Fitting Sigmoidal Equations to Mammalian Growth Curves

Abstract

Three sigmoidal growth equations were tested for their usefulness in fitting mammalian growth curves. The von Bertalanffy, Gompertz, and logistic equations were fitted to growth data by nonlinear regression techniques. The residual sum of squares and deviations of predicted neonate, weaning, and adult masses from observed values were used as criteria to choose among the models. The von Bertalanffy equation provided the smallest residual sum of squares, while the Gompertz equation fitted equally well by this criterion. The logistic equation overestimated neonate mass and underestimated adult mass, the Gompertz overestimated neonate mass, and the von Bertalanffy overestimated weaning mass. The Gompertz model was chosen to fit a sample of 331 species; growth rate constants, K, for these species are presented. The relationship of K to adult mass was calculated and was found to have a slope similar to that of altricial birds. K-values were found to be comparable to those reported by Case (1978) and consistently higher than those of Millar (1977). Species having differing rates from those reported by Case or Millar were also identified; ground squirrels had faster growth rates and seals had slower growth rates when rate of growth was estimated by the Gompertz equation.