Discovery of the smallest Fayum Egyptian primates (Anchomomyini, Adapidae)

Abstract

Two new adapiform primate species from locality 41, Jebel Qatrani Formation, Egypt, are described. The first, here named Wadilemur elegans genus novum species novum (holotype Cairo Geological Museum 42211), consists of a right mandible with P₃–M₃. The second is assigned to Anchomomys milleri species novum, with a holotype Cairo Geological Museum 42842, that includes the right mandible with lower canine to M₃. Both species are allied closely with genera that are known to be from Eocene deposits either in Europe, Tunisia, or the Sultanate of Oman (Arabia), thus enhancing earlier paleomagnetic evidence that locality 41 was deposited in Eocene times.

Collections made in the last 5 or 6 years have yielded more new information about early anthropoidean evolution than finds of any previous decade of the century, and new descriptions have greatly added to knowledge of Paleogene prosimians in Africa. Most of this information has come from the Fayum, Egypt (1–5). There have been other finds further west in North Africa and in the Sultanate of Oman, Arabian peninsula—a biogeographic part of Africa (6). Godinot and Mahboubi (7–9) also described new primate genera and species from Algeria thought to be earliest anthropoideans. Together with their descriptions, these authors clearly disallowed any affinity between the new Eocene anthropoideans or “near anthropoideans” from Africa and the Chinese Eosimias (10)—a point also stressed by Simons (11). Djebelemur from Tunesia was figured and analyzed by Hartenberger and Marandat (12). They considered this find to be an adapid, but others thought it may be an anthropoidean (9). Lastly, Ducrocq et al. (13) described a new possible anthropoidean relative, Wailekia, from the Paleogene of Thailand.

The transition between prosimians and anthropoideans is being narrowed by these new discoveries. There also have been striking advances in understanding other early Tertiary prosimian groups found in the African zoogeographic region. One group of these prosimians, the cercamoniine Adapidae, was first reported from Oman, Arabia, by Gheerbrant et al. (14). One year later, Simons and Rasmussen (15) described mandibles and a remarkable cranium of a rather large prosimian, Plesiopithecus teras, from the Fayum late Eocene locality 41 (L-41). Subsequently, Simons et al. (5) reported on an even larger new genus and species of cercamoniine also from the late Eocene Fayum quarry L-41, Aframonius dieides. In this paper, another new genus and species of a diminutive adapoid found in L-41 is reported, Wadilemur elegans, as well as a new species of Anchomomys, Anchomomys milleri. All of these fossils represent an important addition to understanding the radiations that may have produced the Anthropoidea. The Anchomomys group may be considered either a tribe under the subfamily Ceracmoniinae of the Adapidae (as here) or it may prove to be a separate subfamily of its own.

Most of the newly described species from sites in the Afro-Arabian plate are discussed variously by those who named them as being either adapoid or anthropoidean or, in one case, basal omomyid. Nevertheless, with the exception of Djebelemur, which consists of a five-toothed, mandibular fragment, the classification of these new Afro-Arabian primates is based on isolated teeth, mostly from the lower dentition. The newly described Paleogene African prosimians mentioned above join an enigmatic, earlier Algerian find, Azibius trerki, reported by Sudre (16). In all cases, prosimians found outside the Fayum lack the critical parts of the cranial anatomy that would confirm definite ranking, whether with adapoids or with anthropoideans (3, 17).

The diversity of primates at L-41 rivals any site in the upper sequence of the Fayum, for example Quarry M, where six primate species have been found. The new genus and species here described as well as the new species of Anchomomys makes eight primate genera so far described from L-41. Together, these constitute a distinct, much earlier, and different Fayum lower-sequence fauna in which both the prosimians and probably the anthropoideans represent more diversified primate groups than occur in the Fayum upper sequence.

The age of Fayum site recently was reviewed by Kappleman et al. (18) and Rasmussen et al. (19). In brief, the upper- sequence quarries appear to be about 33 million years ago (Ma) and the L-41 level about 36 Ma, the latter being correlated as Priabonian late Eocene.

Both of the new species described below are clearly related to the Djebelemur–Anchomomys–Omanodon–Schizarodon group. It seems, on the other hand, that the idea that the latter two of these genera are distinct from Anchomomys is not correct. However, it also seems that the described material of Schizarodon is inadequate to sustain a valid generic diagnosis.

