a new pliocene xerine sciurid (rodentia) from kossom ... - BioOne

3 downloads 41 Views 1MB Size Report
terrestrial squirrel belonging to the genus Xerus was unearthed at the Pliocene site of Kossom Bougoudi in Chad. Xerus daamsi, sp. nov. is characterized by a ...
Journal of Vertebrate Paleontology 23(3):676–687, September 2003 q 2003 by the Society of Vertebrate Paleontology

A NEW PLIOCENE XERINE SCIURID (RODENTIA) FROM KOSSOM BOUGOUDI, CHAD CHRISTIANE DENYS1, LAURENT VIRIOT2, REMMERT DAAMS3,5, PABLO PELAEZ-CAMPOMANES3, PATRICK VIGNAUD2, LIKIUS ANDOSSA4, and MICHEL BRUNET2 1Laboratoire Mammife ` res et Oiseaux, Muse´um National d’Histoire Naturelle, 55 rue Buffon, 75005 Paris, France, [email protected]; 2 Laboratoire de Ge´obiologie Biochronologie et Pale´ontologie Humaine, CNRS UMR 6046, Universite´ de Poitiers, 40 avenue du Recteur Pineau, 86022 Poitiers Cedex, France; 3 Departamento de Paleobiologı´a, Museo Nacional de Ciencias Naturales (C.S.I.C.), Jose´ Gutierrez Abascal 2, Madrid 28006, Espana; 4 Universite´ de N’Djamena, BP 1117, N’Djamena, Tchad; 5 Deceased

ABSTRACT—A very well preserved, incomplete, articulated skeleton with nearly complete skull and mandible of a terrestrial squirrel belonging to the genus Xerus was unearthed at the Pliocene site of Kossom Bougoudi in Chad. Xerus daamsi, sp. nov. is characterized by a narrow nasal associated with medium size. The phylogenetic position of the new species among African Sciuridae was determined using cladistic analysis of craniodental characters. It is most similar to extant Xerus rutilus and Xerus erythropus, currently found in Ethiopia and Chad, respectively. Cladistic analysis also supports the monophyly of the African members of the tribe Xerini, and a sister group relationship between X. daamsi and X. rutilus whose position within the Xerini is poorly supported. The North African genus Atlantoxerus is valid and distinct from Xerus, emphasizing the faunal differences between North African and sub-Saharan regions of the continent. The close affinities of Xerus daamsi with xerines which are currently living in northern savannas, more precisely in the Horn of Africa, suggests the presence of the Somali-Masai vegetation type in Chad by 5 Ma.

INTRODUCTION A number of Pliocene sites in Chad in Central Africa are well known because they have produced the first australopithecines west of the Rift Valley (Brunet et al., 1995, 1996, 1997). In the area south of Djourab Erg, the MPFT (Mission Pale´oanthropologique Franco-Tchadienne) has discovered many sites with abundant, well preserved vertebrate fossils (Brunet et al., 1998). Kossom Bougoudi (KB) produces one of the richest faunas. It is dated by biochronology close to the Miocene-Pliocene boundary, and has yielded a well preserved skeleton and skull of a sciurid (Brunet and MPFT, 2000). This specimen is described here for the first time and compared with East African fossil forms and extant relatives. Taphonomic and phylogenetic analyses are made to test several biogeographic and paleoecological hypotheses. Ethiopian ground squirrels of the tribe Xerini are not uncommon in East African Plio-Pleistocene deposits. Although these sciurids are known from the early Miocene in East and North Africa, this is the oldest find in Central Africa. The new record from Chad raises interesting questions about their biogeographic distribution around 5 to 4 Ma and about within-tribe phylogenetic relationships especially between modern and fossil species. REVIEW OF THE AFRICAN XERINI RECORD Moore (1959) described the tribe Xerini as being composed of three genera: Xerus, Atlantoxerus, and Spermophilopsis. The last is the curious ‘‘prairie dog’’ of Eastern European deserts, whose phylogenetic relationships are still unclear. Xerus is found today only in tropical Africa, while Atlantoxerus is a monotypic genus from North Africa. Sciurids are well known from Miocene deposits of Europe, the Siwalik Range (Pakistan), and North Africa. They are known in East Africa only by Vulcanisciurus (Lavocat and Mein, 1973) found at Songhor, Koru, Napak, and Fort Ternan

