HERPETOLOGICALJOURNAL, Vol. 6, pp. 125-132(1996) LIFE-HISTORY VARIATION IN A COMMUNITY OF LACERTID LIZARDS FROM THE LAKE SKADAR REGION (MONTENEGRO) D. BEJAKOVICI*,I. ALEKSIC',A. TARASJEV1**,J. CnNOEnNJA-ISAILOVIC'. G. DZUKIC', M. L. KALEZIC2 t Institutefor Biological Research "sinisa Stankovic", 29 Novembra 142, 11000 Beograd, Yugoslavia 2lnstitute of Zoology, Faculty of Biolog, Studentski trg 3, 11000 Beograd, Yugoslavia * Present address: MAF Quality Management,Plant Protection Centre, PO Box 24, Lincoln, New Zealand a* Author for corresPondence

We compared life-history attributes of Algyroides nigropunctatus, Lacerta oxycephala, Podarcis melisellensis andPodarcis muralis populations in the Lake Skadarregion (mainland and islands). These lacertids are moderately sized with differing morphology (flattened vs. cylindrical). They have similar duration ofegg incubation, size at onsetofsexual maturity, size of hatchlings, and rate of juvenile growth. Clutch size is variable between populations and species:P. muralis producedlarger clutchesthan the other species,especially ,4. nigropunctatus andL. oxycephala.A significant differencein egg sizewas apparentat the population level only. Smaller clutch size and elongatedeggs were associatedwith a flattened body in L. oxycephala. A canonicaldiscriminant analysis ofreproductive variables showed that speciessimilarities in ecologicalcharacteristics reflect phylogenetic relationships.

