O. Gy. Dely szerk.: Vertebrata Hungarica 22. (Budapest, 1984)

Korsós, Z.: Comparative niche analysis of two sympatric lizard species (Lacerta viridis and Lacerta agilis) 5-14. o.

The third and perhaps most important factor allowing coexistence is the sharing of food. The two mainly insectivorous lizard species have apparently very similar prey base (Coleoptera, Orthoptera, Lepldoptera larvae, Arachnoidea), but they differ with respect to species and sizes. Such a segregation may only be clarified by stomach content analyses. However, these studies are usually limited to small sample sizes (20-50 specimens) because of the methodological difficulties and the drastic interference with the native populations, especially in the case of lizards (SMITH & MILSTEAD 1971, VERON 1969). The identification of the partly digested preys may also be rarely carried out to taxa lower than families. With respect to taxonomic composition there are 12 common food categories of the two species. These are usually various genera of mainly Insecta and Arachnoidea. Diversity of stomach contents has been calculated according to SIMPSON (1949): L. viridis 0.088, L. agilis 0.141. The weighted frequency distributions of the prey groups are shown in Fig. 4. The overlap of the 68 distinct categories is only 18.5 %. While the sand lizard has a greater Orthoptera, Diptera and Arachnoidea consumption, the green lizard selects more of caterpillars and beetles like Cara- bus or Cetonia. The latter preys however, lead us to the problem of food size segregation. Fig. 5. Food composition of lizards according to the prey size The frequency distributions in Fig. 5 have been obtained by pooling the prey groups with equal average body length. The overlap between the food sizes taken by the two lizard populations has been found as 35.2 %. On the specimens from which the stomach contents had been analysed, four head measurements were taken. These were suspected to be correlated with the size of prey consumed. A discriminant analysis on the four variables has completely separated the two populations: the overlap (i.e., probability of misclasslflcatlon) was equal to 0 (Fig. 6).