Genotype x Environment Interaction in Pinus patula and its implications in South Africa

Abstract

The Genotype x environment interaction was examined using 81 provenance and / or progeny tests of Pinus patula over 54 sites in Southern Africa (South Africa and Swaziland). Type B genetic correlation estimates were calculated for all possible pairs of tests amongst these trials. The mean Type B genetic correlation for all tests across all sites was 0.69. The difference between the imported Mexican provenance material and a range of South African genetically improved populations was minimal with estimates of 0.68 and 0.70, respectively. Standardization of the data had a small but significant effect on the estimates of genotype x environment interaction with increasing significance as growth differences amongst tests increased. The proportion of interactive genotypes was examined using the joint regression method, and varied markedly between populations. Amongst the imported Mexican material, the most interactive populations were found to be those originating from the northern part of the distribution of P. patula in Mexico. The proportion of interactive genotypes varied from 9 – 40% amongst the Mexican material and 10 – 16% for the South African families. Interacting environments were not restricted to a small number of sites and moderate levels of genotype x environment interaction were found across most sites. A range of climatic factors was examined in an attempt to define regions of predictable, minimal genotype x environment interaction. Some variables related to the spring rainfall and winter potential evaporation were identified and utilized in provisional models that were able to differentiate a difference in Type B genetic correlation estimates of 0.11. Small predicted genetic gains of between 0.5 – 1.5 % were estimated when utilizing the benefits of regionalization as proposed in these models. The breeding strategy and testing procedures are discussed in the context of P. patula within the region.