Rule #1: More taxa more precisely nest all taxa
Once again, latest Cretaceous Cimolestes goes under review, this time with many more candidate sister taxa. At present the best material for this genus is a single mandible with a complete set of teeth (Fig. 1). Rook and Hunter 2013 nest Cimolestes as the direct outgroup to the tradtional clade Taeniodonta.
According to Wikipedia,
“[Members of the genus Cimolestes] were once considered to be marsupials, then primitive placental mammals, but now are considered to be members of the order Cimolesta (which was named after the genus), outside of placental mammals proper (Rook and Hunter 2013). Before they were determined to be non-placental eutherians, the cimolestids were once considered the common ancestral group of the clades Carnivora and the extinct Creodonta.”
With so few traits to score,
Cimolestes is difficult to nest and generally causes loss of resolution, especially when other taxa data include skulls without mandibles (so, no comparable traits = loss of resolution). When deleting taxa without preservation of the mandibles the best match is with the extant marsupial, Dasyurus (Fig. 1), but with a narrower coronoid process and larger incisors, and therefore a likely shorter and smaller cranial region.
Cimolestes is considered a basal taeniodont and all taeniodonts are considered eutherians (placentals). Other traditional taeniodonts include Protictis, Onychodectes and Stylinodon. The Rook and Hunter cladogram of eutherian relationships nests only one traditional taeniodont alongside Cimolestes (Fig. 3) and the basalmost member of the tenrec-odontocete clade (in the LRT), Anagale.
In the large reptile tree (LRT, 1362 taxa, subset Fig. 4) none of these taxa nest with one another. Their previous joining may be due to eyeballing, a reliance on dental traits and taxon exclusion. That’s all Cope had available at the time. Modern workers appear to have followed traditional taxon lists and convergent dental traits without testing a wider gamut of taxa. The LRT includes more taxa and does not emphasize dental traits.
When tested with additional taxa,
(Fig. 4) the traditional eutherian clade Taeniodonta is polyphyletic and should be abandoned. Only a few traditional members are closely related to one another.
taxa in the LRT provide and document a gradual accumulation of derived traits that competing cladograms cannot match.
More former taeniodonts to come.
Fox RC 2015. A revision of the Late Cretaceous–Paleocene eutherian mammal Cimolestes Marsh, 1889. Canadian Journal of Earth Sciences (advance online publication) doi: 10.1139/cjes-2015-0113.
Marsh OC 1889. Marsupialia, Cimolestidae. American Journal of Science and Arts 3d ser., XXXVIII, 89, pl. iv, figs. 8–19.
Rook DL and Hunter JP 2013. Rooting around the eutherian family tree: the origin and relations of the Taeniodonta. Journal of Mammal Evolution