INTRODUCTION lacertidlizard community of the Lake Skadarregion (CrnobrnjaJsailovic& Dzukic, 1995). In thisreport we Componentsof an organism'slife history (reproduc- comparethe life-historyattributes of four sympatric(in tive characteristics,behaviour, longevity, growth rate, somecases syntopic) species, both in a coastalarea and etc.) may be viewed as co-adaptedtraits that in concert on islandsin LakeSkadar. In additionto differencesin produce the best strategy for survival in a particular en- the geneticbackground and meaningof the species, vironment (e.g. Stearns, 1976). Most studies on theselacertids also differ in morphology, life-historiesof lizards show large interspecificdiffer- distribution andecological traits. ences and considerable intraspecific temporal and geographic variations of the traits (reviewed in Dunham et al., 1988). Thesevariations have been at- MATERIALSAND METHODS tributedprimarily to the local environment(e.g. Tinkle SPECIESSTUDIED and Dunham, 1986;Abts, 1987).However, foraging mode, phylogenetic inertia and morphological con- The speciesstudied here represent the closestgenera straints (body size and shape) were also evoked as within Old-World lacertids,with affinities as follows: factorspotentially influencing variation of life-history (P odarc is-Lac erta) (Al gr oides) (Arnold, I 989). The traits (e.g.Vitt & Congdon,1978; Vitt, 1981;Vitt & Dalmatian Algtroides, Algtroides nigropunctatus, Price, 1982).In order to obtain empiricaldata essential ranges along the coastal region of the Adriatic and Io- to discernthe relative importanceof all thesedetermi- nian seas,from Italian and Sloveniankarst regionsto nants for the evolution of life-history traits in lizards, the Akarnanien region (Greece), including some north- simultaneousstudies on a setof syntopicpopulations of ern Adriatic islands. There are a few significant range speciesdiffering in the degree of their phylogenetic penetrationsinland. This speciesis usually encountered relatedness,ecological traits and morphology, are ap- in degradedscrub, bushes between fields and rocky parently needed(Vitt, 1986;Tinkle & Dunham, 1986; cliffareas, usually nearwater. The sharp-snoutedrock Henle, 1990).Studies on life-history traits of lacertids lizard,,Lacerta oxycephala, is a steno-endemicBalkan (Lacertidae)in comparisonwith other lizard families species,restricted in its range to a narrow zone along are still scarce(see Dunham et al., 1988).This is par- the Adriatic coast,from the river Krka to the northwest- ticularly true for the Balkan'lacertids(Podarcis spp., ern part of Albania. L. orycephala differs distinctly in Lacerta spp.,Algtroides spp.,Ophisops spp.). Also, al- morphology from all other lacertidsby having a flat- most all previousstudies of theselizard populations tenedbody, with a very pointedsnout. In an ecological havebeen conducted in locationslacking diverselizard sense,this speciesis peculiar among the Balkan's communities.Out of 19 speciesthat canbe encountered lacertidsbecause it is the most specialistpetricole, in- throughout the Balkans, nine speciescompose the habiting mostly sunny cliff faces,rock-pavements, 126 D. BEJAKOVICET AL. stone piles and screes. The Dalmatian wall lizard, sides of railroad tracks and nearby road (Podgorica- Podarcis melisellensis,is a polymorphic species,with a Bar). P. melisellensis shares this habitat with p. considerableabundance in diverse habitats, but it is muralis. Malo Starcevois a small island (0.9 ha) with mainly encountered in densely vegetated places. It densepopulations of P. muralis only. Beska Velja is ranges from Monfalcone (Italy) to Skadar (Albania) one of the biggestislands (15.9 ha), with Z. orycephala along the coastalzone of the Dinaric Alps, extending as the only lizard species.The island Malo Besko (1.7 inland in Lika, Hercegovina,Montenegro and Albania ha) is characterizedby a dense tree vegetation and (Tiedemann& Henle, 1986).The common wall lizard, grassypatches with bouldersof varioussizes, inhabited Podarcis muralis, has a wide geographicdistribution by a relatively robustpopulation of P. melisellensis.ln (central and southern Europe), with immense geo- nearby waters, mainly on rocks, many P. muralis indi- graphic variation in morphology and habitat viduals can be encountered.Mali Moracnik is a characteristics(Gruschwitz & Bohme, 1986).Thus, moderate-sizeisland (0.7 ha), with a sparsetree vegeta- this lizard is thought to be more opportunisticwith re- tion and grassyfields among large stony patches.Here gard to microhabitatpreference than its relatives. the largeP, melisellensripopulation inhabits grassy ar- eassyntopically with L. orycephala.A few individuals of P. muraliswere noticedon this islandas well. Bisage is an island of 2.8 ha, with P. melisellensisand A. HABITAT AND SPECIES COMPOSITION nigropunctatls as the lizard inhabitants. Lake Skadaris the largestof the Balkan lakes.The lake is locatedin a karst terrain in the outer part ofthe DATA COLLECTION AND STATISTICAL ANALYSES south-easternDinaric Alps. Mountains rise steeply from the lake'ssouth-western shore along which an ar- Details about field and laboratorymethods (body chipelagoof about forty large and small islandsexists. and egg measurements,egg-laying process, egg incu- Vegetationof this part, including the islands,is charac- bation period, hatchlingsraised), including data on terizedby maquisand guarigue.Scattered xeromorphic clutch and egg size,size at sexualmaturity for females treesare separatedby patchesof dry grassand herbs. of L. oxycephala, P, melisellensis and P. murahs, can Larger islandsusually have a water's-edgerocky zone be found in our previouspapers on reproductivechar- with scatteredtrees only, and a cenffalgrassy zone with acteristicsofthe lacertidsfrom the Lake Skadarregion bushy vegetation. The northern and north-eastern (Aleksic& Tucic, 1994;Bejakovic et al.,1995,1996). shores are flat, providing an extensive semi-littoral Snout-ventlength (SVL) was taken as a measureof zone.This formerly inundatedterrain with recentallu- overall size. Head dimensions(length, height, width), vial depositsis now mostly under cultivatedsoil, with relative to SVL, expressedthe body shape.The length fragmentsof common oak forests.The climate of the of hind limb relativeto SVL was consideredas a mor- region is characterizedby a large amountof precipita- phologicalindicator of mobility capacitiesof the lizard tion andhigh summertemperatures (Lasca et al.,1981). species(see Pianka, 1986). Egg volume(V) wastaken The shoreregion of the lake hastwo distinctzones with asthe overall measureof sizeof eggslaid. RCM values respectto lizard distribution, but on the islandsthere were calculatedas the ratio of clutchmass to body mass are speciesin various combinationsof allotopic and after egg laying (see Shine, 1980) and were used as a syntopicoccurrence (Crnobrnja-Isailovic & Dzukic, measureof reproductiveinvestment per clutch. The 1995). P. muralis and L. orycephala appearto be the sizeof the smallestfemale with enlargedvitellogenic most common species,probably due to their better follicles was consideredthe minimum size at maturity. colonizing abilities. They are the only specieswith Femalereproductive characteristics were studiedon allotopicaldistributions. L. oxycephalaoccurs as the a number of individuals,except for A. nigropunctatus only lizard on 19, and P. muralis on 10,of the 40 is- whereonly ten sexuallymatured females were avail- lands.These islands are the most denselypopulated (up able, of which three laid eggs. Nevertheless,we to one individualper 2 m2,unpublished data.). When includedthis speciesin this study, though we were speciesare syntopic, spatial partitioning occurs; in such awarethat the paucity of samplesize precludedwell- casesthey generally follow their preferedhabitats. supportedconclusions. Thus, P. melisellensisis mostly encounteredin veg- Statisticswere calculatedwith SAS package(SAS etated areas, P. muralis and L. oxycephala incline to Institute,1989). Mean valuesand standarderrors of shorelinerocky zones,while l, nigropunctatasshow life-history traits were weighted for population sizes. habitatpreferences between these two groups.Data on SVL, egg volume and clutch size were logarithmically morphologicaltraits relevantto this study came from transformed.Life-history traits (estimatedon females) the completelizard samplestaken from the overall were analysedwith populationsnested within species. Lake Skadarregion. Specimens included in the analysis For eachspecies, one-way ANOVA's were also per- of life-history traits were collectedfrom the following formed. Pair-wisecomparisons between species and six localities.Bistrica is located on the northernlow- populationswere done by Scheffetest which controls land bank.Lizards were collectedthere from the grassv an experimentaltype I error. Morphologicaltraits were LIFE-HISTORY OF LACERTIDS t27