(Kenya). Senut et al. (1992) also recorded Vulcanisciurus from Namibia (Berg Aukas sites 47 and 63 dated at 14 Ma and Harasib 3a dated at 10.5 Ma). No sciurids are known from late Miocene or Pliocene sites in Namibia, nor from Langebaanweg (South Africa) at about 5 Ma (Hendey, 1981). Munthe (1987) reported Atlantoxerus cf. getulus from the Pliocene of Sahabi, Libya. Currently, the oldest record of Xerus is from the 3.7 Ma Laetolil Beds (Denys, 1987). Xerus cf. janenschi and Xerus sp. were both reported from the Laetolil and Upper Ndolanya beds (Denys, 1987). Wesselman (1984) attributed teeth from Omo sites members B and C (3 to 2.5 Ma) to Xerus erythropus and others from Omo member F to Xerus sp. For Olduvai Bed I, Denys (1990) reported the oldest occurrence of Xerus cf. inauris. Xerine sciurids are absent from South African Plio-Pleistocene cave sites (Denys, 1990) and the Humpata deposits in Angola (Pickford et al., 1992). Cuenca Bescos (1988) described the sciurids from the Catalayud Basin (early Miocene of Spain) and included two more genera in the tribe Xerini (Aragoxerus and Heteroxerus) on the basis of dental characters and by comparison with Atlantoxerus. Xerus is a diurnal, terrestrial sciurid found in the SaheloSudanian, Guinean, Somali-Masai, and Zambezian savannas of tropical Africa and in primary forest regions, but is absent from the mountain forest zone (above 1,500 m). It is also present in the temperate subtropical grassland of the highveld in South Africa and in the Kalahari-SW arid vegetation zone, but is absent from the Namib and the Cape regions (Denys, 1999). Four extant species are currently recognized (Hoffmann et al., 1993): Xerus rutilus Cretzschmar, 1826; Xerus erythropus Geoffroy, 1803; Xerus inauris Zimmermann, 1780; and Xerus princeps Thomas, 1929. MATERIALS AND METHODS Terminology for Xerus molars follows Cuenca (1986). Measurements of the skulls were made with calipers with a preci-

676

DENYS ET AL.—NEW XERUS FROM CHAD sion of 0.01 mm. Images were digitized using a CCD camera and MTV 1.3. Phylogenetic analysis were performed with PAUP 3.1 and tree branch support was estimated using the decay index of Bremer (1994). SYSTEMATIC PALEONTOLOGY Order RODENTIA Bowdich, 1821 Family SCIURIDAE Fischer de Waldheim, 1817 Subfamily SCIURINAE Fischer de Waldheim, 1817 Tribe XERINI Murray, 1866 Genus XERUS Hemprich and Ehrenberg, 1832 XERUS DAAMSI, sp. nov. Holotype Specimen KB03-97-162, an incomplete, articulated skeleton consisting of a skull with mandible, two distal humeri, proximal radii and ulnae, a single caudal vertebra, a partial pelvis, and two nearly complete posterior limbs without metapodials. Type Locality Site 3 of Kossom Bougoudi, south of Djourab Erg, northern Chad (168 199 N, 188 429 E). Age and Horizon Early Pliocene, close to the MiocenePliocene boundary (Brunet and MPFT, 2000); collected from green sandstone. Repository The specimen is housed in the Centre National d’Appui a` la Recherche, N’Djamena, Chad. Etymology Named in honor of Remmert Daams (who died suddenly in 1999), former eminent specialist on fossil rodents at the Museo Nacional de Ciencias Naturales, Madrid. Remmert found the holotype at the height of a sandstorm during the 1997 field trip. Diagnosis Xerus of medium size with bunodont teeth lacking visible roots; anterior cingulum linked to M2 protocone; no anterolophid on lower molars. Xerus daamsi differs from X. rutilus in having a P3, a distinct hypocone on M3, more alternating and unfused protoconids and metaconids, and a more prominent mastoid. The talonid of m3 is clearly separate from the rest of the tooth and has three distinct cusps, unlike X. rutilus. X. daamsi differs from Xerus erythropus in smaller size, and by having an anterior cingulum that is not connected to the protocone on M1, a stout mesostyle, a hypocone and protocone not connected by a loph on M3, and a longitudinal entoconid on m3. The lack of any connection between the anterior cingulum and protocone on M1, a loph connecting protocone and hypocone on M3, an anterolophid, and an ectolophid on m3, the separation of the protoconid and metaconid, and the presence of a hypoconulid on m3 are among the features that differentiate X. daamsi from Atlantoxerus getulus. Xerus daamsi differs from Xerus inauris and Xerus princeps in smaller size, lack of interparietal and loph connecting protocone and hypocone on M3, and in having a P3, a bunodont hypocone, and a hypoconulid. X. daamsi differs from Xerus janenschi by the presence of a P3, a hypoconid on m3, and a crest-shaped entoconid on m2–m3, and the absence of a protocone-anteroloph connection on the upper molars and a metaloph. DESCRIPTION OF THE KOSSOM BOUGOUDI SPECIMEN Skeletal Preservation and Taphonomy The Kossom Bougoudi deposits form a regular pattern of sand-clay sequences sometimes with discontinuous layers of diatomite. The skeleton was found in a weakly consolidated bed of fine-to-medium grit. The sequences very probably reflect episodic climatic variations during which ephemeral stream and lake environments alternated. The adult skull was nearly complete except for the left zygomatic arch and bulla (Fig. 1A, B). The mandible was attached

