Genotype-environment interaction in wood basic density of Eucalyptus clones

Wood basic density was determined in samples taken from eight-year-old trees of 26 Eucalyptus clones growing in tests at four sites in Brazil. Within-clone variation in basic density was low. The broad sense heritability of basic density was 0.96 when estimated from five clones growing on two sites and 0.64 when estimated from 26 clones on four sites. Analysis of variance showed that there were significant differences in basic density between clones and sites, and that the clone x site interaction was also significant. Regression analysis was used to examine the stability of clones over environments. There were marked differences in stability between clones (linear regression coefficients 0.03 to 1.93). Interactions were variable (coefficients of determination zero to 93%), and in most cases it was not possible to predict the basic density of wood produced at one site from its value at another site. However, interactions accounted for less than 4% of the total variance in basic density; rank correlations between sites were all significant, and it was possible to identify clones which produced wood of consistently high or low basic density on the four test sites.