Comparative morphology of Western Australian varanid ......Comparative Morphology of WesternAustralian Varanid Lizards (Squamata: Varanidae) GRAHAMG.THOMPSON1*AND PHILIPC.WITHERS2

Comparative Morphology of Western Australian Varanid Lizards(Squamata: Varanidae)

GRAHAM G. THOMPSON1* AND PHILIP C. WITHERS21Centre for Ecosystem Management, Edith Cowan University,Joondalup, 6027 Western Australia2Department of Zoology, University of Western Australia,Nedlands, 6907 Western Australia

ABSTRACT Varanid lizards, which vary considerably in body mass bothinterspecifically and intraspecifically, are generally considered to be morpho-logically similar. However, significant and non-isometric variation in therelative appendage dimensions for 17 species of Western Australian goannassuggest that these lizards are not morphologically conservative. The first andsecond canonical variates clearly distinguish the two subgeneral Odatria andVaranus, and species are generally sexually dimorphic. The morphologicalvariation observed among these 17 species of goanna is associated withforagingmode and ecology. However, no single or small group of morphologicaldimensions discriminates phylogenetic groups, sexes, or ecological groups,and body size is an important component in these analyses. J. Morphol.233:127–152, 1997. r 1997 Wiley-Liss, Inc.

Themorphology of a lizard is largely deter-mined by its ancestry, ecological niche, bodysize, and development (Peters, ’83; Calder,’84; Schmidt-Nielsen, ’84). In addition, somespecies of reptile are also sexually dimorphicin body shape or size (Vitt and Cooper, ’85;Shine, ’92). The lizard family Varanidae pro-vides an excellent opportunity to study theinterrelationships of body size and shapewith ecology. Varanidae consists of only asingle extant genus, Varanus, and containsabout 45 species. Themass range of Varanusis more than three orders of magnitude,ranging from <20 g (V. brevicauda; personalobservations) to <54 kg (V. komodoensis;Auffenberg, ’81). There are a variety of eco-logical specializations, including tree climb-ing, rock scampering, and swimming. Never-theless, a numbers of authors (Shine, ’86;Greer, ’89; King and Green, ’93b; Pianka,’95) have suggested that their body form isconservative compared with the variation inother families of lizards.The genus Varanus is considered to be

monophyletic (Baverstock et al., ’93), andthus comparison of varanid species is notcomplicated by higher level phylogenetic dif-ferences. Baverstock et al. (’93) summarizedthe phylogeny ofVaranus and suggested fourclades based on immunogenetic and karyo-typic studies: an Asian clade, an African

clade, anAustralian/S.E.Asian clade of largegoannas (subgenus Varanus), and a cladeof Australian pygmy goanna (subgenusOdatria). Nearly all of the members of theVaranus clade (except V. komodoensis andV. salvadorii) and all of the members of theOdatria clade are found in Australia; V. er-emius probably belongs to theOdatria group,although it was initially placed outside theseclades (Pianka, ’95). Morphometric examina-tion of the 18 species/subspecies of goannafound inWesternAustralia allows a compari-son of the Varanus andOdatria subgenera.Others (e.g., Snyder, ’54; Collette, ’61; Ball-

inger, ’73; Laerm, ’74; Moermond, ’79;Pianka, ’86; Losos, ’90a–c; Miles, ’94) havesuggested that there aremorphological char-acteristics that can be associated with habi-tat and performance traits. Pianka (’68, ’69,’70a,b, ’71, ’82, ’86, ’94) provides most of thelimited ecological and behavioral data, andsome additional general information on theirecology is provided by Storr et al. (’83) andWilson and Knowles (’92). Greer (’89) groupsall Australian goannas into four broad eco-logical categories (ground, rocky outcrop, ar-boreal, and aquatic/arboreal). The only obvi-

*Correspondence to: Graham G. Thompson, Edith Cowan Uni-versity, Joondalup Drive, Joondalup, Western Australia, Austra-lia, 6027. E-mail: G. [email protected]

JOURNAL OF MORPHOLOGY 233:127–152 (1997)

r 1997 WILEY-LISS, INC.

ous morphological adaptation of any of thesegroups is the laterally compressed tail anddorsal placement of the nostrils for the semi-aquatic species V. mertensi.Varanus, being a speciose genus of lizard

with a very wide range in body size, providesan ideal opportunity to explore variation inbody size and shape as these may relate todifferences in phylogeny, ecology, and habi-tat, although the present lack of a rootedphylogenetic tree for Varanidae limits ourcapacity to account for lower order phyloge-netic effects. The objectives of this studywere to examine the allometry of WesternAustralian goanna morphology and to deter-mine whether there are size or shape differ-ences among these goannas which are as-sociated with species ecology, habitat, orphylogeny, particularly comparing the sub-genera Varanus andOdatria.

MATERIALS AND METHODSMeasurements

Various morphological dimensions (width,depth, and length) were measured for 17species (including two subspecies of Varanuspanoptes; see Results for a list of species) ofgoanna specimens from theWestern Austra-lian Museum (WAM). Unfortunately, thenumber of specimens of V. kingorum in theWAM collection was too small to enable anymeaningful analysis and so this species wasnot included in the study. The nomenclatureused for V. gouldii and V. panoptes is that ofStorr (’80). It is our view that the use of thealternative names as suggested by Bohme(’91) will lead to further confusion until thetaxonomy of both species is further clarified.Total length (TL), snout-to-vent length

(SVL), tail length (TAIL), head length (HL),neck length (NECK), head width (HW), headdepth (HD), fore-limb length (FLL), upperfore-limb length (UFL), lower fore-limblength (LFL), hind-limb length (HLL), upperhind-limb length (UHL), lower hind-limblength (LHL), and thorax-abdomen length(TA) were measured (Fig. 1). All measure-ments were made to 61 mm, after position-ing the body in the approximate shape shownin Figure 1. The recommendation of Rey-ment et al. (’84) was adopted in selecting 12measurements (since 10 is considered opti-mal). Only dimensions likely to show mini-mal shrinkage after preservation (i.e., notsoft tissue) were included. The sex of eachgoanna was determined by examination ofthe gonads.

Body sizeTo determine whether SVL, TL, or TAbest

characterized ‘‘body size,’’ the ratio of anyparticular dimension to either TL, SVL, orTA was compared by analysis of variance(ANOVA) for adult specimens, using theminimum SVL adopted for each species toaccount for affects of growth on changingbody proportions, to best discriminate be-tween species (Bookstein, ’91: pp. 83–87).Both raw and logarithmically transformeddata were analyzed in this fashion.The overall best separation of species by

TL, SVL, or TA determined that TA was thebest index of ‘‘body size.’’ The highest num-ber of significant F-ratio values was ob-tained for the body dimension divided by TL,when logarithmically transformed (Table 1).However, during the measurement of goan-nas, it was apparent for numerous individu-als that the end of their tail had been bro-ken. Where the end of the tail had obviouslybeen broken off the measurement was notused, but for many individuals it was notpossible to determine if a small tip of the tailhad been broken off. Therefore, as theF-ratios for values of the logarithmicallytransformed appendage length divided bylogarithmically transformed TAwere gener-ally only slightly less than those for TL (whenTL was the higher; Table 1), we chose TA asthe best measure of overall body size tosubsequently discriminate between the rela-tive appendage dimensions of the variousspecies, and TA was used to determine therelative appendage proportion dimension foreach species.An outline of the body shape was obtained

by scanning a photograph of an adult of eachspecies. These body outlines were adjustedto the same relative TA using Corel Draw(V6.0), to enable a visual inspection of rela-tive body proportions (Fig. 2).

Isometric similarityThe extent of isometric similarity among

goanna species was determined by examin-ing slopes of the regression lines for themeans of each species of logarithmicallytransformed body appendage dimensionswith the logarithmically transformed TA,i.e., untransformed relationships, append-age (mm) 5 aTAb. If the body proportionsare isometrically similar, then b 5 1.0 (or,statistically, b is not significantly differentfrom 1.0).

128 G.G. THOMPSON AND P.C. WITHERS

Fig.1.

Measurementsforgoannas.

Sexual dimorphismPossible morphological differences be-

tween sexes were determined separately foreach species, using discriminant analysis ofthe logarithmically transformed data for HL,HW, NECK, UFL, LFL, UHL, LHL, TAIL,and TA for each specimen. Stepwise discrim-inant analysis was then used to determinewhich dimensions contributedmost to sexualdimorphism.Analysis of covariance (ANCOVA)was also

used to determine whether any single vari-able (logHL, logHW, logFLL, logHLL, or log-NECKwith logTAas the covariate to removethe effect of size) could be used to separatesexes, for each species.

Body appendage dimensionsLogarithmically transformed body append-

age dimensions were regressed against thelogarithmically transformed TA for the vari-ous species, and standardized residual val-ues were used to examine the extent thatrelative appendage dimension was not ex-plained by size. ANOVA was used to exam-ine differences between species for thesestandardized residuals, and t-testswere usedto determine if the residual appendage di-mensions of individual species differed sig-nificantly from zero.

Morphometric analysisIn our analyses to determine the extent to

which morphological characteristics catego-rized individuals by their correct species orspecies by their correct subgenus, we havebeen heavily influenced by the view of Book-stein et al. (’85: p. 27) that size ought not tobe removed from observed measures as it

often explains meaningful covariance in themorphological variance, and the commentsof Klingenberg (’96), who reports that Burna-by’s (’66: p. 35) procedure to eliminate theeffects of growth only works when all groupsshare a common allometric pattern. Forgroups that differ in their size vectors, as weknow goannas do (unpublished observa-tions), removing all size vectors from thedata may leave non-meaningful variation(Humphries et al., ’81). Canonical variateanalysis using the within groups covariancematrix (SPSS-PC) was used to determinethe interrelationships among all species ofgoannas for the logarithmically transformeddata (Reyment et al., ’84). Eigenvalues .1.0were used to indicate which canonical vari-ate functions were significant; confidencelimits in all tests were P , 0.05.

RESULTSBody and appendage dimensions

For the 17 species (and 2 subspecies) ofWesternAustralian goanna that were exam-ined, there was a predominance of sexuallymature individuals with only a small num-ber of juveniles and subadults for each spe-cies. Mean values for the linear dimensionsof each species are given in Table 2 alongwith the range in SVL.Table 3 summarizes the appendage dimen-

sions in proportion to TA, our estimate ofbody size, for adult individuals of each of the17 goanna species. Our estimate of mini-mum SVL for adults of each species is in-cluded in Table 3. Table 4 summarizes thestandardized residuals (from regressionanalysis of those log-transformed measure-ments against logTA) for each appendagemeasurement for the 17 goanna species. Ingeneral, if a species has one short append-age (e.g., HL), then the other appendagesare also short (e.g., V. brevicauda); or if oneappendage is long, then the other append-ages are also long (e.g., V. glauerti). Thereare highly significant intercorrelations be-tween residuals of all length measurements,except HW and HD, for all individual speci-mens (Table 5).

Discriminant classification of individualsby species

Canonical variate analysis, using a withingroups covariance matrix, correctly classi-fied 81.3% of the 562 individual goannas tospecies, using the logarithmically trans-formed values of HL, HD, HW, NECK, UFL,

TABLE 1. F-ratio values determined by ANOVAfor relative appendage dimensions of goannas divided

by TL, SVL, or TA*

Variables

Denominators

TL SVL TA logTL logSVL logTA

TAIL 102.1 110.4 118.4 198.0 120.3 108.3HL 86.8 34.1 33.9 188.6 188.2 150.3HW 91.6 30.2 24.3 71.1 60.5 71.6HD 48.0 13.7 12.1 49.5 42.4 50.6FLL 30.1 31.8 55.9 195.6 167.4 154.4UFL 24.0 43.5 70.0 199.3 195.9 213.1LFL 39.3 37.2 59.6 260.4 231.0 214.5HLL 35.2 71.2 85.8 265.1 226.8 184.1UHL 34.7 68.4 86.7 256.2 257.9 254.6LHL 47.5 90.9 105.6 384.7 335.4 287.4NECK 28.3 54.3 82.9 156.1 154.6 165.5

*The highest P-value is underlined.


Fig.2.

