Sampling methods for assessing the diets of Atlantic puffin chicks

Three methods for sampling the diets of Atlantic puffin Fratercula arctica chicks were compared and evaluated. Prey-carrying puffins were captured in mist-nets, or observed with a telescope or with unaided eyes. Prey samples collected in mist-nets yielded higher proportions of gravid capelin Mallotus villosus than did samples obtained visually. Telescopic samples produced higher estimates of the numbers of larval fish landed by puffins (larval fishes are often lost when food-laden birds are caught in mist-nets), and yielded underestimates of the lengths of male capelin. Slight correction factors made the 3 prey sampling methods comparable. Visual sampling procedures had the advantages of creating less disturbance and allowing puffins to deliver food to chicks. Use of telescopic sampling is, however, dependent on colony topography. Forage fishes, heavily dominated by 2-yr-old capelin, made up 88% and 98% of the frequency and mass, respectively, of the prey delivered to puffin chicks. Small prey (larval capelin and sandlance Ammodytes spp., hake Urophycis tenuis, crustaceans) were most commonly fed to young chicks. Gravid and male capelin tended to be more commonly fed to chicks midway through the nestling period, a time of rapid growth. Spent female and immature capelin were fed more commonly to older chicks. The lengths of gravid capelin tended to increase through the nestling period. Gravid and spent female and male capelin and sandlance fed to puffin chicks were shorter in 1992 than in 1993 or 1994. Successful puffin chicks on Great Island, Newfoundland, Canada, were fed an average of 1105 (SD = 435) prey items weighing 2812 g (SD = 579) equivalent to 12506 kJ (SD = 2391). Extrapolation of these findings suggests that the population of Atlantic puffins in Newfoundland consumes about 8000+ t of forage fishes, mostly 2-yr-old female capelin during a breeding season. Data on age-class and gender status of prey harvests by marine birds will provide needed input in energetics and trophic models.