As has been noticed by Gheerbrant et al. (13), Schizarodon, Omanodon, and by inference, Anchomomys and Djebelemur bear a strong resemblance to the cheirogaleid lemur Microcebus, as can be seen in Fig. 1 by comparison with a stereophotograph of a Microcebus right mandible. Although the molar resemblances are striking, at least one of the species described below, A. milleri, which preserves the canine, shows that these Paleogene prosimians did not have a tooth comb. The origin of the tooth–comb prosimians is still a mystery.

Figure 1.

Stereopair of right dentary of a mouse lemur (Microcebus murinus). Duke University Primate Center (DPC) osteo.035, crown view. (×11.) Note caniniform P₂ molars with broad talonids and mesially extended entoconids.

SYSTEMATICS

Order Primates Linnaeus, 1758; Suborder Prosimii, Illiger, 1811; Family Adapidae, Trouessart, 1879; Family Cercamoniinae, Gingerich 1977; Tribe Anchomomyini Szalay and Delson, 1979 (sensu Godinot 1988).

Wadilemur, New Genus

Generic diagnosis.

Distinctly larger than Omanodon and slightly larger than Shizarodon, Waldilemur shows typical adapoid lower molars with crescentic ridges connecting the cusps. Lower first and second molars are subequal in size and length, with M₃ longer and showing central hypoconulid—no hypoconulid on M_1-2. Lower molars differ from those of O. minor in that the cristid obliqua does not run up to join the metaconid cusp crown as in both Omanodon and Shizarodon; also, the entoconid of O. minor is shifted further lingually than in Wadilemur, and the high point of the paraconid crest (presumably the actual paraconid) in Omanodon is shifted more laterally than in Wadilemur. Also, in Wadilemur, a continuous crest or ridge runs from the tip of the protoconid around to the metaconid encircling the anterior fovea without a break; in Omanodon, there is a deep, central notch in this crest. Wadilemur differs from Shizarodon in being slightly larger, with lower molar metaconid shifted further anteriorly and with a paraconid cusp located more posteriorly in relation to the protoconid so that the trigonid is less open anterolaterally. In both D. martinezi and Algeripithecus minutus, the cristid obliqua of M₁ runs up into the apex of the metaconid whereas in Wadilemur, this cristid runs into the middle of the wall between the protoconid and metaconid. Algeripithecus also shows a more crenulate talonid with possible traces of a twinned entoconid—characteristics not seen in Wadilemur. Close to Djebelemur but much larger and apparently considerably younger, Wadilemur has lower molars of subequal length that narrow posteriorly whereas, in contrast, Djebelemur has a larger M₁ and a smaller M₃ compared with M₂. In the M₁ and M₂ of Djebelemur, the crest running posteriorly from the hypocone is relatively higher above the basin and has a slightly developed cuspule, which may be the analog of the M₃ hypoconulid, whereas in Wadilemur, this crest is lower and less well defined and has no cuspule. The latter also differs from Djebelemur in having a more mesiodistally compressed trigonid, a comparatively larger M₂ relative to M₁, a relatively smaller M₃ talonid basin, and a more distinctly set off hypoconulid (Fig. 2). In addition, the lower molars of Wadilemur are more “notched” externally between the talonid and trigonid, especially in M_2-3, whereas in Djebelemur, the lateral or buccal outline of these teeth is smoothly rounded.

Figure 2.

Stereopair of W. elegans type CGM Cairo Geological Museum (CGM) 42211. (×11.) Right mandibular fragment with P₃–M₃. Note that the large and lingually directed entoconids are broken off on both M₁ and M₂.

Type species. W.

elegans.

Distribution.

The Fayum, Egypt, quarry L-41.

Etymology.

From Arabic Wadi, meaning a valley or wash, plus Lemur/Lemures L.

W. elegans, New Species

Holotype.

CGM specimen 42211, right mandible with P₃–M₃.

Hypodigm.

Type and DPC 13439 left dentary with P₄–M₃ and partial ascending ramus.

Species diagnosis.

Same as generic diagnosis.

Etymology.

From Latin elegans, meaning neat, elegant.

Type species. W.

elegans.

Anchomomys, Stehlin 1916

Type species.

Anchomomys gaillardi.

Distribution.

Middle Eocene site at Lissieu, France.

Etymology.