Artist’s

draw

ingofgoannas

scaled

tothesameTA(top

)andarrangedinascendingorderbasedon

TA(bottom).

LFL, UHL, LHL, TA, and TAIL (Table 6).The analysis correctly classified 100% of theindividuals ofV. glauerti, V. glebopalma, andV. brevicauda. The order of all 18 species/subspecies in Table 6 has been arrangedaccording to morphological affinity using themisclassification of species from the canoni-cal variate analysis and the data in Tables 3and 4.V. acanthuruswas themostmisclassi-fied species of the subgenus Odatria, beingmiscategorized as six other species (includ-ing one subgenus Varanus). V. gouldii wasthe most misclassified of the subgenusVaranus, beingmiscategorized as seven otherspecies (including three subgenus Odatria).Although the classification of individuals intotheir correct species was not always 100%,we considered the discrimination sufficientto justify further examination of the canoni-cal variate analysis.The first three canonical variates had eig-

envalues .1.0 and accounted for 89.7% ofthe total variance in morphology for all spe-cies measured, whereas the fourth and sub-sequent canonical variates had eigenvalues,1.0 (Table 7). The first and second canoni-cal variates clearly separated the two sub-genera, Odatria and Varanus (Fig. 3; thisfigure shows males and females separately,because there is sexual dimorphism). TAILwas the measurement most highly corre-lated with the first canonical variate (Table7), but residual TAILwas significantly corre-lated with all other variables except HD andHW (Table 5). The first canonical variatedoes not seem to be a pure indicator of size,although there is a general size effect espe-cially when considering the subgenera sepa-rately. That is, the mean TA of each speciesdoes not rank strictly in accordance with thefirst canonical variate score, although thereis a general trend for small species to anegative first canonical variate score andlarge species to a positive score. The secondcanonical variate correlated best with HD(Table 7; Fig. 3). The third canonical variatedoes not separate the species into the twosubgenera (Fig. 4); it is correlated best withNECK (Table 7).Because the first two canonical variates

clearly separated the subgeneraOdatria andVaranus, as proposed by Mertens (’42) andsupported by Baverstock et al. (’93), we ana-lyze these subgenera separately below.

Subgenus OdatriaTAIL correlated best with the first canoni-

cal variate (Table 7) for the Odatria group.

TABLE2.

Numberofspecim

ensexam

ined,m

axim

umandminim

umSVL,andthemean(6SD)forotherbodyappendagedim

ensions(m

m)for17

speciesofgoannas

Varanus

N

Mini-

mum

SVL

Maxi-

mum

SVL

SVL

HL

NECK

HW

HD

FLL

UFL

LFL

HLL

UHL

LHL

TA

TAIL

brevicauda

4051

126

99.1

616.3

17.1

62.2

14.9

62.7

8.9

61.4

6.2

61.3

19.7

62.6

5.7

61.2

15.6

61.9

23.5

63.2

7.3

61.3

17.4

62.1

66.9

612.5

85.5

625.7

caudo-

lineatus

6773

131

106.8

610.0

19.3

61.8

20.6

63.0

10.5

61.2

6.6

60.8

25.6

63.1

7.5

61.1

19.4

62.5

33.1

63.6

10.9

61.6

24.2

62.5

67.6

67.0

106.9

652.5

storri

2492

134

114.5

612.7

23.4

62.6

20.9

62.7

12.0

61.8

7.8

62.4

29.6

63.7

9.5

61.5

22.5

62.7

40.2

64.7

12.7

61.9

29.7

63.2

70.7

69.4

151.2

672.4

gilleni

26103

175

134.6

621.1

23.0

62.9

26.6

64.2

12.2

62.0

7.5

61.5

31.3

64.9

8.9

61.5

23.2

63.5

39.6

66.7

11.7

62.0

29.0

64.3

84.3

614.4

163.7

646.5

pilbarensis

1067

180

137.4

631.9

26.9

65.5

29.7

67.2

12.9

62.5

7.4

61.7

39.6

69.9

12.8

63.2

28.4

66.4

52.6

613.6

18.0

64.9

38.0

69.9

79.2

620.7

215.1

693.8

erem

ius

5468

185

139.4

623.7

26.7

63.8

25.5

66.1

13.0

62.2

9.3

61.6

33.9

65.4

10.5

62.1

24.9

63.9

51.5

68.4

15.8

62.7

37.6

65.7

86.8

65.2

229.0

663.0

scalaris

5672

268

170.6

634.5

29.9

64.7

35.7

67.6

14.0

62.8

10.3

62.1

41.6

68.0

12.4

63.1

31.9

65.6

55.5

610.6

18.1

63.9

41.9

67.5

104.7

622.5

181.5

6119.2

acanthurus

3690

220

178.6

630.6

31.9

64.1

35.7

66.9

15.6

62.6

10.7

61.8

44.8

67.8

13.8

62.7

34.4

66.1

62.2

612.8

19.8

64.1

45.5

67.8

111.2

621.5

226.1

6132.5

mitchelli

23118

253

193.3

639.8

33.8

65.7

41.3

68.6

15.1

63.4

9.9

62.2

46.0

69.2

13.1

62.6

36.2

67.4

63.7

611.8

20.2

64.3

48.3

68.9

117.6

626.9

303.7

6116.6

glauerti

2890

239

199.6

635.8

37.5

65.7

48.8

69.8

15.9

62.7

9.3

61.9

53.7

610.4

18.2

64.2

40.9

68.0

72.5

613.3

25.6

65.0

53.7

610.1

116.0

623.5

369.1

6174.3

tristis

5368

290

208.7

651.4

37.5

67.6

46.7

612.0

17.2

63.6

11.8

62.8

54.4

612.5

17.1

64.5

41.7

69.7

76.9

618.2

24.5

66.2

56.7

613.1

124.8

631.7

353.8

6111.7

p.panoptes

12143

510

252.0

6122.0

47.5

616.9

53.8

626.6

21.7

68.7

15.7

66.6

68.5

633.0

23.2

612.1

52.5

626.9

96.5

648.3

31.8

617.8

72.3

66.4

153.1

681.0

251.7

6160.5

gouldii

76107

590

277.3

6106.7

48.0

614.4

56.4

624.0

22.3

67.1

16.5

65.4

72.5

626.5

23.7

68.9

55.2

621.2

104.5

638.4

34.7

613.8

77.8

629.5

171.7

667.9

374.1

6184.9

glebopalma

31152

397

297.9

649.1

51.1

66.9

82.6

615.8

23.2

63.7

16.4

62.7

79.8

613.6

30.2

65.0

59.3

69.3

115.9

620.4

43.8

68.3

82.6

613.1

166.8

627.4

428.1

6246.3

rosenbergi

38150

422

319.2

665.5

58.2

610.3

66.8

614.7

28.1

65.8

21.1

64.5

84.2

617.3

28.2

66.3

67.7

614.3

116.8

622.6

40.9

69.1

89.7

617.4

193.7

642.1

423.3

6170.3

mertensi

26150

460

320.6

678.0

49.6

69.4

76.1

621.9

24.6

65.3

17.1

63.1

78.3

621.0

25.6

67.3

61.3

616.9

111.3

629.3

37.6

611.6

85.2

621.6

193.8

650.6

435.0

6143.5

p.rubidus

17141

535

350.8

6132.9

61.0

617.3

77.7

632.9

26.9

68.5

21.1

66.7

100.4

639.0

34.2

615.0

75.8

630.0

147.1

657.9

48.9

620.8

110.2

644.4

216.6

683.2

564.0

6218.1

giganteus

25159

660

442.2

6124.3

82.0

620.1

114.2

636.0

34.0

69.1

25.1

67.9

130.0

637.7

43.4

613.0

97.1

626.1

174.4

648.9

60.1

618.3

129.0

635.7

245.1

668.9

467.4

6306.6


TABLE3.

Bodyappendageratiosfortheadultindividualsforeach

ofthe17

speciesofgoanna1

Varanus

Minimum

adultSVL

(mm)

TAIL/TA

HL/TA

HW/TA

HD/TA

NECK/TA

FLL/TA

UFL/TA

LFL/TA

HLL/TA

UHL/TA

LHL/TA

brevicauda

901.34

60.109

0.25

60.024

0.13

60.015

0.09

60.012

0.22

60.036

0.29

60.031

0.09

60.012

0.23

60.024

0.35

60.035

0.11

60.016

0.26

60.028

caudolineatus

100

1.93

60.158

0.28

60.021

0.15

60.014

0.10

60.010

0.30

60.032

0.38

60.038

0.11

60.016

0.28

60.027

0.49

60.042

0.16

60.020

0.36

60.032

storri

100

2.55

60.269

0.33

60.022

0.17

60.017

0.11

60.031

0.30

60.028

0.42

60.032

0.13

60.015

0.32

60.021

0.56

60.041

0.18

60.017

0.41

60.029

erem

ius

130

2.76

60.232

0.31

60.018

0.15

60.010

0.10

60.008

0.30

60.034

0.39

60.029

0.12

60.017

0.28

60.021

0.59

60.038

0.18

60.016

0.43

60.027

gilleni

130

2.08

60.157

0.27

60.019

0.14

60.010

0.09

60.008

0.31

60.022

0.37

60.027

0.10

60.011

0.27

60.023

0.47

60.047

0.14

60.013

0.34

60.027

pilbarensis

130

2.95

60.436

0.33

60.033

0.16

60.015

0.09

60.011

0.37

60.035

0.49

60.088

0.16

60.031

0.35

60.054

0.66

60.109

0.23

60.044

0.47

60.066

scalaris

150

2.38

60.167

0.28

60.020

0.13

60.012

0.10

60.010

0.34

60.034

0.39

60.026

0.12

60.013

0.30

60.023

0.53

60.042

0.17

60.016

0.40

60.028

acanthurus

160

2.63

60.418

0.28

60.023

0.14

60.013

0.09

60.009

0.32

60.029

0.40

60.033

0.12

60.010

0.31

60.027

0.56

60.058

0.18

60.024

0.41

60.044

tristis

200

2.93

60.212

0.29

60.015

0.13

60.009

0.09

60.006

0.38

60.030

0.43

60.032

0.14

60.012

0.33

60.021

0.61

60.048

0.20

60.015

0.45

60.026

mitchelli

200

2.73

60.141

0.27

60.013

0.13

60.013

0.08

60.010

0.34

60.038

0.38

60.022

0.11

60.010

0.30

60.200

0.53

60.045

0.17

60.015

0.40

60.023

glauerti

200

2.78

60.282

0.31

60.016

0.13

60.008

0.08

60.007

0.42

60.035

0.46

60.035

0.16

60.016

0.35

60.023

0.62

60.050

0.22

60.019

0.46

60.026

p.panoptes

200

2.17

60.414

0.29

60.030

0.13

60.015

0.09

60.017

0.34

60.039

0.44

60.032

0.15

60.017

0.33

60.027

0.62

60.058

0.22

60.027

0.47

60.037

glebopalma

300

3.26

60.318

0.30

60.012

0.14

60.007

0.10

60.005

0.50

60.029

0.48

60.025

0.18

60.011

0.35

60.014

0.70

60.025

0.26

60.017

0.49

60.023

rosenbergi

300

2.45

60.107

0.30

60.015

0.14

60.010

0.11

60.009

0.35

60.038

0.43

60.036

0.14

60.013

0.45

60.021

0.60

60.046

0.21

60.019

0.46

60.031

gouldii

300

2.33

60.139

0.27

60.014

0.13

60.009

0.09

60.008

0.33

60.030

0.42

60.026

0.14

60.015

0.32

60.019

0.60

60.038

0.20

60.016

0.45

60.026

mertensi

300

2.34

60.276

0.25

60.015

0.12

60.009

0.09

60.007

0.39

60.041

0.40

60.028

0.13

60.011

0.32

60.021

0.58

60.044

0.20

60.017

0.44

60.020

p.rubidus

300

2.63

60.187

0.26

60.021

0.12

60.008

0.09

60.006

0.37

60.029

0.47

60.046

0.16

60.021

0.36

60.026

0.70

60.050

0.23

60.021

0.52

60.029

giganteus

400

2.59

60.216

0.33

60.021

0.14

60.011

0.10

60.009

0.47

60.045

0.53

60.035

0.18

60.013

0.40

60.027

0.71

60.058

0.25

60.026

0.52

60.039

rwithTA,allspecies(P)

0.22

(0.37)

0.18

(0.47)

20.54

(0.02)

0.01

(0.97)

0.61

(0.01)

0.59

(0.01)

0.64

(0.01)

0.69

(0.01)

0.67

(0.01)

0.68

(0.01)

0.73

(0.01)

rwithTA,O

datria(P)

0.70

(0.01)

0.01

(0.99)

20.49

(0.11)

20.28

(0.38)

0.84

(0.01)

0.50

(0.10)

0.61

(0.04)

0.58

(0.05)

0.61

(0.03)

0.66

(0.02)

0.62

(0.03)

rwithTA,Varanus(P)

0.97

(0.01)

20.31

(0.55)

0.04

(0.94)

0.15

(0.77)

0.76

(0.08)

0.80

(0.05)

0.77

(0.07)

0.37

(0.47)

0.88

(0.02)

0.78

(0.07)

0.87

(0.03)

1 Correlation

coefficientsareforproportionalappendage

lengthwithTA;valuesaremeans

6SE.

