To assess the possibility of discarding production variables in 942 laying hens by principle component analysis to eliminate unnecessary and difficult to measure characteristics, data were obtained from laying hen lines of the genetic breeding program of the Universidade Federal de Viçosa. The traits analyzed were egg production rate (TP) from the 26th to the 58th week, individual mean weight at the 34th (PMI1), 42th (PMI2), 50th (PMI3), 58th (PMI4) and 66th week (PMI5); egg mean weight at the the 34th (PMO1), 42nd (PMO2), 50th (PMO3), 58th (PMO4) and 66th weeks of age (PMO5). Eight of the 11 principal components showed variance lower than 0.7 (eigenvalue lower than 0.7), suggesting 8 variables to discard. The discarded variables were those that showed the highest coefficients, in absolute value, from the last principal component because variables highly correlated with the principal components of smaller variance represent practically insignificant variation. The discarded variables presented significant simple linear correlation with the others, therefore, they were redundant. Based on these results, the following variables are recommended for use in future experiments: TP, PMI1, PMO4.
correlations; discarding of variables; Leghorn; multivariate analysis
