Moeritherium

Moeritherium ("Lake Moeris' beast") is an extinct genus of basal proboscideans from the Eocene of North and West Africa. The first specimen was discovered in strata from the Fayum fossil deposits of Egypt. It was named in 1901 by Charles William Andrews, who suggested that it was an early proboscidean, perhaps ancestral to mastodons, although subsequent workers considered it everything from a relative of manatees to a close relative of both clades' common ancestor. Currently, Moeritherium is seen as a proboscidean that, while fairly basal, diverged before the split between elephantiforms and deinotheres. Seven species have been named, though only three (M. lyonsi, M. gracile, and M. chehbeurameuri), are currently considered valid. The name comes from Lake Moeris, and the Ancient Greek θηρίον (thēríon), meaning "beast".

Moeritherium is unusual even among basal proboscidean standards. Like many later members of the group, it had two sets of tusks: the ones on the upper jaw pointed downwards, while those of the mandible (lower jaw) were flat and formed a spade shape. In addition to these tusks, it retained its upper canines, though had lost the lower set. The morphology of the skull, particularly the nasal cavity (which was only slightly retracted), suggests that Moeritherium lacked a trunk. It may have instead possessed a small, tapir-like proboscis, formed from the fusion of the upper lip and the nose, an evolutionary precursor of trunks. Though poorly described in the literature, Moeritherium's torso is known to have been very long, and its limbs were short. These divergent traits have led to comparisons with desmostylians, a lineage of extinct mammals formerly believed to have been relatives of manatees.

Moeritherium has been suggested to have led a semi-aquatic lifestyle. While this originally stemmed from perceived similarities to sirenians (manatees and dugongs), morphological data and isotope analysis has since lent it a great deal of support. The elongated body of Moeritherium, and the high position of its eyes and ears, are likely a result of its lifestyle, and its unusual dentition is likely an adaptation for feeding on water plants.

The type species of Moeritherium, M. lyonsi, was discovered in strata belonging to the Qasr el Sagha Formation in the Fayum fossil deposits of Egypt. The type specimen (CGM C.10000) consists of an almost complete mandible. It was described in 1901 by Charles William Andrews, who proposed two hypotheses for its phylogenetic position: either Moeritherium was part of the obsolete order Amblypoda, or it was an early proboscidean, perhaps "a generalised forerunner of the Mastodon type". In any case, he regarded it as an ungulate.

In 1902, after conducting a more thorough examination of specimens collected by himself and his colleague, Hugh John Llewellyn Beadnell, he named a second species from the Qasr el Sagha, M. gracile; a third was recognised in the same paper, though he did not provide a name, and referred to it simply as M. sp. The two species were distinguished from M. lyonsi by a more gracile build and a larger body size respectively. The lack of material overlap has made it difficult to determine how M. gracile actually relates to M. lyonsi, as their holotypes consist of different skull elements; the type specimen of the former (CGM C.10003) is a palate with no associated lower teeth. Regardless, they are treated as belonging to the same genus, and are likely separate. Two years later, a fourth taxon, M. trigodon, was described, also by Andrews, based on remains recovered from the "fluvio-marine beds" (equivalent to the Jebel Qatrani Formation) around the lake Birket-el-Qurun. In 1955, over half a century after the genus' initial naming, Sri Lankan artist and palaeontologist Paulus Edward Pieris Deraniyagala named two additional species, P. latidens and P. pharaonensis, based on isolated mandibular fragments.

In 1911, German zoologist Max Schlosser divided M. lyonsi into two species: M. lyonsi, restricted to the Qasr el Sagha Formation, and M. andrewsi, restricted to the Jebel Qatrani. This classification, however, has been rejected. In 1971, German zoologist Heinz Tobien opted to synonymise the entire genus with M. lyonsi, though he chose to altogether disregard, Deraniyagala's species, likely as they were poorly diagnostic. In 2006, Cyrile Delmer et al. published a paper describing a new Moeritherium species, M. chehbeurameuri, from Bir El Ater, Algeria. In their paper, they treated most of the above species (with the exception of M. latidens and M. pharaonensis) as valid. While the paper was not intended as a systematic revision, they chose to treat at the very least three species as valid: the type species M. lyonsi, M. gracile, and M. chehbeaurameuri.

Henry Fairfield Osborn, in 1909, suggested that Moeritherium was more similar to sirenians (manatees and dugongs, and their extinct kin) to any living or extinct proboscidean. In 1921, however, he rejected this view, and divided Proboscidea into four suborders or superfamilies: Moeritherioidea, Deinotherioidea, Mastodontoidea, and Elephantoidea. In a 1988 paper discussing the systematics of proboscideans, Pascal Tassy abandoned this system and neglected to provide any superfamily-rank clades. Erecting the suborder Elephantiformes, Tassy placed Moeritherium outside it, alongside Barytherium, Numidotherium, and the Deinotheriidae. He considered Moeritherium among the most basal proboscideans, with Numidotherium being the most basal and Barytherium being only slightly less basal than that. In a 2021 paper describing a new genus (Dagbatitherium tassyi) Lionel Hautier et al. ran a phylogenetic analysis which recovered Moeritherium as sister to a clade including deinotheres and elephantiforms.

A cladogram of Proboscidea based on the phylogenetic analysis of Hautier et al. 2021 is below: