The proposed objective of this study was to validate the prediction of linear regression models of the first degree, the estimated values of metabolizable energy (ME) regarding the observed ME values of maize obtained in biological assays with swine. Seventy four records of chemical and energetic composition of maize were obtained from literature and used to estimate the ME of 41 prediction models depending on the chemical composition. The significance of the regression parameters (β0and β1)was evaluated by partial t test and the prediction validation of first degree models was obtained by accepting the null hypothesis β0=0 and β1=1. The ME7= 1.099 + 0.740GE - 5.5MM - 3.7NDF; ME9= 16.13 - 9.5NDF + 16EE + 23CP*NDF - 138MM* NDF and ME13= 5.42 - 17.2 NDF - 19.4MM + 0.709GE models are the most adequate to estimate the metabolizable energy of maize and can be used as a tool to formulate diets for swine.
chemical composition; corn; linear equations; validation
