Using a target-neighborhood approach with six annual dicot species in the Negev Desert of Israel, we tested whether neighborhood biomass constrained the upper limit of plant performance and if the slope of the upper boundary was correlated to species trait means, such as relative growth rate (RGR) and seed mass. Target individuals were measured in early spring and then collected at the onset of the dry season along with all naturally occurring neighbors within a 5-cm radius of the target. Using a minimum of 50 samples for each of the six target species, we found no significant relationships between target performance and either density or aboveground biomass of neighbors, when including all target individuals in the regressions. However, aboveground neighbor biomass did influence the maximum potential size of targets. When regressions were restricted to the largest target plant within classes of neighbor biomass, significant negative relationships were found for all six species, with neighbor biomass explaining 55 to 94% of the variance in maximum target biomass. The slopes of the regressions were used as an index of competitive response. The correlations between competitive response and estimates of species traits were not significant with the possible exception that species with lower RGRmax may be better response competitors (i.e. were less sensitive to competition). These findings indicate the usefulness of the boundary regression technique for describing competitive interactions among neighbors, particularly in low productivity environments.