| Summary: | Growth of morphological traits in Atlantic puffin (Fratercula arctica) offspring has typically been characterized by linear models, despite clearly displaying nonlinear patterns. We assessed the fit of six typical avian growth models to measurements of puffling mass, wing length, and tenth primary feather length. Across all three biometrics, the first- and second-best performing models were nonlinear, and the worst performing model was the linear model. Specifically, the preferred model for mass, wing, and tenth primary growth was the quadratic model, logistic model, and extreme value function model, respectively. The preferred models were used to generate separate growth curves for individual chicks, from which parameter estimates for growth rate, normalized growth rate, inflection value, and asymptotic value can be extracted. These parameter estimates are easily interpretable and comparable between several of the nonlinear models. We recommend the reported models in future studies incorporating metrics of puffling growth, and encourage the use of this methodology to explore nonlinear growth patterns across avian species.
|
|---|