The desmostylian ?Behemotops is basal to the Mysticeti (baleen whales)

Backstory:
Traditional paleontology nests desmostyians with proboscidians (elephants) and sirenians. By contrast, the large reptile tree (LRT) nests desmostylians with hippos, anthracobunids, mesonychids and, pertinent to the present discussion, mysticetes (baleen whales) like Balaenoptera, the blue whale (Figs. 1, 2).

Figure 1. ?Behemotops cf. proteus as originally published, after lifting the posterior maxilla, with the addition of a warped Balaenoptera mandible and compared to Balaenoptera.

Figure 1. ?Behemotops cf. proteus as originally published, after lifting the posterior maxilla, with the addition of a warped Balaenoptera mandible and compared to Balaenoptera. Note the holotype of Behemotops, the short, toothy mandible at upper left, would never fit the long rostrum of the cf. specimen. The long and large humerus does not indicate that flippers were present here.

More backstory:
The desmostylian Behemotops proteus (Domning, Ray and McKenna 1986, Beatty and Cockburn 2015; Oligocene, 34-23 mya) was originally described on the basis of short, deep, but largely complete (missing the retroarticular process) and immature (unerupted molars) mandible USNM 244035 (Fig. 1). Another short, deep mandible with a distinctly different architecture and preserving the retroarticular process, USNM 244033 was also also assigned to this genus (Fig. 1), IMHO in error. Neither of these short mandibles appear to fit the long skull of another specimen tentatively assigned to this genus, Behemotops cf. proteus RBCM.EH2007.008.0001 (Fig. 1), which preserves the left half of much of the skull, but no mandible or cranium. An imaginary mandible has been supplied here (Fig. 1) based on phylogenetic bracketing and… it’s a good fit when warped to fit.

We need to come up with new genus
for the RBCM specimen.

This is more evidence
that extant whales are not monophyletic and that desmostylians are not extinct. They live on as baleen whales.

Figure 3. Blue whale (Balaenoptera musculus) skull and skeleton. Note the lack of a thumb goes back to Mesonyx.

Figure 3. Blue whale (Balaenoptera musculus) skull and skeleton. Note the lack of a thumb goes back to Mesonyx.

The resemblance of Behemotops
to the blue whale, Balaenoptera (Figs. 1, 2), is striking. Available post-cranial material for Behemotops (one relatively big humerus in Fig. 1) does not indicate flipper development (Fig. 2).

A phylogenetic fix is needed.
Earlier I reported on Janjucetus, which was considered the most basal mysticete whale by Fitzgerald 2006. Unfortunately revisiting the data indicates that leaf-toothed Janjucetus is now more closely related to Anthracobune, a small taxon known from more plesiomorphic teeth and most of a skull (no post-crania). Even so, with long-legged desmostylians now nesting between Janjucetus and Balaenoptera (Fig. 4), phylogenetic bracketing indicates that, despite their dorsal nares, Janjucetus and Anthracobune both had legs, like desmostylians, not flippers. So Janjucetus is not the most basal mysticete and its very distinct teeth are not precursors or placeholders to baleen.

Figure 5. Subset of the LRT, higher placental mammals with a focus on whales (yellow) and their ancestral clades, the Tenrecidae and Mesonychidae. Both are a fair distance from artiodactyls.

Figure 3. Subset of the LRT, higher placental mammals with a focus on whales (yellow) and their ancestral clades, the Tenrecidae and Mesonychidae. Both are a fair distance from artiodactyls. Note the new nesting of Janjucetus with Anthrobune and Behemotops with Balaenoptera.

References
Beatty BL and Cockburn TC 2015. New insights on the most primitive desmostylian from a partial skeleton of Behemotops (Desmostylia, Mammalia) from Vancouver Island, British Columbia. Journal of Vertebrate Paleontologoy 35(5):e979939: 15 pp.
Cockburn TC and Beatty BL 2009. A Partial Skeleton of Behemotops (Desmostylia, Mammalia) from Vancouver Island, British Columbia. Journal of Vertebrate Paleontology. 29 (3, Supplement): 1A–211A.
Domning DP, Ray, CE and McKenna, MC 1986. Two new Oligocene desmostylians and a discussion of Tethytherian systematics. Smithsonian Contributions to Paleobiology. 59. pp. 1–56.
Fitzgerald EMG 2006. A bizarre new toothed mysticete (Cetacea) from Australia and the early evolution of baleen whales. Proceedings of the Royal Society B 273:2955-2963.

wiki/Janjucetus
wiki/Behemotops

Paleoparadoxia is just a long-legged sea hippo!

Unfortunately,
this appears to be just one more example of paleontologists wearing blinders, evidently waiting for an amateur without access to the fossil itself to reveal the answer to this long-standing enigma/paradox. It just took a little phylogenetic analysis.

According to Wikipedia
Paleoparadoxia tabatai (Reinhart 1959, 2.2m long; Miocene, 20-10 mya) “is a genus of large, herbivorous aquatic mammals that inhabited the northern Pacific coastal region. Originally interpreted as amphibious, Paleoparadoxia is now thought to have been a fully marine mammal like their living relatives, the sirenians, spending most of their lives walking across the sea bottom like marine hippos.”

Paleoparadoxia
(“ancient paradox”) is considered a member of the clade Desmostylia. Wikipedia reports, “The relationship between Desmostylia and the other orders within Tethytheria (elephants, hyraxes, sirenians) has been disputed. That assignment has been seriously undermined by a 2014 cladistic analysis that places anthracobunids and desmostylians, two major groups of putative non-African afrotheres, close to each other within the laurasiatherian order Perissodactyla” (odd-toed ungulates). That’s strange because Paleoparadoxia had four fingers and toes (Fig. 1), like a hippo (Fig. 2).

Carroll (1988) considered the hands and feet specialized as paddles and seal-like in their abilities, but the jaws were similar to those of primitive elephants with the exception that the canines were also tusk-like. He considered desmostylians close to proboscidians, even though hippos have a similar set of tusks that include the canines.

Figure 1. Paleoparadoxia turns out to be a long-legged sea hippo in the large reptile tree.

Figure 1. Paleoparadoxia turns out to be a long-legged sea hippo in the large reptile tree. Note the four fingers and toes on each extremity and those hippo-like tusks. The naris has shifted to the dorsal rostrum and a premaxillary ascending process has redeveloped.

Here
in the large reptile tree Paleoparadoxia nests strongly with Hippopotamus (Fig. 2) and Mesonyx members of the Artiodactylia (even-toed ungulates). I don’t think any prior studies have recovered this pretty obvious relationship, but several prior papers have noted the convergent size, body type and lifestyle. The nares have moved to the dorsal skull on Paleoparadoxia and a premaxillary ascending process reappears. Note the vertical orientation of the pelvis. The elevation of the ankles likely signaled the placement of hippo-like pads beneath the feet, rather than spreading toes creating paddles (contra Carroll 1988).

Figure 2. Hippopotamus, the living sister to the extinct Paleoparadoxia.

Figure 2. Hippopotamus, the living sister to the extinct Paleoparadoxia.

This cladistic nesting
keeps the topology of the large reptile tree simpler than traditional trees.

References
Carroll RL 1988. Vertebrate Paleontology and Evolution. W. H. Freeman and Co. New York.
Reinhart RH 1959.
A review of the Sirenia and Desmostylia. University of California Publications in Geological Sciences 36(1):1–146.