Understanding the nature of competitive interactions among trees is critical to understand forest community structure and dynamics, and to predict their responses to global change. However, predicting the outcome of competitive interaction between numerous pairs of tree species and across large gradients of abiotic conditions remains challenging. A classical hypothesis -the competition-relatedness hypothesis- assumes that because of niche and trait conservatism, closely related species tend to be ecologically more similar and thus compete more severely than distantly related species. However, this hypothesis has remained largely untested with direct measurements of competition. Rather, it was only evaluated from simple patterns of species co-occurrence in the past. Recent publications have challenged this hypothesis on the ground that coexistence requires more complex explanations than simply niche differentiation. Coexistence is also driven by differences in species' competitive ability, and this could have effects opposed to the competition- relatedness hypothesis. For instance, pairs of species sharing similar values of trait responsible for a high competitive ability should compete less strongly than species with more distinct trait values. Using data on individual tree radial growth from more than 17000 forest plots in the French Alps, we explore the relative role of these two contrasting views of competitive interactions for 130 pairs of species. We developed a Bayesian neighbourhood growth competition model to estimate the intensity of competition between pairs of species. We then tested the relation between strength of competition between pairs of species and i)their phylogenetic distance or ii)their functional similarity for several key functional traits.