We looked at this genus earlier
when Rose et al. 2014 first described Cambaytherium from some bits and pieces of this sheep-sized mammal from the Early Eocene (Fig. 1). Back in 2017 the large reptile tree (LRT; subset Fig. 5) nested Cambaytherium within the oreodont + mesonychid + hippo (Fig. 6) + anthracobunid (Fig. 2) + desmostylian (Figs. 3, 6) + mysticete (Fig. 3) clade. In the LRT that clade is far from Perissodactyla (= tapirs + horses + rhinos).
Figure 1. Cambaytherium with a an alternate rostrum reversing taphonomic shifts.
Rose et al. 2020 update their discoveries
with new data (= more bones from more specimens) of Cambaytherium, which they still insist is a perissodactyl (rhino + horse) ancestor) ancestor.
taxon exclusion was the problem then and remains the problem now. The authors do not include enough taxa that nest closer to Cambaytherium in LRT (subset Fig. 5), other than adding Anthracobune (Fig. 2) and Behemotops (Fig. 3).
Figure 2. Anthracobune reconstructed with a larger skull to match the teeth on the mandible.
Rose et al. 2020 employed 72 taxa.
That is twenty more than in Rose et al. 2014. Their more recent strict consensus resulted in 408 shortest trees. If that many shortest trees happened in the LRT using only 72 taxa, I would dive back in and fix what was wrong. In this case, taxon exclusion is still the problem with Rose et al. 2020.
Figure 3. Rorqual evolution from desmostylians, Neoparadoxia, the RBCM specimen of Behemotops, Miocaperea, Eschrichtius and Cetotherium, not to scale.
Rose et al. 2020 explain their 20 taxon additions:
“The Eocene desmostylian Behemotops proteus was included for two reasons. First, there are interesting similarities between the teeth of this taxon and those of cambaytheres and anthracobunids (Ray et al., 1994). Second, the analysis of Cooper et al. (2014) placed Desmostylia in a polytomy that included most perissodactyls and anthracobunids (but not cambaytheres). Historically, paleontologists have considered Desmostylia (like anthracobunids) to be closely related to tethytheres (Domning et al., 1986; Ray et al., 1994).”
Rose et al. 2020 cherry-picked some of the right taxa, but not enough of the right taxa, according to the LRT. They are “Pulling a Larry Martin” (= putting too much attention on a few ‘key’ traits). These mammal experts still have no idea what the actual outgroups and relatives of Cambaytherium are because they don’t include enough pertinent taxa (see Fig. 5).
In the LRT, force-shifting Cambaytherium
to the base of the Perissodactylia added 18 steps to the LRT. That’s not a large number considering the phylogenetic distance, so convergence is an issue here. Convergence is something the LRT handles pretty well due to its wide gamut of taxa.
Rose et al. 2014 put their faith in genomic results with unclear and mixed results. Never put your faith in genomic results in deep time studies (especially fossil taxa). Also, don’t trust the work of others. Don’t borrow cladograms. Do your own phenomic analysis with a wide gamut of taxa, as in the LRT. Then you’ll have a good model for actual evolutionary events.
From the Rose et al. 2020 abstract:
“The anatomy of Cambaytherium, a primitive, perissodactyl-like mammal from the lower Eocene Cambay Shale Formation of Gujarat, India, is described in detail on the basis of more than 350 specimens that represent almost the entire dentition and the skeleton.”
Many of the specimens are small disarticulated and broken pieces of bone. At least one skull is pretty well fossilized in 3D (Fig. 1).
“Cambaytherium combines plesiomorphic traits typical of archaic ungulates such as phenacodontids with derived traits characteristic of early perissodactyls.”
The authors cannot be more specific due to taxon exclusion.
“Cambaytherium was a subcursorial animal better adapted for running than phenacodontids but less specialized than early perissodactyls. The cheek teeth are bunodont with large upper molar conules, not lophodont as in early perissodactyls; like perissodactyls, however, the lower molars have twinned metaconids and m3 has an extended hypoconulid lobe. A steep wear gradient with heavy wear in the middle of the tooth row suggests an abrasive herbivorous diet.
Homologous desmostylian and hippo teeth also show heavy wear.
“Three species of Cambaytherium are recognized: C. thewissi (∼23 kg), C. gracilis (∼10 kg), and C. marinus (∼99 kg). Body masses were estimated from tooth size and long bone dimensions. Biostratigraphic and isotopic evidence indicates an age of ca. 54.5 Ma for the Cambay Shale vertebrate fauna, the oldest Cenozoic continental vertebrate assemblage from India, near or prior to the initial collision with Asia.
“Cambaytheriidae (also including Nakusia and Perissobune (Fig. 4) and Anthracobunidae are sister taxa, constituting the clade Anthracobunia, which is sister to Perissodactyla.
You heard it here first: The LRT nested Cambaytherium with Anthracobune (Fig. 3), but far from Perissodactyla due to taxon exclusion.
Perissobune (Fig. 4) has not been tested in the LRT, but closely resembles the basal perissodactyl, Protapirus (Fig. 4), with a taller coronoid process and only two lower molars. Compare to the distinctly different mandible of Cambaytherium (Fig. 1).
“We unite them in a new higher taxon, Perissodactylamorpha.
This is invalid, inappropriate, premature and based on cherry-picked taxa.
Figure 4. Perissobune mandible compared to diagram of a primitive perissodactyl, Protapirus. Compare to figure 1.
Continuing from the Rose et al. 2020 abstract:
“The antiquity and occurrence of Cambaytherium—the most primitive known perissodactylamorph—in India near or before its collision with Asia suggest that Perissodactyla evolved during the Paleocene on the Indian Plate or in peripheral areas of southern or southwestern Asia.”
Figure 5. The oreodont-mesonychid-hippo-desmoystlian-mysticete clade subset of the LRT
Figure 6. Merycopotamus, Hippopotamus, and Paleoparadoxia compared to scale.
Back in 2017 readers learned the following
from Rose et al. 2014, “Cambaytherium was found on the marine coastline of island India.”
Then I added, “close to where odontocete whales were also evolving from Tenrec relatives like Pakicetus. Based on its relationships and geography, Cambaytherium was likely much more aquatic than is typical for perissodactyls.”
Figure 7. Simbakubwa from Broths and Stevens 2019, colors added, and compared to a lion mandible. Note the two medial views of the mandible with different shapes. Dorsal view of indented mandible and palate is similar to hippos.
For some reason no one wants to test enigma taxa with hippos
and mesonychids. We looked at this same problem earlier with the putative carnivore, Simbakumba, which also turned out to be a hippo relative.
Rose, KD et al. (8 other authors) 2014. Early Eocene fossils suggest that the mammalian order Perissodactyla originated in India. Nature Communications. 5 (5570). doi:10.1038/ncomms6570.
Rose KD et al, 2020. Anatomy, Relationships, and Paleobiology of Cambaytherium (Mammalia, Perissodactylamorpha, Anthracobunia) from the lower Eocene of western India, Journal of Vertebrate Paleontology (2020). DOI: 10.1080/02724634.2020.1761370
“New research published today in the Journal of Vertebrate Paleontology describes a fossil family that illuminates the origin of perissodactyls—the group of mammals that includes horses, rhinos and tapirs. It provides insights on the controversial question of where these hoofed animals evolved, concluding that they arose in or near present day India.”