TUBULIDENTATA, Huxley, 1872

Lehmann, Thomas, 2009, Phylogeny and systematics of the Orycteropodidae (Mammalia, Tubulidentata), Zoological Journal of the Linnean Society 155 (3), pp. 649-702 : 650-651

ORDER TUBULIDENTATA

As shown in the introduction, the study of the aardvarks has its roots in the 18 th century. However, Tubulidentata was not recognized as an order on its own until the early 20 th century. Furthermore, most of the studies dealing with fossil aardvarks were limited to the description of one species. Patterson (1975, 1978) is the first author to have suggested intraordinal relationships ( Fig. 1 ). In his review, Patterson split the sole family of the order – the Orycteropodidae Gray, 1821 – into two subfamilies. On one hand, the Orycteropodinae , which consists of the African and Eurasian forms ( Leptorycteropus , Myorycteropus , and Orycteropus Geoffroy St Hilaire, 1796 ), and on the other hand, the Plesiorycteropodinae Patterson (1975), represented by a single genus Plesiorycteropus Filhol, 1895 . The latter taxon is a subfossil from Madagascar known by numerous but fragmentary remains (for instance, no dental remains have been found so far). Following Filhol (1895), Lamberton (1946) and Patterson (1975) suggested that Plesiorycteropus was a relative of the aardvarks. However, MacPhee (1994) showed the ambiguities around the phylogenetic relationships of the genus in a cladistical analysis at a high hierarchical level. He concluded his extensive study ( MacPhee, 1994: 201) and his ‘... recognition of the distinctiveness of Plesiorycteropus by the erection of a new higher taxon to receive it’: the new order Bibymalagasia . Thus, the order Tubulidentata is considered in the present study to comprise only one subfamily: the Orycteropodinae .

Within this subfamily, three genera have been described so far. Myorycteropus is a Kenyan genus from the Early Miocene of Rusinga and Mfwangano. This form is based on an incomplete skeleton of a juvenile individual, and is monospecific. Leptorycteropus is also a Kenyan genus, but from the Late Miocene of Lothagam. Very few specimens of this taxon are known, and they are mostly post-cranial remains. Finally, Orycteropus is known from the Early Miocene up to the present day. The extant aardvark belongs to that genus as well as, supposedly, the oldest known fossil Tubulidentata : Orycteropus minutus Pickford, 1975 . This genus shows the largest biodiversity, with 13 described species widespread from Africa to Eurasia ( Lehmann, 2006a). Although they clearly belong to the Orycteropodinae , the relationships between the genera of this subfamily are still poorly known. For instance, Patterson (1975: 216) speculated that the Orycteropus and Leptorycteropus lineages may ‘... be more closely related to each other than either is to the Myorycteropus one’ ( Fig. 1 ). Moreover, he agreed with MacInnes (1956) that Myorycteropus is too specialized to be involved in the ancestry of Orycteropus . Conversely, van der Made (2003) proposed that Myorycteropus could be the stem group of the two other genera ( Fig. 2 ). This author also suggested that some Orycteropus species should be attributed to Leptorycteropus and Myorycteropus , thereby demonstrating the absence of well-established systematics for the order. Finally, Pickford (1975, 2004, 2005) simplified the question by assigning all Orycteropodinae to the sole genus Orycteropus , without giving substantial evidence to justify this. Taking into consideration all currently available material, the intergeneric relationships are investigated in the present study.

A preliminary cladistic analysis set by Lehmann et al. (2005), mostly based on characters found in the literature, demonstrated that the monophyly of the genus Orycteropus should be re-examined. Moreover, this analysis was the first to clearly show affinities between an African species, Orycteropus abundulafus Lehmann et al., 2005 , and a Eurasian species, O. gaudryi ( Fig. 3 ). van der Made (2003) and Lehmann et al. (2005) performed the only studies that tried to distinguish the relationships between species of Tubulidentata . However, they did not examine the complete fossil record, as it is scattered in many different international institutions, and also because some of the holotypes were inaccessible at the time.

Some fossil specimens not taken into consideration in the present study deserve comments, nonetheless. First, some of the material recovered from the old collection of the Quercy ( France) was previously assigned to Tubulidentata : Palaeorycteropus quercyi, Filhol (1894) , Archaeorycteropus gallicus Amaghino, 1905 , and Leptomanis edwardsi Filhol, 1894 . The two former specimens (two humeri and a fragment of tibia, respectively) show affinities with the pangolins, but are now usually considered to be both indeterminate eutherians ( Patterson, 1975). The latter, a dorsal part of a skull, is usually considered ‘faute de mieux’ as a manid ( Patterson, 1975). However, Thewissen (1985) suggested that Leptomanis could be the oldest tubulidentate known so far. It is not the aim of this study to determine the status of Leptomanis ; a more comprehensive study must be undertaken, as this specimen might speak against the isolated evolution of the Tubulidentata , along with the other Afrotheria , in Africa during the Paleogene. Jepsen (1932) referred a mandible fragment from the Eocene of Wyoming, that he called Tubulodon taylori Jepsen, 1932 , to the order Tubulidentata . However, further discoveries revealed that Tubulodon is rather a palaeanodont ( Gazin, 1952; Simpson, 1959). Finally, Alferez et al. (1988) described fossil remains at Corcolès ( Spain) that they believed were the oldest known Orycteropodidae from Eurasia. However, Pickford & Morales (1998) identified other Spanish remains as belonging to a peculiar lineage of Tayasuidae with tubulidentate microstructure in their cheek teeth roots. They refer the material described by Alferez et al. (1988) to that suiform lineage ‘being its most derived known member in which the roots of the posterior premolars and molars have become fully tubulidentate, while the anterior premolars have retained their enamel cap and “normal” roots’ ( Pickford & Morales, 1998: 286).

Figures

Figure 1. Suggested relationships of the known Tubulidentata genera (modified after Patterson, 1975).

Figure 2. Stratigraphic distribution of different tubulidentates (modified after van der Made, 2003). Localities and their approximate ages in Mya on the left. Thick lines indicate possible ancestor–descendant relationships and arrows indicate possible dispersal events.

Figure 3. Strict consensus tree of eight most parsimonious trees. Tree length = 45, consistency index (CI) = 0.69, and retention index (RI) = 0.60 (modified after Lehmann et al., 2005).