677

to the skull but not in its original anatomical position because the upper and lower incisors were not in contact. To study the dental morphology, the mandible was separated from the skull. The left side of the cranium is poorly preserved as the skull was asymmetrically compressed, meaning only the right side is truly complete although slightly deformed. Bone sutures are generally clearly visible except in the frontonasal and orbital region. The frontal has numerous fine cracks. The nasal is slightly upturned and overrides the frontal, the front of which is depressed. The front of the right zygomatic plate is broken and points outward. The subzygomatic sulcus is compressed on the right side toward the inside of the skull. Fine striations are observed on the dorsal surface where the bone looks smooth and polished. The right tympanic bulla is almost complete. The enamel of the molars bears traces of iron or manganese, but these encrustations are not extensive, and the bone surface is mainly brown with gray patches. Molar enamel shows no sign of deterioration. The incisors have some slight transverse cracks but the enamel and dentine have not been affected by digestion. On the right parietal crest (Fig. 1A, arrow) is a circular depression with a 4.17 mm diameter hole in its center that could be a puncture mark made by a mammalian carnivore (canid or felid). A second poorly preserved puncture mark is found at the base of the left tympanic bulla and could have been made by a predator picking up the skull from behind to carry the carcass to its den. Xerus is not a common prey for owls because of its large size and diurnal habits but is captured by small carnivorans and diurnal raptors (Kingdon, 1997). Except for the hind limbs, skull, and elbow joints, the skeleton is incompletely preserved. The ribs, vertebrae (all but one), scapulae, metapodials, and phalanges are missing. All of these belong to less dense parts that can be easily destroyed or transported (Behrensmeyer, 1975). The limb bones are in articulation as are those of the rear of the foot. The right femur has a root trace in the middle of the diaphysis (Fig. 1D, arrow). All of the long bones have longitudinal cracks and desquamation marks, possibly caused by weathering (Fig. 1D–F). The spongy structure of the bone, the surface of which has a patina, can be seen at the bone joints. In the absence of any marks of digestion on the teeth and in view of the large number of articulated bones, it seems more likely that the surface deterioration was caused by intense modern eolian erosion that currently characterizes this part of the Djourab Desert rather than by weathering or digestion. Some of the deformation of the skull may be related to the bite marks. Compression of the gritty sediment may also have contributed to crushing of the skull, although very few fossils from the site exhibit such deformation. The fact that portions of the postcranial skeleton were found with the skull is an argument in favor of transport of the entire body into a cache where it was abandoned. Another explanation might be that the animal became buried in a burrow during flooding or drought, with low energy flooding removing just the lighter bones. Skull The skull (Fig. 1A, B) is slightly compressed on its left side, preventing accurate measurement of its zygomatic and frontal width. Nevertheless, it is characterized by a narrow skull with a small braincase. Other measurements indicate the new species had a medium-sized skull compared with other xerines (Table 1). In dorsal view, the nasal is broken anteriorly but is narrow and does not widen posteriorly. The postorbital process of the frontal is well preserved on the right side of the cranium and there is no indication of an interparietal. In lateral view, the region of the auditory foramen has a well developed mastoid process and very prominent parieto-occipital crests, the upper incisors are broken but their orientation indicate a typical opisthodont disposition. The sharp, blade-like anterior edge of the zygoma overhangs the zygomatic plate so far as to create a deep

678

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 23, NO. 3, 2003

FIGURE 1. Xerus daamsi, sp. nov., holotype, KB03-97-162. A, dorsal view of the skull (black arrow indicates a possible puncture mark made by a mammalian carnivore). B, ventral view of the skull. C, vestibular view of the left dentary. D, medial view of the right hind limb showing a broken pelvis and a complete femur in articulation with the tibia (the black arrow indicates a root trace). E, medial view of the broken distal part of the femur (black arrow) in connection with a complete tibia and the proximal fragment of the fibula. F, anterior view of the left proximal femur. Scale bars equal 1 cm.

DENYS ET AL.—NEW XERUS FROM CHAD

679

TABLE 1. Skull dimensions of modern and fossil representatives of the tribe Xerini. Abbreviations: LGT, total length of the skull; LNAS, nasal length; WNAS, nasal width at the largest point; LBT, greatest length of tympanic bulla; LP4–M3, length of the upper tooth row; Lp4– m3, length of the lower tooth row; Sd, Standard deviation; Min, Minimum value; Max, Maximum value; n, number of specimens. Species

LGT

LNAS

WNAS

LBT

LP4–M3

Lp4–m3

Xerus daamsi, sp. nov.

