Genetic gains expected from phenotypic and combined index selection in Pinus elliottii

In this study, genetic gains expected per gener~lfil)n and per year, and accur~,cies of selection, were used as criteria to compare the effectiveness of phenotypic and combined index selection for the single breeding objective of maximization of saw-timber production at 15 years; selection was based on stem diameter at breast height at five years. The data used were generated in a progeny test of P;nus eU;ott;; var. eU;ott;; Engelm. which contained 157 out of a possible 216 controlled crosses in a factorial mating design between 24 female and 9 male parent~. The 33 parents were selected as superior phenotypes from unimproved plantations in both Zimbabwe and South Africa. A combined selection index that included all available genetic information resulted in higher genetic gain (52.5% ) and accuracy of selection (0.83) than vas achieved with phenotypic selection (46.3% and 0.72 respectively), providing an economic justification for the breeding programme. However, the combined selection index resulted in a greater reduction in the additive genetic variance (down to 50% compared with 83% for phenotypic selection, after one generation of selection), the lowest number of families from which individuals were selected (15 as opposed to 32), and the highest number of individuals selected per family (a maximum of 12 as opposed to 6). This could have serious implications for advanced generations through increased risk of inbreeding depression and a reduction in the genetic variation required to make further gains in response to unpredictable future needs. With most forestry breeding programmes nmv entering their second cycle, it will be crucial to continue returning near maximum gains in the short-term. nut, at the same time, selection within the parental population for advanced generation breeding will have to be performed with care to preserve genes that may be needed in the future. In the P. elliott;; programme in Zimbabwe, a variant of the multiple- population breeding strategy is seen as providing the most appropriate solution to the problem. The best compromise may be to use combined index selection based on precise testing and accurate measurement in an intensively bred nucleus, and phenotypic selection of individuals for infusion into it from a much larger, less costly main population. This could either be maintained locally or consist of direct imports from other breeding programmes.