Seasonal changes in carbon isotope discrimination (Δ) and gas exchange traits were assessed in four Populusxeuramericana clones differing in growth potential. Measurements were made during the second year after establishment in the field under two watering regimes, which were defined by the time-span between flood irrigations, hence resulting in different dry-down cycles: high irrigation (conservative schedule currently applied in the Ebro Valley, Spain) and low irrigation (equivalent to about a one-fourth reduction in water inputs). Net CO2 assimilation rate (A), stomatal conductance (gs), intrinsic water-use efficiency (A/gs) and other related photosynthetic traits (leaf nitrogen concentration, leaf greenness and leaf mass per area) were measured prior to watering, and Δ was analysed in water-soluble leaf extracts (Δs) and bulk leaves (Δl). Stem growth was monitored over 3 years starting at the year of establishment (1998). Data were subjected to a repeated measures ANOVA over time for a randomised block split-plot design across watering regimes. Significant differences between watering regimes were detected using a long-term estimate of photosynthetic performance such as Δl, in agreement with changes in soil water status and evapotranspirative demand. However, the lack of significant genotypexwatering regime interactions for gas exchange traits and Δs suggested that water shortage imposed by low irrigation was not sufficient to reveal physiological adaptations to drought. In this regard, the reduction in water inputs brought about by low irrigation did not reduce tree growth for any of the clones, suggesting that the current irrigation scheme employed in the region is superfluous to the water consumption needs of poplars. Genotypic variation was detected in gas exchange traits, Δs, Δl and stem growth under both watering treatments. Significant correlations with stem volume for Δs (r = -0.60, p<0.05) and A (r = + 0.61, p<0.05) suggested that growth was improved by higher water-use efficiency (the ratio of carbon fixed to water lost, as inferred by Δs) due to variation in A rather than in gs. This observation corroborated the expectation derived from current theories that a lower Δ is related to higher stem volume, as a result of changes in net CO2 assimilation rates.