Index Introduction Paleophylogeography and Primate Dispersals Conclusions Acknowledgments References

Introduction

The first discoveries of primates in Africa date back to the beginning of the 20th century. Osborn (1908), then Schlosser (1911), described the first specimens from the famous Oligocene Fayum deposits of Egypt. The great interest in these Paleogene African primates arose only at the end of the 1950s when new fossil discoveries revealed their affinities with higher primates (Simons, 1959, 1961). Since then, numerous field expeditions to the Fayum led by Elwyn Simons have substantially improved our knowledge of the evolutionary history of African primates, notably that of early anthropoids. To date, the late Eocene–early Oligocene Fayum deposits have yielded more than 20 distinct primate forms, including anthropoids and prosimians (see Simons, 1995, 1998 for an overview). The high degree of anthropoid diversity in Egypt at the end of the Paleogene and the fact that some of the Fayum taxa may represent stem lineages of modern anthropoids have led several authors to regard Africa as the ancestral homeland of the anthropoid clade. This scenario was strengthened by the discovery of Oligocene anthropoids, closely related to the Egyptian taxa, in the Sultanate of Oman in the Arabian peninsula (Thomas et al., 1999), and by the discovery of primitive anthropoids in the early–middle Eocene of Algeria (de Bonis et al., 1988; Godinot and Mahboubi, 1992). Analyzing the paleontological evidence along with paleogeographic data and intercontinental faunal resemblances, Holroyd and Mass (1994) suggested that an endemic African origin of anthropoid primates was the most likely scenario.

Paradoxically, despite the suggested endemism of Paleogene African fauna (such as Proboscidea and Hyracoidea), subsequent fossil discoveries in the Paleogene of the Maghreb (1980–1990s) have extended the paleo-distributions of some of the North Tethyan mammalian groups to Africa (e.g. Mahboubi et al., 1997). Such widespread distribution of certain taxa undoubtedly indicates terrestrial faunal interchanges between regions, demonstrating in turn that the Arabo-African continent was not strictly isolated in the Paleogene. Recent discoveries of Eocene primates in East Asia (e.g. Beard et al., 1994, 1996; Jaeger et al., 1998, 1999), notably anthropoids (e.g. Chaimanee et al., 1997, 2000a, b; Takai et al., 2001, 2003; Shigehara et al., 2002; Marivaux et al., 2003), and discoveries of adapids with holarctic affinities in the Fayum (e.g. Simons et al., 1995; Simons, 1997, 1998) and Oman (Gheerbrant et al., 1993) challenge the ‘orthodox’ view of African anthropoid endemism.

In the following discussion, we consider three significant Paleogene events: (1) the origin of the order Primates in the Paleocene or earlier; (2) the origin of anthropoids in the early Eocene; and (3) the origin of catarrhines in the middle Eocene. Each of these events has been the subject of long-standing debate. To date, in the light of available paleontological evidence, Africa and East Asia are the landmasses most often considered in the various scenarios. The purpose of this paper is to assess the current state of knowledge regarding mammalian faunal interchanges between Africa and Eurasia during the early Paleogene as a means of analyzing temporal constraints on anthropoid dispersals between Asia and Africa (Figure 1).

View Details

Figure 1. Mammalian paleobiogeographic relationships between Africa and Eurasia during the early Paleogene. Three phases of faunal exchange are depicted, related with scenarios concerning the origin of (a) primates during the Paleocene/Eocene boundary, (b) anthropoids during the early Eocene, and (c) catarrhines during the middle Eocene (see the text for details). The question marks and dotted lines indicate that phylogenetic and biogeographic relationships are conjectural, and the double arrow indicates unknown direction of migration.

Paleophylogeography and Primate Dispersals

The geographic origin of the order Primates and its higher-category phylogenetic relations are subject to ongoing debate. The oldest known Asian primates are Teilhardina and Altanius from the early Eocene of China (Ni et al., 2004) and Mongolia (Gingerich et al., 1991), respectively. Both genera are positioned near the root of the primate radiation. Rose and Krause (1984), however, regarded Altanius as a probable plesiadapiform, although Altanius does not unequivocally exhibit the developed incisors diagnostic of the group. A plesiadapiform status was also advocated by Hooker et al. (1999) for the oldest known African primate, Altiatlasius, from the late Paleocene of Morocco (Sigé et al., 1990). Moreover, even if some authorities consider Altiatlasius as the oldest true primate (‘euprimate’), the affinities of this taxon within the order remain uncertain; it has been originally considered as an early omomyid (Sigé et al., 1990), more cautiously as a basal primate (Gingerich, 1990), and subsequently as seemingly the closest out-group of anthropoids (Godinot, 1994). If Altiatlasius represents the oldest known primate in the fossil record, the possibility of an African origin for the order and a subsequent northward dispersal at the Paleocene–Eocene boundary could be envisaged. In contrast, Beard (1998) hypothesized that an Asian origin is more parsimonious because the closest out-group of Primates (i.e. Plesiadapiformes, Scandentia, and Dermoptera) originated in the holarctic, notably in Asia.