TABLE l.Comparison of male and femalemorphology among lacertid speciesfrom the Lake Skadarregion. HL (head length), HW (headwidth), HH (headheight), HLL (hind limb length) are presentedas a proportion of snout-ventlength (SVL), Means (+S.E.) and samplesizes (n) are given. ANOVA for HL,HW, HH and HLL was performedon residualsfrom regressionon SVL.

Species SVL HL/SVL HW/SVL HH/SWL HLL/SVL

A. nigropunctatus Males 27 61.27+1.06 0.26+0.00 0.16+0.00 0.l0+0.00 0.56+0.01 Females l0 56.71+l.30 0.23+0.00 0.l4+0.00 0.08+0.00 0.51+0.01

L. oxycephala Males 292 63.01+0.32 0.2710.00 0.1610.00 0.09+0.00 0.59+0.00 Females t96 61.21+0.37 0.25+0.00 0.14+0.00 0.08+0.00 0.55+0.00

P. meliselensis Males 92 59.76+1.20 0.24+0.00 0.l4+0.00 0.10+0.00 0.54r0.00 Females 77 59.70+0.80 0.21+0.00 0.12+0.00 0.09+0.00 0.48+0.00

P. muralis Males 396 70.7 4+0.35 0.26+0.00 0.15+0.00 0.12r0.00 0.57+0.00 Females Jb/ 67.93t0.32 0.22r0.00 0.13+0.00 0.09+0.00 0.52r0.00

Significancelevels from ANOVA

Sex 0.0003 0.0001 0.0001 0.0001 0.0001 Species 0.0039 0.0002 0.0079 0.0005 0.0080 Locality (Species) 0.2183 0.001I 0.0961 0.133 l 0.0554 Sex x species 0.0001 0.0001 0.000l 0.0001 0.0001 Sex x locality (species) 0.0136 0.0001 0.0001 0.0001 0.0027