54.59

15.64

7.06

13.17

11.62

11.08

Xerus erythropus Mean Sd Min Max

59.13 3.26 52.11 65.06

18.07 1.21 16.38 20.33

7.91 0.6 6.86 8.99

13.3 0.75 11.99 14.52

11.84 0.73 10.69 13.27

12.01 0.46 10.92 12.68

Xerus rutilus Mean Sd Min Max

51.07 2.48 45.44 56.36

13.26 5.59 12.86 17.64

6.74 0.5 6 7.72

12.14 0.7 11.11 14.08

9.39 0.48 8.27 10.32

9.65 0.42 8.71 10.5

Xerus inauris Mean Sd Min Max

55.65 2.86 51.91 58.67

17.35 1.31 15.47 18.97

8.44 0.62 7.72 9.45

13 0.57 12.17 13.85

10.97 0.32 10.58 11.33

11.89 0.4 11.14 12.24

Xerus princeps Mean Sd Min Max

58.5 1.87 56.6 60.75

19.89 1.56 17.82 21.19

7.74 0.2 7.44 7.88

13.72 0.57 13.03 14.36

11.16 0.48 10.57 11.69

12.01 0.45 11.52 12.6

Xerus janenschi Berlin Gadj. 45 Berlin Gadj. 100 Laetolil 18/4875

53.4 44.8 53

16.4 — —

7 — —

12.8 13.2 12.7

11.4 12.2 11.4

— — —

fossa. The lacrimal region is impossible to see on the lateral and dorsal sides of the skull because of deformation. The squamosal extends up to the base of the postorbital process of the frontal. In ventral view, the tympanic bullae are rather small and the right bullae show traces of three transbullar septa. The masseteric knob is well developed. The tooth rows are parallel and the bony palate extends beyond the end of the molars for a distance at least equaling the anteroposterior width between the M3s. Dentition Upper Dentition The new species is characterized by the presence of a minute P3 and molars with bunodont, well individualized cusps (Fig. 2), but very low crowns. The P3 is bicuspid. The P4 has neither anterior nor posterior cingulae, but has a large parastyle and a small posterostyle. The metacone and hypocone are distinct, and there is no metaloph. The hypocone and protocone on the P4 are not linked by a loph. On the M1, a large parastyle is extended by an anterior cingulum that terminates at the base of the protocone without being connected to it. The protoloph is transversely oriented but not strongly connected with the protocone. The protocone and hypocone are connected by a loph which continues distally as a short posteroloph that ends in the middle part of the posterior wall. There is no metaloph, but a large metaconule occurs between the highly bunodont hypocone and metacone. Small mesostyle and posterostyle are present. The M2 is very similar to the M1, but its anteroloph weakly joins with the protocone and is more crest-like. The protoloph is nearly complete, but narrows near the paracone. There is no metaloph, but a small intermediate cusp lies near the metacone. There is a mesostyle and a longer posterior cingulum than in M1, but no posterostyle. The M3 is relatively large posteriorly because of a prominent hypocone and small metacone. Posterior cingulum and mesostyle are absent. The mesiolingual corner of the M3 is

occupied by a large protocone that is isolated from the hypocone. A long, rather crescentic anteroloph extends from the protocone and occupies the cingular anterior margin of the tooth, ending at the anterolabial corner. The protocone is connected to the paracone by an oblique protoloph. Lower Dentition The incisors are smooth. The p4 protoconid and metaconid are very close together (Fig. 2B). The metaconid is twice as large as the protoconid, and it alone occupies the anterior part of the p4. The hypoconid is connected to a small entoconid by a very low cingulum. The m1 has a typically xerine morphology with a longitudinally oriented protoconid. The metaconid is oriented transversely and is twice as high as the protoconid. A small crest runs anteriorly from the protoconid toward the base of the crown. Although the cusps are clearly abraded, there is no trace of an ectolophid. The posterolophid is connected to a crest-shaped entoconid by an oblique loph on which a small cusp develops, here termed the ‘entoconulid.’ As with all sciurids, the occlusal morphology of the m2 is very similar to that of the m1. However, in contrast to the m1, the m2 entoconulid is isolated in the central sulcus. The m3 is long and narrow. The protoconid describes a crescent shape in a much more oblique position than in the preceding teeth. A small cusp occupies an anteromedian position between the protoconid and metaconid. The hypoconid is also very oblique and ends at the back of the tooth in a small longitudinal sulcus. On the other side of the sulcus is a clearly individualized hypoconulid, itself separated from the entoconid by another small oblique sulcus. The entoconid forms a longitudinal crest that is the posterolingual wall of the crown, but remains separated from the metaconid by a valley. An unobtrusive mesostylid lies at the front of the entoconid crest. The entoconulid is very small and forms a ridge within the posterolingual wall. Postcranial Skeleton Forelimbs The right and left forelimbs are represented by the distal ends of the humeri articulated with the proximal ends

