Test Site Allocation within and among Stages of a Sugarcane Breeding Program

Abstract

The Louisiana Sugarcane Variety Development Program (LSVDP) currently uses up to 13 yr from crossing to develop a new sugarcane (Saccharum sp. hyb.) cultivar. The outfield trials constitute the final testing stage. Recent changes in pre‐outfield testing warrant examination of resource allocation among and within these two testing stages. Sucrose yield variance components were obtained from outfield tests harvested from 1987 to 1990 and from a published report concerning pre‐outfield tests. Sucrose yield repeatability estimates for three crops were made with varying numbers of test locations and replications for the two test stages. Three replications at each of 10 locations resulted in 95.5% of the repeatability of using three replications at 13 plant‐cane outfield tests. Plant‐cane pre‐outfield results with eight locations and three replications effected 94.7% of the repeatability of using three replications and 13 locations. Repeatability slightly increased in the pre‐outfield tests and slightly decreased in the outfield tests with older crops for a given number of locations and replications. Correlation of genotype‐location mean yields among test locations indicated no redundancy of test sites. The outfield tests are customarily analyzed as either heavy or light textured soil tests. The efficacy of this soil classification was tested with three statistical models. The results did not support partitioning of the outfield tests by soil texture and indicated approximately as much genotype × location interaction within soil groupings as among the groupings. The conclusion was that a reduction in the number of outfield test sites coupled with an increase in pre‐outfield test number should render a more efficient breeding program.