estimatedon both malesand femalesand analysedby RESULTS partialhierarchical ANOVA on residualsfrom regres- MORPHOLOGY sion of each trait on SVL (both logarithmically transformed).Sex and specieswere consideredas fixed Significantdifferences in SVL were found among (P. main factorsand populationsas randomfactors nested species muralis vs. P. melisellensis and L. (males within species.Effect of sexwas tested over errormean oxycephalo),and betweensexes are largerthan square(MS), effectof speciesover population MS, and females;Table1). Males and females of allspeciesdif- sexx speciesinteraction over sexx populationinterac- fer significantlyin headdimensions relative to body (SVL), (Table tion. Growth rates were estimated by regressing size maleshaving larger heads l). Spe- (2. juvenilelength on months.Comparison of growthrates cies differ in headheight oxycephalavs. Podarcis with of was done by analysis of covariance(ANCOVA). species).Head dimensions correlated the size prey (length Month x speciesinteraction (measure of slope'shetero- items and width) are considerablydiffer- geneity)was testedover month x femaleinteraction. ent betweenspecies. Relative head length discriminates vs. Sinceproduct-moment correlations estimated on popu- long-headedL. oxycephalaand A. nigropunctatus width lationmeans were insignificant due to thesmall number Podarcisspecies. Differences in the relativehead a significantly of populations,correlations betweQn life-history traits are lessapparent; only I orycephalahas species wereestimated on the whole sample.As we dealthere wider head than P. melisellensis.Males of all (relativeto SVL) with structureddata sets, a canonicaldiscriminant have statisticallylonger hindlimbs (Table which indicatestheir higher analysis,which maximizedvariation between groups, than females l), morphologicaltrait, lo- wasused for joint analysisof reproductivetraits instead locomotorcapabilities. For one (within highly of a principalcomponent analysis (James & McCulloch cality effect species)is statistically 1990).We turned to populationsas groupsrather than significant. the species.l. speciesbecause of the smallnumber of speciesused in Egg shape also varies between elongatedeggs (width/length: this study. oxycephalahas the most 128 D. BEJAKOVICET AL.

TABLE 2. Comparisonof reproductivetraits amonglacertid species from SkadarLake region.Mean values(+S.E.) and significancelevels from nestedANOVA's are given. Species (populations) that share the same letter do notdiffer significantly. r- samplesize. Species Population FemaleSVL Numberof eggs Eggvolume RCM (mm) perfemale (mm,)

A. nigropunctatus Bisage 57.84 3.33AB 360.64 0.29 n=3 +2.94 r0.33 30.08 +0.08

L. oxycephala 62.76 3.40A 272.38 0.22 +0.27 +0.10 +38.45 *0.02 Beska 62.49 3.50 310.83a 0.24 n:10 *0.62 +0.27 +16.09 *0.01 Mali Moracnik 63.03 J.JU 233.94b 0.20 n=30 +0.40 +0.11 + 10.00 +0.01

P betweenpopulations 0.4838 0.6323 0.0024 0.0672

P. melisellensis 62.9r 4.05AB 257.02 0.22 12.52 +0.43 +23.78 +0.01 Mali Moracnik 63.38 a 4.90a 217.12a n rl n:20 *0.67 +0.26 r15.50 +0.02 Mala Beska 67.02a 3.60b 299.38b 0.21 n:10 +0.88 +0.16 + 15.l8 +0.02 B istrica 58.33c 3.64b 254.56ab 0.23 n:|4 +0.84 +0.25 +22.19 +0.02

P betweenpopulations 0.000r 0.0007 0.0285 0.7836

P. muralis 70.'t1 5.51B 266.89 0.24 +2.76 +0.49 +39.56 +0.02 M. Starcevo 74.59a 4.60 206.67ab 0.20 n:5 +1.03 r0.24 +54.53 r0.01 Mala Beska 72.18a 5.67 341.44 a 0.25 n:18 +0.9'7 r0.4l L19.66 r0.02 Bistrica 65.36b 6.26 252.56b 0.26 n:19 +0.82 +0.42 *7.99 *0.01

P betweenpopulations 0.0001 0.1642 0.0010 0,3340 P betweenspecies 0.2031 0.071l 0.7867 0.5080

