The paper addresses the question: how does asymmetric competition for light aoeect the spatial pattern of trees? It is based on an individual­ based spatially explicit model of forest dynamics, whose growth equations are derived from gap models. The model is calibrated on a stand of natural rainforest in French Guiana, where the tree pattern exhibits regularity at short distances (< 10 m) and clustering at medium distances (~ 30 m). The model reproduces the regularity but not the clustering. As mortality and recruitment have been modelled so as to favor a random pattern, we conclude that regularity emerges from the asymmetric competition in the growth submodel. Also the scale at which regularity appears is linked to the range of interactions between trees.