Reinterpretation of cranial materials of the diadectomorphs Limnoscelis and Diadectes has prompted a reconsideration of the origin and early evolution of the amniote occiput. The basic approach is a phylogenetic study of major groups of Paleozoic tetrapods based on the occiput and closely associated elements of the skull roof. A lack of adequate anatomical data has forced the elimination of only a few relevant higher-level taxa from consideration, and, using Acanthostega as the reference outgroup, a cladistic analysis of the interrelationships of the Lepospondyli, Temnospondyli, Seymouriamorpha, Baphetidae (=Loxommatidae), Anthracosauria, Diadectomorpha, Synapsida, and Reptilia has produced the following results: 1) the ingroup taxa exhibit a basal dichotomy in which one division consists of the unresolved relationships of Lepospondyli, Temnospondyli, and Seymouriamorpha; 2) the pattern of relationships of the second division of the ingroup taxa is a series of nested clades, terminating with the Diadectomorpha and Synapsida as sister taxa sharing a more recent common ancestor than either does with Reptilia. This relationship requires assignment of Diadectomorpha to Amniota; and 3) the Anthracosauria and Baphetidae are progressively more distant clades or sister taxa. On the basis of the cladistic analysis, the attainment of the amniote occiput is described as passing through four morphological grades of organization. Each grade of the series is characterized by a set of derived character states that defines the progressively more-derived nodes and from which branch a clade containing the unresolved trichotomy of Lepospondyli, Temnospondyli, and Seymouriamorpha; the Baphetidae clade; the Anthracosauria clade; and the Diadectomorpha+Synapsida and Reptilia clades, respectively.