The logarithmically transformed residualsfor TAIL are significantly and positively cor-related with UHL and LHL (r 5 0.79 and0.87, respectively), suggesting that thelength of the hind appendages is the pri-mary determinant in separating species onthe first canonical variate (Fig. 5). NECK isthe measurement most highly correlatedwith the second canonical variate (Table 7).The first canonical variate is not a pure sizeindicator as the species mean TA is notstrictly ranked in accordance with the firstcanonical variate. The third canonical vari-ate (Fig. 6) has the strongest negative corre-lation with TA (although relatively weak,r 5 20.24); it separates V. caudolineatusfrom V. gilleni, and V. glebopalma from V.glauerti (which are closely aligned on thefirst and second canonical variates), and V.mitchelli fromV. scalaris andV. tristis (whichare closely associated by the second canoni-cal variate). There is considerable overlap in‘‘morphological space’’ for individuals of mostspecies of the Odatria clade based on thefirst three canonical variates (Figs. 5, 6).Nevertheless, V. brevicauda is clearly sepa-rated from the other species, primarily bythe first canonical variate, while V. glebo-palma and V. glauerti form an overlappinggroup toward the other end of this axis. V.eremius is separated from most of the otherspecies by the second canonical variate, over-lapping slightlywithV. storri andV. acanthu-rus, and to a much lesser extent with V.mitchelli (Fig. 5). V. caudolineatus and V.gilleni form a group located between themajority of the Odatria species and V. brevi-cauda, being separated from the other spe-cies primarily by the first canonical variate.

Subgenus VaranusThe highest correlation with the first ca-

nonical variate for the subgenus Varanus isHL (Table 7). However, Tabachnick and Fi-dell (’89: p. 539) suggest that correlationslower than 0.30 cannot be interpreted ad-equately, although the interpretation of HLas the primary discriminator is in accor-dance with the univariate data (Tables 3, 4),which indicate thatV.mertensi andV. gigan-teus have the smallest and largest relativeHL for this group. The second canonical vari-ate is most strongly correlated (r 5 0.29)with NECK, indicating that V. giganteusand V. mertensi have relatively longer necksthan the other species (Fig. 7), a result con-cordant with data in Tables 3 and 4. Eigen-values are ,1.0 for subsequent functions

TABLE4.

Speciesstandardized

residualappendagedeviationsfrom

theregression

lineofallVaranusspecieswithTA1

Varanus

Mean

TA

(mm)

HL

NECK

HW

HD

FLL

UFL

LFL

HLL

UHL

LHL

TAIL

brevicauda

66.90

21.58

60.82

21.72

61.17

21.30

60.99

20.63

60.88

22.01

60.87

21.65

60.91

21.76

60.96

22.33

60.77

22.07

60.96

22.36

60.91

22.27

60.39

caudolineatus

67.54

20.68

60.63

20.03

60.65

0.10

60.91

20.26

60.80

20.24

60.75

20.25

60.78

20.30

60.91

20.39

60.57

20.12

60.71

20.47

60.61

20.72

60.38

storri

70.71

0.64

60.56

20.20

60.62

1.02

60.93

0.96

61.04

0.46

60.59

0.71

60.69

0.49

60.58

0.49

60.56

0.43

60.69

0.48

60.59

0.49

60.49

gilleni

84.27

20.88

60.61

20.01

60.56

20.21

60.62

20.94

60.67

20.54

60.58

20.74

60.59

20.75

60.72

20.84

60.62

21.12

60.59

20.91

60.59

20.67

60.42

pilbarensis

79.20

1.04

60.93

1.07

60.67

1.02

61.02

20.45

60.72

1.72

61.29

1.63

61.13

1.43

61.28

1.37

61.11

1.54

61.09

1.21

60.97

1.05

60.69

erem

ius

86.85

0.19

60.51

20.52

60.56

0.12

60.69

0.56

60.60

20.17

60.68

20.08

60.76

20.43

60.79

0.57

60.55

0.25

60.57

0.49

60.59

0.66

60.38

scalaris

104.71

20.28

60.60

0.18

60.63

20.68

60.80

0.02

60.73

20.17

60.64

20.38

60.67

20.09

60.77

20.28

60.61

20.24

60.59

20.16

60.63

20.03

60.36

acanthurus

111.17

20.20

60.79

20.24

60.65

20.10

60.91

20.06

60.82

20.12

60.70

20.19

60.57

20.05

60.75

20.06

60.63

20.14

60.74

20.10

60.78

0.45

60.75

mitchelli

117.61

20.14

60.58

0.26

60.79

20.89

60.91

21.08

60.87

20.35

60.59

20.77

60.74

20.11

60.70

20.22

60.77

20.35

60.70

20.10

60.69

0.56

60.41

glauerti

115.96

0.53

60.71

1.26

60.66

20.23

60.64

21.45

60.81

0.88

60.72

1.00

60.59

0.92

60.67

0.64

60.69

0.96

60.74

0.63

60.59

1.89

60.45

tristis

124.83

0.36

60.77

0.53

60.61

0.04

60.86

20.15

60.69

0.43

60.78

0.26

60.74

0.48

60.72

0.49

60.67

0.25

60.60

0.47

60.69

0.79

60.44

p.panoptes

153.10

1.10

61.22

0.12

60.88

0.71

61.43

0.75

61.50

0.61

60.81

0.67

60.77

0.61

60.88

0.55

60.83

0.36

60.77

0.63

60.83

20.22

60.81

gouldii

171.75

0.14

60.97

20.54

60.66

0.01

60.83

0.30

60.96

0.06

60.82

0.08

60.91

20.02

60.78

0.22

60.80

0.12

60.84

0.23

60.81

20.28

60.51

glebopalma

166.81

0.61

60.34

1.72

60.36

0.42

60.47

0.22

60.45

0.81

60.43

1.38

60.40

0.63

60.43

0.87

60.37

1.35

60.34

0.65

60.36

1.02

60.44

rosenbergi

193.66

0.59

60.42

20.43

60.67

0.99

60.61

1.09

60.59

0.03

60.62

0.09

60.58

0.40

60.53

20.08

60.59

0.08

60.53

0.12

60.58

20.20

60.38

mertensi

193.77

20.72

60.72

0.28

60.64

20.21

60.74

20.44

60.67

20.51

60.57

20.42

60.57

20.41

60.60

20.41

60.51

20.42

60.46

20.21

60.44

20.51

60.57

p.rubidus

216.60

20.12

60.86

20.37

60.54

20.15

60.67

0.46

60.59

0.45

60.72

0.34

60.70

0.32

60.67

0.53

60.57

0.26

60.53

0.56

60.53

20.14

60.36

giganteus

245.10

1.73

60.65

0.97

60.53

0.95

60.76

0.66

60.79

1.32

60.57

0.88

60.57

1.21

60.69

0.73

60.74

0.58

60.74

0.70

60.74

20.03

60.58

1 SpeciesarerankedaccordingtoincreasingTA.


and are therefore not considered to be signifi-cant. Figure 7 suggests that the morphologyof bothV. mertensi andV. giganteus is appre-ciably different from that of the other spe-cies in this clade, with both species beingseparated from the others primarily by thefirst canonical variate. There is considerableoverlap in the morphology of V. rosenbergi,V. panoptes, andV. gouldii (Fig. 7).V. p. panop-tes and V. p. rubidus are separated by V.rosenbergi (Fig. 7).

Sexual dimorphismThe specimens measured in the WAM col-

lection were predominantly male (Table 8).There was significant sexual dimorphism formost species in body shape, as determinedby a discriminant analysis for all specimensof each species (that were able to be sexed)using a combination of logarithmically trans-formed appendage dimensions (Table 8). The

combined standardized canonical discrimi-nant function coefficients that separated thesexes for all species of goannas are shown inTable 9.Stepwise discriminant analysis was subse-

quently used to determine which variable(s)contributed most to the separation of sexes.F-values were insufficient to isolate signifi-cant individual variables for some species.However, logHL and logTA correctly classi-fied 82.5% of V. caudolineatus (standardizedcanonical discriminant function coefficients[scdfc] 2 1.346logTA 1 1.790logHL); log-NECK correctly classified 69.7% of V. brevi-cauda; logUFL correctly classified 75.5% ofV. eremius; logUFL correctly classified 73.1%of V. glauerti; logTAIL, logUFL, and logUHLcorrectly classified 100% of V. glebopalma(scdfc1.109logTAIL21.396logUFL11.665log-UHL); logHD and logUHL correctly classified68.1% of V. gouldii (scdfc 3.913logHD 2

TABLE 5. Correlation coefficients (r) between residual appendage lengths for all specimens*

HW HD NECK FLL UFL LFL HLL UHL LHL TAIL

HL 0.80 0.48 0.60 0.89 0.85 0.91 0.80 0.64 0.86 0.62HW 0.64 0.39 0.75 0.77 0.74 0.71 0.64 0.70 0.36HD 20.18 0.25 0.29 0.26 0.33 0.06 0.35 20.09NECK 0.77 0.75 0.76 0.70 0.76 0.67 0.74FLL 0.95 0.98 0.94 0.83 0.93 0.73UFL 0.93 0.92 0.82 0.89 0.72LFL 0.91 0.75 0.91 0.73HLL 0.80 0.99 0.82UHL 0.76 0.71LHL 0.81

*Values underlined are P , 0.05.

TABLE 6. Proportion of individuals correctly allocated to species using canonical variate analysis (logarithmicallytransformed variables of HL, HW, HD, NECK, UFL, LFL, UHL, LHL, and TA) for Odatria and Varanus subgenera

Varanus/subspecies

Speciesno.

Predicted species/subspecies%

Correct1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

glauerti 1 24 100.0glebopalma 2 24 100.0pilbarensis 3 7 1 1 77.8mitchelli 4 18 3 1 85.7tristis 5 2 5 39 1 2 1 1 76.5eremius 6 46 3 2 1 88.5acanthurus 7 3 2 2 12 1 7 1 42.9storri 8 1 1 15 1 1 1 78.9scalaris 9 4 36 1 87.8gilleni 10 1 23 1 92.0caudolineatus 11 1 9 45 81.8brevicauda 12 38 100.0

mertensi 13 2 21 1 1 84.0p. rubidus 14 13 3 1 1 64.3gouldii 15 1 1 2 2 6 45 5 8 64.3rosenbergi 16 3 2 27 2 79.4p. panoptes 17 2 2 6 60.0giganteus 18 1 18 94.7

WESTERN AUSTRALIAN VARANID MORPHOLOGY 135

3.140logUHL); logTA, logLHL, and logTAILcorrectly classified 94.4% ofV.mertensi (scdfc9.032logTA 2 12.889logLHL 1 4.375log-TAIL); logHD and logTA correctly classified100% of V. pilbarensis (scdfc 2 4.578logTA14.784logHD); logLFL, logLHL, logUHL, andlogUHL correctly classified 90.0% ofV. p. pan-optes (scdfc 2 30.234logLFL 1 12.606log-LHL 2 4.117logLFL 1 21.948logUHL); andlogTA, logHL, logLHL, and logTAIL cor-rectly classified 94.7% of V. storri (scdfc5.381logTA2 4.833logHL 2 2.927logLHL 12.426logTAIL). No single logarithmicallytransformed variable could separate thesexes for each of the species (Table 10).Where a significant difference existed be-tween sexes,males had proportionally longerappendages or, alternatively, the TA of thefemales was proportionally longer than formales.

