GENETIC IMPROVEMENT VIA SELECTION IN MAIZE

Abstract

A total of 144 S1 maize progenies (derived from the CIMMYT's DTP-1 population) were made and evaluated (in 2001 and 2002 seasons) under water-stress and non-stress conditions. The best 10 % of them were selected under each environment and separately intercrossed (in early 2003 season). The two developed (Pop-1 and Pop-2) along with the original (Pop-0) populations were evaluated under both environments (stressed and unstressed) in late 2003 season. The objectives were to develop new maize populations with increased tolerance to drought, test which selection strategy is the best for achieving maximum progress and identify the secondary traits strongly associated with grain yield. Results

,...... ( 0"' 0 0on ""'" ".....'. N""'" o...n... .,; J":' - :::> 2 •;; u '-""

indicate the existence of wide genetic diversity among the 144 S1 's. Broad sense heritability {11,) for grain yield showed a general tendency to decrease with imposing drought stress. Predicted selection gain in grain yield/fad (GYPF) supported the 1" selection strategy (genotypes should be selected under water-stressed environment for use under the same environment). Results indicated that kemels!rrl, ears/plant, barrenness, AS! and stay green traits are valuable adjunct in increasing the efficiency of selection for GYPF under water-stress. One cycle of S 1 recurrent selection using the well-watered as a selection environment caused a significant improvement in GYPF of Pop-1 over Pop-0 under water deficits by 30.8 % (1.34 ard/fad) and under well-watered conditions by 16.0% (2.57 ardlfad). This progress assures the efficiency of the selection procedure used in this study. The actual progress confirms the 2"d • strategy of selection (genotypes should be selected under well-watered for use under water-stressed environment). Moreover, using a combination of stressed and unstressed as a selection environment in this study also gave the best actual progress (30.8 %) for use under water-stress as a target environment, confirming the 3'd strategy (genotypes should perform well under both stressed and unstressed selection environments).