ABSTRACT: It is important to understand the population structure and space use of sea turtles because of their potential effects on the dynamics of nearshore ecosystems. Much of our current understanding is skewed because the ecology of sea turtles on their foraging grounds is known best from areas where there are, or have been, major human impacts. We studied green Chelonia mydas and loggerhead Caretta caretta sea turtles on the relatively pristine feeding ground of Shark Bay, Western Australia, where tiger sharks Galeocerdo cuvier are a major mortality agent. Recapture rates were lower for green than for loggerhead turtles, possibly reflecting a larger population rather than low site fidelity for greens. The sex ratio of loggerhead turtles was not significantly different from 1:1, whereas green turtles showed a strong female bias. Size distributions of both species were skewed towards larger (and presumably older) individuals relative to a human-impacted feeding ground ineastern Australia. Body condition varied temporally for green turtles, but not for loggerhead turtles, possibly due to longer distances traveled to nesting beaches by green turtles. Rates of shark-inflicted injuries were higher for loggerhead turtles, especially males, than for greens. Sublethal effects of these injuries were not evident. There are notable differences between sea turtles threatened primarily by tiger sharks in Shark Bay and populations on feeding grounds where historical and current mortality causes are anthropogenic. We conclude that without baseline data from relatively pristine habitats our understanding of human impacts on sea turtle populations suffers from a limited scope.