Bender’s Cave on the Edwards Plateau of Texas contains evidence of Late Pleistocene biodiversity that contrasts with the record from 17 regional sites dating to the last glacial interval, Marine Isotope Stage 2 (MIS 2). Bender’s Cave is a groundwater conduit system with an underground stream. Fossils occur in the cave primarily as an underwater lag assemblage and represent taxa that are typical of the Rancholabrean Land Mammal Age and common to central Texas (e.g., Bison, Mammuthus, Camelops). Megalonyx jeffersonii and Mammutidae also occur in Bender’s Cave but are rare elsewhere in the region. Other fossils provide the first regional records of Holmesina septentrionalis and a species of giant Hesperotestudo. The paleoecology of those novel taxa is inconsistent with regional paleoenvironmental proxies for MIS 2, which document a relatively open, dry grassland and cool climate. The novel composition of the assemblage may be the product of sample bias, and the fossils may be vastly time-averaged. However, the identified taxa frequently co-occur in other Texas sites interpreted as dating to interstadial/interglacial intervals of the late Pleistocene (i.e., MIS 3 or MIS 5), suggesting that fossils in Bender’s Cave may also date to one of those earlier warm periods.

The deposits in Cumberland Bone Cave (Allegany County, Maryland) preserved one of the most taxonomically diverse pre-radiocarbon Pleistocene faunas in the northeastern United States. The site has long been recognized as an important record of Pleistocene life in the region, but numerical age control for the fauna was never developed, and hypotheses for its age have been based upon biochronological assessments of the mammalian fauna. We used fossil teeth and preserved sediment housed in museum collections to obtain the first numerical age assessment of the fauna from Cumberland Bone Cave. Coupled U-series Electron Spin Resonance (US-ESR) was used to date fossil molars of the extinct peccary, Platygonus sp. The age estimates of two teeth gave ages of 722 ± 64 and 790 ± 53 ka. Our results are supported by previously unpublished paleomagnetic data generated by the late John Guilday, and by plotting length-width of the first molar (m1) of Ondatra (muskrats) from Cumberland Bone Cave on the chronocline of Ondatra molar evolution in North America. Our age assessments are surprisingly close to the age estimate previously proposed by Charles Repenning, who based his age on a somewhat complicated model of speciation and morphotype evolution among arvicoline rodents.

The fossil vole Microtus deceitensis occurs in Early Pleistocene deposits at Fort Selkirk, Yukon Territory, and Late Pliocene beds at the type locality, Cape Deceit, Alaska. Analyses of simple vs complex morphotypes in the cheek teeth, and of differentiation of tooth enamel, show that the Cape Deceit sample of M. deceitensis is less derived, and thus appears to be older, than the Fort Selkirk sample. The fossiliferous deposits at Fort Selkirk are well constrained by fission-track and radiometric dates and are 1.5 to 1.7 myr. Sediments at Cape Deceit bear a normal magnetic polarity, are correlated with the Olduvai subchron, and probably are latest Pliocene.

Speleological, stratigraphic, paleomagnetic and faunal data is presented for the Buffalo Cave fossil site in the Limpopo Province of South Africa. Speleothems and clastic deposits were sampled for paleomagnetic and mineral magnetic analysis from the northern part of the site, where stratigraphic relationships could be more easily defined and a magnetostratigraphy could therefore be developed for the site. This is also where excavations recovered the fossil material described. A comparison of the east and South African first and last appearance data with the Buffalo Cave fauna was then used to constrain the magnetostratigraphy to produce a more secure age for the site. The magnetostratigraphy showed a change from normal to reversed polarity in the basal speleothems followed by a short normal polarity period in the base of the clastic deposits and a slow change to reversed directions for the remainder of the sequence. The biochronology suggested an optimal age range of between 1.0 Ma and 600,000 yr based on faunal correlation with eastern and southern Africa. A comparison of the magnetobiostratigraphy with the GPTS suggests that the sequence covers the time period from the Olduvai event between 1.95 and 1.78 Ma, through the Jaramillo event at 1.07 Ma to 990,000 yr, until the Bruhnes–Matuyama boundary at 780,000 yr. The faunal-bearing clastic deposits are thus dated between 1.07 Ma and 780,000 yr with the main faunal remains occurring in sediments dated to just after the end of the Jaramillo Event at 990,000 yr.

The paraphyletic genus ‘Paleorhinus’ is understood currently as a cosmopolitan phytosaur taxon from the Late Triassic. There is no consensus regarding the number of species of ‘Paleorhinus,’ with multiple species and genera synonymised into a single genus or even a single species at various points in its published history. The taxonomy is confounded by historical descriptions without the benefit of comparisons to more recently collected specimens, emphasis on plesiomorphic cranial morphology as diagnostic features of the genus, and lack of cladistic analyses. When included in a recent explicitly cladistic phylogenetic analysis, the holotype of ‘Paleorhinus’ scurriensis (TTU P-00539) was found to be the earliest-branching phytosaur with respect to other North American specimens previously referred to ‘Paleorhinus,’ and is generically distinct from Paleorhinus. ‘Paleorhinus’ scurriensis differs from all known phytosaurs in five unambiguous characters: basitubera widely separated mediolaterally; ridge present on lateral surface of jugal; thickened shelf present along posteroventral edge of expanded pterygoid-quadrate wing; ‘septomaxillae’ separated and excluded from internarial septum; and nasal swelling present posterior to posterior borders of nares. This detailed morphological description of an early-branching phytosaur taxon is a first step towards resolving long-standing issues surrounding specific anatomical features and relationships among early members of the clade.

Recent stratigraphic revisions of the Upper Triassic Chinle Formation of Petrified Forest National Park, in conjunction with precise and accurate documentation of fossil tetrapod occurrences, clarified the local biostratigraphy, with regional and global implications. A significant overlap between Adamanian and Revueltian faunas is rejected, as is the validity of the Lamyan sub-land vertebrate faunachron. The Adamanian–Revueltian boundary can be precisely placed within the lower Jim Camp Wash beds of the Sonsela Member and thus does not occur at the hypothesised Tr-4 unconformity. This mid-Norian faunal turnover, may coincide with a floral turnover, based on palynology studies and also on sedimentological evidence of increasing aridity. Available age constraints bracketing the turnover horizon are consistent with the age of the Manicouagan impact event. The rise of dinosaurs in western North America did not correspond to the Adamanian–Revueltian transition, and overall dinosauromorph diversity seems to have remained at a constant level across it. The paucity of detailed Late Triassic vertebrate biostratigraphic data and radioisotopic dates makes it currently impossible to either support or reject the existence of globally synchronous Late Triassic extinctions for tetrapods.