0.51+.0.01),followed by P. melisel/ensls(0.53+0.01), averagelaid-egg female sizesin all speciesstudied. P. muralis (0.57+0.04) and A. nigropunctatus (For the sake of comparison,most female squamates (0.59+0.02)with the most roundedeggs' Differences grow between15 and 30o/oafter maturation; Andrews, amongspecies appeared non-significant (P:0.260), but 1982.)P. muralis producedthe largestclutches and l. differenceswithing species(interpopulation variabil- nigropunctatasthe smallestclutches (Table 2). The ity) were significant (P<0.05). clutch size of P. muralis and L. oxycephals were statis- tically different,though overall interspeciesdifferences in clutch sizeswere not. A significantintraspecies dif- LIFE.HISTORYTRAITS ference in clutch size was found only in P. Size of "laid-egg" femalesdid not differ between melisellensis.The number of femalesin clutch size species, but there was significant variation in classesshowed quite different distributionpatterns, interpopulationcomparisons in P. melisellensisand P' whereP. muralis was distinguishedby the largestvari- muralis(Table 2). Femalesmature at minimumSVL of ation in clutch size (Fig. l). Egg size did not differ 5l .0 mm (P. melisellensis)' 53'3 mm (A' significantlyamong species, but therewere remarkable nigropunctatus), 53.3 mm (2. oxycephala) and 55'0 differenceswithin species(Table 2). The measureof mm (P. muralis). Thus, femalesat the onset of repro- the reproductiveinvestment per clutch, RCM, was ductive life have alreadyreached more than 80% of the nearly uniform for the Skadar Lake lacertidsranging LIFE-HISTORYOF LACERTIDS t29

A nigopunc{alusTABLE 3. Betweenpopulation canonical structure from canonicaldiscriminant analysis performed on reproductive traits measuredin four lizard soeciesfrom SkadarLake region.

CAN I CAN 2

Number of eggs per female 0.91I -0.400 Egg volume 0.551 0.800 RCM 0.685 0.015 L, oryeptnla Eigenvalue r.53t 7 0.4979 Proportion 0.'1256 0.2359 Cumulative 0.7256 0.9615 P 0.0001 0.000I o 20 CORRELATIONBETWEEN LIFE.HISTORY TRAITS tto t'" As expected,RCM was significantlycorrelated with g the number of eggsper brood, and with the volume of t(rr |lt eggs(r:0.471, P<0.001,and r:0.593, P< 0.001,re- t spectively).The body size (SVL l.10 values) also was significantly correlatedwith thesevariables of repro- ductive potential (r0.424, P<0.0001, r0.288, P<0.01; respectively),as was found in most reptiles (Fitch, 1970;Dunham et al.,1988). There was no rela- tionship betweenthe number of eggs and the size of hatchlings (r-0.24, P:0.439), or between RCM and the sizeof hatchlings(r:-0.167, P:0.585). It is possible that small samplesize precluded confirming suchrela- tionships. Egg shape (width/length ratio) was significantly positively correlatedwith the number of eggsper female,egg volumeand RCM (r : 0.47,0.35 and 0.39, respectively),and significantlynegatively 345 6 78 correlated with total length of hatchling (r:-0.70). Therefore,more round egg shapeassociates with larger CLUTC H stzE brood, bigger egg size and, concomitantly,higher re- productive investment(RCM), while newborn lizards FIG. l. Clutchsize distribution in fourlacertid soecies from are bigger when they hatch from more elongatedeggs. Skadarlake region. MULTIVARIATE COMPARISoNS

