a matrix of 14 variables by 31specimens that had reliable measurements for allvariables (no missing data). For a given variable,allometry is the statistical deviation of its corre-sponding eigenvector element with respect to a hy-pothetical isometric value that represents pure sizechange. That expected value under isometry is cal-culated as 1/p0.5 with p equal to the number of vari-ables; i.e., 0.267 for the current study based on 14variables.

Statistical departures from isometry were esti-mated using the application of jackknife (Que-nouille, 1956; Tukey, 1956; Manly, 1997) developedby Giannini et al. (2004; see Weston [2003] for abootstrap alternative). Briefly (but see Giannini etal., 2004), the goal of the technique is to generate

427POSTNATAL ONTOGENY OF D. ALBOPUNCTATUS

Journal of Morphology DOI 10.1002/jmor

confidence intervals for the empirically derived eig-envector elements. The confidence interval may beinclusive of the null value 0.267 and therefore indis-tinguishable from isometry, or may exclude suchvalue and therefore can be considered significantlyallometric: either positive if the observed element is�0.267, or negative if the observed element is�0.267. To that end, pseudosamples are generated

such that a new first unit eigenvector is calculatedfrom a matrix with one individual removed at atime. Each time a pseudovalue for each element iscalculated using the traditional formulation for thefirst-order jackknife:

e*j � ne�(n � 1)e�j

Fig. 1. Cranial measurements of Dasyurus albopunctatus used in this study. BB, breadth of braincase; BP, breadth of palate; BZ,zygomatic breadth; HD, height of mandibular body; HM, height of muzzle; HO, height of occipital plate; LC, length of coronoid process;LD, length of mandible; LN, length of nasals; LO, length of orbit; LP, length of the lower postcanine row; PAL, length of palate; TL,total length of the skull; UP, length of upper postcanine row.

428 D.A. FLORES ET AL.


where one pseudovalue e *j corresponds to the re-moval of specimen j from the sample of size n, e isthe observed element of the unit eigenvector thatcorresponds to the multivariate coefficient of allom-etry of the skull variable x, and e –j is the value of thecoefficient obtained with specimen j removed (termi-nology follows Manly, 1997). From the collection of31 pseudovalues per variable, a mean is calculatedfor each variable, representing the jackknife esti-mate of the multivariate allometry coefficient forthat variable. The difference between that estimateand the actual value from the complete sample is ameasure of bias. The standard deviation and thecorresponding 99% confidence interval (for n – 1degrees of freedom) are calculated for each allometrycoefficient.

Giannini et al. (2004) followed Manly (1997) inusing trimmed values for the calculation of pseudo-values. For Dromiciops gliroides, trimming the mlargest and m smallest pseudovalues (with m � 1)for each variable significantly decreased the stan-dard deviations and allowed for more realistic allo-metric estimations (see Giannini et al., 2004). Herewe report untrimmed as well as (m � 1) trimmedcalculations, opting for the results with either loweraverage standard deviation or bias (see Allometry inResults and Discussion). The jackknife procedurewas done partly manually and partly with the helpof an NTSYS-pc 1.6 (Rohlf, 1990) batch file.

RESULTS AND DISCUSSIONQualitative Trends

We identified 31 structures exhibiting ontogeneticchanges between youngest and oldest specimens(summarized in Table 1). These include changes inthe rostrum (4 characters), palatal region (5), orbi-totemporal region (6), occiput (2), basicranial andauditory region (7), and mandible (7). Almost a thirdof those changes represent the appearance of struc-tures absent in juveniles (characters 3, 6, 10, 12, 13,15, 16, 18, 20, and 30 in Table 1). Another third ofthe changes imply completion or enlargement ofstructures already present in juveniles (characters5, 8, 11, 17, 19, 21, 23, 25, 26, 27, and 28). A fewchanges imply the relative decrease in size or disap-pearance of some structures or relationships withadvancing age (characters 2, 7, 9, 14, 31). Finally,some reorganizations occur (characters 1, 4, 22, 24,and 29).

A series of changes are indicative of an ontoge-netic strengthening of the masticatory apparatustowards adulthood. These include changes related tothe dentition, to the masticatory muscles, and to thetemporomandibular joint (see in Table 2 andchanges in articulations, below). For example, theparacanine fossa enlarges to receive the lower ca-nine (character 3 in Table 1, Fig. 2B). When thistooth is fully emerged it barely diverges from thevertical and therefore must fit in the paracanine

