Background/Question/Methods: Mutualisms, by definition, are interactions that benefit partner species, but there is increasing evidence that the magnitude of gain in fitness depends on partner genotype. Using a legume-rhizobia mutualism, we sought to understand additive genotype-by-genotype interactions by examining whether a single rhizobia strain could, on average, systematically affect productivity in a population of competing legumes comprised of multiple plant genotypes. In a greenhouse, we created replicates of a subsampled population (n=20) of Medicago lupulina, each containing the same representation of plant genotypes and exposed each ‘population’ to single inoculations of eight rhizobia genotypes (Ensifer meliloti). Following two months of growth, plant biomass, nodule number and nodule mass per plant was measured.

Results/Conclusions: Our results indicate that single strains did not have a significant average effect on population productivity- micro-environmental variation had a larger influence on plant productivity.  Plant genotype and nodule number was the strongest predictor of individual plant biomass. Strains significantly varied in nodule size, but nodule size was contingent on plant genotype. Our results provide increasing evidence of non-additive genotype*genotype interactions, which is one mechanism that maintains genetic variation in natural populations of rhizobia.