Methods for Calculating and Evaluating a General Selection Index Obtained by Pooling Information from Two or More Experiments.

ELECTION indexes have been proposed in animal and plant breeding programs S as a criterion for selection when certain quantitative characters are considered. Such characters present particular difficulty in selection programs since heritable variations are often masked by nonheritable variations. Investigations have demonstrated that a selection index based on a number of measurements could be used effectively to increase the probability of selecting desirable genotypes as compared to that if selection were practiced with an independent culling level for each measurement. The pertinent question is whether a selection index determined from a source of data can be used as a general index for a breeding program. The success of such a general index depends upon the sampling errors expected for the estimated weights of the index and the errors involved in the estimation of the genotype by environment interaction. The effects of genetic source sampled upon the expected genetic correlations between the factors involved in the index are of critical interest. Experimental data are necessary to adequately evaluate this problem; therefore, the approach considered in this paper is to present arguments for the combining of information from a series of experiments, to obtain a general index and to present basic relationships involved in the application of a selected set of phenotypic weights to specific sources of data.