Clonal growth and reproduction in tristylousLythrum salicariaL. were examined experimentally, using cloned genotypes of each of the three flower morphs, in field studies involving four moisture and three nutrient treatments. Clonal growth was measured in terms of diameter of clones, number of ramets per clone, and total length of ramets, and an index of reproduction was recorded as the total length of infructescence per clone. Neither clonal growth nor reproduction differed significantly among flower morphs, but both differed significantly as a consequence of both moisture and nutrient treatments. The pattern of seasonal growth indicates that ramet production was restricted mainly to the beginning of the season following vigorous vegetative growth. Although flowering began in June, it was restricted to plants in drier treatments in the water-gradient experiment. Characters intrinsic to tristyly (such as lengths of styles and stamens, and allocation of biomass to stamens and pistil) differed significantly among morphs. Soil moisture levels but not fertilizer treatments significantly affected the size of floral structures and biomass. Although absolute levels of biomass allocation to whole flowers and to attractive structures did not differ significantly among morphs, relative allocation to stamens increased progressively from long morph to mid-morph to short morph, with a corresponding decrease in relative mass of pistil. Although proportional allocation differed significantly among morphs, it was unaffected by moisture treatment, suggesting tight genetic control of herkogamy (spatial separation between anther and stigma). This should maintain the floral polymorphism in different ecological conditions.Key words:Lythrum salicaria, nutrient and water gradients, heterostyly, floral morphometry, floral allocation, clonal growth, sexual reproduction.