Primary succession is a fundamental process in macroecosystems; however, if and how soil development influences microbial community structure is poorly understood. Thus, we investigated changes in the bacterial community along a chronosequence of three unvegetated, early successional soils ( approximately 20-year age gradient) from a receding glacier in southeastern Peru using molecular phylogenetic techniques. We found that evenness, phylogenetic diversity, and the number of phylotypes were lowest in the youngest soils, increased in the intermediate aged soils, and plateaued in the oldest soils. This increase in diversity was commensurate with an increase in the number of sequences related to common soil bacteria in the older soils, including members of the divisions Acidobacteria, Bacteroidetes, and Verrucomicrobia. Sequences related to the Comamonadaceae clade of the Betaproteobacteria were dominant in the youngest soil, decreased in abundance in the intermediate age soil, and were not detected in the oldest soil. These sequences are closely related to culturable heterotrophs from rock and ice environments, suggesting that they originated from organisms living within or below the glacier. Sequences related to a variety of nitrogen (N)-fixing clades within the Cyanobacteria were abundant along the chronosequence, comprising 6-40% of phylotypes along the age gradient. Although there was no obvious change in the overall abundance of cyanobacterial sequences along the chronosequence, there was a dramatic shift in the abundance of specific cyanobacterial phylotypes, with the intermediate aged soils containing the greatest diversity of these sequences. Most soil biogeochemical characteristics showed little change along this approximately 20-year soil age gradient; however, soil N pools significantly increased with soil age, perhaps as a result of the activity of the N-fixing Cyanobacteria. Our results suggest that, like macrobial communities, soil microbial communities are structured by substrate age, and that they, too, undergo predictable changes through time.