In spite of the ecological and economic importance of temperate bamboos, relatively little is known about their population biology or evolutionary history. Recently, hybridization has emerged as a potential source of diversity in this group, as well as an underlying cause of taxonomic problems. As part of a broader phylogenetic study of the temperate bamboos, we report the results of an analysis of the North American Arundinaria gigantea species complex, including estimates of genetic variation and molecular evidence of natural hybridization among A. gigantea, A. tecta, and A. appalachiana. The study involved a comparative analysis of amplified fragment length polymorphisms (AFLPs) and chloroplast DNA sequences representing diversity within and among all three species plus individuals with intermediate or unusual morphological characteristics (putative hybrids). Molecular results support the recognition of three species previously defined on the basis of morphology, anatomy, and ecology, with most of the molecular variance accounted for by among-species variation. Molecular evidence also demonstrates that A. tecta and A. appalachiana are sister species, forming a clade that is significantly divergent from A. gigantea. The role of hybridization in the phylogenetic history of Arundinaria is discussed along with implications for the evolution and taxonomy of the temperate woody bamboos.