ESTIMATION OF LIPIDS AND LEAN MASS OF MIGRATING SANDPIPERS

Estimation of lean mass and lipids levels in birds involves the deriva tion of predictive equations that relate morphological measurements an d, more recently, total body electrical conductivity (TOBEC) indices t o known lean and lipid masses. Using cross-validation techniques, we e valuated the ability of several published and new predictive equations to estimate lean and lipid mass of Semipalmated Sandpipers (Calidris pusilla) and White-rumped Sandpipers (C. fuscicollis). We also tested ideas of Morton et al. (1991), who stated that current statistical app roaches to TOBEC methodology misrepresent precision in estimating body fat. Three published interspecific equations using TOBEC indices pred icted lean and lipid masses of our sample of birds with average errors of 8-28% and 53-155%, respectively. A new two-species equation relati ng lean mass and TOBEC indices revealed average errors of 4.6% and 23. 2% in predicting lean and lipid mass, respectively. New intraspecific equations that estimate lipid mass directly from body mass, morphologi cal measurements, and TOBEC indices yielded about a 13% error in lipid estimates. Body mass and morphological measurements explained a subst antial portion of the variance (about 90%) in fat mass of both species . Addition of TOBEC indices improved the predictive model more for the smaller than for the larger sandpiper. TOBEC indices explained an add itional 7.8% and 2.6% of the variance in fat mass and reduced the mini mum breadth of prediction intervals by 0.95 g (32%) and 0.39 g (13%) f or Semipalmated and White-rumped Sandpipers, respectively. The breadth of prediction intervals for models used to predict fat levels of indi vidual birds must be considered when interpreting the resultant lipid estimates.