DETERMINING TEST AND BLOCK NUMBERS IN CONSIDERATION OF PROGENY TESTING QUALITY AND COST: AN EXAMPLE OF HALF SIB FAMILIES WITH RCB DESIGN AND SINGLE-TREE PLOTS

摘要

Cost-effective test and block numbers for tree progeny testing using randomized complete block designs (RCB) with single-tree plots were investigated through a generalization of the relationships between progeny testing quality criteria and test design parameters. Optimal test and block numbers were defined as those that achieve an adequate level of testing quality using the minimum resources. Results indicated that the desirable test and block numbers were functions of the target quality criteria, genetic architecture of a character and costs associated with each additional test site. A trait with low heritability and strong G x E interaction generally required a larger number of test sites than one with high heritability and weak G x E interactionto reach the same level of testing quality. Higher cost for each additional site favored using fewer test sites and more blocks per site. A testing quality standard higher than 90 % in the reliability of breeding value prediction consumed substantiallymore resources than one of 75-85 %. Desirable test and block numbers were provided to satisfy different progeny testing quality requirements under various scenarios of genetic architecture of a trait and the cost associated with each additional site.