Graduation date: 2007This research was designed to evaluate the Fatty Acid Signature (FAS) technique as a non-lethal alternative to more traditional, and sometimes destructive, methods of studying the diet composition of piscivorous birds. Specifically we tested the technique with Caspian terns (Hydroprogne caspia) which currently nest in large numbers in the Columbia River estuary and are known to consume juvenile salmonids (Oncorhynchus spp.) listed under the U.S. Endangered Species Act. From captive feeding trials conducted with Caspian tern chicks, we determined that FASs of the birds reflected differences in their diets. After 20 days of being fed consistently mixed or monotypic diets of two fish types, chicks displayed different adipose tissue FASs between all 4 diet treatments. When diets were changed, adipose tissue FASs reflected the shift in diet treatments within two weeks. Fatty acid (FA)- specific calibration coefficients (FA level in the consumer divided by FA level in the food) were calculated for Caspian terns fed monotypic diets for 34 days; some calibration coefficients varied in association with diet and age of the terns, and also differed between terns and common murres (Uria aalge), whose calibration coefficients were measured in a separate study. Variation in FA-specific calibration coefficients may be problematic for obtaining accurate estimates of diet composition in piscivorous birds using the Quantitative Fatty Acid Signature (QFASA) technique. We advocate sensitivity analysis to test whether the QFASA models are robust to the magnitude of variation in calibration coefficients detected in this study. FASs differed among the 3 major fish prey types observed in diets of Caspian terns nesting in the Columbia River estuary during the 2003 breeding season: juvenile salmonids, surf smelt (Hypomesus pretiosus), and northern anchovy (Engraulis mordax). We detected differences in FASs of nesting Caspian terns between early and late in the nesting season of 2003; these differences were associated with a shift in diet compostion from a diet dominated by juvenile salmonids to a diet dominated by northern anchovy. The FASs of several species of juvenile salmonids, however, exhibited little inter-specific variation, especially between species raised in hatcheries, which comprise the majority of smolts consumed by Caspian terns in the estuary. We found levels of highly-unsaturated FAs (HUFAs) to be higher in wild steelhead smolts than in hatchery-reared steelhead smolts, but HUFA levels in terns did not reflect the changing prevalence of wild steelhead in their diets. This is likely due to contribution of HUFAs to the diets of terns from marine forage fishes. Thus, HUFAs do not appear to be useful indicators of wild steelhead in the diets of these birds. If the QFASA technique can be validated, it has the potential to provide general information on diet composition for piscivorous birds foraging on broadly different prey types over extended periods. However, due to similarities in FASs of key prey types consumed by Caspian terns in the Columbia River estuary, namely salmonids, obtaining the precise estimates of diet composition and consumption of different species of salmonids that are requested by resource managers does not seem feasible using the QFASA technique alone. Consequently, more traditional methods of diet composition analysis for Caspian terns (bill load identification, stomach contents analysis) can not be replaced by the FAS analysis technique.