Both African and Asian early Paleogene primate fossils undoubtedly indicate a trans-Tethyan distribution of primates at around the Paleocene–Eocene boundary or even earlier. Dispersal events are substantiated by various groups of placental mammals. For instance, the late Paleocene Moroccan fauna from the Adrar M’Gorn 1 locality has yielded, in addition to Altiatlasius, a family incertae sedis represented by the genus Tinerhodon and some possible miacid carnivores and hyaenodontid creodonts of North Tethyan affinities (Gheerbrant, 1995), suggesting a northward mammalian dispersal from Africa to Eurasia at around the Paleocene–Eocene boundary. However, because of the fragmentary nature of the Moroccan material and the undoubted presence of a hyaenodontid creodont in the late Paleocene of China (Meng et al., 1998), immigration from Asia to Africa seems more likely. Furthermore, a southward dispersal from Europe to Africa was proposed by Tabuce et al. (2001) to explain phylogenetic relationships between the early Eocene African Macroscelidea and the Paleocene European louisinine condylarthrans. From the foregoing, it is obvious that faunal exchange between Africa and the North Tethyan continents, including primates, occurred at the Paleocene to early Eocene interval. Nevertheless an understanding of the polarity of such dispersal events requires more paleontological evidence than is currently available.

The second event, the origin of Anthropoidea, can also be discussed from a paleophylogeographical perspective. On the basis of fossil evidence, the ‘Afrocentrist model’ would suggest that all Paleogene anthropoids descended from a common African ancestor. The earliest anthropoids found on this continent are known from the early–middle Eocene of Algeria and Tunisia. According to Godinot and Mahboubi (1994), Algeripithecus, Tabelia, and two other indeterminate anthropoids from the Glib Zegdou locality (early–middle Eocene, Algeria), as well as Djebelemur from Chambi (early–middle Eocene, Tunisia), would be the earliest members of that group. Nevertheless, because of the fragmentary nature of the fossil remains, generally limited to isolated teeth, no consensus has been established concerning the phylogenetic relationships of these small-bodied primates. For instance, Hartenberger and Marandat (1992) and most researchers argue for an adapid status for Djebelemur, whereas Godinot (1994) suggests that Algeripithecus and the upper teeth from Chambi (named cf. Djebelemur) are related to Parapithecidae, Tabelia to the catarrhine Propliopithecidae, and the lower jaw from Chambi (Djebelemur) to Oligopithecidae. Godinot’s point of view would extend all anthropoid families represented in the Oligocene of Egypt and Oman back to the early–middle Eocene chronostratigraphic range. The above dental evidence is, however, tenuous, and the earliest well-established members of the anthropoid families are known only from the middle–late Eocene of Bir El Ater (Algeria); Biretia, described as a possible primitive catarrhine (de Bonis et al., 1988; R. Tabuce, in progress), and an unpublished oligopithecid.

The African-origin hypothesis of anthropoids was recently challenged by discoveries in East and South Asia: the Eosimiidae from the middle Eocene of China and Myanmar (e.g. Beard et al., 1994; Jaeger et al., 1999), additional Amphipithecidae from Myanmar (e.g. Chaimanee et al., 2000a; Takai et al., 2001; Marivaux et al., 2003) and from the late Eocene of Thailand (e.g. Chaimanee et al., 1997, 2000b; Ducrocq 1999a). The eosimiids are considered as one of the first offshoots of the anthropoid clade (Beard et al., 1996), having diverged before the African oligopithecids (Kay et al., 1997, 2004). Although eosimiids and amphipithecids are currently unknown outside of Asia, if Asia was the ancestral homeland of the anthropoid clade (the ‘Asiocentric model’), the Maghreb middle Eocene anthropoids (parapithecids and oligopithecids) might be descended from an early Asian anthropoid ancestral form. However, oligopithecids and parapithecids were not direct descendants of eosimiids, as they appear to be paraphyletic (Kay et al., 1997). The facts are that oligopithecids first occurred in the early–middle Eocene or more probably in the middle–late Eocene, while eosimiids first appeared in the middle Eocene. Whatever the phylogenetic relationships between these basal Asian and African anthropoid primates might be, the hypothesis of an Asian anthropoid origin implies a dispersal from Asia toward Africa during the early Eocene.

However, the possibility of such an early dispersal event of anthropoid primates from Asia to Africa is not clearly documented by other mammalian groups. In fact, only faunal exchanges between Africa and Europe have been evidenced for this period. For example, immigration from Europe is considered for some African marsupials (Crochet et al., 1992). The same can be said of Djebelemur if this is really an African representative of cercamoniine adapid (Hartenberger and Marandat, 1992), the latter known in Europe since the beginning of the Eocene. Alternatively, some African mammals could have immigrated to Europe. Sigé (1991) and Hartenberger et al. (1997) considered a possible African origin for Chiroptera of the European middle Eocene.