IsometryAlthough all body appendage dimensions

are significantly and positively correlatedwith TA, only HL, HD, and TAIL scaledisometrically with TA for all species (i.e., theslope b was not significantly different from1.0; Table 10). For Odatria, FLL as well asHL and HD scale isometrically with TA. Thesmaller sample size and hence higher stan-dard error of the slope for the subgenusVaranus probably account for the highernumber of appendage dimensions that scaleisometrically with TA (Table 10). The re-sidual body appendage dimensions, other

than HD, are significantly correlated; all ofthe significant correlations were positive,i.e., any goanna species with a relativelylong (or short) appendage dimension alsohad relatively long (or short) other append-ages (except HD; Table 4).

Grouping of species based on the relativelength of appendages

Figures 8–11 show the relative appendagesizes as a proportion of TA for each Varanusspecies/subspecies.ANOVAfor the standard-ized residual values (from the regressionwith logTA) of appendage dimensions showeda significant difference between species forHL (F17,624 5 37.52,P, 0.001), HW (F17,624 519.74, P , 0.001), HD (F17,623 5 23.57, P ,0.001), NECK (F17,624 5 46.49, P , 0.001),FLL (F17,624 5 37.59, P , 0.001), UFL(F17,624 5 34.73, P , 0.001), LFL (F17,624 529.03, P , 0.001), HLL (F17,624 5 50.25, P ,0.001), UHL (F17,624 5 43.05, P , 0.001),LHL (F17,624 5 46.27, P , 0.001), and TAIL(F17,545 5 117.51, P , 0.001). Those relativeappendage dimensions for which the residu-als were significantly different from 0 arecircled in Figures 8–11.A visual inspection of Figures 8–11 when

interpreted in conjunction with Tables 2–4indicates the following:1. V. brevicauda has a relatively shorter

head, neck, fore- and hind-limbs, and tailthan any other species (Tables 3, 4; Figs.8–11).

TABLE 7. Pooled within group correlations between discriminant variables and canonical discriminant functionsfor all 17 species, and separately for the Odatria and Varanus subgenera

Variables

All species Subgenus Odatria Subgenus Varanus

Function1

Function2

Function3

Function1

Function2

Function3

Function1

Function2

Function3

logNECK 0.408 0.205 0.332 0.418 0.342 20.148 0.136 0.286 0.304logTA 0.289 0.270 0.154 0.258 0.147 20.239 0.074 0.117 0.236logUFL 0.449 0.280 0.225 0.424 0.217 0.024 0.170 0.121 0.323logHW 0.354 0.312 0.157 0.296 0.116 20.030 0.146 0.122 0.175logHL 0.463 0.348 0.186 0.423 0.127 20.203 0.220 0.136 0.232logTAIL 0.494 0.094 0.101 0.556 0.076 20.218 0.124 0.115 0.311logUHL 0.461 0.257 0.172 0.459 0.164 0.034 0.142 0.100 0.310logHD 0.350 0.389 0.051 0.260 20.021 20.064 0.149 0.046 0.213logLHL 0.482 0.300 0.121 0.487 0.079 20.122 0.136 0.118 0.340logLFL 0.422 0.300 0.228 0.410 0.228 20.183 0.159 0.128 0.288

Eigenvalue 12.977 4.172 1.813 15.14 2.31 1.11 4.15 1.34 0.66% Variance 61.38 19.73 8.58 74.83 11.41 5.50 63.70 20.59 10.04Wilks’ lambda 0.012 0.061 0.173 0.038 0.13 0.27 0.186 0.436 0.721Chi-square 2,424 1,525 959 1,226 777 496 279 137 54df 144 120 98 90 72 56 36 24 14P ,0.0001 ,0.0001 ,0.0001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001


Fig. 3. The first and second canonical variates for morphometrics of Western Australian goannas, based onlogarithmically transformed data for different sexes. The values marked with an F are for females of each species.The first standardized canonical vector for the datanot separated by sex is:23.505logTA20.517logHD12.291logHL20.277logHW 2 0.776logLFL 1 1.497log 2 LHL 2 0.182logNECK 1 0.285logUFL 1 0.187logUHL 1 1.414logTAIL.The second standardized canonical vector is: 2 0.004logTA 1 1.156logHD 1 1.994logHL 2 0.638logHW 10.418logLFL 1 1.649logLHL 2 0.631logNECK 2 0.097logUFL 2 0.409logUHL 2 3.203logTAIL.

Fig. 4. The first and third canonical variates for Western Australian goannas based on logarithmically trans-formed data. See Figure 3 for the first canonical vector. The third canonical vector is: 20.627logTA2 1.496logHD 10.840logHL 1 0.501logHW 1 2.253log 2 LFL 2 3.186logLHL 1 2.246logNECK 1 0.673logUFL 1 0.160logUHL 21.146logTAIL.


Fig. 5. The first and second canonical variates for morphometrics of the Odatria clade of goannas based onlogarithmically transformed data showing the extreme values for different individuals as a convex polygon. The firststandardized canonical vector is: 22.231logTA 2 0.866logHD 1 0.947logHL 1 0.034logHW 2 0.466log 2 LFL 10.878logLHL1 0.091logNECK 1 1.698logTAIL1 0.139logUFL1 0.195logUHL. The second standardized canonicalvector is: 20.024logTA 2 1.419logHD 2 0.031logHL 1 0.486logHW 1 1.875logLFL 2 2.430 log 2 LHL 11.606logNECK2 0.785logTAIL1 0.659logUFL1 0.319logUHL.

Fig. 6. The first and third canonical variates for the Odatria clade of goannas based on logarithmicallytransformed data showing the extreme values for different individuals as a convex polygon. The first standardizedcanonical vector is: 22.231logTA 2 0.866logHD 1 0.947logHL 1 0.034logHW 2 0.466logLFL 1 0.878log 2 LHL 10.091logNECK 1 1.698logTAIL 1 0.139logUFL 1 0.195logUHL. The third standardized canonical vector is:21.316logTA 1 0.550logHD 2 2.618logHL 1 1.375logHW 2 0.499logLFL 1 0.706logLHL 1 0.062logNECK 20.725logTAIL 1 1.051logUFL 1 1.468logUHL.


2. V. glebopalma, V. giganteus, V. glauerti,and V. pilbarensis have an appreciablylonger neck (0.6 SD above themean) thanother species, whereas V. giganteus, V.pilbarensis, and V. p. panoptes have anappreciably longer head (0.6 SD abovethe mean). The head length of V. gilleniand V. mertensi is appreciably shorter(0.6 SD below the mean) than for allother species, except V. brevicauda. Thehead of V. storri, V. pilbarensis, V. rosen-bergi, V. giganteus, and V. p. panoptes isappreciably wider (0.5 SD above themean) than for other species, while thatof V. mitchelli and V. scalaris is apprecia-bly narrower (0.5 SD below the mean)than for all other species, except V. brevi-cauda. The head depth for V. storri, V.rosenbergi, V. p. panoptes, V. giganteus,and V. eremius is appreciably deeper (0.5SD above the mean) than for other spe-cies, while that of V. glauerti, V. mitchelli,V. gilleni, and V. brevicauda is apprecia-bly shallower (0.5 SD below themean) thanthat of other species (Tables 3, 4; Fig. 8).

3. The fore-limb lengths of V. pilbarensis, V.giganteus, V. glebopalma, and V. glauerti

are appreciably longer (0.6 SD above themean) than for other species, while for V.mertensi and V. gilleni the fore-limb isgenerally shorter (0.3 SD below themean)than for other species, except V. brevi-cauda (Tables 3, 4; Fig. 9).

4. The hind-limbs of V. pilbarensis, V. glebo-palma, and V. glauerti are appreciablylonger (0.6 SD above the mean) than forother species, while V. gilleni has appre-ciably shorter hind-limbs (0.6 SD belowthe mean) than other species, except V.brevicauda (Tables 3, 4; Fig. 10).

5. The tail length of V. glauerti, V. pilbaren-sis, and V. glebopalma is appreciablylonger (0.8 SD above the mean) than forall other species. V. caudolineatus and V.gillenihave anappreciably shorter tail thanany other species (0.5 SD below the mean),exceptV. brevicauda (Tables 3, 4; Fig. 11).

DISCUSSION

Morphometrics is the study of covariancesof biological form. Bookstein (’91) has sug-gested that, for many biological investiga-tions, the most effective way to analyze theform of an organism is to record the geomet-

Fig. 7. The first and second canonical variates for the Varanus clade of goannas based on logarithmicallytransformed data showing the extreme values for different individuals as a convex polygon. The first standardizedcanonical vector is: 25.541logTA 2 0.261logHD 1 7.390logHL 2 1.122logHW 1 1.017logLFL 2 1.274 2 logLHL 21.044logNECK 1 0.413logTAIL 1 0.323logUFL 1 0.219logUHL. The second standardized canonical vector is:20.399logTA 2 2.160logHD 2 0.116logHL 1 1.593logHW 2 0.473logLFL 2 1.514logLHL 1 4.873logNECK 20.739logTAIL 2 0.345logUFL 2 0.600logUHL.


TABLE8.

Determinationofmorphologicaldifferencesformaleandfemalegoannas

of18

species/subspeciesusingdiscrim

inantanalysis,

basedon

logarithmicallytransformed

morphom

etricdata

Varanus

Sex

%Predicted

correctly

Eigenvalue

Wilks’

lambda

score

Chi-

square

P

Group

centroids

Pooledwithingroupcorrelationsbetweendiscriminating

variablesandcanonicaldiscriminantfunctions

MF

MF

logH

LlogH

WlogH

DlogN

ECK

logU

FL

logLFL

logU

HL

logLHL

logTAIL

logTA

brevicauda

1814

90.6

1.67

0.375

,0.01

1.10

21.42

0.126

0.290

0.160

0.322

0.160

0.202

0.025

0.204

0.091

0.014

caudolineatus33

1986.5

1.09

0.478

,0.01

0.78

21.35

0.608

0.296

0.391

0.290

0.255

0.292

0.407

0.489

0.349

0.133

storri

108

100.0

12.9

0.072

0.01

3.03

23.79

0.139

0.128

0.154

20.002

0.055

0.094

0.087

0.158

0.141

20.005

gilleni

129

100.0

10.2

0.089

,0.01

22.63

3.51

20.076

20.039

20.043

20.016

20.009

20.016

20.094

0.003

0.008

0.053

pilbarensis

62

100.0

29.3

0.033

0.11

22.70

8.11

20.047

20.116

20.159

0.057

20.150

20.070

20.056

20.065

20.031

20.053

erem

ius

3017

76.6

0.4

0.700

0.16

0.48

20.85

0.574

0.335

0.412

0.293

0.651

0.513

0.631

0.625

0.347

0.375

scalaris

2310

87.9

0.9

0.518

0.07

20.62

1.42

0.231

0.128

0.142

0.144

0.154

0.235

0.228

0.218

0.155

0.302

acanthurus

136

100.0

5.7

0.149

0.01

1.54

23.33

20.015

0.043

0.058

20.096

20.087

20.032

20.005

20.029

20.047

20.147

mitchelli

118

94.7

2.5

0.283

0.13

21.28

1.77

0.171

0.120

0.145

0.256

0.295

0.243

0.206

0.272

0.220

0.228

glauerti

175

94.4

1.8

0.357

0.12

0.69

22.36

0.271

0.155

0.202

0.266

0.374

0.198

0.248

0.231

0.349

0.169

tristis

2023

76.7

0.2

0.813

0.68

0.50

20.44

0.220

0.149

0.306

0.052

0.235

0.120

0.037

0.199

0.257

0.015

p.panoptes

53

100.0

38.3

0.025

0.09

4.15

26.92

20.132

20.078

20.059

20.151

20.381

20.123

20.198

20.095

20.161

20.157

gouldii

4324

79.1

0.7

0.573

,0.01

20.64

21.14

0.367

0.372

0.417

0.371

0.361

0.337

0.274

0.330

0.309

0.293

glebopalma

174

100.0

8.1

0.110

,0.01

1.31

25.57

0.457

0.348

0.335

0.472

0.377

0.486

0.600

0.466

0.376

0.457

rosenbergi

1911

86.7

0.8

0.560

0.20

0.65

21.13

0.214

0.203

0.276

0.217

0.394

0.266

0.233

0.319

0.205

0.161

mertensi

108

100.0

4.4

0.186

0.05

21.77

2.21

0.187

0.145

0.171

0.144

0.171

0.177

0.166

0.183

0.272

0.216

p.rubidus

112

76.9

0.6

0.616

0.96

0.31

21.70

0.107

0.054

0.116

0.004

0.196

0.117

0.154

0.088

0.128

0.112

giganteus

97

100.0

8.5

0.105

0.03

22.41

3.10

0.066

0.059

0.071

0.035

0.052

0.049

0.092

0.053

0.052

0.055

ric locations of landmark points. Then, themeasurement of shape configurations oflandmark locations can be reduced to mul-tiple vectors of shape coordinates. There-fore, the study of covariance of landmarkconfigurations (triangles) can then be under-taken independent of size. However, the na-ture of the available data (measurements forpreservedmuseum specimens) and the focuson appendage lengths do not enable use ofthis multiple vector approach in this study.Rather, canonical variate analysis, ratios,and residuals from regression equations ofappendage dimensions with TA have been