fossa. The upper alveolar line is below the level ofthe root of the zygomatic arch (character 4, Fig. 2B),whereas a rearrangement occurs in the lower alve-olar line: the lower molars are elevated with respectto the lower premolars (character 29, cf. Fig. 3A,B),the latter no longer occluding with the upper premo-lars (character 31). The canine is clearly the mostdeveloped element in the anterior dentition of theadult and there exist small spaces between canineand first premolar and also between the premolars.Osteological changes related to the muscle systemare basically enlargements of areas of origin andinsertion of the temporalis, masseter, and pterygoidmuscles. In relation to the origins of the temporalis,the postorbital process appears in adults (character10, Fig. 4B), as well as the sagittal (Figs. 2B, 4B),nuchal (Figs. 2B, 4B), and infratemporal crests (Fig.5B; characters 15, 16, and 12, respectively). Ofthese, the nuchal crest experiences the most ex-treme change. This likely reflects the significance ofthe canine bite in Dasyurus, as the posterior fibers ofthe temporal muscle (which originate in the nuchalcrest) are more important than anterior temporalfibers (which originate in the sagittal crest) in gen-erating forces transmitted to the canines—a patternonly comparable to Lutreolina among the marsupi-als hitherto described by us (Flores et al., 2003). Inrelation to the masseter, its origin along the masse-teric line of the jugal becomes strongly marked(character 13, Fig. 2B). Its insertion in the mandibleis on the widened shelf of the masseteric fossa (char-acter 27, Fig. 3B), which is reinforced by a lateralcrest that becomes prominent in adults (character28, Fig. 3D). Finally, the angular process markedlyincreases its robustness in adults, thereby increas-ing the surface for the insertion of the superficialmasseter and internal insertion of the medial ptery-goid muscles (character 25, cf. Fig. 3C,D). Also re-lated to the feeding function, the paracondylar pro-cess of the exoccipital, origin of the posterior belly ofthe digastric muscle, becomes increasingly promi-nent in adults (character 17, Fig. 2A,B).

Modifications in the palate are noteworthy. First,the palatine openings (also known as “vacuities”)either appear (character 6 in Table 1, palatinefenestra, cf. Figs. 5B, 6B), increase in size (character5, major palatine foramen, cf. Fig. 6A,B), or arereshaped into an almost closed structure (character7, minor palatine foramen, cf. Fig. 6A,B) with ad-vancing age. In general, the palatal openings are notas developed as in many marsupials (e.g., Didelphisalbiventris; Abdala et al., 2001). Second, the palateis greatly strengthened caudally in adults by thedevelopment of a thickened postpalatine torus (char-acter 8, Fig. 6).

We detected ontogenetic changes in most sutures,synchondroses, and synovial joints (Table 2). Preco-cially mature joints exhibit no change from the ju-venile to the adult, presumably because their func-tion is attained early in ontogeny. Precocially



mature joints are those designated as 3, 4, 12, 16, 17,20, and 36 in Table 2. The overall trend of theontogenetic transformations in the sutural patternis obviously to increase structural strength and ri-gidity. First, in the cranium the synchondroses ei-ther fuse or their components come to a closer con-tact (joints 31–35), and the plane sutures form highcrests (joints 5 and 7) or become serrated in theborder (joint 1). Second, in the rostrum the border ofthe squamous sutures changes from simple to ser-rated (joints 22), and the plane sutures become ser-rated (joints 9, 10, 21, 24, 25, and 26). The palatinesutures (sutura palatina mediana et transversa) are aremarkable example of the latter (cf. Fig. 6A,B). And

third, in the union of rostrum and cranium (Fig. 4), thesutures involved become either serrated (joints 9, 10,and 23) or with serrated borders (joints 11 and 30).Finally, components of synovial joints (37 and 38) arereshaped to provide a firmer anchorage via transverse(in 37, Fig. 3C,D) or longitudinal (in 38, Fig. 2) enlarge-ment of the mandibular condyles.

The changes in the temporomandibular joint (num-ber 37 in Table 2) are noteworthy. The glenoid fossa isreshaped such that the contribution of the jugal to thefossa greatly increases in adults, whereas the postgle-noid process and the mandibular condyle expand lat-erally (cf. Fig. 7A,B, Fig. 3C,D). In our interpretation,these changes in the temporomandibular joint, as well

TABLE 1. List of morphological changes detected in a comparison of smallest (youngest) and largest (oldest)specimens of Dasyurus albopunctatus

Characters Figure Juveniles Adults

Rostrum1. Caudal extension of nasals 4 Reaching the lacrimal Surpassing the lacrimal2. Lacrimal foramen Proportionally large Proportionally small3. Paracanine fossa 2 Absent Large4. Level of upper alveolar line 2 At zygomatic root Ventral to zygomatic root

Palatine region5. Major palatine foramen 6 Rounded, small Elongated, enlarged6. Palatine fenestrae 6 Absent Present, irregular7. Notch for minor palatinenerve and vessels

6 Widely open Notch almost closed ventrally

8. Postpalatine torus 6 Smooth Sculptured ridge9. Nasal spine 6 Prominent Reduced

Orbitotemporal region10. Postorbital process 4 Absent Prominent11. Ethmoidal foramen Incompletely enclosed by the frontal ventrally Completely enclosed by the frontal12. Infratemporal crest 5 Absent Present13. Masseteric line of jugal 2 Absent Marked14. Suprameatal foramen 2 Large Minute15. Sagittal crest 2, 4 Absent Well developed

