From the text...'It is abundantly clear that for all land tortoises, environmental limitation, so obvious in the xeric habitats in which these reptiles live, is more related to density than to total tortoise population. This is clearly shown in transects through Gopherus polyphemus populations (Table 8) in several Florida and Georgia localities. Differences in density are related to two major factors: removal rate by humans, and habitat quality (of which food resource is most important). In this species the limiting food resource is largely grass, whereas in other tortoise species herbs and fruits may be important. Table 8 shows that in G. polyphemus habitat grass cover is clearly correlated with tortoise density and is probably the major factor regulating density in undisturbed populations. In low-lying areas, particularly near the coast, groundwater level may be a determining factor in both spatial distribution and population density, for this species tends to be absent from areas where deeper soils are permanently flooded. Figure 2 shows that even within the same community, density of resident tortoises may vary considerably throughout the population, dependent on several ecological factors, of which grass biomass is the most important. The distribution of grass type and density varies on the hillsides, based on edaphic factors, such as soil chemistry, moisture, and shade. In areas with less relief, edaphic features often vary less markedly, so that density zonation is a less common phenomenon. However near the boundary of differing soil types, zonation of this type is sometimes readily obvious. Density change in time that is not due to mortality has not previously been demonstrated in land tortoise populations. Long-term studies on G. polyphemus in Florida now make it clear that population change due to natural processes are common in this species. For the most part such changes are due to primary community succession. The best data are from a study site in Lake County, Florida, where a single population has been under observation since 1960. During the intervening years the study area changed from an old field community to mature sand pine scrub. During this time, gopher tortoise density in the area changed from one tortoise per 1330 m2 in 1960 to one tortoise per 26,230 m2 in 1971; a reduction of 1800% in 11 years. There was no evidence of increased mortality, but only of emigration, increasing in intensity as open habitat grasses were replaced with subcanopy herbs and finally a thick pine needle duff.'