used here to analyze the size shape of goan-nas. Some authors have been critical of theuse of ratios in the analysis of body shape.Reyment et al. (’84) criticized the use ofratios since 1) the ratio will not be constantfor organisms of the same species by virtueof the almost universal occurrence of differ-ential growth rates; 2) ratios contain onlytwo variables and this affords a poor appre-ciation of contrast between forms with mul-tidimensionality; and 3) to compound twocharacters into a ratio implies that only onecontrast of the form is studied. These criti-cisms have been addressed here in that mul-

TABLE 9. Standardized canonical discriminant function coefficients for distinguishing sex differences in goannas1

Varanus

Standardized canonical discriminant function coefficients

logHL logHW logHD logNECK logUFL logLFL logUHL logLHL logTAIL logTA

brevicauda 20.413 1.753 0.321 0.796 0.282 1.578 20.696 20.225 20.313 22.558caudolineatus 2.027 20.420 0.620 20.098 0.118 20.454 20.135 0.157 20.189 21.274storri 4.891 0.139 20.309 20.452 20.820 2.244 1.359 2.581 22.657 26.936gilleni 26.460 1.092 21.528 0.564 0.623 0.560 22.874 3.555 0.328 4.291pilbarensis 22.900 1.693 211.818 — — 9.962 — 28.150 — 12.345eremius 0.792 20.591 20.270 20.088 0.312 20.018 20.070 2.140 20.920 20.763scalaris 20.989 21.073 20.644 21.406 21.686 3.125 0.957 20.093 22.796 4.705acanthurus 0.444 2.300 3.452 20.323 2.481 21.715 0.059 21.506 1.783 25.880mitchelli 25.146 22.763 0.417 2.262 1.117 1.636 21.716 3.741 0.063 0.721glauerti 20.434 0.048 2.129 1.740 2.428 20.869 20.025 22.553 1.755 23.176tristis 20.269 0.144 1.762 20.355 0.241 20.323 21.168 1.925 0.762 22.454p. panoptes 3.885 23.773 5.498 213.988 — 22.842 — — — 214.258gouldii 4.811 0.652 2.116 0.407 0.805 2.401 22.943 20.629 20.076 27.176glebopalma 20.510 0.759 20.491 20.299 21.972 0.151 2.475 20.633 1.667 0.299rosenbergi 0.110 21.079 1.479 20.296 1.430 1.484 0.184 2.279 21.010 24.162mertensi 21.277 3.770 1.057 20.109 2.009 24.147 22.073 215.058 7.900 8.664p. rubidus 4.407 24.348 3.792 23.570 20.036 8.533 — 212.456 1.616 2.114giganteus 13.114 1.534 2.654 215.436 23.288 218.469 10.620 218.608 23.612 4.721

1Missing values failed the tolerance test.

TABLE 10. Slope (b) of the regression line for appendage dimension and logTA using only values of adultsfor each of the 18 species/subspecies of goanna*

HL NECK HW HD FLL UFL LFL HLL UHL LHL TAIL

All Species (n 5 18)

Slope 0.98 1.27 0.88 0.99 1.18 1.29 1.19 1.25 1.32 1.27 1.106SE 0.048 0.077 0.037 0.047 0.059 0.082 0.049 0.070 0.088 0.066 0.146P NS ,0.05 ,0.05 NS ,0.05 ,0.05 ,0.05 ,0.05 ,0.05 ,0.05 NSr2 0.97 0.94 0.98 0.97 0.97 0.94 0.98 0.97 0.94 0.96 0.77

Odatria (n 5 12)

Slope 1.01 1.54 0.84 0.88 1.23 1.38 1.24 1.36 1.48 1.35 1.596SE 0.084 0.128 0.069 0.087 0.123 0.178 0.105 0.152 0.191 0.147 0.236P NS ,0.05 ,0.05 NS NS ,0.05 ,0.05 ,0.05 ,0.05 ,0.05 ,0.05r2 0.94 0.94 0.94 0.90 0.90 0.86 0.94 0.88 0.86 0.90 0.81

Varanus (n 5 6)

Slope 1.61 1.18 0.97 1.13 1.50 1.54 1.41 1.45 1.46 1.42 1.606SE 0.290 0.33 0.228 0.248 0.20 0.24 0.16 0.154 0.162 0.132 0.693P NS NS NS NS ,0.05 ,0.05 ,0.05 ,0.05 ,0.05 ,0.05 NSr2 0.88 0.77 0.81 0.85 0.94 0.90 0.94 0.96 0.96 0.96 0.58

*P-values for the t-value testing the difference of the slope from 1.0 (isometry) and the coefficient of determination between thelogarithmically transformed appendage length and logTAare also given. NS, not significant.


tiple ratios have been used in comparison ofbody forms, and they have been used inconjunction with a suite of other approaches,most often as collaborating data. Juveniles

and subadult specimens were not includedin the calculation of appendage dimensionratios, thereby minimizing proportionalvariation due to non-isometric growth.

Fig. 8. Head and neck dimensions expressed as apercentage of TA, with species arranged by habitatutilization; a value enclosed in circle represents a mean

residual value from the regression equation with logTAthat is significantly different from the mean for allspecies.

Fig. 9. Upper and lower fore-limb lengths as a per-centage of TA, with species arranged by habitat utiliza-tion; a value enclosed in circle represents a mean re-

sidual value from the regression equation with logTAthat is significantly different from the mean for allspecies.


Size, phylogeny, habitat or ecological niche,growth, and sex are all related to body pro-portions (Peters, ’83; Calder, ’84; Schmidt-Nielsen, ’84; Miles, ’94). There is no estab-

lished and rooted phylogenetic tree forvaranids (Baverstock et al., ’93) and it hastherefore not been possible to properly ac-count for phylogenetic effects in the analysis

Fig. 10. Upper and lower hind-limb lengths as apercentage of TA, with species arranged by habitatutilization; a value enclosed in circle represents a mean


Fig. 11. Fore- and hind-limbs, and tail lengths as apercentage of TA, with species arranged by habitatutilization; a value enclosed in circle represents a mean



of morphology of these goannas, other thanat the subgeneric level. The basic ecology isalso not known for many Western Austra-lian goanna species, and so inferences drawnfrom this ecomorphological analysis ofVaranus must therefore be considered pre-liminary.

Phylogenetic patterns of morphologyThe first and second canonical variates

clearly separate the two goanna subgeneraOdatria and Varanus, consistent with thephylogeny of King and King (’75) and Baver-stock et al. (’93). Our morphological group-ing of species also provides supporting evi-dence that those species not included in theimmunogenetic and karyotypic studies ofKing and King (’75), Holmes et al. (’75), andBaverstock et al. (’93) (e.g., V. pilbarensis, V.storri, V. glebopalma) have been correctlyclassified by King and Green (’93a) and Pi-anka (’95). The present analysis indicatesthat the Odatria clade, with generallysmaller species, is nevertheless morphologi-cally different (in shape) from the Varanusclade, with generally larger species. The twovariables that are best correlated with thefirst and second canonical functions separat-ing the two clades were logTAIL and logHD,respectively.The first canonical variate is often inter-

preted as ‘‘size,’’ but this is not necessarily so(see Klingenberg, ’96). From ourmorphomet-ric analysis, it is clear that the first canoni-cal variate is not just a pure size indicatorfor Western Australian goannas. Figure 3indicates a general size affect (e.g., the small-est V. brevicauda and largest V. glebopalmabeing at either end of the first canonicalvariate scale) but V. pilbarensis, V. acanthu-rus, V. mitchelli, and V. scalaris are all out of‘‘size’’ sequence (based on TA) for the subge-nus Odatria. The overlap of the two subgen-eraOdatria and Varanus on the first canoni-cal variate is also not ‘‘size’’ related. Similarly,the first canonical variates for the two sub-genera when considered separately (Figs. 5,7) are not a pure size indicator.

Sexual dimorphismThe sex of varanids is reported as being

difficult to determine from external morpho-metric characteristics or scalation (Greenand King, ’78; King and Green, ’93b), al-though Auffenberg (’81), Yadav and Rana(’88), andAuffenberg et al. (’91) report differ-ences in cloacal scales for sexes of V. benga-

lensis and V. komodoensis, and Yadav andRana (’88) and Auffenberg (’94) report thatbody micropores can be used to distinguishthe sexes for V. bengalensis. Adult males aregenerally larger than females (Shine, ’86;Auffenberg, ’94; Pianka, ’94), but this is nota useful discriminator for juvenile and small-adult specimens. Using a suite of logarithmi-cally transformed appendage lengths, it waspossible to correctly determine the sex of anumber of species of Varanus with a rela-tively high level of accuracy (Tables 8, 9),although no single character or set of charac-ters was a useful discriminator for all spe-cies of the genus. There were significantdifferences in logarithmically transformedappendage lengths between sexes for someof the dimensions commonly used (HL, HW,NECK, FLL, and HLL) when the effects ofTA were removed for some species (Table11). In all cases, the appendage length wasproportionally longer in males than females,or alternatively, the TA was proportionallylonger in females than males. For most spe-cies, there was little separation between thesexes when the first and second canonicalvariates were plotted (Fig. 3). The greatestseparation of sexes with the first and secondcanonical variates was for V. caudolineatus,V. gilleni, and V. storri. Generally, the maleswere displaced to the more positive value onthe first canonical variate and negatively onthe second canonical variate, suggesting thatmales have longer tails and deeper heads. Itmight be speculated that fighting betweenmales (Thompson et al., ’92; Horn et al., ’94)may have resulted in the evolution of rela-tively larger heads and longer limbs or, alter-natively, females have evolved relativelylonger bodies to maximize fecundity. How-ever, there appears to be no sexual dimor-phism in the number of presacral vertebraefor goannas (Greer, ’89; p. 208) and so thevertebrae are presumably more elongate infemales. FemaleV. gilleni andV. caudolinea-tus appear to be morphologically more simi-lar to each other than to males of their ownspecies (Fig. 3). If these two species areclosely related and the sexual dimorphismevolved prior to the evolution of separatespecies, then this might explain the closemorphological affinity of the sexes of sisterspecies. There also appears to be an appre-ciable difference between male and femaleV. storri, which may account for their morpho-logical overlapwith other species (Table 2).


