Abstract:

Objective: The aim of this study was to screen out superior clones of Eucalyptus from five locations in Guangxi, Guangdong and Hainan by analyzing the genetic diversity of growth traits and evaluating the interaction between genotype and environment. Method: The data of tree height, DBH and volume, preserving rate and forest stock of 3.5-year-old Eucalyptus clones was collected, The single factor analysis of variance and multiple comparison were conducted for each trait with the collected data from the five locations. Then, a linear mixed model was established for single-location and multi-locations analysis of variance. The variance components of each trait and the BLUP values of each clone were obtained by taking clones and clones×repeats (locations) as random effects, so as to calculate the genetic parameters of each trait. Finally, AMMI model and BLUP-GGE model were integrated to select elite clones and suitable environment. Result: 1) This study showed that there were significant differences in each growth trait among different parents, clones and locations (P < 0.01). Among the different parents, the traits of E. pellita×E. wetarensis were relatively better, and its volume and annual average forest stock were 0.069 0 m³ per plant and 26.6 m³·hm^-2a^-1, respectively. Among the different clones, EC245 was the best, and the volume and annual average forest stock were 0.090 9 m³ per plant and 118.46 m³·hm^-2a^-1, respectively, but the preserving rate (77.8%) was relatively lower. Among the different locations, the clones in Leizhou area had the best growth performance with the average volume and preserving rate of 0.083 6 m³ per plant and 85.5%, respectively, while those in Haikou area had the worst growth performance with the average volume and preserving rate of 0.023 8 m³ per plant and 55.5%, respectively. 2) The analysis of variance with data from multiple locations and genetic parameters showed that the interaction between each trait and the environment reached a significant level (P < 0.01), and the repeatability of each trait ranged from 0.717 4 to 0.869 5. Among the effect sizes of different traits, the environmental effect size (43.2%) and interaction effect size (17.0%) of tree height were much greater than those of other traits, and the environmental effect size (26.5%) and interaction effect size (8.8%) of DBH were the lowest, indicating that tree height was susceptible to environmental effects, while DBH was relatively stable. In different locations, tree height repeatability ranged from 0.653 9 to 0.901 8, DBH repeatability ranged from 0.679 1 to 0.930 1, volume repeatability ranged from 0.680 8 to 0.936 0, and annual average forest stock repeatability ranged from 0.481 1 to 0.925 0, indicating that volume was suitable as the target trait in selection. 3) The interaction analysis between genotype and environment showed that the result of AMMI and BLUP-GGE biplots based on volume traits were slightly different. Both models agreed that Leizhou was the ideal environment and EC245 was the high volume clone, followed by EC234 and CM1509. AMMI model showed that EC244, EC228 and EC238 had the higher stability, and EC245 was the lowest, and EC245 was screened as the clones with high volume and low stability. However, the BLUP-GGE model showed that EC244, EC232 and EC224 had higher stability, and EC238 was the lowest, and EC245 was screened as a high volume and medium stability clone. Conclusion: According to the comprehensive evaluation, EC245 was a clone with high volume and medium stability, and it has good growth performance in all locations.

Key words: Eucalyptus, repeatability, multi-area test, AMMI model, BLUP-GGE Model