ASSESSING BIAS IN STUDIES OF BALD EAGLE FOOD-HABITS

Although studies of bald eagle (Haliaeetus leucocephalus) food habits are numerous, few authors have quantified biases inherent in the techniques used. In our study of food habits of nonbreeding bald eagles on the northern Chesapeake Bay, we examined biases associated with pellet analysis, food remains analysis, and direct observation. We assessed these biases through controlled feedings of 2 captive bald eagles and through observations of free-ranging eagles. Fish fed to 2 captive bald eagles were underrepresented (P < 0.001) in egested pellets. Most bird and mammal food items were detected in greater-than-or-equal-to 1 pellet; however, species and carcass condition affected frequency of occurrence. Analysis of captive eagles' food remains overrepresented birds, medium-sized mammals, and large bony fish (P < 0.001); small mammals and small fish were underrepresented (P < 0.001). This bias increased over time due to greater persistence of some remains in shoreline plots. Direct observations of free-ranging eagles resulted in biases toward easily identified species such as eels and catfish, but also documented the use of small, soft-bodied fish, which were not well documented by the other techniques. Because of the variety of biases present, accurate assessment of bald eagle foods requires use of multiple techniques.