An evaluation of regression methods to estimate nutritional condition of canvasbacks and other water birds

摘要: Abstract : Regression equations that use mensural data to estimate body condition (i.e., the general health of a bird based on reserves fat or protein) have been developed for several water birds. These often represent different sexes, age classes, seasons, without being adequately tested intergroup differences. When compared methods using total moisture content, frequently provide poor fits measures condition, measurements do not appreciably increase model's precision. We used proximate carcass analysis 538 adult and juvenile canvasbacks (Aythya valisineria) collected during fall migration, winter, spring migrations in 1975-76 1982-85 test regression estimating condition. weighed, measured, analyzed each canvasback fat, protein, ash. Analyses covariance provided estimates extractable fat(R2 = 0.71), index (total divided by fat-free dry mass; R2 0.64), protein (R2 0.74) could be accounted explanatory variables. adjusted equation significant effects age-sex classes seasons. Body mass, season, class were useful explaining but characters added only 1-3% accountable variance.