Flowering components and survival are particularly important to Eucalyptus cladocalyx in the arid zones of Chile. The aim of this study was to examine the population variability for flowering components and survival in eight year old trees grown in two sites of the southern Atacama Desert, Chile. A Bayesian generalized lineal modeling approach (implemented using an independence chain algorithm) was used in data analysis. The deviance information criterion (Delta DIC), implemented using Gibbs sampling, evidenced significant differences (Delta DIC>5) for population effect, for each binary trait analyzed. The population-site interaction effect was significant (Delta DIC>5) for the variables: presence/absence of maximum flowering intensity (MI) and presence/absence of flowers or flower buds (FOY) in the beginning of growing season 2009/2010. This interaction was not statistically significant (Delta DIC<5) for presence/absence of capsules and/or flower (or buds) (EF) and survival. The results of this study emphasize the importance of the seed source selection. For example, 45% of the trees from Cowell, Australia, evidenced maximum flowering intensity in a determined site, while the trees from Flinders Chase, Australia, evidenced only about 3%. Information generated on inter-population variability could be used to improve traits of interest, although the genotype-environment interaction should be regarded for some flowering components.