Gastroliths or ‘stomach stones’ occur frequently in some, but not all, groups of fossil and living m arine tetrapods. C om parative analysis of gastrolith distribution suggests a role in buoyancy control rather than food processing. Once accidental ingestion by bottom-feeding animals is excluded, gastroliths occur in most tetrapods which ‘fly’ underw ater with hydrofoil limbs, including plesiosaurs, penguins, and otariid pinnipeds, but not the m arine chelonians. They do not usually occur in cetaceans, ichthyosaurs, mosasaurs, and odobenid and phocid pinnipeds, which swim with a caudal fin or the equivalent. Occurrence in amphibious forms is variable; crocodilians often have gastroliths, but nothosaurs and placodonts do not. T he correlation of gastroliths and underw ater flight is corroborated by a com parative analysis which takes phylogenetic factors into account. There is no correlation with diet. Consideration of function and occurrence in terrestrial forms suggests that the use of gastroliths in digestion would not be useful, and might even be harmful, to a carnivorous m arine tetrapod. Gastroliths are more efficient than skeletal bone (as in pachyostosis) in terms of sinking force per unit of added mass or volume. As well as driftwood and ice, m arine tetrapods should be considered as a potential source of erratic stones in freshwater and marine sediments. Gastroliths may have evolved by the accidental ingestion of stones, the retention into adulthood of stones used by juveniles to process insect or plant food, or as a com pensatory replacem ent for dense bones habitually filling the stomach. Their presence or absence should be more carefully recorded and further studies should be carried out on their function.
