Japanese Journal of Herpetology 18(2): 57-65., Dec. 1999 (C)1999 by The HerpetologicalSociety of Japan Spatial Utilization and Social Interactions in Oplurus cuvieri cuvieri (Squamata, Opluridae) in Madagascar HERILALA J. A. R. RANDRIAMAHAZO AND AKIRA MORI Abstract: The spatial distribution pattern and associated social interactions of Oplu- rus cuvieri cuvieri were studied in a deciduous dry forest of Ampijoroa, Madagascar. Home range sizes were significantly larger in males than in females. In both sexes, snout-vent length was not correlated with home range sizes. Home range of males overlapped both inter-and intrasexually. Female home ranges rarely overlapped in- trasexually. Consexual aggressive interactions indicated the presence of territoriality. Distribution pattern and observed intersexual interactions suggested a polygynous mating system. Site fidelity was observed for both home range and shelter tree levels. Tail breakage was moderately frequent, implying high predation pressure and the im- portance of safety refuge. No sexual differences were observed in thermal environ- ment, perch height, or perch diameter, whereas sexual dimorphism in body size and dorsal color pattern was found. More than 80% of the lizards of both sexes performed "resting" . Key words: Oplurus cuvieri cuvieri; Opluridae; Spatial distribution; Social interac- tion; Madagascar Stamps (1977) reported a rich variety of spac- area of Madagascar. Populations of Oplurus ing systems ranging from sharp territoriality to cuvieri on Grand Comoro have evolved differ- broadly overlapping home ranges in lizards. ently from the Malagasy counterparts and are When accurately estimated, distribution patterns considered as a subspecies, O. c. comoriensis, of home ranges can lend insight into the evolu- (Angel, 1942; Blanc, 1977; Meirte, 1992). All tion of the mating system (Rose, 1982). Most other species are endemic to Madagascar. The lizard species have adopted the polygynous mat- systematics of this group has already been ing system (Stamps, 1983). Several studies in revised (Titus and Frost, 1996), but ecological lizards have focused on the role of ecological and behavioral information is scarce. factors in determining spatial distribution pat- We studied the spatial distribution and the as- tern, which accordingly affect the mating system sociated habitat use in a Malagasy oplurid, (e.g. Davies and Lundberg, 1984; Hews, 1993; Oplurus cuvieri cuvieri (Gray, 1831). We exa- Colli et al., 1997). For instance, multi-purpose mined the sexual differences in habitat utilization territories in Sauromalus obesus offering food, as well as basic morphometric characters to shelter, and access to females are often main- clarify the influence of morphological characters tained by males, but in terms of importance, and sex on lizard activities. The mating system food defense is suggested as a consequence of a of this lizard is theorized on the basis of the ob- territorial defense (Berry, 1974). Similarly, in served social interactions and spacing patterns. Uta palmeri, the availability of resources was suggested to be a more important factor than MATERIALS AND METHODS male phenotypic traits in determining females to Oplurus c. cuvieri is a medium-sized, diurnal reside, and thus their mating partners (Hews, lizard which is basically arboreal having a tree 1990, 1993). hollow especially selected for a night shelter The Opluridae was recently elevated to the fa- (Blanc, 1977). Brillet (1982) reported that mily rank from a subfamily of the Iguanidae Oplurus lizards were basically insectivorous, but (Frost and Etheridge, 1989). Two genera fruits and leaves were also accepted in captivity. (Chalarodon and Oplurus) compose the Opluri- The present study was carried out during the dae and have successfully invaded the dry forest early parts of their breeding season (from Sep- tember 1997 to January 1998 and from Septem- Accepted 1 Dec. 1999 ber 1998 to November 1998), when the lizards 58 Jpn. J. Herpetol. 18 (2) 1999 actively courted and mated. toe clipping, color painting, and passive in- The study area (16.6ha) was a quadrat plot of tegrated transponders (PIT), and then released a dry deciduous forest located at the Jardin at the site of capture. Botanique A of Ampijoroa, Madagascar Home range size was calculated by the poly- (16°15'S and 46°48'E). The elevation is ap- gon convex method (Jennrich and Turner, 1969) proximately 200m a. s. l. The vegetation was on the basis of more than six observations for dominated by three species of Strichnos general- each adult lizard. By tracing individual home ly 9-12m in height (Razafy, 1987; Urano et al., ranges on paper, we used the cut and weigh 1994), and scattered shrubs and open areas with method to measure the individual area and the abundant leaf litter were present. The quadrat overlapping regions of other lizards. Three plot was subdivided into 5m×5m subplots by overlap measurements (percent overlap, overlap trails. All trails were human-made and regular- pressure, and