In southern and western Wisconsin, oak savanna was once the dominant vegetation type, but less than one percent remains.  The lower Chippewa River corridor contains perhaps 40% of the remnant oak savanna in the state.   This unique and important vegetation type also contains steep gradients in soil quality, fire and management history, and tree canopy cover.  One of our principal interests is in understanding the relative importance of these gradients on plant diversity, especially in terms of alternative measures of diversity, such as phylogenetic diversity and functional trait diversity.  The second principal interest is in understanding the relative importance of the gradients and how they affect functional trait composition of the plant communities.   Toward these ends, we have recently collected 150 quadrat samples from 30 sites in which we also collected functional trait data from each species.  We used several methods of analysis.  We measured components of phylogenetic diversity using the Phylocom program.  We also measured several aspects of functional diversity and functional trait composition using the FD package in R.  Data analysis has focused on two methods: structural equation modeling and indirect gradient analysis.  
Results/Conclusions

Structural equation models for phylogenetic and functional diversity are somewhat equivocal, especially in terms of their comparison to taxonomic measures such as species richness.   Phylogenetic and functional diversity indices that are closely tied to richness are under similar controls, i.e., fire and land management have the strongest effects which are unimodal in nature, while soil effects are weaker and more complex, and are partially mediated by production.  Diversity indices that are not tied to taxonomic richness are only very weakly explained by the underlying gradients.   Key questions are whether these aspects of diversity are important and if they are, what is affecting them? 

Indirect gradient analysis of functional traits reveal strong effects of fire and management on traits (decreasing mean SLA and seed mass, while increasing leaf DMC), while soil quality principally affected plant size, but not leaf economic traits.  A key difference between taxonomic and functional composition was that the disturbance gradient and the soil quality gradient had largely orthogonal effects taxonomic composition, while they had more similar effects in the functional composition.  This suggests a curious trap for advocates of either approach.