680

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 23, NO. 3, 2003

FIGURE 2. Left upper (A) and lower (B) cheek tooth rows of Xerus daamsi, sp. nov., holotype, KB03-97-162. Scale bar equals 2 mm; anterior to top.

of the radii and ulnae. The connection appears to be natural, corresponding to anatomical position. Left Hind Limb The proximal part of the femur is well preserved (Fig. 1F). The broken distal part of this femur is in connection with a complete tibia and the proximal fragment of the fibula (Fig. 1E). Parts of the pes are connected to the distal end of the tibia (calcaneum, astragalus, cuboid, navicular, and the three cuneiforms). The end of the tibia forms an angle of 1288 with the long axis of the calcaneum, which appears to be in a normal position. The distal end of the tibia has an extended internal malleolus beside a deep longitudinal sulcus. Right Hind Limb The femur is continuous with a broken pelvis but they are not fully articulated (Fig. 1D). Relative to the femoral head, which fits into the acetabulum of the left pelvis, the pubic symphysis is found facing posteriorly, as if the pelvis had pivoted through 1808. The femur is complete and articulated with the tibia and probably with the fibula, which is in matrix. The calcaneum, astragalus and navicular are found in continuity with the distal end of the tibia. COMPARISON WITH MODERN AND FOSSIL XERINI Xerus daamsi exhibits some of the skull characters that define the Xerini sensu Moore (1959): the squamosal extends to the base of the postorbital process of the frontal; three transbullar tympanic septae in each auditory bulla; a bony palate extending posteriorly beyond the ends of maxillary tooth rows; opisthodont upper incisors; and a thick, prominent masseteric tubercle. However, Xerus daamsi is characterized by a short nasal that is not broadened anteriorly and by relatively large size (Table 1).

FIGURE 3. Scatterplots showing differences in size proportions among the compared species of Xerini (Xerus rutilus: filled diamonds, Xerus erythropus: filled circles, Xerus daamsi KB03-97-162: filled triangle, Xerus inauris: open squares, and Xerus princeps: stars). In A (total skull length versus nasal length), Xerus daamsi fits with the largest individuals of Xerus rutilus. In B (length of the tympanic bullae versus length of upper tooth row), Xerus daamsi plots among Xerus erythropus.

Although the amount of intraspecific variation is great, on average modern Xerus erythropus skulls are larger than those of Xerus rutilus. X. daamsi is of intermediate size between those two species, but towards the upper end of the range for modern X. rutilus (Fig. 3). Xerus inauris and Xerus princeps are also intermediate in size, but distinguished by a very wide nasal. X. daamsi has bunodont teeth and a hypocone and entolophid on m1, which are dental characters of the Xerini noted by Cuenca Bescos (1988). But X. daamsi does not have a metaloph, while Cuenca Bescos (1988) noted the presence of a complete to incomplete metaloph on M1 or M2 in the definition of the Xerini. Comparison with Atlantoxerus getulus Although Moore (1959) voiced doubts about the position of Atlantoxerus within the Xerini and it is impossible to verify on our fossil the key character of the jugolacrimal suture differ-

DENYS ET AL.—NEW XERUS FROM CHAD

FIGURE 4. Upper (top) and lower (bottom) cheek tooth rows of Atlantoxerus getulus specimens at different wear stages (A, NHM 21.5.3015, slight wear; B, MNHN CG 1974-259, moderate wear; C, MNHN CG 1959-145, heavy wear). Scale bar equals 2 mm.

entiating Atlantoxerus from Xerus, we prefer to include this genus in the comparison. Atlantoxerus getulus of Morocco has a smaller cranium, and smaller teeth and limbs than X. daamsi. An interparietal can be seen clearly at the back of the skull of A. getulus. The braincase is wide and the occipital region not inflated. Above the auditory foramen, the ectotympanic-parietal suture forms a right angle and then rises to follow the outline of the tympanic bulla. The mastoid process is posterior to the auditory foramen, but less developed than in X. daamsi, as is the masseteric tubercle. The number of transbullar septa varies from three to four. The molar crowns are lower in A. getulus at the same stage of wear. P4 has no hypocone (Fig. 4). The metaconid and protoconid of p4 are of the same size and in the same alignment at the front of the tooth. The lower molars have no entoconulid, but the trace of an ectolophid can be detected on m2 and m3 even at very slight wear stages. There is also a clearly individualized mesostylid. The limbs are more robust and longer. There is no trace of the longitudinal sulcus at the distal end of the tibia as observed in X. daamsi. Resemblances between the two species include the presence of a P3, a me-

