Aims: Spatial environmental heterogeneity has been considered an important co-existence mechanism because environmental variation enables different species to co-occur. We predict that if functional differences are important for co-existence, then both species and functional diversity should be positively related to environmental heterogeneity. Location: Thirty three dry calcareous grassland sites in Estonia. Methods: In each site, we established a transect (10.0 m × 0.1 m), consisting of 100 quadrats (10 cm × 10 cm). In each quadrat, we recorded species richness and composition, and measured soil depth, moisture and light availability. We collected data on eight traits from most of the species found across the sites. We calculated functional diversity (FD) at the quadrat scale, and compared the observed FD to that expected at random using two null models. The first null model used all the species that occurred in the transect in the randomizations to determine environmental filtering from the transect to the quadrat. The second null model restricted the species used in the randomizations to those species whose trait range was within the range of the observed values in the quadrat, to detect evidence for niche partitioning. Quadrat mean trait values and variability in functional composition were then related to small-scale mean environmental conditions and heterogeneity, respectively. Results: We found convergent patterns in biomass, specific leaf area, specific root length and clonality compared to randomized communities that included all species occurring in the transect, and found divergence in height and leaf area using the range-restricted null model. Hence, we found patterns consistent with environmental filtering and niche partitioning depending on the traits considered. Quadrat mean traits were significantly correlated with the measured environmental variables, with the main trend being a positive relationship between size-related traits and soil resources, and a negative relationship between these traits and light availability. Trait variability in height, leaf area, leaf dry matter content, specific root length and clonality was positively related to soil depth heterogeneity. There was a trend for species richness to be negatively related to environmental heterogeneity. Conclusions: Co-occurring species were more similar in several traits partly due to micro-environmental filtering in response to variability in soil depth. While environmental heterogeneity increased trait variability, negative relationships with species richness were observed. Hence, niche partitioning does not appear to be important for the maintenance of high small-scale species richness in these grasslands.