encroachment) described by ly maintained for tourism purposes. In this Sinervo (1994) and Abell (1999) were calculated. region, the climate is mainly dominated by a hot Basically, the former two measurements indicate rainy season from November to April alternat- how much of the home range of a given individ- ing with a cold dry season from May to October. ual is overlapped by other individuals, and the To find lizards, we walked slowly along the latter one indicates to what extent the lizard edge of a path or trail on a daily basis approxi- overlaps the home ranges of other lizards. mately between 0800h and 1800h. We started The mean Ta of resighting was used to indicate the census from different locations, but all trails the thermal habitat of a given individual, and were censused in every session of a daily survey. comparison was made among individuals of the Only the first sighting per lizard per day was used same sex and between the sexes. Variations in in the following data analysis even when more perch height and diameter between the sexes (us- than one sighting was recorded. When a lizard ing mean values) were also compared to examine was sighted, we recorded the posture and behav- the stratification in habitat utilization. ior of the lizard, whether the lizard was in sun- Tail break frequency was determined for the light or shade, and whether it was in the shelter population over a two-year period. Because (1) or out of the shelter. We also recorded the lo- intense intrasexual interactions deduced from cation of the site on the map and measured the the presence of fighting marks or recent scars on perch height (using a stick ruler to the nearest the head of male lizards did not manifest any in- 5cm), perch diameter (to the nearest 1mm), and cidence of tail break and (2) no biting tail at- ambient temperature (Ta). Direct neighboring tempt was observed during experimental male- lizards were recorded if also sighted, and social male contests (Randriamahazo, unpubl. data), interactions were described. Lizards observed the tail break frequency is considered to be an for the first time were captured by noose. At indirect measure of predation. Besides, shelter- each capture, we measured the cloacal body ing lizards have often a portion of their tails left temperature (Tb; to the nearest 0.1℃), body outside the hollow (Blanc, 1977), and if a preda- mass (BM; to the nearest 1g), snout-vent length, tion attempt aimed at lizards in refuge is more tail length, and regenerated tail length (to the likely to result in tail break, the lizards would nearest 1mm for measured lengths). Occur- show behavioral attachment to such shelters as rence of regenerated tail was confirmed by the provide more safety to the tail. We estimated disarrangement of the spines on the tail (Blanc, the site fidelity in this species for home range 1977). We also described the dorsal pattern of level and for shelter level. Lizards recaptured each captured lizard using the number of dorsal and resighted within the home range of the transverse black bands (DBB) and the number of previous reproductive season were defined as scales in a row of the 2nd DBB, which was al- prossessing a certain degree of home range fidel- ways present. Sex was determined by the ity. When a given lizard actively used a shelter presence (male) or absence (female) of a yellow- for the two consecutive years, we considered that ish ventral patch on the cloacal scales (Blanc, sheltering fidelity was present. 1977). Based on preliminary observations and To estimate abundance following Bailey's this study, male lizards that were larger than triple-catch method (Bailey, 1952; Marvin, those which performed sexual behaviors 1996), we divided mark-recapture data of each (SVL>133mm) were considered mature lizards. individual during the first study period (Sep 1997 In females, the smallest size found laying eggs to Jan 1998) into three equal capture intervals of (SVL=120mm) was used to determine the 25 days, assuming that capture and resight maturity. Lizards were individually marked by probabilities were constant. Recruitment and RANDRIAMAHAZO AND MORI-SOCIAL BEHAVIOR IN LIZARD 59 TABLE 1. Body measurements, cloacal body temperature, and related ambient temperature at capture in Oplurus cuvieri cuvieri. *Variables tested with one -way ANOVA for sexual differences . loss rates were also estimated using the same 1315.88±336.55m2, N=17; females=256.43± method. 70.65m2, N=14). The SVLs of females used in the home range estimation were significantly RESULTS smaller than those of males (F1,29=106.73, Among captured lizards, males were larger in p<0.0001; males=156.35±1.49mm, N=17; overall body size and had fewer mean dorsal females=133±1.66mm, N=14).Neither black bands than females (Table 1). However, males nor females showed significant correlation the number of scales in a row of the 2nd band between SVL and home range size (Kendall rank and the regenerated tail length showed no sexual correlation: males, τ=0.29,N=17, p=0.1;fe- difference (Table 1).