Greater white clover (Trifolium repens L.) competitive ability can increase the forage quantity and quality of clover-grass mixed stand (MS) cropping. The present work assessed the relationship of clover competitive ability with morphophysiological traits, and the efficiency of direct and indirect selection procedures for increasing clover dry matter (DM) yield in MS. Some 165 large-leaved clover genotypes representative of Italian Ladino germplasm and 13 clover genotypes from an elite small-leaved population were clonally evaluated for one year in pure stand (PS) and in MS environment with Italian ryegrass (Lolium multiflorum Lam.) and tall fescue (Festuca arundinacea Schreb.) that reproduced the level of interspecific interference met by the clover over a multi-year crop cycle. Results for all genotypes and their subsets indicated that the level of leaf size variation among tested genotypes can affect the results. Wider variation decreased the relative importance of clover genotype×evaluation environment interaction, because competitive ability was mainly correlated with traits (longer petiole; larger leaflet) contributing to scale foraging in MS and yielding ability in PS. Competitive ability tended to be more closely associated with other traits (high stolon density; plasticity of petiole elongation) contributing to foraging ability, i.e. the ability to make fine-scale exploitation of light and nutrients in MS, within germplasm of similar leaf size. Leaf size variation also affected the relative efficiency of selection procedures estimated from expected responses to clonal selection. Results for all entries indicated direct selection in MS as 23% more efficient than selection in PS based on a selection index including three traits associated with competitive ability (petiole length; stolon density; leaflet size) which, in turn, was predicted to be 16% more effective than selection in PS based on DM yield. The advantage in relative efficiency of direct selection was almost doubled when considering only large-leaved germplasm. In a second experiment, evaluating 16 clover half-sib progenies of elite large-leaved parents for 2 years in PS and in MS with Italian ryegrass, tall fescue and cocksfoot (Dactylis glomerata L.), direct selection was estimated to be over three-fold and 68% more effective than indirect selection in PS based on DM yield and the selection index, respectively, based on expected responses to genotypic selection. The selection index may improve the efficiency of preliminary, PS-based clonal selection. Further evaluation of elite genotypes and/or their polycross progenies in MS is recommended for final selection.