02054nam a2200229 a 450000100080000000500110000800800410001910000190006024501240007926001380020352012700034165300280161165300260163965300230166565300140168870000180170270000210172070000220174170000220176370000170178570000220180210461572026-03-18       2002    bl uuuu     u00u1 u    #d1 aMARTINS, E. N.  aBayesian analyses of heterogeneous variances among generations of brazilian Canchim beef calves.h[electronic resource]  aIn: WORLD CONGRESS APPLIED TO LIVESCTOCK PRODUCTION, 7., 2002, Montpellier, France. Abstracs... Montpellier: INRA, 2002. p.197.c2002  aIn the genetic evaluations, the models usually assume homogeneous variance through subclasses of identified factors. However, several studies have reported heterogeneous genetic and residual variances in analyses of many animal production traits (Weigel and Gianola, 1993). Different regions, birth years, feeding systems, genetic variability within herds and production levels can have heterogeneous variance among their subclasses (Ibáñez et al., 1996; De Veer and Van Vleck, 1987). Synthetic breeds are usually formed from selection of crossbred animals. Crossbred populations in general have relatively high genetic variability. However, selection changes gene frequencies and consequently it tends to reduce the genetic variability. Furthermore, the herds have had their environment modified over time by improvement in feeding and management. Therefore, data from different generations of a single population can have heterogeneous variances. Ignoring heterogeneity of variance can reduce the efficiency of genetic evaluation procedures and consequently response to selection (Hill, 1984). Thus, the aim of this study was to evaluate the occurrence of heterogeneous variance of yearling live weight among the first six generations of Brazilian Canchim breed  aBrazilian Canchim breed  aHeterogeneous genetic  aResidual variances  aVariances1 aENGLER, E. O.1 aSAKAGUATI, E. S.1 aFREITAS, A. R. de1 aALENCAR, M. M. de1 aNIETO, L. M.1 aFALCÃO, A. J. S.