RESPONSE TO S(1) PER SE AND RECIPROCAL FULL-SIB SELECTION WHEN TESTED IN A WIDE ARRAY OF ENVIRONMENTS

Two contrasting recurrent selection methods were compared for improvement of index and its components using Nebraska Krug (K), Nebraska Stiff Stalk Synthetic (S), and Nebraska B Synthetic (B) maize populations. Index, the selection criterion, is composed of yield (adjusted to 15.5% moisture) x proportion of upright plants x proportion of retained ears. Progress through three cycles of replicated S(,1) per se and reciprocal full-sib (RFS) selection was evaluated by estimating stability parameters using the Eberhart and Russell (1966) model. Evaluations were made using random-mated populations from the original and base populations together with K, S, and B replicate subpopulations (entries) selected in the third cycle by both methods. Gains in index and its components from S(,1) per se selection were observed in the K and S entries but not in the B selected material. The index increase was due to an increase in proportion of retained ears in the K entries and increase in corrected weight in the S entries. Similar response from RFS selection was observed in the K and S improved entries. Linear regression coefficients increased with both selection methods when calculated across 15 environments. K and S entries had higher index values in both favorable and unfavorable environments, whereas B entries had yield advantage only under favorable conditions. Selection replicates, though selected in different years, showed no discernable pattern in response to environments