Isometry and shapeThere is considerable conservatism in the

general body morphology of the family Vara-nidae.King andGreen (’93b) describeAustra-lian goannas as having long and slendernecks, long bodies with strong, musculartails, and well-developed, pentadactyl limbswith a strong claw on each digit. They wenton to indicate that Australian goannas aremedium to large in size. Greer (’89: p. 195)indicates that, although there is a hugevariation in body size, ‘‘all varanids lookmore or less alike, differing only in detailssuch as relative lengths of snouts and tails,shape and spinyness of the tail, and theposition of the nostrils on the snout.’’A simi-lar view was expressed by Shine (’86). Pi-anka (’95) suggests that Varanus is morpho-logically conservative, despite speciesvarying in mass by five orders of magnitude.Although we concur with the generaliza-tions that goannas aremorphologically simi-lar despite their considerable variation inbody size, we address here the quantitativequestion: Are goanna species isometric?In the absence of changes in basic body

physiology or mechanics, the shape of anorganism must alter with size to preservefunctional equivalence, both in ontogeny andphylogeny (Huxley, ’32; Gould, ’66; Schmidt-Nielsen, ’75, ’79; Sweet, ’80; McKinney, ’90;Losos, ’90a; Swartz andBiewener, ’92). Thereare, however, several examples in whichshape does not vary among members of a

size series (Meunier, ’59; Sweet, ’80), a phe-nomenon termed ‘‘geometric similarity’’(Gould, ’69; Gunther, ’75). Considering thecomments of other researchers on morpho-logical conservatism in Varanus (Shine, ’86;Greer, ’89; Pianka, ’95), with which we con-cur, varanids might be expected to approachgeometric similarity. To have isometric orgeometric similarity, changes in appendagelengths of goannas would need to be propor-tional to changes in body length; i.e., inter-specific variations in appendage dimensionswould scale with TA1.0 since we use TA as alinear measure of body size. The relation-ship between appendage dimensions and TAappears to be generally linear for most mea-surements (Figs. 12, 13), but the 18 species/subspecies of goanna measured are not geo-metrically similar for all measures; only HL,HD, and TAIL were consistently isometricfor most species. HL, HD, and FLL wereisometric for the subgenus Odatria, whileHD, NECK, and TAIL were isometric for thesubgenus Varanus (Table 10). Limb dimen-sions of the larger species are generally pro-portionally larger than for the smaller spe-cies, as changes in appendage dimensions(residuals) are significantly and positivelycorrelated with changes in TA. The scalingexponent with TA varies from about 0.88 forHW to 1.32 for UHL for the combined 18species/subspecies (Table 10). Even for asingle limb, the upper and lower proportionsvary with TA for different-sized species. Thisgeneral lack of isometry, and body propor-tions which are correlated with size in goan-nas, are in accord with the generally ac-cepted dogma of non-isometric body scaling(Ricklefs et al., ’81; McKinney, ’90; Losos,’90a,c; Swartz and Biewener, ’92). However,there are no comparable morphometric ana-lyzes for other groups of lizards to indicatewhether goannas are nevertheless verymor-phologically similar (as we and others sug-gest) or whether goannas are in fact just asmorphologically variable as other families oflizards.

Habitat and morphology of species basedon relative appendage length

Variations in appendage length ofVaranusthat are not correlated with body size orgrowth nor found in sister taxa may, assuggested by King (’91), be related to theirhabitat and niche utilization, or perfor-mance traits. This view is supported byLaerm (’74) for basilisk lizards (Basiliscusspp.) and Losos and Sinervo (’89) and Losos

TABLE 11. F-values and significance levels for sexualdifferences for selected body dimensions*

Varanus dflogHL

logHW

logFLL

logHLL

logNECK

brevicauda 1,30 4.69 19.87 4.28 9.70 8.20caudolineatus 1,60 50.69 4.85 21.99 23.20 5.09storri 1,20 38.04 12.43 8.32 25.38 0.26gilleni 1,19 32.47 13.81 4.04 1.33 4.33pilbarensis 1,6 0.00 3.10 0.37 0.56 0.86eremius 1,46 6.64 0.05 5.88 2.71 0.00scalaris 1,44 3.88 7.31 1.26 1.03 5.01acanthurus 1,24 13.01 9.68 6.70 8.49 0.16mitchelli 1,18 2.09 2.21 1.89 2.00 0.24glauerti 1,23 3.47 0.01 4.15 0.62 1.58tristis 1,42 4.21 2.06 1.26 1.67 0.45p. panoptes 1,7 7.53 1.55 1.61 0.40 0.04gouldii 1,69 23.79 9.63 7.36 0.00 13.11glebopalma 1,25 1.93 0.01 0.39 0.07 2.38rosenbergi 1,30 2.27 0.36 2.33 4.70 0.41mertensi 1,16 0.40 4.63 2.87 0.08 2.09p. rubidus 1,10 0.00 0.62 0.51 0.40 1.70giganteus 1,18 0.02 0.04 1.25 1.86 1.24

*Values are F-ratios and P-values. Values underlined are P ,0.05.


(’90a–c) for Anolis lizards. Consequently,some morphological attributes of goannasare likely to correspond to various ecologicaland environmental variables, as suggestedby Collette (’61), Moermond (’79), Ricklefset al. (’81), Pianka (’86), and Losos (’90a) for

other lizards. With few exceptions (Arnoldand Bennett, ’88; Losos, ’90a; Miles, ’94), theassumption that similarmorphological char-acteristics among species grouped by habi-tat or niche are based on common functionsor performance attributes has seldom been

Fig. 12. The relationship between the logarithmically transformed dimensions of tail, fore-and hind-limbs, and TA for 17 species of goannas.

Fig. 13. The relationship between the logarithmically transformed dimensions of head andneck and TA for 17 species of goannas.


tested. Unfortunately, little is known of thehabitat, foraging areas and retreats, or loco-motor performance attributes of manyWest-ernAustralian goannas.To evaluate possible links between mor-

phological features and habitats, the goan-nas measured in this study have been classi-fied into four broad categories (widelyforaging terrestrial, sedentary terrestrial,arboreal/rock scamperers, and semiaquatic;Table 12) according to the best availabledata for habitat and ecology. Discriminantanalysis of the goanna morphometrics byhabitat and ecology classified 82% of theindividuals into the correct habitat/ecologygroup. The first canonical variate had aneigenvalue of 2.0 and accounted for 74% oftotal variance, whereas the second had aneigenvalue of 0.6 and accounted for a further22% of the total variance. On the first canoni-cal variate, the widely foraging goannas ofthe subgenus Varanus are clearly separatedas a group from the odatrians (Fig. 14).However, the widely foraging odatrian V.eremius was closer to this group than to theother odatrians. The dimension best corre-lated with the first canonical variate waslogHD. Many of the odatrians clustered to-gether, but the second canonical variate sepa-rated the sedentary V. brevicauda and to alesser extent the arboreal/rock scamperingV. glauerti and V. glebopalma, and the semi-aquatic V. mertensi. The second canonicalvariate was best correlated with logNECK.

Widely foraging terrestrial goannasVaranid lizards are generally considered

to be large, widely foraging, active, terres-

trial lizards, corresponding to our categoryof widely foraging terrestrial goannas. Thespecies we place in this category are V. pan-optes, V. gouldii, V. rosenbergi, V. giganteus,and V. eremius.Of the goannas in the subgenus Varanus,

V. panoptes, V. gouldii, and V. rosenbergi arethe most similar morphologically, with V.giganteus and V. mertensi having higher andlower first canonical variates (Fig. 7). A sur-prising result of this morphological study isthe considerable difference between the twosubspecies ofV. panoptes, although it is notedthat there was appreciable overlap in themorphological space of V. panoptes, V. goul-dii, and V. rosenbergi based on the first twocanonical variates (Fig. 7). The V. p. panop-tes has a head that is longer, wider, anddeeper, and a neck that is longer than V. p.rubidus; however, limb and tail lengths arevery similar. These two subspecies mightreasonably have been expected to be moresimilar in appendage dimensions than whencomparedwith other species. There is consid-erable overlap in the morphological shape ofV. gouldii, V. panoptes, and V. rosenbergi(Tables 3, 4; Fig. 7) as might be expected forthese three species, which are all medium-to-large-sized, widely foraging, terrestrial, andcarnivorous predators of the subgenusVaranus (Pianka, ’70a, ’82; Green and King,’78; King and Green, ’79; Shine, ’86) (per-sonal observations). Perhaps a revision ofthe taxonomy of the gouldii/panoptes/rosen-bergi complex may explain these differencesbetween the two subspecies of V. panoptesand the misclassification of V. gouldii (Table6) for other species. If, e.g., V. gouldii was

TABLE 12. General habitat and foraging groups for Western Australian Varanus1

Habitat type/foraging mode Varanus Reference

Widely foraging eremius Thomson and Hosmer, ’63; Pianka, ’68, ’82terrestrial rosenbergi Green and King, ’78; King, ’80

panoptes Shine, ’86gouldii Zietz, ’14; Thomson and Hosmer, ’63; Pianka, ’70a, ’82, ’94; Shine, ’86giganteus Zietz, ’14; Cogger, ’65; Brunn, ’80; Pianka, ’82, ’94; King et al., ’89

Sedentary terrestrial brevicauda Pianka, 70b, ’94; James, ’94, ’96storri Zietz, ’14; Peters, ’73; Stirnberg and Horn, ’81acanthurus Thomson and Hosmer, ’63; Brunn, ’82

Arboreal/rock caudolineatus Pianka, ’69, ’94; Thompson, ’93scamperers gilleni Zietz, ’14; Thomson and Hosmer, ’63; Martin, ’75; Delean, ’80; Pianka, ’82

pilbarensis Pianka, ’95; Kendrick (personal communication)scalaris Pianka, ’95mitchelli Shine, ’86glauerti Sweet (personal communication); Pianka, ’95tristis Pianka, ’71, ’82, ’94; Brunn, ’82; Fitzgerald, ’83glebopalma Sweet (personal communication); Horn and Schurer, ’78; Swanson, ’79

Semiaquatic mertensi Hermes, ’81; Shine, ’86

1Data from a variety of specific references as listed, plus more general descriptions of habitat type selected as reported in Storr et al.(’83), Greer (’89), Cogger (’92), Wilson and Knowles (’92), and personal observations.


subsequently divided into more than onespecies (it already has two subspecies), thiswould alter the basis of the canonical variateanalysis and may reduce the degree of mis-classification.V. giganteus, the largest Australia go-

anna, differs appreciably from the otherwidely foraging terrestrial species (Tables 3,4; Fig. 7). Head length is the primary charac-teristic that separates the five species in thissubgenus (Table 7; Fig. 7), with V. giganteushaving the longest relative head length. Inaddition to head length, V. giganteus has alonger neck, wider head, and longer fore-limbs than any of the other species in theVaranus clade (Tables 3, 4). V. giganteus, V.rosenbergi, and V. p. panoptes have rela-tively wider and deeper heads than othergoanna species. However, there is no evi-dence to suggest that their foraging behav-ior or prey, a possible factor in influencinghead dimensions, differs from those of V.gouldii or V. p. rubidus, and so differences inhead dimensions in this category are diffi-cult to interpret from an ecological perspec-tive.V. eremius is a small, widely foraging ter-

restrial odatrian goanna (Pianka, ’68, ’94;

Storr et al., ’83) which differs from the otherOdatria (except for V. brevicauda) by itsrelatively short neck (Table 4; Fig. 5). In thisrespect, V. eremius is morphologically simi-lar to thewidely foraging terrestrialsV. goul-dii, V. rosenbergi, and V. p. rubidus of thesubgenus Varanus, although it clearly fallswithin the odatrian group (Fig. 3).

Sedentary terrestrial goannasWe place V. acanthurus, V. storri, and V.

brevicauda in a sedentary terrestrial cat-egory rather than the widely foraging terres-trial category, but only for V. brevicauda arethere substantial morphological differences.V. acanthurus is the most generalized

odatrian species measured here; its bodyappendages appear to deviate least from theregression lines for all body dimensions withTA (Table 4) for the Odatria group. This isalso supported by the data in Table 6, show-ing that V. acanthurus is the odatrian mostoften incorrectly classified. V. acanthurus isa ground-dwelling monitor that retreats torock crevices or holes under medium-sizedboulders (Storr et al., ’83; Dryden et al., ’90)(personal observations). The most morpho-logically similar species to V. acanthurus is

Fig. 14. The first and second canonical variates forgoannas classified by habitat and ecology: circles, widelyforaging terrestrial; diamonds, sedentary terrestrial;squares, arboreal/rock scamperers; triangle, semiaquatic.The first standardized canonical vector is: 2 0.280logTA1 0.840logHD 1 2.961logHL 2 0.192logHW 2 2.970log-

LFL 1 5.251logLHL 2 2.879logNECK 2 1.900logTAIL 20.471logUFL 2 0.072logUHL. The second standardizedcanonical vector is: 2 0.788logTA 2 1.374logHD 22.321logHL 1 0.950logHW 2 1.649logLFL 2 4.046lo-gLHL1 2.726logNECK 2 0.972logTAIL2 0.591logUFL 20.454logUHL.


