While many migratory marine organisms converge at breeding areas, identifying foraging strategies away from these reproductive sites can be challenging. Adult female green turtles Chelonia mydas regularly migrate thousands of kilometers between nesting and foraging areas, making it difficult to identify foraging habitats that support nesting populations and to understand their feeding strategies. In this study, we use stable isotope analysis to investigate the trophic ecology and spatial distribution of foraging green turtles in the Greater Caribbean. Further, we explore the possibility that adult green turtles, originally considered to be herbivores, may, like their counterparts in the Pacific Ocean, display carnivorous feeding strategies. The wide range of carbon and nitrogen isotope values in bulk epidermis observed in the nesting population at Tortuguero, Costa Rica, could indicate that these turtles feed over several trophic levels. Isotopic niches -or the range of δ 13 C and δ 15 N values, which can be used as a proxy for ecological niche -varied among the 5 green turtle foraging aggregations sampled. Similarly, the isotopic composition of the primary producer Thalassia testudinum also varied substantially with geographic location. However, compound-specific stable isotope analysis of amino acids (AA-CSIA) indicated that individuals in the nesting population with different bulk δ 15 N values feed at the same trophic position. The combined results suggest that spatial differences in the isotopic composition of seagrass at the base of the food web, rather than differences in turtle foraging strategy, contribute to the isotopic variation in the nesting population. This study improves understanding of the foraging ecology of a highly dispersed and migratory species.

No abstract

Management of the critically endangered hawksbill turtle in the Wider Caribbean (WC) has been hampered by knowledge gaps regarding stock structure. We carried out a comprehensive stock structure re-assessment of 11 WC hawksbill rookeries using longer mtDNA sequences, larger sample sizes (N = 647), and additional rookeries compared to previous surveys. Additional variation detected by 740 bp sequences between populations allowed us to differentiate populations such as Barbados-Windward and Guadeloupe (F (st) = 0.683, P < 0.05) that appeared genetically indistinguishable based on shorter 380 bp sequences. POWSIM analysis showed that longer sequences improved power to detect population structure and that when N < 30, increasing the variation detected was as effective in increasing power as increasing sample size. Geographic patterns of genetic variation suggest a model of periodic long-distance colonization coupled with region-wide dispersal and subsequent secondary contact within the WC. Mismatch analysis results for individual clades suggest a general population expansion in the WC following a historic bottleneck about 100 000-300 000 years ago. We estimated an effective female population size (N (ef)) of 6000-9000 for the WC, similar to the current estimated numbers of breeding females, highlighting the importance of these regional rookeries to maintaining genetic diversity in hawksbills. Our results provide a basis for standardizing future work to 740 bp sequence reads and establish a more complete baseline for determining stock boundaries in this migratory marine species. Finally, our findings illustrate the value of maintaining an archive of specimens for re-analysis as new markers become available.