Despite numerous descriptive studies, the embryonic origin of vertebrate taste buds has never been experimentally determined. A number of different alternatives have been suggested for taste bud origins, including epibranchial placodes, the neural crest, and the local epithelium of the oropharyngeal cavity. The role of a series of epibranchial placodes and the cephalic neural crest, which together give rise to the cranial nerves innervating taste buds, was examined with regard to the development of oropharyngeal taste buds in an ambystomatid salamander, the axolotl. When pigmented placodal ectoderm or neural folds were grafted isotopically and isochronically into nonpigmented host embryos, known derivatives of each tissue contained pigmented cells, but labeled taste buds were never encountered. Thus, neither epibranchial placodes nor neural crest contribute cells to taste buds during embryogenesis. The majority of the oropharyngeal cavity of ambystomatid salamanders is lined by an endodermal epithelium. In order to demonstrate conclusively that taste buds arise from this local epithelium, the presumptive cephalic endoderm of early axolotl gastrulae was microinjected with the lipophilic dye, DiI. In the oropharyngeal epithelium of all larvae examined, both taste buds and general epithelial cells were labeled with DiI, indicating their common endodermal origin. Our findings are novel in that this is the first experimental demonstration of the endodermal origin of a vertebrate sensory receptor cell class.