Article: New stem elephant-shrews (Mammalia, Macroscelidea) from the Eocene of Dur At-Talah, Libya

Two new primitive macroscelideans, Eotmantsoius perseverans gen. et sp. nov. and Nementchatherium rathbuni sp. nov. from the late Middle or Late Eocene locality of Dur At-Talah, Libya, are described. Eotmantsoius is known from a single tooth and is characterized by low crown height, isolated conules, and lack of a preprotocrista and prehypocrista. Based on these characters, it appears to be more plesiomorphic than Chambius (late Early Eocene, Chambi, Tunisia), which was, to date, universally accepted as the oldest and most basal macroscelidean. Eotmantsoius appears, however, somewhat derived in having an entostyle, a pericone, a shorter mesial cingulum and larger intercusp valleys. The phylogenetic position of Eotmantsoius is unclear. The second new species, N. rathbuni, allows a reassessment of Nementchatherium. We suggest that it is the most derived of the herodotiines in having long preprotocrista and prehypocrista, reduced M2/m2, a robust entostyle twinned with the hypocone on M1 and the posterior aspect of the root of the maxillary jugal process placed along the length of M2. Nementchatherium rathbuni appears more derived than N. senarhense, from the Eocene locality of Bir El Ater, Algeria, in having smaller m2-3 with no paraconid, a smaller hypoconulid on m2 and a more reduced talonid on m3 (with smaller hypoconid and entoconid). Nementchatherium rathbuni also has, distal to the m3, a coronoid canal. This character has commonly been proposed as a synapomorphy for paenungulates (e.g. proboscideans, sirenians and hyracoids); we investigate the possibility that it unites macroscelideans and paenungulates, as the superordinal clade Afrotheria suggests. Comparisons reveal that a coronoid canal also occurs in several nonafrotherian orders (lagomorphs, xenarthrans, artiodactyls and perissodactyls), suggesting that is not an exclusive synapomorphy for paenungulates plus macroscelideans. Moreover, the homology of the character within paenungulates is doubtful.