Occipital region16. Nuchal crest 2, 4 Absent Well developed17. Paracondylar process 2, 7 Blunt and short Prominent

Basicranial and auditoryregions18. Posttympanic process 2 Absent Prominent19. Basioccipital crest 7 Faint Marked20. Muscular tubercles onbasicranium

7 Absent Strong

21. Ectotympanic 7 Not widened laterally Widened laterally22. Alisphenoid-ectotympanic

contact7 Both bones juxtaposed Alisphenoid abuts ectotympanic

ventrolaterally23. Alisphenoid strut 5 Present, thin Present, well developed24. Contact between

postglenoid process andectotympanic

7 Absent Present

Mandible25. Angular process 3C, D Thin Robust26. Mandibular foramen Near caudal edge of ramus In middle of ramus27. Posterior shelf of the 3C, D Narrow Greatly expanded laterally

masseteric fossa28. Masseteric line of

mandible3A, B Low, not marked Forming a prominent ridge

29. Alveolar line of mandible 3A, B Molars placed ventral to premolars Molars displaced dorsally withrespect to premolars

30. Retromolar space 3 Absent Present31. Occlusion of upper and

lower premolarsPresent Absent

The characters are grouped by skull regions and numbered successively. When available, a reference is made to the figure(s) of thepresent study that show the structure.



as the presence of the paracanine fossa, limit the ros-trocaudal displacement and the transverse rotationalmovement of the mandible. A similar but less markedpattern has been reported for the large didelphids(e.g., Didelphis [Abdala et al., 2001] and Lutreolina[Flores et al., 2003]).

Allometry

The results of our multivariate analysis of allo-metric coefficients are given in Table 3. The meandifference in standard deviations of jackknife esti-mates between untrimmed (0.091) and trimmed

(0.071) appears to be negligible. However, the meandifference in absolute bias clearly favors untrimmedover trimmed values, with 0.004 average absolutebias for the former, and 0.017 for the latter (4.25times higher). Therefore, as the jackknife was orig-inally intended as a technique for bias removal(Quenouille, 1956; Manly, 1997), our interpretationsare based on untrimmed values (cf. our results forDromiciops gliroides; Giannini et al., 2003).

Three variables (length of palate, breadth ofbraincase, and height of occipital plate) are nega-tively allometric; three (length of coronoid process,height of mandible, breadth of zygoma) are posi-

TABLE 2. Ontogenetic changes in skull joints between smallest (youngest) and largest (oldest) specimens of Dasyurus albopunctatus

Joints No. Juveniles Adults

Suturesoccipitomastoidea 1 plane slightly serratedoccipitoparietalis 2 plane squamouscoronalis 3 squamous squamoussquamosal 4 squamous squamoussagittalis 5 plane raised in crestsphenoparietalis 6 plane squamousinterfrontalis 7 plane raised in crest caudallysphenofrontalis 8 slightly squamose squamousfrontonasalis 9 plane serratedfrontomaxillaris 10 plane serratedfrontolacrimalis 11 squamous squamous, caudal border foliatefrontopalatina 12 squamous squamoussphenopalatina 13 squamous, little overlap squamous, large overlappterygosphenoidalis 14 squamous, little overlap squamous, large overlaptemporozygomatica 15 squamous, little overlap squamous, large overlapincisivomaxillaris 16 foliate foliatenasoincisiva 17 plane planevomeroincisiva 18 foliate, unfused fusedinterincisiva 19 plane, loose contact plane, close contactinternasalis 20 plane planenasomaxillaris 21 plane serratedlacrimomaxillaris 22 squamous, border simple squamous, border serratedzygomaticomaxillaris 23 foliate, border simple foliate, border serratedpalatomaxillaris 24 plane serratedintermaxillaris 25 plane serratedpalatine transversa 26 plane serratedvomeropalatina dorsalis 27 squamous, loose contact squamous, close contactpterygopalatina 28 squamous, unfused squamous, fusedpalatolacrimalis 29 squamous, border simple squamous, border serratedzygomaticolacrimalis 30 squamous, border simple squamous, border serrated

Synchondrosesspheno-occipitalis 31 unfused fusedpetro-occipitalis 32 unfused close contactintersphenoidalis 33 loose contact close contactintraoccipitalissquamolateralis

34 close contact seemlessly fused

intraoccipitalisbasilateralis

35 close contact seemlessly fused

intermandibularis 36 close contact close contactSynovial joints

Temporomandibular 37 Mandibular condyle rounded Mandibular condyle laterally expandedGlenoid process of jugal small Glenoid process of jugal large, contributing

up to one-third of joint surfacePostglenoid process tapering to a point Postglenoid process greatly expanded

mediolaterallyLateral notch of postglenoid process absent Lateral notch of postglenoid process present

Atlanto-occipital 38 Occipital condyle not elongated caudally Occipital condyle elongated caudally

Articulations are grouped into sutures, synchondroses, and synovial joints, and numbered for reference in the text. Articulation namesare in Latin following NAV (1994).



tively allometric; and the remaining eight vari-ables scale isometrically with increasing generalsize. On the basis of these trends, we describe thepostnatal skull growth in Dasyurus albopunctatusas follows.

