Poplar hybrids were grown with irrigation in a large-scale plantation to investigate the mechanisms underlying clonal differences in drought resistance. Beginning in spring 1992, Populus trichocarpa x P. deltoides (TD) and P. deltoides x P. nigra (DN) cuttings received 46, 76, or 137 cm year(-1) of irrigation to supplement the 18-20 cm of annual precipitation, and all trees received the same fertilization regime. Stem volume, assessed as the square of stem diameter at breast height times tree height (D(2)H), and water relations of the trees were studied from the end of their second growing season until the end of their fifth growing season. By the end of the second growing season, stem volume of Clone TD was 40-146% larger than that of Clone DN, but stem volume growth was independent of irrigation in excess of 46 cm year(-1) in both clones. During the third growing season, stem volume growth of both clones was limited by both the 46- and 76-cm irrigation treatments, so that by the end of the third growing season trees in the 46-cm irrigation treatment were only half the size of trees in the 137-cm irrigation treatment. These treatment differences were maintained through the fifth growing season. Although stem volumes of Clone TD trees in the 76- and 137-cm irrigation treatments were larger than the corresponding values for Clone DN trees at the end of the third growing season (1994), these clonal differences gradually decreased in subsequent years and were not detectable after 5 years, because stem volume relative growth rate of Clone DN was greater than that of Clone TD in all treatments. Although both clones exhibited similar predawn leaf water potentials, Clone DN typically maintained higher midday leaf water potentials, suggesting better stomatal control of water loss. Clonal and treatment differences in osmotic potential at full turgor were minimal and could not explain the clonal differences in drought resistance. Root density and root density to stem volume ratio increased more in response to moderate drought in Clone DN than in Clone TD, resulting in enhanced drought resistance (high stem volume growth rate under moderate drought conditions) and an increased capacity to withdraw water from the soil. We conclude that the greater drought resistance of Clone DN compared with Clone TD was the result of the maintenance of a more favorable water balance by stomatal regulation and greater carbon allocation to roots during the early stages of drought. However, the low root density to stem volume ratio in Clone DN growing in the 46-cm irrigation treatment suggests that severe water limitation restricted the preferential allocation of carbon to belowground tissues, so that both root and shoot growth were constrained by severe drought.