Species' poleward ranges are thought to be primarily limited by climatic constraints rather than biotic interactions such as competition. However, theory suggests that a species' tolerance to competition is reduced in harsh environments, such as at the extremes of its climatic niche. This implies that under certain conditions, interspecific competition near species' range margins can prevent the establishment of populations into otherwise tolerable environments and results in geographic distributions being shaped by the interaction of climate and competition. We test this prediction by challenging an experimentally-parameterized mechanistic competition model to predict the poleward range boundaries of two widely co-occurring and ecologically-similar aquatic duckweed plants. We show that simple, mechanistic ecological niche models which include competition and thermal response terms best predict the northern range limits of our study species, outperforming competition-free mechanistic models and matching the predictive ability of popular statistical niche models fit to occurrence records. Next, using the theoretical framework of modern coexistence theory, we show that relative nonlinearity in competitors' responses to temperature fluctuations maintains coexistence at the subordinate competitor's poleward range boundary, highlighting the importance of this underappreciated fluctuation-dependent coexistence mechanism. Our results demonstrate the predictive utility of mechanistic niche models and support a more nuanced, interactive role of climate and species interactions in determining range boundaries, which may help explain the conflicting results from previous tests of classic range limit theory and contribute to a more mechanistic understanding of range dynamics under global change.K eywords range limit | competition | coexistence | ecological niche model | thought to have primacy over interspecific interactions in determining species' poleward boundaries. However, empirical support for this century-old hypothesis -variously called stress gradient hypothesis or the species interactions-abiotic stress hypothesis (SIASH) [5]-remains mixed [6,7,8,9,10].These hypotheses commonly posit that the per capita magnitude [11,5] or relative frequency [12,13] of negative interactions, such as competition, should decrease with environmental stress resulting in weaker competitive regulation of populations in harsher environments. While this prediction often holds (reviewed in [5]), a population's tolerance of competition can also decrease in harshening environments, where even weak competition can drive an already-low per capita growth rate, r (= dN/N dt, where N is population size), below zero [14,15,16,17,18,19]. The relative importance of competition or other negative interactions in shaping range margins may therefore hinge on a balance between the overall intensity of competition experienced by a marginal population and the extent to which competition suppresses its environmentallydetermined per capita growth rate.Assumin...