Condylobasal length, a commonly employed in-dex of cranial size, is ontogenetically isometric.The braincase grows at a slower pace than theskull, as do other measured neurocranial compo-nents, such as the height of the occipital plate.Thus, the two main dimensions of the braincase(height and breadth) share a markedly negativeallometric growth of similar magnitude (jackknifecoefficient for breadth of braincase is 0.111, andfor height of occipital plate is 0.164, as comparedwith the null isometric value, 0.267). By contrast,

length of the orbit scales isometrically (see below;note that the orbit is negatively allometric whenconsidering trimmed values; Table 3). The tempo-ral space, which contains the temporalis muscle,expands both inward and outward, by the combi-nation of the positive allometry of the zygomaticbreadth with the negative allometry of the brain-case breadth (cf. Fig. 4A,B). In relative terms, thepalate shortens as the animal grows as the resultof isometric width scaling and negative lengthscaling. However, two other dimensions of the fa-cial skeleton, rostral height and length of nasals,scale isometrically. Both tooth rows grow at aboutthe same relative rate as the skull, as does man-dibular length. By contrast, vertical dimensions ofthe mandible grow relatively faster than the skull;

Fig. 2. Lateral view of the skull of young Dasyurus albopunctatus AMNH 151982 (A) and adult AMNH 157081 (B). astp,alisphenoid tympanic process; ju, jugal; M2, second upper molar; M3, third upper molar; M4, fourth upper molar; mlj, masseteric lineof jugal; mx, maxilla; nc, nuchal crest; oc, occipital condyle; pa, parietal; pap, paracondylar process; pcf, paracanine fossa; pmx,premaxilla; ptp, posttympanic process; sc, sagittal crest; smf, suprameatal foramen; sq, squamosal; ual, upper alveolar line. Scalebars � 10 mm.



in particular, the mandible becomes relativelymore robust, and the coronoid process relativelylonger (cf. Fig. 3A,B) in larger animals.

Although sexual dimorphism is statistically non-significant in our small sample (see Materials andMethods), adult males tend to be larger than adult

Fig. 3. Mandible of young Dasyurusalbopunctatus AMNH 151982 in lateral(A) and occlusal view (C), and mandibleof adult AMNH 157081 in lateral view(B), and occlusal view (D). an, angularprocess; c, lower canine; cmlm; crest ofmasseteric line of mandible; con, man-dibular condyle; cor, coronoid process;m1, first lower molar; m2, second lowermolar; m3, third lower molar; m4,fourth lower molar; mf, masseteric fos-sa; mlm, masseteric line of mandible;psmf, posterior shelf of masseteric fos-sa; rms, retromolar space. Scale bars �10 mm.



females (Fig. 8). We deemed it unnecessary to sta-tistically test for a change in slope or position be-tween curves for males and females given, first, theclose fit of the common curve to the data points fromthe pooled sexes and, second, our limited samplesize. In effect, we assume that Dasyurus albopunc-

tatus exhibits parallel growth trajectories for bothsexes without vertical curve displacement (i.e., iden-tical y-intercept), such that there are no growth-invariant shape differences between males and fe-males. Of course, there may be differences inaverage shape between like-aged males and fe-

Fig. 4. Dorsal view of the skull of young Dasyurus albopunctatus AMNH 151982 (A) and adultAMNH 157081 (B). fr, frontal; na, nasal; nc, nuchal crest; pa, parietal; poc, postorbital constriction;pop, postorbital process; sc, sagittal crest; sco, sutura coronalis; sfn, sutura frontonasalis; so,supraoccipital; ss, sutura sagittalis. Scale bars � 10 mm.



Fig. 5. Ventral view of the skull of young Dasyurus albopunctatus AMNH 151982 (A) and adult AMNH 157081 (B). as, alisphenoid;astp, alisphenoid tympanic process; bo, basioccipital; bs, basisphenoid; cf, carotid foramen; ec, ectotympanic; eo, exoccipital; fm,foramen magnum; fo, foramen ovale; gf, glenoid fossa; hf, hypoglossal foramen; ips, foramen for the inferior petrosal sinus; itc,infratemporal crest; jf, jugular foramen; lpc, lateral palatine crest; M2, second upper molar; mapf, major palatine foramen; mtmx,masseteric tuberosity of maxilla; mx, maxilla; pal, palatine; pap, paracondylar process; pgp, postglenoid process; pr, promontorium ofthe petrosal; ps, presphenoid; ptp, posttympanic process; sfo, secondary foramen ovale; tc, transverse canal. Arrows indicate the courseof the mandibular division of the trigeminal nerve (V3) between the primary and secondary foramen ovale. Scale bars � 10 mm.



males, but these appear to be due primarily to theincreased size of males. Either females grow moreslowly than males, or males prolong growth, or both.

