Spatial patterns of functional traits have received little attention in community ecology but have the potential to provide insights into the processes that structure communities. In this study, I used semivariograms to describe spatial patterns of functional traits and evaluate processes (niche differentiation, environmental filtering, and dispersal limitation) driving functional divergence in old-field plant communities. I collected spatially explicit data on key plant functional traits (vegetative height, specific leaf area [SLA], and leaf dry matter content [LDMC]) and environmental variables (soil depth and soil moisture) across a range of spatial scales (< 1–1500 m) in old fields in central New York. All traits displayed nonrandom spatial patterns consistent with the environmental filtering hypothesis, but patterns differed among traits. Height had strong spatial dependence at scales congruent with spatial heterogeneity of soil depth, indicating that soil depth acted as a spatial template for divergence in height. SLA and LDMC had much weaker spatial dependence, with > 90% of total divergence occurring within 1-m2 plots, demonstrating that high levels of functional divergence may occur at very fine spatial scales. Spatial patterns of intraspecific functional divergence differed among four common species (Solidago canadensis, Bromus inermis, Poa pratensis and Galium mollugo), indicating that species differed in their trait responses to environmental variation. This study provides novel descriptions of spatial patterns of functional traits in plant communities and demonstrates how these patterns can help understand the processes driving functional divergence across spatial scales.