A putative dispersal event between Asia and Africa during the early Eocene could be evoked from interpretations regarding the geographical origin of the early–middle Eocene Zegdoumyidae of North Africa. These peculiar rodents are clearly distinct from all known European Paleogene rodents, but in contrast show significant affinities with the North American Eocene Sciuravidae (Vianey-Liaud et al., 1994; Marivaux et al., 2004). With the absence of sciuravid rodents in the Paleogene of Europe to date, a dispersal route from North America via Europe is highly unlikely. On the other hand, a migration route via Asia might be suspected since some early Eocene rodents from Asia, especially those from China, have been tentatively referred to as sciuravid (Li, 1963; Shevyreva, 1971; Wang and Li, 1990). However, the status of some of these rodents has been revised (Dawson, 1977; Wood, 1977; Korth, 1994; Dawson et al., 2003a) and the proposed Asian hypothesis requires better paleontological support—the dispersal history of the zegdoumyid rodents remain unresolved.

Concerning the middle Eocene, the fossil record of both South Asia and North Africa is substantially more adequate for understanding patterns of mammalian faunal exchange between the two continental masses. Ducrocq (2001) hypothesized that the middle Eocene Asian amphipithecids could be the sister group of the African propliopithecid catarrhines of the early Oligocene. Even if phylogenetic relations between these two primate groups are not in consensus (e.g. Kay et al., 2004), a middle Eocene Asian origin of catarrhines can be discussed from a paleophylogeographical point of view. In the early–middle Eocene of Africa, no stem catarrhine has been firmly identified, although Tabelia and Biretia are potential candidates (de Bonis et al., 1988; Godinot, 1994). On the other hand, the occurrence of late Eocene catarrhines in Africa is still disputed since oligopithecids from the Fayum (L41 locality) are viewed either as stem catarrhines (e.g. Seiffert and Simons, 2001) or basal anthropoids (Kay et al., 2004). Consequently, if amphipithecids are catarrhines (Ducrocq, 2001), a middle Eocene immigration event from Asia could be considered for African catarrhines. The catarrhine status of amphipithecids remains controversial, however, and more fossils of these enigmatic primates are needed. A competing hypothesis argues that catarrhines originated in Afro-Arabia during the Paleogene, and were restricted to this zoogeographic province until the early Miocene (Harrison and Gu, 1999).

In any case, during the middle Eocene, several instances of faunal exchange between Asia and Africa have been demonstrated. For instance, Ducrocq (1997, 1999b) claimed that anthracotheriid artiodactyls of the Fayum have their origin in Asian Eocene forms. A recent discovery of an anthracotheriid at Bir El Ater (Mahboubi et al., 2003), showing stronger affinities with Eocene Asian forms than with the advanced Oligocene Egyptian fossils, constrains the immigration event of anthracotheriids into Africa at least back to the middle Eocene. The Bir El Ater site has also yielded rodents, Protophiomys and Nementchamys (Jaeger et al., 1985), that strongly support such dispersal events. Protophiomys is considered as the only African representative of the well-diversified hystricognathous ‘baluchimyine’ rodents of South Asia (Marivaux et al., 2000, 2002; Marivaux and Welcomme, 2003). Moreover, the anomaluroid Nementchamys displays unusual affinities with a recently discovered new rodent from the late middle Eocene of Myanmar (Dawson et al., 2003b; Marivaux et al., 2004).

Conclusions

The Paleogene fossil record of Eurasia and Africa testifies to several mammalian dispersal events between the continental landmasses. Even if such faunal exchanges are evident for several modern placental orders (such as chiropterans, carnivores, rodents, and ungulates), the primate paleobiogeographic history is still conjectural because of the lack of phylogenetic resolution of some groups, notably anthropoids. As a consequence, the various and competing paleobiogeographic scenarios proposed for primates are all speculative. The polarities of the suggested dispersals for primates between Africa and Asia are unresolved to date. Such migrations must have concerned the whole circumference of the South Tethys, which during the early Tertiary, extended from southern Asia (Indian subcontinent, Myanmar, and Thailand) to northwestern Africa. These circum-Tethyan landmasses, which correspond to low-latitude regions, certainly had rather stable ecosystems with minor climatic and ecological changes that conditioned the emergence and the development of anthropoid primates.

Acknowledgments

The authors would like to greatly thank Professor Shigehara, Dr Huffman, Dr Takai, Dr Hongo, and Dr Kunimatsu (Primate Research Institute of the Kyoto University, Japan) for their kind invitation to the international symposium “Asian Paleoprimatology: Evolution of the Tertiary Primates in Asia.” This is publication no. 2004-010 of the Institut des Sciences de l’Évolution de Montpellier (ISEM, CNRS—UMR 5554).