Plasma metabolite levels predict bird growth rates: A field test of model predictive ability

Abstract

Bird growth rates are usually derived from nonlinear relationships between age and some morphological structure, but this procedure may be limited by several factors. To date, nothing is known about the capacity of plasma metabolite profiling to predict chick growth rates. Based on laboratory-trials, we here develop predictive logistic models of body mass, and tarsus and wing length growth rates in Gull-billed Tern Gelochelidon nilotica chicks from measurements of plasma metabolite levels at different developmental stages. The predictive model obtained during the fastest growth period (at the age of 12 days) explained 65-68% of the chicks' growth rates, with fasting triglyceride level explaining most of the variation in growth rate. At the end of pre-fledging period, β-hydroxybutyrate level was also a good predictor of growth rates. Finally, we carried out a field test to check the predictive capacity of the models in two colonies of wild Gull-billed Tern, comparing field-measured and model-predicted growth rates between groups. Both, measured and predicted growth rates, matched statistically. Plasma metabolite levels can thus be applied in comparative studies of chick growth rates when semi-precocial birds can be captured only once.