V. baritji from the Northern Territory ofAustralia (King and Horner, ’87), which hasa similar ecological niche (Sweet, personalcommunication).Storr et al. (’83) report V. storri to be very

like, but smaller than, V. acanthurus. How-ever, our morphological analysis indicatesthat V. storri has a relatively longer, wider,and deeper head than V. acanthurus andappreciably longer limbs (Table 4). As littleis known of the ecology or the preferredhabitat of V. storri (possibly under rocks andlarge objects lying on the ground; Sweet,personal communication), it is not possibleto infer a relationship between its grossmor-phology and habitat use.V. brevicauda is the smallest and most

morphologically different goanna measured(Table 4; Figs. 3–6, 8–11). This species, rela-tive to its thorax-abdomen length, has theshortest head and neck, narrowest head,shortest fore- and hind-limbs, and shortesttail. However, Greer (’89: p. 208) reportsthat V. brevicauda has more presacral verte-brae than any of the other goannas mea-sured and so this might reflect an evolution-ary pressure for relative elongation of thebody length per se. James (’96) reports that73% of V. brevicauda movements in a pit-trapping program were less than 20 m andthat these goannas were very seldom seensurface-active, a view supported by Pianka(’94). The relatively short limbs and shorthead of V. brevicauda, and the evidence ofJames (’96), suggest that this terrestrial go-anna is not widely foraging and may catchmost of its prey by using a sit-and-wait strat-egy or by searching in burrows locatedwithina relatively small home range. Vitt andCong-don (’78) suggest that widely foraging liz-ards have a relatively low clutch mass,whereas the sit-and-wait foragers have arelatively high relative clutch mass. A rela-tively high clutch mass can also be associ-ated with a larger abdominal cavity (Shine,’92). Although James (’96) concluded thatthe absolute clutch size of two eggs for V.brevicauda was consistent with that pre-dicted from correlations between SVL andclutch size in other goannas, a few V. brevi-cauda were recorded with clutches of 4 and5. If, however, V. brevicauda had a relativelylarge clutch mass for its body size, due to thesize of its eggs, then this goanna’s morphol-ogy would support the established dogmathat taxawith relatively smaller limbs, tails,and heads have relatively larger volumes

available in their body cavity to carry eggsand are generally sit-and-wait foragers (Vittand Congdon, ’78; Shine, ’92). At a casualglance,V. primordius appears to be a similarsize and have a similar morphology to V.brevicauda. A comparative analysis of thesetwo small goannas’ morphology and ecologymight assist in explaining why the morphol-ogy of V. brevicauda is appreciably differentfrom the rest of the genus.

Arboreal/rock scampering goannasThis category includes a continuum of

scampering goannas from typically arborealspecies (V. scalaris, V. gilleni) to typicallysaxicolous species (V. pilbarensis, V. glebo-palma), with V. caudolineatus, V. tristis, V.mitchelli, and V. glauerti being found in bothtrees and rocks. Within this category, mor-phology seems to be related more to bodysize or performance traits than habits. V.mitchelli, V. tristis, and V. scalaris have afairly generalized odatrian morphology, likeV. acanthurus and V. storri, which are gen-eralized sedentary terrestrial goannas. V.glebopalma, V. pilbarensis, and V. glauertihave relatively long appendages and a flat-tened head. In contrast, V. caudolineatusand V. gilleni have relatively short append-ages.V. scalaris is morphologically very similar

to V. acanthurus, having a relatively similarlength head and neck, head depth and fore-and hind-limb length. It, however, has anarrower head and shorter tail. Little isknown of its ecology or behavior other thanit is arboreal (Sweet, personal communica-tion). The larger species, V. tristis, appearsto be a ‘‘stretched’’ version of V. acanthurus,with all dimensions being slightly longer,except it has a shallower head (Table 4).Pianka (’71) reportsV. tristis as moving fromtree to tree as it mostly climbs in search ofprey. Baverstock et al. (’93) report that V.scalaris and V. tristis are closely related;this, together with their preference for arbo-real habitats, would explain their similarbody shape. The high degree of morphologi-cal similarity between these two arborealspecies and a terrestrial species (V. acanthu-rus) is unexpected. Baverstock et al. (’93)indicate that V. acanthurus is more closelyrelated to V. gilleni and V. brevicauda (andtwo goannas not measured in this study, V.kingorum and V. primordius). The general-ized body form of V. acanthurus may simplyindicate that it is similar to the ancestralform of the subgenus Odatria, and this may


account for the similarity in body shape be-tween V. acanthurus, V. scalaris, and V. tristis.V. mitchelli, which is grouped with V.

acanthurus, V. scalaris, and V. tristis on thefirst and second canonical variates, differsmarkedly on the third canonical variate (Fig.6). Table 4 indicates that V. mitchelli has arelatively longer neck but narrower and shal-lower head, and shorter limbs than V.acanthurus, or alternatively it has a rela-tively longer body and similar-sized append-ages. V. mitchelli’s relative appendagelengths are similar to those of V. scalaris,having proportionally the same head andneck length, head width, and fore- and hind-limb length. It differs from V. scalaris in thatit has amuch shallower head.V. mitchelli, V.tristis, and V. scalaris are all known to bearboreal (Pianka, ’71; Storr et al., ’83; Shine,’86), whereasV. acanthurus is a groundmoni-tor retreating to rock crevices or holes undermedium-sized boulders (Storr et al., ’83;Dryden et al., ’90) (personal observations).The most noticeable difference between V.acanthurus and these three arboreal moni-tors is that V. acanthurus has an apprecia-bly shorter neck (Table 4), which might sug-gest that small-to-medium-sized arborealspecies of goannas have longer necks com-pared with similar-sized terrestrial species.The relative neck length of these three arbo-real species is, however, not as long as thoseof the rock scampering/arboreal V. pilbaren-sis, V. glebopalma, and V. glauerti. Addi-tional arboreal species (e.g.,V. prasinus) needto be measured before a more conclusivestatement is possible.V. glebopalma, V. glauerti, andV. pilbaren-

sis, from the Odatria clade, have similarlylong tails, long necks, and long FLL, al-though V. glebopalma differs from V.pilbarensis and V. glauerti in having a pro-portionately longer neck, and V. pilbarensishas a proportionately longer HLL than V.glebopalma and V. glauerti (the large V. gi-ganteus from the Varanus clade also has along NECK, HL, FLL, and HLL; see below).The long heads and necks of V. glebopalma,V. glauerti, and V. pilbarensismay be associ-ated with searching for prey in rock or treecervices in the medium (and large)-sized go-annas. The longer hind-limbs of V. glebo-palma andV. pilbarensismay be a specializa-tion that separates these ‘‘rock scramblers’’from V. glauerti, which is primarily arboreal(this suggestion is supported by the longerHLL of the sometimes rock-dwelling, widelyforaging, but never arboreal, V. giganteus ofthe Varanus clade). Losos and Sinervo (’89)

report that species of Anolis lizards withlonger hind-limbs run faster on thicker rodsand that their speed is more affected onthinner rods than short-legged lizards. Simi-larly, very long hind-limbs may be disadvan-tageous for medium-sized arboreal goannas,if they impede movement on narrowbranches or in tree hollows.V. caudolineatus and V. gilleni are two

small, arboreal goannas from central inlandWesternAustralia. They are, as Pianka (’69)suggests, morphologically similar with com-paratively short heads, tails, and hind-limbs. However, they can bemorphologicallyseparated, with V. gilleni having a shallowerhead and slightly shorter fore- and hind-limbs. The relatively shorter limbs and shal-lower head may enable V. gilleni, the largerof the two pygmy goannas, to move morefreely in tree hollows (Pianka, ’69) (personalobservations). These two small, arboreal go-annas, however, do not conform with theslender body, long tail, long hind-toes pat-tern suggested by Snyder (’54), Ballinger(’73), Ricklefs et al. (’81), and Pianka (’86) forarboreal species. Behavioral observations ofV. caudolineatus suggest that this goannaseldom jumps from a tree to the ground,even when fleeing a predator, which gener-ally fits Moermond’s (’79) category of ‘‘run-ners and crawlers’’ having more equally pro-portioned fore- and hind-limbs than thejumpers. Arboreal species may need to befurther subdivided into jumpers (slenderbody, long tail, long hind-toes) and runnersand crawlers (shorter fore- and hind-limbs)before links can be drawn between theirmorphology and habitat-performance traits.Arboreal lizards have been described as

having a slender body, long tail, and longhind-toes (Snyder, ’54; Ballinger, ’73; Rick-lefs et al., ’81; Pianka, ’86), a hypothesissupported by the body form of V. glauerti,although it is not supported by the bodyshape of V. caudolineatus, V. gilleni, V. sca-laris, and V. tristis. Moermond (’79) assertsthat for Anolis lizards, jumpers tend to havelonger tails, a view generally supported byPianka (’86). It would be interesting to quan-tify the extent to which V. glebopalma, V.glauerti, and V. pilbarensis jump from rockto rock, tree to tree, or tree to ground com-pared with other goannas, as this may belinked with the length of their tail and hind-limbs and may be a performance trait thatseparates these three species from other ar-boreal species such as V. scalaris and V.tristis.The total TAofV. tristis andV. glauertiis similar; they are closely related (Baver-


stock et al., ’93) and from the limited dataavailable they both are arboreal. It is there-fore surprising that their relative append-age lengths are appreciably different (Table4). This paradox is only likely to be resolvedwith a better understanding of their behav-ior and their microhabitat.

Semiaquatic goannasV.mertensi has the narrowest and shallow-

est head, and shortest limbs and tail, ofmembers the Varanus clade. V. mertensi isthe only semiaquatic goanna of this cladethat was measured here. V. mitchelli, of thesubgenus Odatria, is sometimes seen in wa-ter but is mostly arboreal (Shine, ’86); it alsohas short limbs. This may suggest a linkbetween semiaquatic habits and appendagelength. Measurement of other semiaquaticvaranids would obviously assist in the inter-pretation of these variations in body append-age length.

LITERATURE CITED

Arnold, S.J., andA.F. Bennett (1988) Behavioural varia-tion in natural populations. V. Morphological corre-lates of locomotion in the garter snake (Thamnophisradix). Biol. J. Linn Soc. 34:175–190.

Auffenberg, W. (1981) The Behavioral Ecology of theKomodo Monitor. Gainesville: University Presses ofFlorida.

Auffenberg, W. (1994) The Bengal Monitor. Gainesville:University Presses of Florida.

Auffenberg, W., Q. Nazar, and N. Khurshid (1991) Pre-ferred habitat, home range and movement patterns ofVaranus bengalensis in southern Pakistan. Mertensi-ella 2:7–28.

Ballinger, R.E. (1973) Experimental evidence of the tailas a balancing organ in the lizard Anolis carolinensis.Herpetologica 29:65–66.

Baverstock, P.R., D. King, M. King, J. Birrell, and M.Krieg (1993) The evolution of species of the Varanidae:Microcomplement fixation analysis of serum albu-mins. Aust. J. Zool. 41:621–638.

Bohme, W. (1991) The identity of Varanus gouldii (Gray,1983), and the nomenclature of the V. gouldii-speciescomplex. Mertensiella 2:38–41.

Bookstein, F.L. (1991) Morphometric Tools for Land-mark Data. Cambridge: Cambridge University Press.

Bookstein, F., B. Chernoff, R. Elder, J. Humphries, G.Smith, and R. Strauss (1985)Morphometrics in Evolu-tionary Biology: The Geometry of Size and ShapeChange, With Examples From Fishes. Philadelphia:Academy of Natural Sciences of Philadelphia.

Brunn, W. (1980) Goannas caught on trips. S.A. Herp.Group Newsl.Oct:2–3.