681

FIGURE 5. Upper (top) and lower (bottom) cheek tooth rows of Xerus rutilus specimens at different wear stages (A, MNHN CG 1978-254; B, MNHN CG 1974-2; C, NHM 9.6.1.16). Scale bar equals 2 mm.

sostyle on the upper molars, and the absence of a metaloph on the upper molars even at stages of advanced wear. Comparison with Modern Xerus rutilus The skull of Xerus rutilus is shorter than that of X. daamsi (Table 1). Some specimens of X. rutilus exhibit no traces of an interparietal. There is no swelling of the mastoid process in the zone lateral to the auditory foramen. The teeth are smaller and the P3 is absent (Fig. 5). The crowns of Xerus daamsi are taller than in X. rutilus. The molars are very variable in X. rutilus. In a tooth row where P4 is erupting, the lophs and cusps are arranged much as in X. daamsi, with, in particular, a small posteroloph and no metaloph. There is a significant difference in the M3, which in X. rutilus has a crest-shaped lateral wall ending at the back of the tooth with no trace of posteroloph nor separate hypocone, whereas in X. daamsi this is not such a lateral continuous wall. An anteroloph and a protoloph extend from the protocone in both species, but the protoloph is more crest-shaped in X. rutilus than in X. daamsi. On the other upper

682

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 23, NO. 3, 2003

FIGURE 7. Upper (A–B) and lower (C–D) cheek tooth rows of Xerus inauris specimens at different wear stages (A, TM 5395; C, TM 25631; B and D, TM 25632). Scale bar equals 5 mm.

FIGURE 6. Upper (A–D) and lower (E–H) cheek tooth rows of Xerus erythropus specimens at different wear stages (A and E, MNHN CG 1973-170; B and F, MNHN CG 1904-1985; C and G, MNHN CG 1986-582; D and H, MNHN CG 1904-1985). Scale bar equals 2 mm.

molars of X. rutilus, the cusps become broader and bunodont at more advanced stages, and in modern specimens a posteroloph occupies the entire posterior wall of the molar, in contrast to X. daamsi where it is very short (Fig. 5). The anterior cusps of the lower molars seem more closely spaced in X. daamsi. A very slight transverse connection can be seen as the metaconid is much taller in X. rutilus, there being no sign of this connection in X. daamsi. The m3 of X. daamsi has a crest-shaped entoconid oriented longitudinally, a feature found only in very young specimens of X. rutilus. As with the upper molars, the back of the m3 of X. rutilus forms an unbroken high crest with poorly individualized cusps whereas in X. daamsi, the cusps are more distinct although they are joined. No ectolophid is found in X. rutilus or X. daamsi. X. daamsi has longer and wider limb bones than X. rutilus. Comparison with Xerus erythropus The skull has the same general aspect in X. erythropus and X. daamsi, in particular its narrow, elongate shape and the absence of an interparietal with the parietal crests meeting in the middle of the occipital suture. The tympanic bullae are larger

in X. erythropus. At the back of the auditory foramen, X. erythropus and X. daamsi have a small swelling at the end of the occipital crest, however it is clearly more swollen in the latter (Fig. 3). The molars are larger and more hypsodont in X. erythropus (Fig. 6, Table 1). P3 is found in most specimens of X. erythropus, but is monocuspid. X. daamsi has a similar molar arrangement to that of a X. erythropus from Niger (MNHN CG 1986-582, Fig. 6), although the latter has a much more marked posterior cingulum and a nearly formed metaloph on the M2. However, the metaloph is not developed in other specimens of X. erythropus. The M3 of X. erythropus from Niger, which appears a little less worn than that of X. daamsi, has a double hypocone with a posterior cingulum and a large metacone; whereas the metacone is very small and there is no posterior cingulum in X. daamsi. On the lower molars, the size of the hypoconid in X. daamsi is striking, occupying as it does the rear half of the m3. It is much smaller in X. erythropus where the three cusps of the talonid are of similar size and are interconnected. However, they never form a high wall bounding a deep basin and connecting the talonid to the front of the tooth as in X. rutilus. X. erythropus also tends to have many accessory tubercles. In X. erythropus, the entoloph is clearly visible, and there is a well developed mesostyle not found in X. daamsi. For the m3, a longitudinal, crest-shaped entoconid, a protoconid, and a hypoconulid are clearly visible in Xerus daamsi as in X. erythropus. There is no ectolophid, nor any transverse connection between the anterior cusps of any of the molars in X. erythropus. On the upper molars, however, there are clear traces of lophs with a complete metaloph. Comparisons with Xerus inauris and Xerus princeps There are very few differences between extant Xerus inauris and Xerus princeps (Denys, 1987). The skull of X. inauris is slightly smaller or of similar size to that of Xerus daamsi, but narrower. At least four septae are found in the tympanic bulla. There is also an interparietal and a slight mastoid process. In contrast, X. princeps is larger than X. daamsi with no interparietal, a slight mastoid process, and a tympanic bulla with three septae (Fig. 7, Table 1). However, the molar crowns are taller in X. princeps than in X. daamsi. No ectolophid is found in X. inauris and X. daamsi, whereas this feature is reported in two specimens of X. princeps. The m3 of both X. princeps and X. inauris is like that of X. daamsi in having a prominent hypo-