Reproductive variables were estimated for fiom 0.29 in A. nigropunctatusto 0.22 in L. oxycephala populations of P. muralis, P. melisellensis and L. and P. melisellensis.Time to hatching and size of oxycephala, and for one population of A. hatchlingswere studied in three populationsof P. nigropunctatrs.A canonicaldiscriminant analysis was melisellensis, and in one population of P. muralis performed on this data set. CANI axis took72.60/oof (Malo Starcevo)and L. oxycephala(Beska). Minimum variation in the original variableswhile 23.6Yoof that and maximum time necessaryfor newbornsto hatch variation could be describedby CAN2 axis (Table 3). fiom eggsin incubatorswere found in P. melisellensis CANI describeda positiverelationship among all three and rangedfrom 33.4*2.8days (Bistrica) to 38.0+0.4 variables,while the egg number was negativelycorre- (Mali Moracnik).The smallest,newbornswere those of latedwith CAN2 axis and, at the sametime, contrasted P. melisellensrs(Bistrica, 65,5+3.3 mm) and the big- strongly with the egg volume. Specieswere mostly gestthose of P. muralis(Malo Starcevo,7l.2+3.0 mm). separatedby the egg numberon CAN I axis and by the We found no significantdifferences between species egg volume on CAN2 axis. In the CANI and CAN2 for theselife historytraits, nor betweenpopulations of scatterdiagram, populations separated well by their P. melisellensrs.The somewhatfaster juvenile growth speciesdesignations (Fig. 2). Specieswere differently rate of P. muralis (Malo Starcevopopulation) com- distributedalong the first axis. P. muralis populations pared to L. oxycephala(Beska) was insignificant,too had positive values of CANI axis, while the (ANCOVA, P:0.124). populations of L. oxycephala and A. nigropunctatus 130 D. BEJAKOVICET AL

