Background/Question/Methods There is continued interest in the roles of community attributes, such as biodiversity, in affecting ecosystem processes. Controlled field experiments have shown that diversity can enhance primary production, while uncontrolled field studies have shown that diversity and plant biomass production covary largely because they are affected by same environmental drivers, and any direct effects are weak and from production to diversity, not vice versa. In addition, most of these studies have used biomass as an indicator of productivity, which adds to the confusion. We conducted a field experiment where diversity was indirectly manipulated by altering the number of species that were initially planted (6 to 30 species, with 6 grasses, 0-16 forbs, and 0-8 legumes, from a pool of 40 species). We followed the communities for four years, allowing both species gain and loss to occur. Within the diversity treatments, we nested nitrogen addition (to eliminate N-limitation) and fungicide (to suppress mycorrhizal fungal colonization). In year 4, we collected biomass two weeks after burning in the spring, and then again in August. Above-ground Net Primary Productivity was determined from dry biomass as (1) g m-2 d-1, and as (2) g g-1 d-1 (which is analogous to a relative growth rate).

Results/Conclusions Nested anova showed that in year four, the diversity manipulations increased species richness, and that fungicide reduced MF colonization of plant roots by about 40%. The diversity manipulations significantly increased ANPP g-1 but not ANPP m-2. Fungal suppression sharply reduced ANPP, and the reductions were slightly larger if Nitrogen was also added. The curious negative effect of N was due to enhanced early spring growth and equivalent production. The diversity manipulations and chemical treatments may have indirectly affected ANPP via community composition. ANPP per g was more strongly associated with community attributes, but both showed similar patterns. ANPP was positively correlated with plant richness (r = .53), mycorrhizal fungal richness (r = .37), but not with estimated prokaryotic diversity (r = -.18). ANPP was correlated with community functional parameters (abundance-weighted mean leaf dry matter content r = -.44, height = .56) and simple measures of functional diversity (range of plant height r = .46, SLA r = .33). These relationships may be causal or they may be coincidences due to shared common causes. Structural equation modeling may be useful to untangle this mess.