Loss of Superior Genotypes in Early Stages of Sugarcane Varieties Selection in Kenya

Abstract

Sugarcane selection is used for the identification and release of elite varieties from the original seedling population. In Kenya, only 9 varieties that out-perform the checks have been released since 1980. The varieties are tested in five stages with stage 1 being the original seedlings population consisting of about 35,000 individual seedlings. Moderate selection intensity is performed among families and individual seedlings. This results to about 3,500 (10%) seedlings being selected to stage 2 where the test varieties are tested on a single site single row trial. We hypothesized that the large number of varieties tested in stage 2 results to loss of superior genotypes. To determine the extent of this loss we selected 30 elite varieties from MS 2008 seedling stages for testing in 3 sites on augmented row-column design with single row plot. The 30 elite varieties were also included in stage 2 of the current selection program. Data was analysed and the elites varieties that significantly out-performed the checks were compared with those advanced to stage 3 in the current selection program. The results shows that the current selection method results to loss of superior genotypes since only 4 varieties were advanced to stage 3 out of 12 elite varieties that out-performed the checks. We evaluated analysis of stage 2 data by including data obtained in stage 1 which resulted to improved efficiency with 9 out of 12 elite varieties being selected. We propose the use of Bayesian approach in analysis of sugarcane varieties selection data