Population processes in a large herbivorous reptile: the giant tortoise of Aldabra atoll.

Physical barriers divide the population of giant tortoises (Geochelone gigantea Schweigger) on Aldabra into several sub-populations of different density, which nevertheless are similar genetically. We measured individual growth rates in each sub-population. Mortality was estimated using data from Bourn and Coe (1979). Reproduction and recruitment were studied using data from previous work (Swingland and Coe 1979) and our own estimates of clutch size, egg weight, and laying frequency from 1975 to 1981.Individual growth rates were strongly dependent only on individual size and sub-population density and not on age or sex. Within a sub-population, the relationship between specific growth rate and size (linear measure) was best fitted by a Gompertz model, except for very young tortoises which grew faster in volume, though not in weight, than expected. Animals at high densities grow slowly to a small size whereas those at low densities grow fast to a large size. At very high density many juveniles remain at a small size without growing or maturing.Mortality of larger (> ca. 5 years old) animals was independent of density, but did depend on size in the highest-density sub-population, as predicted by the Gompertz growth model.Reproduction and recruitment were negatively density-dependent over the whole density range (5 to 35 animals ha-1) studied. Clutch size and laying frequency were strongly influenced by sub-population density, but egg weight was not. Laying frequency varied within sub-populations according to rainfall (presumably via annual food supply).All except one sub-population are seen as stages in the development of the same interactive system. Competition between individuals is nearly, but not purely, of scramble type. The remaining sub-population is either a distinct interactive system in which food supply for very young animals is important, or it is a non-interactive system controlled by the effect of natural enemies on very young animals. This suggests that the equilibrium density and/or dynamics of giant tortoise populations are highly sensitive to mortality factors affecting very young animals.In low density sub-populations the animals are large, have many young, low relative reproductive effort, and a short generation time. In high density sub-populations they are small, have few young, high relative reproductive effort, and a long generation time. This variation is largely phenotypic. It is anomalous with respect to r-K life history theory but is a logical consequence of indeterminate growth combined with size-determined risk and benefit functions and may have contributed to the giant tortoises' success as island colonisers.