Genetic homeostasis was studied in different maize (Zea mays L.) genotypes and populations to determine whether selection has affected homeostasis and to determine whether homeostasis is heritable. The maize materials included one group of six inbred lines released before the 1960s (older lines), and another group of six inbred lines released after the 1970s (newer lines). Within each group, single cross, F2, and backcross generations were developed. The experiments were conducted at two Iowa locations in 1985 and 1986. Individual plant measurements were taken for eight maize traits: yield, ear length, ear diameter, kernel-row number, kernel depth, plant and ear height, and number of tassel branches. The objectives of this study were to compare genetic homeostasis of the older versus the newer maize lines to determine whether selection of newer lines has changed the levels homeostasis and to study whether the homeostatic effects are transmitted from the inbred lines to their single cross, FI, and backcross generations;Estimates of genetic homeostasis were obtained using the proportion of the environmental variance contained in the within-plot variance. The significance and magnitude of the genotype x environment interaction mean squares and the coefficient of variation were also used as preliminary estimates of homeostasis. The coefficient of variation, however, was not a good estimator because it had a negative and highly significant correlation with the mean. Genetic homeostasis depended on the level of heterogeneity and heterozygosity of the maize genotypes and populations. The homogeneous and homozygous inbred lines were the least homeostatic. The homogeneous and heterozygous single crosses exhibited good homeostasis. The F2 and backcross generations (heterogeneous and 50% heterozygous) expressed similar homeostasis among them, and better homeostasis than the single crosses;The newer lines were more homeostatic than the older lines. The inbred lines were consistent in transmitting in a heritable manner the homeostatic effects, either good or bad, to their single cross, F2, and backcross generations. In most instances, the eight traits measured followed the same tendency in all comparisons among generations within both groups. Kernel depth, however, did not have a regular pattern of transmission of the homeostatic effects across generations.