Determining whether grain yield is sink- or source-limited for a given environment and genotype is of paramount importance to design management practices and breeding strategies aimed at increasing grain yield. The objectives of this study were: (1) to test the hypothesis that mild water stress around anthesis can reduce wheat individual grain mass by reducing sink size potential, (2) to assess the degree of sink-source limitation to wheat grain yield, and (3) to assess differences in the degree of source-sink limitations of different grains within the spike in spring wheat grown in the dryland Pacific Northwest. Field grown spring wheat was subjected to a combination of two irrigation levels (irrigated from anthesis to maturity and dryland) and two spikelet removal levels (intact and 50% removal at anthesis). Both irrigation and spikelet removal acted additively on individual grain mass. Irrigated, halved spikes had the heaviest (51 mg grain^-1) and dryland-grown intact spikes the lightest grains (40 mg grain^-1), while grains in irrigated intact and dryland halved spikes had intermediate mass (42 and 47 mg grain^-1, respectively). Thus assimilate and water availability affected grain mass independently. A method to assess the degree of sink-source limitation to grain yield is presented, and when applied to this experiment indicated that wheat yield during grain filling was sink-source co-limited. Within the spike, the grain mass increase in halved spikes relative to that of intact spikes was larger the lighter the grain mass in intact spikes. There appeared to be no competition for assimilates within the spike which indicates sink rather than source limitation to grain mass. It remains to be determined if sink strength could be manipulated to increase yield without increasing inputs or use of water, nitrogen, and other resources.