?Phylogenetic relationships were examined within the salamander family Salamandridae using 18 species representing 14 salamandrid genera and six outgroup taxa from the families Ambystomatidae, Dicamptodontidae, Plethodontidae, and Proteidae. Mitochondrial DNA sequences encoding the 12S and 16S ribosomal RNA and the intervening valine transfer RNA provided 431 phylogenetically informative nucleotide sequence positions from a multiple align? ment of approximately 1,000 bases per species. This variation was analyzed in conjunction with 44 previously reported morphological characters representing primarily hyobranchial myology and osteology, cranial osteology, and reproductive biology The molecular and morphological character sets were highly congruent, with only 2.8% of the total character incongruence attrib? utable to conflict between them. Parsimony analysis of the combined molecular and morphological data produced a single most-parsimonious tree whose topology was identical to that of the mostparsimonious tree derived from the molecular data alone. This tree suggests that the true sal? amanders (Chioglossa, Mertensiella, and Salamandra) form a monophyletic sister group to the newts (all remaining salamandrid genera). Within the newts, the first phylogenetic split separates Salamandrina from the remaining genera, within which Pleurodeles and Tylototriton form a monophy? letic sister group to the remaining taxa. The genus Triturus appears not to be monophyletic. Using a phylogenetic reconstruction of character changes, we tested hypotheses of adaptation in the evolution of aquatic suction feeding and terrestrial feeding featuring tongue protrusion. Phylo? genetic trends in the evolution of salamandrid courtship behavior were also examined. [Salamandridae; molecular phylogenetics; mitochondrial DNA; congruence; feeding morphology; court? ship.] Reconstruction of phylogeny is funda? mental to an understanding of the evolu? tion of biological diversity because phylo? genetic trees provide the historical maps along which character evolution is traced. Historical analysis is important for testing hypotheses of adaptive evolution (Baum and Larson, 1991) and for revealing pat? terns of homoplasy that indicate the action of natural selection and developmental constraints (Alberch, 1988; Wake, 1991). Salamanders are a particularly good group for phylogenetic studies of the interactions of design limitations, heterochrony, and se? lection (Wake and Larson, 1987; Wake, 1991). The family Salamandridae exhibits con? siderable morphological and behavioral di? versity. It contains 15 genera and 53 rec? ognized species and is distributed 1 Present address: Department of Biology, Univer? sity of Oregon, Eugene, Oregon 97403, USA. E-mail: titus@darkwing.uoregon.edu. throughout the Holarctic, with the greatest diversity in Europe (seven genera) and Asia (four genera) (Frost, 1985). Salaman? drid evolution has been studied from a va? riety of aspects, including courtship be? havior (Salthe, 1967; Halliday, 1977; Arntzen and Sparreboom, 1989), antipredator behavior (reviewed by Brodie, 1983), tox? icity (Brodie et al., 1974), morphology (Ozeti and Wake, 1969; Wake and zeti, 1969; Zhao and Hu, 1988; Sever, 1992), karyology (reviewed by Macgregor et al., 1990), pro? tein variation (Hedgecock and Ayala, 1974; Rafinski and Arntzen, 1987; Hayashi and Matsui, 1989; Reilly, 1990), and mitochon? drial DNA (Wallis, 1987; Wallis and Arntz? en, 1989, Caccone et al., 1994). Phylogenetic relationships within the Salamandridae have been a source of much conflict. Monophyly of the Salamandridae, although generally accepted, relies largely on interpretation of a single character, the frontosquamosal arch. This character is not present in all salamandrids, and its ab-