Increasing the efficiency of recurrent selection for yield in maize

Abstract EN

The objectives of the present study were to increase the efficiency of recurrent selection within the "Composite-21" variety of maize using selection index to increase the precession of the evaluation of S1-families.

Ninety S1 families were evaluated in six sets, each set included 15 S1 families replicated three times.

Data were collected for yield, yield components characters and agronomic characters to carry out six selection methods.

One method is based on ear yield alone and the other five methods were based on selection index.

Estimation of environmental (σ2 e), phenotypic (σ2 ph) and genotypic (σ2 g) variances and broad - scense heritability (H %) were calculated for each trait to predict the genetic gain from different selection methods.

Phenotypic and genetic correlations among traits were calculated to examine the possibility of indirect selection for yield using some characters which are highly correlated with yield.

Heritability estimates were intermediate (41- 55%) for ear yield, 100-kernel weight and kernel depth and more than 70% for number of rows/ ear, plant and ear heights.

The genetic correlations for yield with the other characters were in general intermediate.

Based upon the expected advance from selection, the expected gain which is coming from the indices ranged between (5.58-19.13 %) comparing with the expected gain which is coming from yield alone (16.43 %).

So, the use of indices was slightly superior to selection based on yield alone.

The actual improvement in yield is 9 % for selection based on yield alone, 8 % for selection index based on yield and all characters, 3 % for the index based on yield and ear characters and 1 % for the index based on yield and vigor characters.

The predicted advance from selection was higher than the actual advance.