It is the combination of large fitness differences and strong stabilizing mechanisms that often constitute niche-based explanations for species abundance patterns. Despite the importance of this assumption to much of community ecology, empirical evidence is surprisingly limited. Empirical tests are critical because many abundance patterns are also consistent with neutral-based alternatives (that assume no fitness differences or stabilization). We quantified interactions of four annual grassland species in two-species mixtures at varying frequencies. We found evidence of strong negative frequency-dependent stabilization, where scaled population growth rates increased with decreasing frequency for all four species. There was also a consistent competitive hierarchy among these species indicative of strong fitness differences that, in most cases, suggested potential competitive exclusion despite the observed strong stabilization.