The use of mathematical models to describe animal growth is not recent. They are able to summarize information on strategic dots of animal growth development and to describe the evolution of weight according to the animal age. It is also possible to compare different individuals in similar physiologic stages. The growth models most commonly used in poultry breeding are derived from Richards function, and they present parameters that provide biological interpretation and knowledge to select a specific shape of growth curve in poultry. However, it is also possible to use segmented polynomials to describe trend changes during the animal growth. One needs to consider important variables affecting the growth curve parameters estimates, such as, production system, specie, sex and their interactions. Model Goodness-of-fit can be based on many criteria such as coefficient of determination (R2), residual mean squared error, (LSe), estimated predicted mean error (PME), the easiness the analysis to reach convergence and the possibility of biological interpretation of parameters. Studies involving modeling and description of growth curve and their components are described in literature, but, there is no selection programs applied to the growth curve shape. The importance of determinating the parameters of growth curve models is more relevant when considering that most of the genetic gains for growth traits are related to selection, on weights near to the inflexion point. Often, selection to fast growth is important in all breeding programs, and could be based on genetic parameters of the growth curve parameters. These parameters are related to important productive and reproductive traits, and present different values, according to specie, sex and models used in evaluation. Alternatively, other methodology used is random regression models, allowing graduation changes in (co) variances between ages during the time and predicting (co)variances during the studied trajectory. The use of random regression models has the advantage to allow the partition of phenotypic growth curve (co)variance in its different genetic additive and the permanent environment effects, using random regression coefficients for each different effect. This review aimed at summarizing the main frequentists mathematical models used in the studies of growth curves in birds, emphasizing those applied to estimate genetic and phenotypic parameters.
growth parameters; growth models; growth rate; poultry