Fig. 6. Posterior portion of the palate of young Dasyurus al-bopunctatus AMNH 151982 (A) and adult AMNH 157081 (B). bs,basisphenoid; lpc, lateral palatine crest; mapf, major palatineforamen; mpn, minor palatine notch; nM3, notch for upper thirdmolar; ns, nasal spine; pf, palatine fenestra; ppt, posterior pala-tine torus; ps, presphenoid; simx; sutura intermaxillaris; spm,sutura palatina mediana; spt, sutura palatina transversa. Scalebars � 5 mm.

Fig. 7. Oblique (caudolateroventral) view of young Dasyurusalbopunctatus AMNH 151982 (A) and adult AMNH 157081 (B).as, alisphenoid; ast, alisphenoid strut; astp, alisphenoid tympanicprocess; bo, basioccipital; boc, basioccipital crest; bs, basisphe-noid; cf, carotid foramen; ec, ectotympanic; eo, exoccipital; fm,foramen magnum; fo, foramen ovale; gf, glenoid fossa; gpju, gle-noid process of jugal; hf, hypoglossal foramen; ips, foramen for theinferior petrosal sinus; jf, jugular foramen; me, mastoid exposure;mp, muscular process; pap, paracondylar process; pgp, postgle-noid process; pr, promontorium of petrosal; ptp, posttympanicprocess; sq, squamosal; tc, transverse canal. Scale bars � 10 mm.



TA

BL

E3.

Res

ult

sof

the

mu

ltiv

aria

tean

alys

isof

cran

ial

allo

met

ryin

Das

yuru

sal

bopu

nct

atu

s

Var

iabl

es

Exp

ecte

dal

lom

etry

coef

fici

ent

Obs

erve

dal

lom

etry

coef

fici

ent

Obs

erve

dde

part

ure

Un

trim

med

valu

esT

rim

med

valu

es

Res

ampl

edal

lom

etry

coef

fici

ent

Bia

s

99%

Con

fide

nce

Inte

rval

Gro

wth

tren

d

Res

ampl

edal

lom

etry

coef

fici

ent

Bia

s

99%

Con

fide

nce

Inte

rval

Gro

wth

tren

d

Tot

alle

ngt

h0.

267

0.26

1–0

.006

0.26

0–0

.001

0.24

1–0.

279

�0.

244

–0.0

160.

229–

0.25

9–

Len

gth

ofn

asal

s0.

267

0.29

10.

024

0.29

80.

006

0.24

7–0.

349

�0.

273

–0.0

250.

228–

0.31

7�

Hei

ght

ofm

uzz

le0.

267

0.29

10.

024

0.29

30.

002

0.24

5–0.

341

�0.

282

–0.0

120.

244–

0.31

9�

Len

gth

ofpa

late

0.26

70.

223

–0.0

440.

226

0.00

40.

199–

0.25

4–

0.21

0–0

.017

0.18

8–0.

231

–B

read

thof

pala

te0.

267

0.23

6–0

.031

0.23

6–0

.001

�0.

195–

0.27

7�

0.22

3–0

.013

0.19

0–0.

256

–L

engt

hof

upp

erpo

stca

nin

ero

w0.

267

0.22

7–0

.040

0.22

2–0

.005

0.15

9–0.

284

�0.

208

–0.0

140.

154–

0.26

2–

Len

gth

oflo

wer

post

can

ine

row

0.26

70.

243

–0.0

240.

236

–0.0

070.

157–

0.31

6�

0.22

0–0

.016

0.18

1–0.

259

–

Len

gth

ofm

andi

ble

0.26

70.

289

0.02

20.

287

–0.0

020.

256–

0.31

8�

0.27

0–0

.017

0.24

3–0.

298

�H

eigh

tof

man

dibl

e0.

267

0.41

30.

146

0.40

9–0

.004

0.35

8–0.

459

�0.

383

–0.0

260.

342–

0.42

4�

Hei

ght

ofco

ron

oid

proc

ess

0.26

70.

327

0.06

00.

340

0.01

30.

279–

0.40

1�

0.31

9–0

.021

0.27

4–0.

365

�

Bre

adth

ofzy

gom

a0.

267

0.31

20.

045

0.30

8–0

.003

0.28

1–0.

336

�0.

290

–0.0

190.

267–

0.31

3(�

)B

read

thof

brai

nca

se0.

267

0.11

1–0

.156

0.10

4–0

.006

0.07

5–0.

134

–0.

098

–0.0

060.

076–

0.12

0–

Hei

ght

ofoc

cipi

tal

plat

e0.

267

0.16

4–0

.103

0.16

50.

001

0.12

7–0.

203

–0.

156

–0.0

090.

