Figure S1 Phenotypic evolution of NGR symbionts; Figure S2 Prevalence of plasmids in metagenome samples from Experimental evolution can enhance benefits of rhizobia to novel legume hosts

Symbiotic effectiveness was measured as shoot biomass (A, C, E, F) and symbiont fitness was estimated using the mean population size of rhizobia within a nodule (B, D, G, H) on har1 (A,B,E,G) and MG-20 (C,D,F,H) hosts. Symbiont effectiveness (A, C) and fitness (B, D) were compared between derived symbionts and their corresponding ancestor using the 4wpi harvest and analyzed using Welch's two sample t-test. Growth rate of host (E, F) and symbiont (G, H) between harvests were compared for ancestral and derived symbionts with an ANCOVA. Warm colors are used for the har1 experiments and cool colors are used for the MG-20 experiments. * indicates P < 0.05 for t-tests. # represents P < 0.05 for ANCOVAs.; Whisker boxplots of relative read depths of each position in replicons from metagenome sequencing samples and the individually sequenced ancestral CE3 strain. Read depth at each base of plasmid replicons was normalized to the mean chromosome read depth for each sample. Each plasmid is predicted to be single copy (~1X read depth relative to the chromosome) in the CE3 WT. Proportions less than one indicate that a subset of a population does not carry a given plasmid. Colors of boxplots indicate experiment group: shades of blue is CE3:har1_a, red is CE3:har1_b, green is CE3:MG-20_a, purple is CE3:MG-20_b, and orange is the CE3 ancestor. Sample numbers (5, 10, 15) indicate generation number that pooled communities were sampled from.