CAN2 Cylindrical body morphology, on the other hand, en- hanceslocomotor capabilities in more open,usually vegetatedhabitats. Predator pressures on surface-active nrSropunclalus lizards(e.g. when basking) in the SkadarLake region tb4o are presumablysimilar. However, potential predators o of thesespecies once they are within crevicesmay be quite different. Their known predatorsin the Skadar Lake region incl'ude snakes (Malpolon od,rt ,* M.Bda monspessulanus,Coluber laurenti, Natrix tessellata), somebirds (e.9.Ardea purpureq, Larus spp.)and rats L. orycephala (Rattusrattus). Becauseof the cylindrical morphology ofsnakes,they arenot capable P. meliselensis H.Sblrrro ofpredating lizardsonce [.lM( o they are in a rock crevice:the flattenedmorphology de- I üsfia 0 P. muralis creasesthe vulnerability to predation. Thus one may expect lower predator pressureon L. orycephala andA. nigropunctatusthan on thosePodarcis speciesstudied and, concomitantly,their higher lifetime reproductive success.Concomitant reproductive adaptations, associ- ated with the effect of habitat selection on flattened morphology,usually include a reducedclutch size and more elongatedeggs compared to other relatives(Vitt, 1981;1993). The consequenceis thatnon-reproductive females are as flat as females carrying eggs. In Z. oxycephala,and in l. nigropunctatzs to a lesserextent, reproductive adaptations that appear to be conse- FIG. 2. Meanpopulation scores on firsttwo cannonicalaxes quencesofthe exclusiveuse ofrock outcropsarejust as for four lacertidspecies from Skadarlake region. Algyroides nigropunctatu.r-solidcircles Lacerta oxycephala-solid notedabove. However, reduced clutch size in this spe- squares;Podarcis melisellensis-empty squares]. Podarcis cieswas not associatedwith reducedclutch volume, as muralis- emptycircles. was found in someflat lizards(Vitt, l98l). Also, con- trary to other lizards in which morphological had the most negativevalues along this axis reflecting adaptationsassociated with the useof crevicesin rocks their tendencyfor the most dissimilarclutch size.P. for predatorescape constrain RCM value (Vitt, 1981), melisellensispopulations took the middle position be- thereproductive investment per clutch of L. orycephala tweenthese species. Two pairsof closelyassociated l. and,A. nigropunctatus is about the same as in other oxycephala and P. melisellensis populations (Mali lacertidsfrom the SkadarLake region. Moracnik- Bistrica,Beska - Malo Besko)emerged due Analyses made to date have suggested that to closesimilarities in the eggsize and the egg volume. phylogeneticrelationships account for a large propor- Syntopicpopulations (Mali Moracnik - P. melisellensis tion of the variationobserved in life-historyvariables and L. oxycephala; Bistrica - P. muralis and P of squamata(e.g. Dunham & Miles, 1985).This seems melisellensis;Malo Besko - P muralis and P to be provenin our studyas well. When reproductive melisellensis)differed considerablyin their positionin- characteristics,which determinatethe coevolutionary framework for other life-history features(see Stearns, 1976; 1977)are compared,populations of P. muralis andP, melisellensisappeared to be closerto eachother dicatingspecies-specific effect on reproductivecharac- than populations of other the two species (L. teristics,rather than site effect. oxycephalaand A. nigropunctatus).The most common DISCUSSION species(P. murolis andL. orycephala)are on the oppo- site sidesin reproductivetraits similarity. Both tactics The following discussionis centeredaround the seemto work well in the SkadarLake lizard commu- main questionin this paper:How do thesevariations in nity. Apparently, species level effects here are life-history traits correlatewith the speciesbod,"' shape. supersededby genus level effects as proposedby phylogeneticrelatedness, distribution areas and eco- Stearns( 1984) for the squamatereptiles. logical demands?Individuals of L orycephala'andto a Speciesdistribution patterns in CAN I andCAN2 re- lesserextent of A. nigropunctatus'are dorsoventrally productive-dependentspace corespond to theirgeneral flattened,showing striking dissimilaritiesto stream- ecologicalrelatedness. Both l. nigropunctatusand L. lined morphology of Podarcis species. Such a oxtcephalaare highly specializedwith respectto their morphologyis adaptiveas it allowsindividuals to seek habitatchoice. Contrary, P. muralisis more generalist refuse in narrow creviceswhen disturbedor inactive' with regardto microhabitatpreference. It seemsthat P. LIFE-HISTORYOF LACERTIDS 131 melisellensisis probably, in this respect,somewhere (Podarcis muralis) from the Skadar lake region, between these two groups. As stated above, Montenegro. Revista Espanola de Herpetologia in intraspecificvariations in reproductivetraits are appar- press. ent, but well within speciesmultivariate framework. Crnobrnja-Isailovic,J. & Dzukic, G. (1995). First report Suchvariation is due to adaptationsto specificenviron- about conservation status of herpetofauna in the mental conditions and plasticity in characteristicsthat Skadar Lake region (Montenegro): current situation reflect immediateresponses to proximateconditions. In and perspectives.In'. Scientia Herpetologica. Llorente this respectinsular populations,at leastthose of P. et