0.17

— — 2.77 3.00 3.50 3.40 3.70

3.30 — — —

2.66

683

— 2.60

3.57 3.35 3.70 0.13

— — 3.50 3.55 3.45 3.60 0.09 — —

3.78 — — — 3.47 3.17 3.77 0.12

2.97 3.17

2.65 2.50

3.63 3.45 3.80 0.15 3.51 3.30 3.80 0.19

— — 3.60 3.70

3.32 2.92 3.72 0.16 3.21 2.96 3.46 0.30

2.76 2.9

2.50 —

3.48 3.30 3.65 0.14 3.07 2.05 3.45 0.51

3.12 3.46 2.76 2.88

3.13 2.99 3.26 0.16 2.68 2.3 3.06 0.36

2.62 2.31

L W

— —

3.13 2.30 3.40 0.42 — — — —

2.50 2.60 — —

2.76 2.38 3.14 0.36 3.10 — — —

3.05

2.41

p4 L W

— —

— — — — — — — —

— — — —

2.83 — — — 3.55 — — —

3.05

2.76

M3 L W

— —

— — — — 3.36 — — —

— — — —

3.08 — — — 3.72 3.08 4.37 0.26

2.86

2.76

M2 L W

— — — Laet 1562 Xerus sp.

3.49 3.40 3.60 0.08 — 2.80 — — — 2.75 — —

— — — 2.98

3.08 2.39 3.78 0.28 2.65 1.66 3.64 0.62 3.13 2.49 3.77 0.40

2.86 2.80 2.50

L

M1

— 2.55

On drawings of X. erythropus from Omo, Wesselman (1984: fig. 55) depicted lower molars with an anterolophid joining the protoconid (not present in Xerus daamsi). The entoconid of the Omo specimens is transverse and not oblique. The posterolophid is long, but the trace of the hypoconulid is visible, which is not true for X. daamsi. X. erythropus from Omo is different

Xerus erythropus Omo 18 members B & C Xerus sp. Omo 33 Xerus cf. inauris Olduvai Mean Min Max Sd

Comparison with Xerus erythropus from Omo B and C

Xerus daamsi, nov. sp. Xerus janenschi UNB-Laetolil Mean Min Max Sd

Xerus cf. inauris from Olduvai Bed I is characterized by tall crowns and a high loph connects the poorly differentiated protocone and hypocone (Denys, 1990). The posteroloph is longer than that of X. daamsi. Talonid cusps of the m3 are arranged differently, with a well developed entoconid oriented transversely to the longest axis of the molar, whereas in X. daamsi it is less developed and oriented longitudinally along the lingual wall of the tooth. This character clearly aligns the Olduvai fossils with Xerus inauris of South Africa. The hypoconid is very prominent and (except on m1) the trace of the hypoconulid, observed only on the left m3 of X. daamsi, is no longer visible. A trace of an ectolophid is found on the worn molars from Olduvai Bed I.

W

Comparison with Xerus cf. inauris from Olduvai

P4

Xerus daamsi has smaller molars than X. janenschi (Table 2). The skulls are about the same size, but the braincase and tympanic bullae are wider in X. daamsi (Table 1). One of the two skulls found at Laetolil (Denys, 1987:figs. 6–7) exhibits no trace of an interparietal and the two parietal crests join in the middle of the occipital crest. There is no P3. As in X. daamsi, the molars attributed to X. janenschi have no ectolophid and the crown height of M1 is similar. The entoconid of X. janenschi is less crestiform, but both have similar orientations. On the upper molars, the anteroloph and a prominent metaloph are well developed in X. janenschi, and the M3 has a trace of an ectoloph.

L

Comparison with Xerus janenschi from Upper Ndolanya Beds at Laetolil

Species

conid and a crestiform but oblique entoconid with an entoconulid. In X. daamsi, the entoconid is more longitudinal and there is a strong hypoconulid while only slightly worn specimens of X. inauris still have a hypoconulid.

TABLE 2.

FIGURE 8. Upper (A–B) and Lower (C–D) cheek tooth rows of Xerus princeps specimens at different wear stages (A and C, TM 38117; B and D, TM 11527). Scale bar equals 5 mm.