Brunn, W. (1982) More of Werner Brunn’s goannascaught on trips series. S.A. Herp. Group Newsl. Mar:6–9.

Burnaby, T.P. (1966) Growth-invariant discriminantfunctions and generalized distances. Biometrics 22:96–110.

Calder, W.A. (1984) Size, Function and Life History.Cambridge: Harvard University Press.

Cogger, H.G. (1965) Reptiles and frogs ofAustralia’s aridregions. Aust. Nat. Hist. 15:128.

Cogger, H.G. (1992) Reptiles andAmphibians of Austra-lia. Sydney: REed.

Collette, B.B. (1961) Correlations between ecology andmorphology in anoline lizards fromHavana, Cuba andsouthern Florida. Bull. Mus. Comp. Zool. 125:137–162.

Delean, S. (1980) A new record of the pygmy mulgamonitor, Varanus gilleni (Lucus and Frost). Herpeto-fauna 12:35.

Dryden, G., B. Green, D. King, and J. Losos (1990)Waterand energy turnover in a smallmonitor lizard,Varanusacanthurus. Aust. Wildl. Res. 17:641–646.

Fitzgerald, M. (1983) A New South Wales record for thefreckled treemonitorVaranus tristis orientalis.Herpe-tofauna 15:23.

Gould, S.J. (1966) Allometry and size in ontogeny andphylogeny. Biol. Rev. 41:587–640.

Gould, S.J. (1969) Size change in vertebrate evolution:Types of scaling and their implications. In: VertebrateEvolution: Mechanism and Process. NATO, Istanbul:Adv. Study Inst., pp. 36–38.

Green, B., and D. King (1978) Home range and activitypatterns of the sand goanna, Varanus gouldii (Rep-tilia: Varanidae). Aust. J. Wildl. Res. 5:417–424.

Greer, A.E. (1989) The Biology and Evolution of Austra-lian Lizards. Sydney: Surrey Beatty.

Gunther, B. (1975) Dimensional analysis and theory ofbiological similarity. Physiol. Rev. 55:659–699.

Hermes, N. (1981) Mertens water monitor feeding ontrapped fish. Herpetofauna 13:34.

Holmes, R.S., M. King, and D. King (1975) Pheneticrelationship among varanid lizards based upon com-parative electrophoretic data and karyotypic analy-ses. Biochem. Syst. Ecol. 3:257–262.

Horn, H.-G., and U. Schurer (1978) Bemerkungen zuVaranus (Odatria) glebopalma Mitchell 1955 (Rep-tilia: Sauria: Varanidae). Salamandra 14:105–116.

Horn, H.-G., M. Gaulke, and W. Bohme (1994) New dataon ritualized combats inmonitor lizards (Sauria: Vara-nidae), with remarks on their function and phyloge-netic implications. Zool. Garten. N.F. 64:265–280.

Humphries, J.M., F.L. Bookstein, B. Chernoff, G.R.Smith, R.L. Elder, and S.G. Poss (1981) Multivariatediscrimination by shape in relation to size. Syst. Zool.30:291–308.

Huxley, J.S. (1932) Problems of Relative Growth. Lon-don: Methuen.

James, C.D. (1994) Spatial and temporal variation instructure of a diverse lizard assemblage in arid Austra-lia. In L.J. Vitt and E.R. Pianka (eds): Lizard Ecology:Historical and Experiemental Perspectives. Prince-ton, NJ: Princeton University Press.

James, C.D. (1996) Ecology of the pygmy goannaVaranusbrevicauda in spinifex grasslands of centralAustralia.Aust. Zool. 44:177–192.

King, D. (1980) The thermal biology of free-living sandgoannas (Varanus gouldii) in southern Australia. Co-peia 1980:755–767.

King, D. (1991) The effect of body size on the ecology ofvaranid lizards. Mertensiella 2:204–210.

King, D., andB.Green (1979)Notes on diet and reproduc-tion of the sand goanna, Varanus gouldii rosenbergi.Copeia 1979:64–70.

King, D., and B. Green (1993a) Goanna: The Biology ofVaranid Lizards. Sydney: New South Wales Univer-sity Press.

King, D., and B. Green (1993b) Family Varanidae. InC.J. Glasby, G.J.B. Graham, and P.L. Beesley (eds):Fauna of Australia: V2a Amphibia and Reptilia. Can-berra: Australian Government Printing Service, pp.253–260.

King, M., and P. Horner (1987) A new species of monitor(Platynota: Reptilia) from Northern Australia and anote on the status of Varanus acanthurus insulanicus


Mertens. The Beagle (Records of the Northern Terri-tory Museum ofArts and Sciences) 4:73–79.

King, M., and D. King (1975) Chromosomal evolution inthe lizard genus Varanus (Reptilia). Aust. J. Biol. Sci.28:89–108.

King, D., B. Green, and H. Butler (1989) The activitypattern, temperature regulation and diet of Varanusgiganteus on Barrow Island, Western Australia. Aust.Wildl. Res. 16:41–47.

Klingenberg, C.P. (1996) Multivariate allometry. In L.F.Marcus, M. Corti, A. Loy, G.J.P. Naylor, and D.E. Slice(eds): Advances in Morphometrics. New York: PlenumPress.

Laerm, J.R. (1974) A functional analysis of morphologi-cal variation and differential niche utilization inbasilisk lizards. Ecology 55:404–411.

Losos, J.B. (1990a) Ecomorphology, performance capabil-ity, and scaling of West Indian Anolis lizards: Anevolutionary analysis. Ecol. Monogr. 60:369–388.

Losos, J.B. (1990b) Concordant evolution of locomotorbehaviour, display rate and morphology in Anolis liz-ards. Anim. Behav. 39:879–890.

Losos, J.B. (1990c) The evolution of form and function:Morphology and locomotor performance in West In-dian Anolis lizards. Evolution 44:1189–1203.

Losos, J.B., and B. Sinervo (1989) The effects of morphol-ogy and perch diameter on sprint performance of Ano-lis lizards. J. Exp. Biol. 145:23–30.

Martin, K. (1975) Reptiles of the Alice Springs Area.Herpetofauna 7:6–7.

McKinney, M.L. (1990) Trends in body-size evolution. InK. McNamara (ed): Evolutionary Trends. London: Bel-haven Press, pp. 85–118.

Mertens, R. (1942) Die Familie der Warane (Varanidae).Abh. Senckenb. Naturforsch. Ges. 465:117–234.

Meunier, K. (1959) Die Grossenabhangigkeit der Korper-form bei Vogeln. Z. Wiss. Zool. 162:328–355.

Miles, D.B. (1994) Covariation between morphology andlocomotory performance in Sceloporine lizards. In L.J.Vitt and E.R. Pianka (eds): Lizard Ecology: HistoricalandExperimental Perspectives. Princeton, NJ: Prince-ton University Press, pp. 207–235.

Moermond, T.C. (1979) Habitat constraints on the behav-ior, morphology, and community structure of Anolislizards. Ecology 60:152–164.

Peters, R.H. (1983) The Ecological Implications of BodySize. Cambridge: Cambridge University Press.

Peters, U. (1973)Acontribution to the ecology ofVaranus(Odatria) storri.Koolewong 2:12–13.

Pianka, E.R. (1968) Notes on the biology of Varanuseremius.West. Aust. Nat. 11:39–44.

Pianka, E.R. (1969) Notes on the biology of Varanuscaudolineatus and Varanus gilleni. West. Aust. Nat.11:76–82.

Pianka, E.R. (1970a) Notes on the biology of Varanusgouldii gouldii.West. Aust. Nat. 11:141–144.

Pianka, E.R. (1970b) Notes on the biology of Varanusbrevicauda.West. Aust. Nat. 11:113–116.

Pianka, E.R. (1971) Notes on the biology of Varanustristis.West. Aust. Nat. 11:180–183.

Pianka, E.R. (1982) Observations on the ecology ofVaranus in the Great Victoria Desert. West. Aust. Nat.15:1–13.

Pianka, E.R. (1986) Ecology and Natural History ofDesert Lizards: Analysis of the Ecological Niche andCommunity Structure. Princeton, NJ: Princeton Uni-versity Press.

Pianka, E.R. (1994) Comparative ecology of Varanus inthe Great Victoria Desert. Aust. J. Ecol. 19:395–408.

Pianka, E.R. (1995) Evolution of body size: Varanidlizards as a model system.Am. Nat. 146:398–414.

Reyment, R.A., R.E. Blackith, and N.A. Campbell (1984)MultivariateMorphometrics. London:Academic Press.

Ricklefs, R.E., D. Cochran, and E.R. Pianka (1981) Amorphological analysis of the structure of communi-ties of lizards in desert habitats. Ecology 62:1474–1483.

Schmidt-Nielsen, K. (1975) Scaling in biology: The conse-quence of size. J. Exp. Biol. 194:287–308.

Schmidt-Nielsen, K. (1979) Animal Physiology: Adapta-tion andEnvironment. Cambridge: CambridgeUniver-sity Press.

Schmidt-Nielsen, K. (1984) Scaling: Why Is Animal SizeSo Important? Cambridge: Cambridge UniversityPress.

Shine, R. (1986) Food habits, habitats and reproductivebiology of four sympatric species of varanid lizards intropical Australia. Herpetologica 42:346–360.

Shine, R. (1992) Relative clutch mass and body shape inlizards and snakes: Is reproductive investment con-strained or optimised? Evolution 46:828–833.

Snyder, R.C. (1954) The anatomy and function of thepelvic girdle and hind-limbs in lizard locomotion. Am.J. Anat. 95:1–41.

Stirnberg, E., and H.-G. Horn (1981) Eine unerwarteteNachzucht im Terrarium: Varanus (Odatria) storri(Reptilia: Sauria: Varanidae). Salamandra 17:55–62.

Storr, G.M. (1980) The monitor lizards (genus VaranusMerrem, 1820) of Western Australia. Rec. West. Aust.Mus. 8:237–293.

Storr, G.M., L.A. Smith, and R.E. Johnstone (1983)Lizards of Western Australia. II. Dragons and Moni-tors. Perth: WesternAustralian Museum.

Swanson, S. (1979) Some rock-dwelling reptiles of theArnhem land escarpment. North. Terr. Nat. 1:14–18.

Swartz, S.M., and A.A. Biewener (1992) Shape and scal-ing. In A.A. Biewener (ed): Biomechanics—Structuresand Systems: A Practical Approach. Oxford: OxfordUniversity Press.

Sweet, S.S. (1980) Allometric inference in morphology.Am. Zool. 20:643–652.

Tabachnick, B.G., and L.S. Fidell (1989) UsingMultivar-iate Statistics. NewYork: Harper & Row.

Thompson, G. (1993) Dailymovement patterns and habi-tat preferences of Varanus caudolineatus (Reptilia:Varanidae). Wildl. Res. 20:227–231.

Thompson, G.G., P.C.Withers, and S.A. Thompson (1992)The Combat Ritual of Varanus caudolineatus andVaranus gouldii (Reptilia: Varanidae).West.Aust. Nat.19:21–25.

Thomson, D.F., and W. Hosmer (1963) A preliminaryaccount of the herpetology of the Great Sandy Desertof CentralWesternAustralia. Proc. R. Soc. Vic. 77:217–237.

Vitt, L.J., and J.D. Congdon (1978) Body shape, reproduc-tive effort, and relative clutch mass in lizards: Resolu-tion of a paradox. Am. Nat. 112:595–608.

Vitt, L.J., and W.E. Cooper (1985) The evolution ofsexual dimorphism in the skink Eumeces laticeps: Anexample of sexual selection. Can. J. Zool. 63:995–1002.

Wilson, S.K., and D.G. Knowles (1992) Australia’s Rep-tiles. Sydney: Cornstalk.

Yadav,A.K., and K.S. Rana (1988) Sexual dimorphism inthe Indian monitor lizard, Varanus bengalensis (Dau-din). Geobios 15:210–214.

Zietz, F.R. (1914) Lacertilia. Trans. R. Soc. South Aust.38:441–444.


Documents

Comparative morphology of Western Australian varanid ......Comparative Morphology of WesternAustralian Varanid Lizards (Squamata: Varanidae) GRAHAMG.THOMPSON1*AND PHILIPC.WITHERS2