Comparative studies have shown that diet, life history, and phylogeny interact to determine microbial community structure across mammalian hosts. However, these studies are often confounded by numerous factors. Selection experiments offer unique opportunities to validate conclusions and test hypotheses generated by comparative studies. We used a replicated, 15-generation selection experiment on bank voles (Myodes glareolus) that have been selected for high swim-induced aerobic metabolism, predatory behavior towards crickets, and the ability to maintain body mass on a high-fiber, herbivorous diet. We predicted that selection on host performance, mimicking adaptive radiation, would result in distinct microbial signatures. We collected foregut and cecum samples from animals that were all fed the same nutrient-rich diet and had not been subjected to any performance tests. We conducted microbial inventories of gut contents by sequencing the V4 region of the 16S rRNA gene. We found no differences in cecal microbial community structure or diversity between control lines and the aerobic or predatory lines. However, the cecal chambers of voles selected for herbivorous capability harbored distinct microbial communities that exhibited higher diversity than control lines. The foregut communities of herbivorous-selected voles were also distinct from control lines. Overall, this experiment suggests that differences in microbial communities across herbivorous mammals may be evolved, and not solely driven by current diet or other transient factors.