Effects of incubation temperature on hatchling phenotypes in an oviparous lizard with prolonged egg retention: are the two main hypotheses on the evolution of viviparity compatible?

Females of several lizard species modify their body temperature during pregnancy, probably in connection with the optimisation of hatchling phenotypes. We studied variations in the temperature selected by gravid females compared with those selected by males and non-gravid females in an oviparous population of Zootoca vivipara (Jacquin, 1797) (Squamata: Lacertidae) of Northern Spain and examined the effects of incubation temperature on the phenotypic variation of hatchlings. Cloacal temperatures of gravid females active in the field were lower than those of males and non-gravid females, as well as the temperatures selected in a thermal gradient created in the laboratory (mean+/-s.d.: 32.33+/-1.27 degrees C for gravid females; 34.05+/-1.07 degrees C for males and non-gravid females). Effects of temperature were assessed by incubating eggs at five constant temperatures (21, 25, 29, 32 and 34 degrees C). Incubation time decreased as temperature increased, following a negative exponential function. Incubation temperatures also affected the hatchlings' morphology: hatchlings incubated at 34 degrees C had shorter heads than those from other temperatures. Survival at 34 degrees C (58%) was significantly lower than at the other temperatures (mean 93%). Pregnant females select lower body temperature, approaching the temperatures that optimise hatchling phenotypes, according to predictions of the maternal manipulation hypothesis on the evolution of viviparity. The shift in preferred temperature by pregnant females would result in only a very short delay, if any, of hatching time and, because the temperature selected by pregnant females is much higher than average temperatures recorded in natural nests of Z. vivipara, egg retention considerably shortens incubation time, according to predictions of the cold-climate hypothesis. Our experimental results indicate that the two main hypotheses on the evolution of viviparity are compatible in our study model.