Competition effects in selection trials

Unselected sugarcane clones (Saccharum spp. hybrids) were evaluated in a replicated experiment in the Burdekin to quantify bias due to interplot competition among neighbouring clones. In plant and ratoon crops, cane yield, sugar content and sugar yield were measured in small plots (single and dual rows) influenced by competition and in pure stands (larger guarded multi-row plots) free of competition. It was found that estimation of cane yield in small plots was biased by competition among elones in neighbouring plots. Genetic and error variance estimates for true yielding ability in small unguarded plots were grossly inflated by competition. The bias due to competition was more serious for yield in the ratoon crop. In contrast to cane yield, sugar content was little affected by competition among neighbouring plots. Greater emphasis should be placed on sugar content as a selection criterion in early stage trials. Such a change may facilitate the use of alternative plot shapes not currently considered. Estimates for sugar yield are biased, largely due to competition affecting cane yield. Simulation studies using these data will investigate alternative selection systems for sugarcane.