124–

0.18

7–

Len

gth

ofor

bit

0.26

70.

220

–0.0

470.

219

–0.0

010.

153–

0.28

5�

0.20

0–0

.019

0.14

8–0.

253

–

Th

efi

rst

thre

eda

taco

lum

ns

show

resu

lts

usi

ng

all

spec

imen

s.T

he

rem

ain

der

ofth

eco

lum

ns

show

jack

knif

ere

sult

sca

lcu

late

dw

ith

un

trim

med

and

(m�

1)tr

imm

edse

tsof

pseu

dova

lues

(see

Mat

eria

lsan

dM

eth

ods)

.All

omet

ryco

effi

cien

tis

the

corr

espo

nde

nt

elem

ent

ofth

efi

rst

(un

it)

eige

nve

ctor

per

vari

able

.Th

eex

pect

edco

effi

cien

t(0

.267

)is

the

valu

eu

nde

ris

omet

ry(e

qual

for

all

vari

able

s).T

he

obse

rved

coef

fici

ent

isth

eva

lue

obta

ined

wit

hal

lsp

ecim

ens

incl

ude

d(n

�31

).T

he

resa

mpl

edco

effi

cien

tis

the

firs

t-or

der

jack

knif

eva

lue.

Bia

sis

the

diff

eren

cebe

twee

nth

ere

sam

pled

and

obse

rved

coef

fici

ents

.Th

eja

ckkn

ife

99%

con

fide

nce

inte

rval

ispr

ovid

ed;a

llom

etri

cva

riab

les

are

thos

ew

hos

eco

nfi

den

cein

terv

alex

clu

deth

eex

pect

edva

lue

un

der

isom

etry

(0.2

67).

Gro

wth

tren

dis

the

sum

mar

yal

lom

etry

ofea

chva

riab

lepr

esen

ted

insy

mbo

ls;�

isom

etry

,–n

egat

ive

allo

met

ry,

�po

siti

veal

lom

etry

,(�

)po

siti

veal

lom

etri

ctr

end.



Whatever the causes of the difference graphed inFigure 8, males clearly proceed further than femalesalong a growth trajectory seemingly common to bothsexes.

Comparative Growth

Allometric trends obtained in the different marsu-pials species sampled to date are compared in Table4. The same trends are shared by all of them in threevariables: the height of the mandible (positive allom-etry) and the breadth of the braincase and height ofthe occipital plate (both with negative allometry).These shared trends can be interpreted as present inthe marsupial ancestor.

Two allometric trends are exclusively shared byDasyurus albopunctatus and Dromiciops gliroides,the only other australidelphian studied using the

same methods to date. These are the isometry of thelength of upper postcanine row (see below) and theisometry of the orbit (Table 4). Regarding the lasttrend, Giannini et al. (2004) reported the surprisingfinding that the orbit in D. gliroides scales isomet-rically with overall cranial growth—a feature in con-trast not only with Didelphis albiventris and Lutreo-lina crassicaudata, but also with most vertebrates,in which the orbit typically is negatively allometric(Emerson and Bramble, 1993). In D. albopunctatusthe orbit is similarly isometric when the allometrycoefficient is calculated using untrimmed values (seeabove; Table 3), suggesting that this unusual fea-ture may be common to other australidelphians aswell.

Dasyurus albopunctatus departs from the schemefound in the other marsupials studied in a verysignificant way, insofar as the palatal length, pala-tal breadth, and zygomatic breath show a markedlydifferent allometric pattern (Table 4). The morpho-functional interpretation that we propose to explainthese changes is as follows. In D. albopunctatus thespace for the m. temporalis increases both by meansof the negative allometry of the braincase (as in theother species compared) and the positive allometryof zygomatic breadth (isometric in the other speciescompared; see Fig. 4). Because of the last trend, thezygomatic arches of adults of D. albopunctatus areclearly more expanded laterally than in the otherspecies studied so far. Regarding the other trends,the combined effect of negative allometry of palatallength (isometric in the other species compared) andisometry of the palatal breadth (negatively allomet-ric in the other species compared) renders the palateshort and broad in older animals, a morphology thatvisually characterizes all species in this genus (Fig.5). These differences point to an emphasis on devel-oping a strong bite, as they favor large temporal andmasseteric muscles loading on a shortened tooth row

Fig. 8. Bivariate scatterplots of mandible length on condylo-basal length in males Dasyurus albopunctatus (circles) and fe-males (squares). The diamond indicates a specimen of unknownsex.

