Premium
Genotype ✕ Environment Interactions for Cane and Sugar Yield and Their Implications in Sugarcane Breeding 1
Author(s) -
Kang M. S.,
Miller J. D.
Publication year - 1984
Publication title -
crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.76
H-Index - 147
eISSN - 1435-0653
pISSN - 0011-183X
DOI - 10.2135/cropsci1984.0011183x002400030002x
Subject(s) - cultivar , saccharum , cane , biology , brix , mathematics , stability (learning theory) , hectare , gene–environment interaction , selection (genetic algorithm) , crop , agronomy , yield (engineering) , sugar , horticulture , statistics , genotype , ecology , computer science , food science , materials science , metallurgy , biochemistry , machine learning , artificial intelligence , gene , agriculture
Genotype ✕ environment (GE) interactions limit the effectiveness of selection when selection is based only on mean yields. This study evaluated three methods of partitioning GE interaction into stability‐variance components assignable to each cultivar in sugarcane ( Saccharum spp.) cultivar evaluation tests. An unbiased stability‐variance parameter (σ̂ 2 1 ) developed by Shukla and an ecovalence stability index (w) developed by Wricke were calculated separately for plant‐cane and ratoon crops for Brix (%), grams of sugar per kilogram of cane, and tons per hectare of cane and of sugar for 11 cultivars. A cultivar ✕ location component of variance (σ̂ 2 c1 ) proposed by Plaisted and Peterson was also calculated. Within each crop, σ̂ 2 1 and w had identical cultivar rankings (r s = 1.00**) (**, Significant at the 0.01 level.) for each of the four traits studied. We propose that w sum of squares for each cultivar be expressed as w mean square and the latter be tested for significance in the same manner as σ̂ 2 1 . Those cultivars with a significant mean square were judged to be unstable. Shukla's method allowed use of a covariate of fertility and cultural practices at different locations to remove heterogeneity variance (nonadditivity) from the GE interaction, and partitioning of the remainder of variance assignable to each cultivar (ŝ 2 1 ). Certain cultivars were judged stable after the covariate adjustment, indicating that the instability Was introduced by the linear effect of the covariate. Methodology for computing the σ̂ 2 CL component is cumbersome and would have little application when a large number of genotypes are evaluated. Repeatability of stability‐variance parameters between crops (individual repeatability) was relatively low for the four traits studied.