COMPARISON OF PREDICTED AND OBSERVED PERFORMANCE FOR 3-WAY AND 4-WAY CROSSES IN EGG-TYPE POULTRY

Nine pure lines and their crossbreds were tested during three years (117 combinations, 46,000 hens). Years, seasons, halls, clusters of cages and numbers of hens in cage were respected in the analysis. The analysed characters were: egg number per test, average egg weight, egg mass per test and body weight. In the diallel design (9 x 9) these crossbreeding parameters were estimated: direct, maternal, residual reciprocal effects; average, line and specific heterosis. The performance of three and four line crossbreds was predicted by means of purebred and single crossbred information (generation means and crossbreeding parameters). The observed means of three and four line crossbreds were compared with the expected means of the same type of crossbreds. The coincidence of observed and expected means was tested by the goodness of fit (chi(2)), correlation and rank correlation. A relatively good coincidence between the observed and expected means was found.