From Greek anchi, meaning near, plus Omomys, name of a North American Eocene primate. This name is from Greek omos, meaning shoulder (many other meanings) combined with Greek mys, -mouse.

A. milleri, New Species

Holotype.

CGM 42847, right mandible with complete horizontal ramus, damaged symphysis, and posterior angular and coronoid processes missing. Preserves right canine through M₂.

Species diagnosis. A.

milleri is slightly smaller in size than W. elegans (Table 1), and the molar metaconids do not seem to project quite as far lingually as in that species. It is slightly larger than Schizarodon and much larger than Omanodon. The paraconid crest is relatively a little longer and directed more in an anteroposterior line, so the angle between it and the protocristid is much wider—a resemblance to other species of Anchomomys. As in Wadilemur and Djeblemur, the paraconid crests slightly overlap the tooth in front of them at a point where the hypoconulid would be were it present, and both M₁ and M₂ are slightly notched on the distal end where this crest overhangs the tooth. In A. milleri, a continuous crest joins the protoconid and paraconid of M₁, which are broken in Schizarodon, and in Omanodon, there is no distinct paraconid cusp on this tooth. A. milleri is morphologically close to A. gaillardi from Lisseau, France, with the anteriorly directed paraconid crest (paracristid) also having a similar-sized metaconid. The cristid obliqua of M₁ in these latter two species is directed toward the apex of the metaconid in the same manner, but A. milleri is slightly larger overall. A. milleri is less close to Anchomomys stehlini from Egerkingen, Switzerland. In the latter species, the positions of cristid obliqua and paracristid more closely resemble the condition in Schizarodon and Omanodon.

Table 1.

Measurement of W. elegans and A. milleri

	C	P2	P3	P4	M1tri	M1tal	M2tri	M2tal	M3tri	M3tal
Breadth (bucco-lingual)
W. elegans
CGM 42211			0.8	0.9	1.3	1.4e	1.4	1.5e	1.3	1.2
DPC 13439			0.9	1.1	1.2	1.2	1.2	1.1	0.9	0.9
A. milleri
CGM 42847	0.85	0.7	0.7	0.85	0.9	1.0	1.1	1.1
Length (mesio-distal)
W. elegans
CGM 42211			1.4	1.4	1.7		1.8		1.9
DPC 13439			1.1	1.2	1.5		1.7		2.0
A. milleri
CGM 42847	1.5	1.0	1.0	1.2	1.4		1.5

Etymology.

The species is named for Ellen Miller, who discovered the type and only specimen.

COMPARISON OF THE NEW SPECIES

Premolars of the type specimen of W. elegans are relatively simple, with a single, principal cusp from which a ridge decends both anteriorly and posteriorly along the anteroposterior axis. Lower P₄ is ∼20% larger than P₃, and both show a slightly developed, lingually tilted basin of the heel. Molars increase in length posteriorly (M₁ < M₂ < M₃), but because of greater trigonid–talonid breadth, M₂ is the largest tooth (M₁ < M₃ <M₂). On all three molars, the metaconids are shifted posteriorly relative to the protoconid, and a cristid obliqua runs forward from the hypoconid toward a line between the protoconid and metaconid. There is no paraconid cusp—only a forward-directed paracristid. The entoconids of the type are broken off in both M₁ and M₂ but are preserved in DPC 13439, where they project lingually with a distinct flare. Hence, in outline viewed from above, all three molars have a distinct notch both lingually and labially between the trigonid and talonid. Neither M₁ nor M₂ of CGM 42211 shows any trace of a hypoconulid, but, in M₃, the hypoconulid is present and well developed and is nearly as large as the hypoconid whereas the entoconid is comparatively much reduced. Although the above features can also be determined in DPC 13439, this particular specimen has considerable damage to the P₃ and some breakage in the trigonid of M₁. In this second specimen also, part of the back of the ramus is preserved and seems rather elongated compared with contemporary primates, inasmuch as the length from in front of P₃ to in back of M₃ is 9.5 mm, and the length from the back of M₃ to the back of the mandible between the articular and angular processes is 10.4 mm. From this specimen, the coronoid process is missing, and the head of the articular process is somewhat damaged. The angular process appears to project sharply backward rather than being rounded off, as in contemporary Fayum anthropoids. Also, unlike these Fayum anthropoids, the depth of the mandible is relatively shallow. For example, under the M₁, this depth is 3.3 mm, and trigonid height is 1.5 mm (or M₁ crown height is 45% of mandibular depth). Comparable measurements in Catopithecus browni DPC 15415 are 7.4-mm depth under M₁ compared with 2.4-mm estimated unworn height of the M₁ trigonid; hence, M₁ crown height is 32% of mandibular depth (Fig. 3).

