Bones of the cranial vault appear to be highly conserved among tetrapod vertebrates. Moreover, bones identified with the same name are assumed to be evolutionarily homologous. However, recent developmental studies reveal a key difference in the embryonic origin of cranial vault bones between representatives of two amniote lineages, mammals and birds, thereby challenging this view. In the mouse, the frontal is derived from cranial neural crest (CNC) but the parietal is derived from mesoderm, placing the CNC–mesoderm boundary at the suture between these bones. In the chicken, this boundary is located within the frontal. This difference and related data have led several recent authors to suggest that bones of the avian cranial vault are misidentified and should be renamed. To elucidate this apparent conflict, we fate-mapped CNC and mesoderm in axolotl to reveal the contributions of these two embryonic cell populations to the cranial vault in a urodele amphibian. The CNC–mesoderm boundary in axolotl is located between the frontal and parietal bones, as in the mouse but unlike the chicken. If, however, the avian frontal is regarded instead as a fused frontal and parietal (i.e. frontoparietal) and the parietal as a postparietal, then the cranial vault of birds becomes developmentally and topologically congruent with those of urodeles and mammals. This alternative hypothesis of cranial vault homology is also phylogenetically consistent with data from the tetrapod fossil record, where frontal, parietal and postparietal bones are present in stem lineages of all extant taxa, including birds. It further implies that a postparietal may be present in most non-avian archosaurs, but fused to the parietal or supraoccipital as in many extant mammals.