TABLE 4. Allometric comparison of Dasyrus albopunctatus (this study) with Didelphis albiventris(Abdala et al., 2001), Lutreolina crassicaudata (Flores et al., 2003), and Dromiciops gliroides

(Giannini et al., 2004)

Variables Didelphis Lutreolina Dromiciops Dasyurus

Length of nasals � (–) � �Height of muzzle (–) � � �Length of palate � � � –Breadth of palate – – – �Length of upper postcanine row – – � �Length of lower postcanine row – ?a – �Length of mandible � � � �Height of mandible � � � �Height of coronoid process � � � �Zygomatic breadth � � � �Breadth of braincase – – – –Height of occipital plate – (–) – –Length of orbit – – � �

aDifferent value depending on method used (see Flores et al., 2003). Symbols for isometry, negativeallometry, and positive allometry, are �, –, and �, respectively. Parentheses indicate allometric trends.Underline indicate, similar allometry trends across species.



(in Dasyurus the canine is, compared to other mar-supials, closer to the fulcrum). This is especiallyremarkable when considered together with thechanges related to the dentition, cranial crests, andtemporomandibular joint described above, and islikely connected with the carnivorous habits re-ported for quolls (Flannery, 1990; Menzies, 1991) aswell as other dasyurids (e.g., Blackhall, 1980; Pellisand Nelson, 1984; Jones and Rose, 2001; Jones et al.,2001).

Finally, it is interesting to discuss the trends ofthe postcanine row lengths. Both upper and lowerpostcanine row lengths are isometric in Dasyurusalbopunctatus. In the first case, the trend is sharedwith Dromiciops gliroides, while the isometry of thelower postcanine length departs from our previousstudies. This measurement was found to be nega-tively allometric in Didelphis albiventris and D. gli-roides, or is inconclusive (i.e., it produced differenttrends with different methods), as in Lutreolinacrassicaudata (see Table 4). The length of the post-canine row is involved in the mutual adjustment ofthe lower and upper tooth rows during growth. Ab-dala et al. (2001) described a pattern in D. albiven-tris by which the upper tooth row has a fastergrowth rate than the lower toothrow, the latter al-ways having one tooth more than the upper too-throw until all teeth emerge. This pattern was con-firmed in L. crassicaudata (Flores et al., 2003).Dasyurus albopunctatus exhibits the same patternof tooth emergence, but, by contrast, showed isome-try of the length of upper and lower tooth rows. Thepostcanine series of D. albopunctatus show only twopremolars, differing from the other marsupials stud-ied so far, which display three premolars. This dis-similarity in premolar numbers (plus the presence ofa long diastema between the first and second pre-molar in D. albiventris) is likely related to the dif-ferent trends of D. albopunctatus for these variablesin relation to didelphids. On the other hand, theisometry in the upper postcanines length of D. gli-roides has apparently no connection with the num-ber of premolars in the dentition.

CONCLUSIONS

Dasyurus albopunctatus exhibits some allometrytrends already observed in three other species ofmarsupials (two didelphids and a microbiotheriid),but it also shares some trends uniquely with theonly other australidelphian examined (the micro-biotheriid Dromiciops), and it has a suite of trendsnot shared with any previously studied marsupials.We relate these allometry trends, together with thechanges in qualitative traits and cranial sutures,with an overall strengthening of the skull and thedevelopment of a skull shape that favors a strongbite. As we are dealing with weaned individuals,those changes can be associated with the shift frommilk suckling to an active feeding. Some of the

trends found in Dasyurus albopunctatus can be in-terpreted as a function of a distinctly carnivore feed-ing system in the adult.

ACKNOWLEDGMENT

We thank R. Voss (American Museum of NaturalHistory) for access to specimens for this study andfor comments on the article.

LITERATURE CITED

Abdala F, Flores DA, Giannini NP. 2001. Postweaning ontogenyof the skull of Didelphis albiventris. J Mammal 87:190–200.

Asher RJ, Horovitz I, Sanchez-Villagra M. 2004. First combinedcladistic analysis of marsupial mammal interrelationships. MolPhylogenet Evol 33:240–250.

Blackhall S. 1980. Diet of the eastern native-cat, Dasyurus viver-rinus (Shaw), in southern Tasmania. Aust Wild Res 7:191–197.

Bryant SL. 1988. Mantenance and captive breeding of the easternquoll Dasyurus viverrinus. Int Zool Yrbk 27:119–124.

Burnett S. 2000. The ecology and endangerment of the spotted-tailed quoll, Dasyurus maculates gracilis. Ph.D. dissertation,James Cook University of North Queensland, Townsville, Aus-tralia.

Collins LR. 1973. Monotremes and marsupials. A reference forzoological institutions. Washington, DC: Smithsonian Institu-tion Press.

Edgar R. 1983. Spotted-tailed quoll. In: Strahan R, editor. TheAustralian Museum complete book of Australian mammals.Sydney, New South Wales: Angus and Robertson. p 18–19.

Emerson SB, Bramble DM. 1993. Scaling, allometry and skulldesign. In: Hanken J, Hall BK editors. The skull. Chicago:University of Chicago Press. p 384–416.

Flannery T. 1990. Mammals of New Guinea. Australia: RobertBrown & Associates.

Fleay D. 1935. Breeding of Dasyurus viverrinus and general ob-servation of the species. J Mammal 16:10–16.

