The superfamily Djadochtatherioidea is a distinctive clade of multituberculates from Upper Cretaceous beds of Mongolia and Inner Mongolia, China. Because many of the 11 included genera are known from skulls, more is known about the cranial anatomy of djadochtatherioids than any other clade of multituberculates. Within Djadochtatherioidea, the most diverse and widely accepted group is the family Djadochtatheriidae. Within the family, the basal genus, KryptobaatarKielan-Jaworowska, 1970, is small with a skull length of about 30 mm, whereas the other four genera, DjadochtatheriumSimpson, 1925, CatopsbaatarKielan-Jaworowska, 1994, TombaatarRougier et al., 1997, and MangasbaatarRougier et al., 2016, have skulls approximately twice as long. Here, we describe a new genus and species, Guibaatar castellanus, based on a single specimen from the Upper Cretaceous Bayan Mandahu Formation, Inner Mongolia that we refer to Djadochtatheriidae. Guibaatar is represented by a relatively complete rostrum, a partial right braincase, and partial lower jaws. As revealed by CT scanning, the specimen is a juvenile, with deciduous enlarged upper and lower incisors with permanent replacements forming, m2 erupting, and M2 forming. Based on the preserved cranial parts, we estimate the skull length to be approximately 50 mm, but as an adult, Guibaatar would have been in the size range of the larger djadochtatheriids.

Phylogenetic analysis including Guibaatar, known djadochtatherioids, and outgroups places Guibaatar within Djadochtatheriidae, as sister to a clade of Mangasbaatar and Catopsbaatar. We suspect the relationships of djadochtatherioids are likely to be refined given the announcements by other researchers that skulls are known for the djadochtatheriids Tombaatar and Djadochtatherium, which were previously represented by incomplete material.

A new small sciuravid rodent, Natrona natronensis, new genus and species, from Uintan and Duchesnean deposits in the Wagon Bed Formation at the Hendry Ranch locality in central Wyoming exhibits some morphological parallelism in the structure of its P4 to the M1 of the more derived cricetid rodents. Other occurrences of this new lineage may be from the Eocene of Trans-Pecos Texas and Lac Pelletier, Saskatchewan, Canada.

Recent claims regarding what is and is not the largest known sauropod dinosaur are tested via dimensional comparisons of the most critical metrics of relative size—especially, when possible, the functional lengths of the dorsal vertebral centra and the articulated length of the combined trunk vertebrae—and analog volumetric models based on technical skeletal restorations. The Cretaceous Argentinosaurus massed 65–75 tonnes, and its dorsal vertebrae and dorsal–sacral series are much larger than those of any other described titanosaur. Specimens of Patagotitan indicate a 50–55 tonne titanosaur, and the less complete Notocolossus, Puertasaurus, and ‘Antarctosaurus’ giganteus appear to have occupied a similar size range. Paralititan weighed between 30 and 55 tonnes. The juvenile Dreadnoughtus, as well as Futalognkosaurus and Alamosaurus, were in the area of 30 tonnes, with the possibility that the last was substantially larger. Entirely analog, skillfully produced, high-anatomical-fidelity skeletal restorations and volumetric models representing a prime-lean condition are approximately as scientifically objective and accurate, as well as more realistic than, analog-digital, crudely-formed convex hull volumetric models, which are based on subjectively and often inconsistently or erroneously mounted skeletons and digitized skeletal reconstructions. The need to ensure that skeletal restorations are as anatomically correct and consistent as the data allow is stressed, which requires that researchers and illustrators be sufficiently skilled in animal and especially dinosaur anatomy, and the procedures and standards for achieving the best possible results are detailed. When properly executed, analog and digital volumetric models produce adequately similar results that can be used to cross-check one another, and both produce accurate masses much more reliably than do methods based on strength factors such as limb bone circumferences or certain other skeletal dimensions that suffer from inherently very high plus-minus factors.