The identification of genotypes more tolerant to water deficit is a challenge to breeding programs. In this research, our objectives were to identify and validate traits for tolerance to water deficit in eucalypts. The estimation of genotypic parameters and early selection are proposed based on mixed models, selection indexes and validation schemes. Seedlings with 110 days were grown in a greenhouse for 12 weeks, and two water deficit treatments were conducted (polyethylene glycol and water limitation). A total of 26 biomarkers were evaluated, and 15 of them were significant, exhibited adequate heritability, and used for screening: final plant height, increment in height, increment in diameter, area of mature and fully expanded leaf, nutrient contents of N, K, Ca, Mg, S, Cu, Zn, Mn and B, photosynthesis (A) and stomatal conductance (gs). Both treatments were adequate to discriminate water deficit-tolerant clones. The ranking of tolerant clones according to their phenotype in the field demonstrates the potential for early selection and is consistent with the maintenance of water-deficit-tolerance mechanisms until adulthood. There is evidence that the choice of biomarker depends on the species involved and different strategies contributing to the tolerance trait.

Water deficit causes a number of changes in plant development and physiology and is reported to be a trigger for eucalyptus dieback disorder. Dieback manifests itself in adult individuals, which hampers the selection of tolerant material and, therefore, means potential jeopardy to commercial plantings. Genetic parameters of phenotypic traits were estimated by the mixed model methodology (REML/BLUP), based on the observation of seedlings from twelve commercial eucalyptus clones grown under water stress. The water deficit was simulated by using polyethylene glycol (PEG 6000) and controlling the water depth administered to the plants for a period of 120 days. Among the evaluated anatomical and physiological characteristics, 10 biomarkers were selected for discriminating the clones, based on their differentiated and significant response among genotypes as well as relevant heritability and accuracy. The best-ranked clones demonstrated tolerance pattern in accordance to the empirical information in the production fields. Our results support the feasibility of early selection for selecting eucalyptus genotypes more tolerant to dieback and to water deficit.