I review and analyze interaction webs involving competition and predation in ensembles of plethodontid salamanders in the southern Appalachian Mountains, USA. The objective is to evaluate hypotheses on interaction webs and population regulation in these ensembles in the context of traditional and recent models of phylogeny and life-history evolution of plethodontids. Two ecological hypotheses evaluated are (1) Hairston's conjecture on differences in the kinds of interactions prevalent among biphasic species with an aquatic larval stage, versus terrestrial species with direct development; and (2) Grover's hypothesis on differences in the factors that limit the distribution of species along the moisture gradient from stream to forest. An evolutionary question concerns the supposition derived from recent phylogenetic reconstructions that the history of one major clade, the supergenus Desmognathus, involved an early transition or reversal from a simple life cycle (direct development) to a biphasic life cycle (larval development, metamorphosis), and that this transition promoted the subsequent adaptive radiation of the taxon. Chippindale et al. proposed that selection for lifehistory reversal in the genus Desmognathus was promoted by competition between direct-developing ancestral desmognathans and other terrestrial plethodontids. Experimental studies of interaction webs in extant ensembles of plethodontids in the southern Appalachians, in concert with data from descriptive studies of feeding and habitat use, demonstrate variation in the relative importance of competition and predation in population regulation of species in the several codistributed lineages of these salamanders. Whereas the existing ecological data do not refute the Chippindale et al. hypothesis, many unresolved questions require additional investigation.