Optimising the number of test sites, crop-years and replications to maximize post-release testing efficiencies

Abstract

The South African sugar industry develops commercial sugarcane cultivars across multiple testing sites, crop-years and replications. The extent of such parameters has not been well quantified for optimum breeding and resource allocation. The objective of this study was to identify the optimum number of testing sites, crop-years and replications needed to maximize efficiency of post-release cultivar evaluations in South Africa. Ten commercial sugarcane varieties were evaluated in randomized block designs across five regions tested over five crop cycles. The study was conducted from the main plant to third ratoon crops across the locations. The general linear model was used to estimate variance components. Genotype × location (G×L) interactions were significant (p ≤ 0.01) for cane yield (TCANE), estimated recoverable crystal content (ERC), sugar yield (TERC) and fibre (F). Tonnes of estimated recoverable crystal were highly influenced by G×L interaction, suggesting that evaluating cultivars across locations is essential. High broad sense heritability estimates were observed for TCANE (71.17%), ERC (97.11%) and F % (93.11%), implying that these traits were genetically controlled. The G×L and G×L×C variances were the largest source of variation, suggesting that the evaluation of sugarcane cultivars across locations as opposed to crop cycles would be essential.