The vertebrate-fossil record in the Karoo Basin has served as the accepted model for how terrestrial ecosystems responded to the end-Permian extinction event. A database of several hundred specimens, placed into generalized stratigraphies, has formed the basis of a step-wise extinction scenario interpreted by other workers as spanning the upper Daptocephalus (=Dicynodon) to Lystrosaurus Assemblage Zones (AZ). Seventy-three percent of specimens used to construct the published model originate from three farms in the Free State: Bethel, Heldenmoed, and Donald 207 (Fairydale). The current contribution empirically tests: (1) the stratigraphic resolution of the vertebrate record on these farms; (2) whether a sharp boundary exists that delimits the vertebrate assemblage zones in these classic localities; and (3) if the Lystrosaurus AZ is of early Triassic age. We have used a multi-disciplinary approach, combining lithostratigraphy, magnetostratigraphy, vertebrate biostratigraphy, and palynology, to test these long-held assumptions.

Previously reported vertebrate-collection sites have been physically placed into a litho- and magnetostratigraphic framework on the Bethel and Heldenmoed farms. The reported assemblage-zone boundary is used as the datum against which the stratigraphic position of vertebrates is compared and a preliminary magnetostratigraphy constructed. We find specimens of the Daptocephalus AZ originate in the Lystrosaurus AZ (as currently defined) and vice versa, and discrepancies between reported and field-checked stratigraphic positions below or above the assemblage-zone boundary often exceed 30 m. Hence, the utility of the data set in defining a sharp or abrupt biozone boundary is questionable. We further demonstrate the presence of a stratigraphically thick reverse polarity magnetozone that encompasses the reported assemblage-zone boundary, implying that these rocks are not correlative with the end-Permian event, which is reported to lie in a normal polarity chron. A latest Permian age is supported by palynological data from the Lystrosaurus AZ on the Donald 207 (Fairydale) farm, with equivalence to Australian (APP602) and Eastern Cape Province assemblages. We conclude that the turnover from the Daptocephalus to Lystrosaurus Assemblage Zones is more protracted than envisioned, it is not coincident with the end-Permian event as recognized in the marine realm, and little evidence exists in support of a three-phased extinction model based on vertebrate assemblages in the Karoo Basin.