Utilização de modelos não lineares para a estimativa da curva de crescimento do acará do congo (Archocentrus nigrofasciatus)

Abstract

The objective of this work was to study the growth of white convict cichlid (Archocentrus nigrofasciatus), defining the mathematical model that best describes the growth curve of this species, using different nonlinear mathematical models. The trial was carried out from January to July 2016, with two hundred fingerlings from two different generations (G1 = generation of parents and G2 = generation of children - inbred) from the Aquarium Laboratory of the Federal University of Pampa, from 9 to 22 weeks of age. Weekly, all fish were measured to obtain body weight, standard length, total length and height. 21 non-linear models were tested for the growth curve; the best fit model was chosen based on the following criteria: residue mean square, determination coefficient, absolute mean deviation of the residues, trajectories of growth curves observed and estimated by the mathematical functions. After choosing the best fit model, the identity and parameter equality test was performed to verify if the same parameters could be used to describe the growth of the two generations. Statistical analyzes were performed using the SAS® statistical program. The best fit models for G1 were Bianchini, Inverse Polynomial and Quadratic Logarithmic, whereas the nonlinear models Brody 23, Nelder and Logarithmic Quadratic fit better the G2 growth curve. It was not possible to use the same parameters for the Logarithmic Quadratic model in order to estimate the growth curve of the two generations. Weight of fish at sexual maturity was lowerr in G2 in comparison to G1, because of consanguinity. The use of nonlinear mathematical models is a suitable tool to describe the growth curve of white convict cichlid.