Figure 3.

Stereopair of right dentary of A. milleri type CGM 42847. (×12.) Note the relatively large canine and the anteriorly directed paraconid crest on the molars, which forms a wider angle with the crest between protoconid and metaconid than in W. elegans.

Both specimens of W. elegans are broken off in front of the P₃, so little can be said of the more anterior teeth. In the type, there is a socket for P₂, which may indicate a two-rooted P₂. In the type and only specimen of A. milleri, the P₂ and lower canine are present, but there is no remaining trace of the incisors. At the base of the symphysis, there is evidence of the inferior transverse torus, which fixes the midline and, in turn, indicates that the incisors were small. The lower canine of the type, CGM 42847 is much taller than the P₂ behind it (there is no trace of a P₁), and, because this canine is not at all directed forward like the lateral component of a tooth comb, the specimen gives no evidence that A. milleri was a species having a close ancestral relationship to the strepsirrhines. The numerous and detailed resemblances between the lower postcanine teeth of Microcebus, Wadilemur, and Anchomomys are consequently hard to evaluate. If there is a close phyletic relationship, it would mean that the prosimian tooth comb arose relatively late in time. If these resemblances are coincidental, the degree of similarity between the lower molars and premolars of the present-day lemur and these Paleogene species is paradoxical.

DISCUSSION

At L-41, the largest prosimian is A. diedes, and P. teras is somewhat smaller. W. elegans and A. milleri are both larger than Djebelemur, Omanodon, and Shizarodon, yet they are the smallest Fayum Egyptian primates. Size comparisons can be drawn as follows from the length of P₃ through M₂ in the types of five of these species:

Aframonius diedes: length of P₃–M₂ is 15.1 mm.

Plesiopithecus teras: length of P₃–M₂ is 11.0 mm.

Wadilemur elegans: length of P₃–M₂ is 8.2 mm.

Achomomys milleri: length of P₃–M₂ is 6.8 mm.

Djebelmur martinezi: length of P₃–M₂ is 6.5 mm.

The L-41 primates are perhaps considerably younger than Djebelemur but are likely to be older than Omanodon and Shizarodon. Hartenberger and Marandat (12) remarked that the type mandible of Djebelemur is very similar to that of Pseudoloris from the Eocene of France but did not mention comparisons with living lemurs. The canine they associated with this type specimen is premolar-like and gracile with an apex located in front of the anterior side of the root, so, from the side, the crown looks somewhat like the letter “P.” In addition to the type jaw of Djebelemur and the isolated lower tooth thought to be a canine, there are two referred upper molars from Chambi, Tunisia. If the canine belongs with this species, it is much more gracile, comparatively, than the canine of A. milleri from L-41. Godinot (20) judged that Djebelemur should be ranked with the anthropoideans. The two referred upper teeth of Djebelemur look anthropoidean-like, but there is no real proof that they actually belong to the same species as the type mandible. Also, these isolated molars are not unlike the molars of Plesiopithecus, an undoubted prosimian. These authors (12) interpreted two small alveolae in front of P₃ as being for a two-rooted P₂, which also may be the case in Wadilemur. Whatever other resemblances may eventually be demonstrated, Hartenberger and Marandat (12) concluded that Djebelemur is a cercamoniine adapid with some characters “reminiscent of anthropoids.”

In 1993, Gheerbrant et al. (14) published a description of two new genera and species that they ranked taxonomically in the tribe Anchomomyini. The first of these, O. minor, they declared the smallest adapid. Although these authors were not certain of the meaning of the similarities, they cited over a dozen dental characters held in common between Omanodon and the cheirogaleid lemur Microcebus. (14, pp. 179–180). These authors stated further that, except for the similarities they itemized between these two Omani genera and the cheirogaleids (as well as some with Djebelemur), the two Oman primates belonged with the adapids and suggested a strong connection with the European species of the genus Anchomomys. They regarded the relationship as close enough to imply a trans-Tethyean dispersion to Afro-Arabia. Finally, Gheerbrant et al. (14) stressed that Omanodon and Shizarodon were among the latest known Paleogene adapids perhaps of Priabonian age. This differs from the range of the anchomomyines in Europe, which do not survive beyond the base of the Priabonian age. They suggested that this seemingly late survival, if real, may be due to less extreme climatic variations in the more southerly regions.

