-Ameiva ameiva, Cnemidophorus lemniscatus, and Kentropyx striatus reproduce seasonally in Amazonian savanna near Alter do Chao, Brazil. Most egg deposition occurs during the wet season but most development of gonads and growth of juveniles occur during the height of the dry season. Relative fat-body size varies little throughout the year and there is no apparent correlation between fatbody size and gonad development. Reproductive cycles of reptiles in seasonal tropical climates have been considered to be correlated with rainfall, most egg deposition occuring during the wet season (Fitch, 1982). However, recent work indicates that the reproductive cycles of tropical lizards show great variability and that reproductive cycles are not always tightly linked to rainfall patterns (Vitt, 1986). Variation in the availability of food is generally given as the reason for cycles. On the other hand, Stamps and Tanaka (1981) have shown that the availability of water may limit the growth of Anolis aeneus independent of the effects of food availability, and Rose (1982) found that supplemental feeding increased fat stores but not reproduction in Anolis acutus. Dunham and Miles (1985) suggest that phylogenetic constraints and biogeographic history may have played major roles in shaping squamate reproductive patterns. James and Shine (in press) suggest that, in northern Australia, lizard species reproduce in the season (wet or dry) that most closely simulates climatic conditions where their antecedents evolved. The data base is insufficient to define clearly the reproductive patterns of lizards from seasonal tropical environments. Here I present data on the reproductive cycles of three sympatric species of teiid lizards (Cnemidophorus lemniscatus, Ameiva ameiva, and Kentropyx striatus) which occur in Amazonian savanna. Seasonal variation in interstitial testicular tissue (Del Conte, 1972) and testis volume (Leon and Cova, 1973) has been reported for Venezuelan C. lemniscatus but female C. lemniscatus with oviductal eggs were found in the garden of the Universidad de Orient en Cumana in most months (Leon and Cova, 1973). Ameiva ameiva in rainforest habitat in Ecuador have strongly seasonal reproduction (Simmons, 1975) but A. ameiva in the arid caatinga of northeastern Brazil breed continuously and show only weak seasonality which is not directly related to rainfall (Vitt, 1982). I know of no studies of the reproductive cycle of K. striatus. Reproductive cycles can be inferred from: (1) longitudinal studies of individual animals; (2) gonadal cycles of males and females; (3) population age structure; (4) the presence of eggs or neonates in the population; or a combination of all four lines of evidence. In this study I present data on aspects (2), (3), and (4) for Cnemidophorus lemniscatus, (3) and (4) for Ameiva ameiva, and (4) only for Kentropyx striatus. During a previous study (Magnusson et al., 1985) we noted that the mean clutch size differed between C. lemniscatus collected in the savanna and on freshwater beaches. The densities of C. lemniscatus in these two habitats also differed significantly (Magnusson et al., 1986). I therefore collected data on possible population differences between habitats for that species. Fat-body cycles are often closely related to reproductive cycles in temperate lizards (Derickson, 1976). However, there may be less relationship between fat-body cycles and reproduction in tropical lizards (DerThis content downloaded from 157.55.39.104 on Sun, 19 Jun 2016 06:07:10 UTC All use subject to http://about.jstor.org/terms WILLIAM E. MAGNUSSON 7 5 7 3 2 7 4 A e. /-Reproductive A Ameiva / /GArovid