Frequency changes of alleles a t eight enzyme loci were monitored in four longterm maize selection experiments. The results indicate that changes in frequencies of the alleles at these loci are associated with changes due to selection for improved grain yield. The frequencies changed more than is consistent with the hyFothesis of selective neutrality. In addition, significant deviations from a random-drift model were nearly always accompanied by significant linear tiends as would result ii allozyme frequencies respond to directional selection. Evaluations of linkages aiid linkage disequilibria in the selected populations indicate that the eight enzyme loci responded independently as selection progressed. HE effects of long-term selection for grain yield or other agronomic traits T o n frequencies of alleles at enzyme loci in experimental populations of maize (Zea n a y s L.) have not been well documented. Only a few experimental programs have been conducted in which the selection procedure has been consistent over an extended period of time. Therefore, few available experimental populations of maize have been suitable for studying the effects of long-term selection on enzyme loci. In a full-sib family selection experiment that originated from the F, between two maize lines (NC7 and CI21), STUBER and MOLL (1972) reported that the frequency of an acid phosphatase allele increased from 0.50 to 0.76 in the first six cycles of selection for increased grain yield. However, after nine cycles of selection, the population appeared to approach a stable intermediate frequency of 0.60 for this allele. BROWN (1971) examined the changes in alleles at six Paper No 532G of the Journal Seiier of the North Carolina \gi icul turd Experiment Station, Raleigh, North Calohna 2 iS50 This Iniestigation was partially supported by iesearch grant GM-11545 from the Public Health Service and a grant from DeKalh kgResearch, Inc , DeKalb, Illmois Genetics 95 : %5-’23S May, 1980 226 C. W. STUBEll et d. enzyme loci (four esterase loci, one alcohol dehydrogenase locus and one peroxidase locus) in the long-term Illinois selection programs for high and low oil and protein in maize. He reported that the changes detected were consistent with the hypothesis of purely neutral drift, but that deviations suggested that isozyme variants were not entirely neutral to selection. BROWN and ALLARD (1971) monitored allele frequencies at nine enzyme loci for two cycles of reciprocal recurrent selection in two populations of maize grown at Davis, California. They reported that the few changes noted could be ascribed to random genetic drift. Although the primary trait under selection was grain yield, their description indicates that selection was based on a number of agronomic and disease-resistance traits in addition to grain yield. Therefore, the selection intensity for any single trait would be low, and two cycles of selection would not be expected to have a significant effect on allozyme hequencies. Long-term selection studies for a single trait, increased grain yield, have been conducted in the maize quantitative genetics program at North Carolina State University. Several selection schemes, both intraand interpopulation, have been used. Remnant seeds representing each of the nine or more selection cycles completed have been maintained in cold storage and were available for growing plants for laboratory analyses of isozymes. Therefore, changes in frequencies of alleles at enzyme loci as selection progressed could be assessed in these selection studies. The questions considered by the investigations reported here are: ( 1 ) Have frequencies of alleles at 20 enzyme loci changed in four long-term selection populations of maize? (2) Are the allozyme-frequency data consistent with the hypothesis of random drift? ( 3 ) Are the allozyme-frequency data significantly correlated with selection gain for grain yield? (4) Can types of gene action be inferred from comparisons of allozyme-frequency data in populations involved in both intraand interpopulation selection schemes? And ( 5 ) is there any evidence that the selection processes have contributed to linkage disequilibria or other correlations among enzyme loci? MATERIALS A N D RIETHODS