al. (eds.),pp.373-380. Barcelona. muralis andP. melisel/enslsspecies, differ from conti- Dunham, A. E. & Miles, D. B. (1985). Patterns of nental populationsin body size and clutch size. Island covariationin life history traits of squamatereptiles: lizardsare largerthan conspecificson the nearbyconti- the effects of size and phylogeny reconsidered.lm. nentalarea. It seemsthat lizardson islands,under high Nat. 126,231-257. intraspecificcompetition (islands in SkadarLake have Dunham,A. E., Miles, D. B. & Reznik,D. N. (1988).Life much denserpopulations than the mainland;unpub- history patterns in squamatereptiles. ln: Biology of lished data),allocate more energyto somaticgrowth Reptilia.C. Gans& R.B. Huey (eds.),16, pp.44l-522. than to reproduction.In contrast,the life-historv strat- Alan R. Liss. Inc., New York. egy of the continentalpopulations could be to invest Gruschwitz,M. & Bohme, W. (1986). Podarcis muralis more in the number of offspring than in growth. (Laurenti, 1768) - Mauereidechse.ln: Handbuch der In conclusion,our analysissubstantiates the claim Reptilien und Amphibien Europas. Band 2/II. W. that much of the patternof covariationin life-history Bohme (ed.),pp. 155-208.Aula-Verlag, Wiesbaden. datacan be explainedby lineage-specificeffects, either Fitch. H. S. (1970). Reproductivecycles in lizards and constrainedby or coadaptedwith morphologicaldiffer- snakes.Univ. Kansas Publ. Mus. Nat. Hist. 13, 85- ences, even if phylogenetically close taxa are 288. compared. Henle, K. (1990). Populationecology and life history of three terrestrial geckos in arid Australia. Copeia ACKNOWLEDGEMENTS 3:759-781 . DraganaCvetkovic helpedin preparationof the final James,F. C. & McCulloch, C. E. (1990). Multivariate version of this manuscript.We wish to thank Laurie J. analysis in ecology and systematics: Panacea or Vitt and an anonymousreviewer for a numberof useful Pandora'sbox? Ann. Rev.Ecol. Syst.2l, 129-166 suggestionswhich improvedthe quality of the paper. Lasca, N., Radulovic, V., Ristic, R. & Cherkauer,D. (1981). Geology, hydrology, climate and bathymetry Lake Skadar. In'. The Biota and Limnology of Lake REFERENCES of Skadar.C. S. Karaman& A. M. Beeton(eds.), pp. l7- Abts.M. L. (1987).Environment and variationin Iife 38. Institut za bioloska i medicinska istrazivanja, historytraits of the Chuckwalla,Sauromalus obesus. Titograd,Smithsonian Institution, Washington. Ecol. Monogr. 57, 21 5-232. Pianka, E. R. (1986). Ecology and Natural History of Aleksic,I., TucicN. (1994).Changes in thecomparisons DesertLizards. Princeton University Press,Princeton. of phenotypicvariance and covariance among traits SAS Institute(1989). SAS/STAT User'sGuide, version 6. during ontogenyin the sharp-snoutedrock lizard Forth E,dition. (Lacertaoxycephala). Amphibia-Reptilia 15, 47 -61. Shine, R. (1980). "Cost" of reproduction in reptiles. Andrews,R. M. (1982).Patterns of growthin reptiles.In: Oec ol ogia (Berl.) 46, 92-100. Biologyof Reptilia.C. Gans& F.fL Pough(eds.), 13. Stearns,S. C. (1976). Life history tactics: a review of pp.273-320.Academic Press, New York. ideas.Q. Rev.Biol. 51,3-4'7. Arnold, E. N. ( 1989). Towardsa phylogenyand Stearns,S. C. (1977). The evolution of life history traits. biogeographyof theLacertidae: relationships rvithin Ann. Rev.Ecol. 9yst.8, 145-171. an Old-World family of lizards derived from Stearns,S. C. (1984).The effectsofsize and phylogenyon morphology.Bull. Br. Mus.nat. Hist. (Zool.) 55.209- patterns of covariation in the life-history traits of 257. lizardsand snakes.Am. Nat. !23,56-'12. Bejakovic,D., Kalezic,M. L., Aleksic,L, Dzukic,G., & Tiedemann,F. & Henle,K. (1986).Podarcis melisellensis Crnobrnja,J. (1995).Female reproductive cycle and (Braun, 1877) -Adriatische-Mauereideschse, clutch traits in the Dalmatianwall lizard(Podarcis Karstlaufer. In: Handbuch der Reptilien und melisellensis).Folia Zoologica44,37 l-380. Amphibien Europas. W. Bohme (ed.), pp. 1I I - 141. Bejakovic,D., Aleksic, I., Crnobrnja-lsailovic.J..Dzukic, Aula-Verlag,Wiesbaden. G. & Kalezic,M. L. (1995).Reproductive cycle and Tinkle, D. W. & Dunham,A. E. (1986).Comparative life clutch size in femalesharp-snouted rock lizard, historiesof two syntopic sceloporinelizards. Copeia Lace r ta oxycep hala.A mp hi b i a- Re ptilia (in press). r986, l-18. Bejakovic,D., Aleksic, 1., Crnobrnja-lsailovic,J., Vitt, L. J. (1981). Lizard reproduction:habitat specificity Dzukic, G. & Kalezic, M. L. (1996). Female and constraintson relativeclutch mass.Am. Nat. ll7, reproductivetraits in the common wall lizard 506-514. r32 D. BEJAKOVICET AL

Vitt, L. J. (1986). Reproductive tactics of sympatric Vitt, L. J. & Congdon, J. D. (1978). Body shape, gekkonid lizards with comment on the evolutionary reproductiveeffort, and relativeclutch mass in lizards: and ecological consequencesof invariant clutch size. resolutionofa paradox.Am. Nat. 112, 595-608. Copeia1986,773-786. Vitt, L. J. & Price, H. J. (1982). Ecological and Vitt, L. J. (1993). Diversity of reproductive strategies evolutionary determinantsof relative clutch mass in among Brazilian lizards and snakes:The significance lizards.Herpetologica38,237-255. oflineage and adaptation.ln ReproductiveBiology of South American Vertebrates.W. C. Hamlett (ed.), pp. 135-149.Springer-Verlag. Accepted: 26'4.96