Flores DA, Giannini NP, Abdala F. 2003. Cranial ontogeny ofLutreolina crassicaudata (Didelphidae): a comparison with Di-delphis albiventris. Acta Theriol (Warsz) 48:1–9.

Giannini NP, Abdala F, Flores DA. 2004. Comparative postnatalontogeny of the skull in Dromiciops gliroides (Marsupialia:Microbiotheriidae). Am Mus Novit 3460:1–17.

Godsell J. 1982. The population ecology of the eastern quoll,Dasyurus viverrinus (Dasyuridae, Marsupialia), in southernTasmania. In: Archer M, editor. Carnivorous marsupials. NewSouth Wales, Australia: Surrey Beatty & Sons. p 198–207.

Godsell J. 1983. Ecology of the eastern quoll Dasyurus viverrinus(Dasyuridae, Marsupialia). Ph.D. dissertation, Australian Na-tional University, Canberra.

Groves CP. 1993. Order Dasyuromorpha. In: Wilson D, Reeder M,editors. Mammal species of the world, a taxonomic and geo-graphic reference. Washington, DC: Smithsonian InstitutionPress. p 29–37.

Hill JP, Hill WCO. 1955. The growth-stages of the pouch-young ofthe native cat (Dasyurus viverrinus), together with observa-tions on the anatomy of the new-born young. Trans Zool SocLond 28:349–425.

Horovitz I, Sanchez Villagra M. 2003. A morphological analysis ofmarsupial higher-level relationships. Cladistics 19:181–212.

Inns RW. 1982. Age determination in the Kangaroo Island Wal-laby, Macropus eugenii (Desmarest). Aust Wild Res 9:213–220.

Jolicoeur P. 1963a. The multivariate generalization of the allom-etry equation. Biometrics 19:497–499.

Jolicoeur P. 1963b. The degree of generality of robustness inMartes americana. Growth 27:1–27.

Jones ME, Rose RK. 2001. Dasyurus viverrinus. Mammal Sp 677.Jones ME, Rose RK, Burnett S. 2001. Dasyurus maculatus. Mam-

mal Sp 676.



Kavanagh JR, Burk-Herrick A, Westerman M, Springer MS.2004. Relationships among families of Diprotodontia (Marsu-pialia) and the phylogenetic position of the autapomorphichoney possum (Tarsipes rostratus). J Mamm Evol 11:207–222.

Manly BFJ. 1997. Randomization, bootstrap, and Monte Carlomethods in biology, 2nd ed. London: Chapman & Hall.

Menzies J. 1991. A handbook of New Guinea, marsupials andmonotremes. Madang, New Guinea: Kristen Press.

Merchant JC, Newgrain K, Green B. 1984. Growth of the EasternQuoll, Dasyurus viverrinus (Shaw), (Marsupialia) in captivity.Aust Wild Res 11:21–29.

Moeller VH. 1973. Zur Kenntis der Schadelgestaltgro�erRaubbeutler (Dasyuridae, Waterhouse, 1838) Eine al-lometrische Formanalyse. Zool Jb Anat Bd 91:257–303.

Nelson JE, Smith G. 1971. Notes on growth rates in native cats ofthe family Dasyuridae. Int Zool Yrbk 1:38–41.

Pellis SM, Nelson JE. 1984. Some aspects of predatory behaviourof the quoll, Dasyurus viverrinus (Marsupialia, Dasyuridae).Aust Mammal 7:5–15.

Quenouille MH. 1956. Notes on bias in estimation. Biometrika43:353–360.

Rohlf FJ. 1990. NTSYS-pc: numerical taxonomy and multivariateanalysis system, v. 1.6. New York: Applied Biostatistics.

Settle GA. 1978. The quiddity of tiger quolls. Aust Nat Hist19:164–169.

Sharman GB, Frith HJ, Calaby JH. 1964. Growth of the pouchyoung, tooth eruption and age determination in the red kanga-roo, Megaleia rufa. CSIRO Aust Wild Res 9:20–49.

Shield JW, Woolley P. 1961. Age estimation and measurements ofthe pouch young of the quokka (Setonix brachiurus). Aust J Zool9:14–23.

Tukey JW. 1956. Bias and confidence in not quite large samples.Ann Math Stat 23:614.

Weston EM. 2003. Evolution of ontogeny in the hipoppotamusskull: using allometry to dissect developmental change. Biol JLinn Soc Lond 80:625–638.

Wible JR. 2003. On the cranial osteology of the short-tailed opos-sum Monodelphis brevicaudata (Didelphidae, Marsupialia).Ann Carn Mus 72:137–202.

Wroe S. 1997. A reexamination of proposed morphology-basedsynapomorphies for the families of Dasyuromorphia (Marsupia-lia). I. Dasyuridae. J Mamm Evol 4:19–52.



comparative postnatal ontogeny of the skull in the ... · comparative postnatal ontogeny of the...

Documents