Godinot (20) commented on the paper of Gheerbrant et al. (14) and questioned their view that the reduced metaconid and anteriorly extended paracristed on M₁ could be derived from an Anchomomys-like morphology, challenging that the anteriorly extended paracristid of Omanadon could be derived from the shorter anchomomyine M₁ paracristid. Godinot stated that “reduction of the metaconid [of Omanodon and Shizarodon] is not found in adapiformes,” and he urged that the two Oman genera, on the basis of the referred P₄ (Taqah, Oman) 260, showed similarities to both Microcebus and some anthropoideans—“simiiformes” in his usage. He considered the low and reduced metaconid more typical of omomyids, lemuriformes, or anthropoideans. Although admitting that Shizarodon and Omanodon are difficult to interpret, Godinot favored ranking them with the anthropoideans. However, I do not see the resemblance in the upper molars of Omanodon with those oligopithecines. Also, it does not seem that the metaconids of these two Omani prosimians are particularly reduced nor do they seem to be particularly small in Wadilemur or A. milleri. Fig. 4, a lateral view of the type of W. elegans, shows well developed, Anchomomys-like metaconids.

Figure 4.

Lateral view of the type mandible of W. elegans CGM 42211. (×5.6.) Right mandibular fragment with P₃–M₃. Note the well developed metaconids.

All of the differences of opinion discussed above raise two important points about these small primates. The first is that when taxa are based only on single teeth, small groups of isolated teeth, or even on mandibles having fewer than five teeth, the very incompleteness of the material weakens the scientist’s ability to achieve a correct taxonomic allocation. The best of modern practice in paleontology mandates that fossils documented by only one to five isolated teeth should not be named at all. This stricture definitely applies to the type and only species of: Biretia, Tabelia, and Shizarodon, for instance. The situation is not much better with the type species of Algeripithecus and Omanodon. Simons et al. (21) dealt with this same issue more appropriately when they reported a few omomyid teeth and a lorisid tooth from the Fayum but did not name them. The assignment of isolated teeth to the hypodigm of a species is an act of judgment by the scientists concerned. Paleontologists well know that, in the past, there have been a number of cases in which even upper and lower jaws were assigned to different genera. At L-41, for instance, at least eight species of small primates exist. The same could be the case for any of the other Afro-Arabian sites discussed here. Assuming a collection of five upper and five lower isolated teeth from L-41, it would be impossible to make a determination, and any grouping one might arrive at would most likely be fictitious. The full extent of this situation for Afro-Arabian Paleogene primates is shown in Table 2. Taking a particular case as an example, Omanodon is supposed to stand as a genus separate from Anchomomys on the basis of upper molar characteristics detected in the referred, isolated upper teeth from Oman. For the reasons stated, it can be very risky to associate such scattered elements in one taxon.

Table 2.

Numbers of teeth described to date that document seven genera of small Paleogene primates from either North Africa or Arabia

Genus	Description
Beritea pivoti	One tooth
Algeripithecus minutis	Several isolated upper and lower teeth
Tabelia hammadae	Apparently three upper teeth and one lower tooth
Azibius trerki	Lower jaw with three teeth
Diebelemur martinezi	Left mandible with P3–P3 and isolated teeth including two uppers
Omanodon minor	Two lower teeth, parts of five upper teeth, and possibly seven other teeth (some not referred with certainty)
Shizarodon dhofarensis	One and a half lower molars

The second point worth stressing is that, taken altogether, the uncertainty of the various authors discussed above as to whether particular primates are adapiformes or anthropoideans may simply be an indication that, with such fragmentary material, it is not easy to distinguish between those groups we know as adapiformes and early anthropoideans. These similarities, if correct, may of course mean that the two groups are closely related. Authorities have less difficulty separating tarsiids and omomyids and distinguishing them from such forms as Catopithecus and Oligopithecus, about which a wide range of paleontologists have said that they look like adapids. In any case, it is through the “window” of the whole range of Fayum anthropoideans that we must look to see the correct picture of what the relatives and ancestors of early anthropoideans should look like.