Dental measurements of some fossil Xerini. Same abbreviations as in Table 1.

m1

W

L

m2

W

L

m3

W

DENYS ET AL.—NEW XERUS FROM CHAD

684

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 23, NO. 3, 2003

from X. daamsi in having an anterolophid and a transverse entoconid. The modern specimens of X. erythropus we examined have no anterolophid on the lower molars, unlike X. rutilus, and their entoconid is clearly connected to the hypoconid, and is slightly oblique. Comparison with Vulcanisciurus Vulcanisciurus was described from lower Miocene deposits of Rusinga (Lavocat and Mein, 1973:239) and the middle Miocene site of Fort Ternan (Denys and Jaeger, 1992). The lower molars have a well developed, longitudinal, crestiform entoconid and a clearly distinct posterolophid. The ectolophid is well developed on all the molars. An anterolophid is found on the m3 but not a hypoconulid. For the upper cheek teeth, Lavocat and Mein (1973) reported a bicuspid P3 and a P4 with a very prominent metaconule and an incipient hypocone. A protoloph and metaloph join with the protocone. Only one M1 is preserved. Vulcanisciurus teeth bear a marked resemblance to those of Xerus rutilus and Atlantoxerus getulus, but nothing is known of their variability. X. daamsi appears to be clearly more bunodont than Vulcanisciurus. With its bunodonty, small hypocone, and cusp arrangement of its lower molars, Vulcanisciurus shares some characters of tribe Xerini sensu Cuenca Bescos (1988). But Vulcanisciurus does not have an entolophid on the m1 or m2, nor a metaloph on the single M1 figured by Lavocat and Mein (1973). PHYLOGENETIC ANALYSIS This analysis includes all extant species of African Xerini. Moore (1959) suggested that African arboreal squirrels evolved from ‘‘ground squirrel stock’’ on the basis of long orbits and short interorbital breadths of the Protoxerini and Funisciurini. Accordingly, this analysis includes representatives of the Protoxerini (Heliosciurus, Epixerus, Protoxerus), Funambulini, Myosciurini, and Funisciurus of the tribe Funisciurini. Sciurus is used as an outgroup. Fossil species represented only by poorly preserved dental material have not been incorporated in the data matrix. Jaeger (1977) reported that Atlantoxerus from the Plio-Pleistocene of Morocco has a smaller anteroconid, wider lower molars, and a larger metaconule than early Heteroxerus. Lavocat and Mein (1973) proposed affinities between Heteroxerus and Atlantoxerus as did Cuenca (1986, 1988) who defined different trends in the evolution of Heteroxerus and Aragoxerus lineages from Spain including enlarged cheek teeth, increased bunodonty, and a slight increase in hypsodonty. Trees and Phylogeographic Hypotheses The first analysis includes only the African Xerini species with Sciurus as an outgroup and using only characters C1 to C17 (Appendices 1 and 2). All 17 characters were considered unordered and of equal weight. A heuristic search based on the criterion of maximum parsimony obtained a single tree with length of 32 steps, consistency index of 0.625, and retention index of 0.52 (Fig. 9A). All the species of the Xerus genus form a monophyletic clade defined by three synapomorphies (C7 bunodonty; C14 nasal not extended; C17 mastoid process marked, but in X. rutilus its absence is interpreted as a reversal). Xerus daamsi is in the same clade as X. rutilus and the two share three apomorphies concerning bunodonty (C4, C6) and m3 hypoconulid differentiation (C13). X. rutilus has four autapomorphies (C1, C12, C16, C17). X. erythropus, X. inauris, and X. princeps form a clade united by two synapomorphies, absence of a mesostyle (C5) and arrangement of the entoconid (C10). The position of Atlantoxerus getulus is not resolved in this analysis. Most of the clades obtained in the most parsi-

monious tree are not well supported and have a decay index of only 1 except the X. inauris–X. princeps clade which has a decay index of 4. When analyzing all the African Sciuridae using the same 17 characters plus four more skull characters adapted from Moore (1959) and taking Sciurus as outgroup, we obtained 6 most parsimonious trees of length 71, with consistency index of 0.394, and retention index of 0.566. The positions of Funambulus and Myosciurus are ambiguous (Fig. 9B). The strict consensus yields a Rohlf consistency index of 0.808 (Fig. 9B), and Xerus daamsi is grouped in 100% of the trees with X. rutilus, while X. erythropus is basal to the monophyletic Xerus clade. The 50% majority rule consensus indicates that only in 33% of the cases is the daamsi-rutilus clade grouped with the inaurisprinceps clade due to relative instability of the position of the daamsi-rutilus clade and a low decay index of 1 for those two nodes. This tree shows a probable monophyly of the Xerini (sensu Moore, 1959) with 5 synapomorphies and a decay index of 2. The African tree squirrels forming the tribe Funambulini sensu Simpson (1945) is paraphyletic as is the Protoxerini. Myosciurus is not close to Funisciurus, contrary to the results of Moore (1959). It is the only African squirrel with two transbullar septae. This may indicate that terrestriality in African squirrels occurred only once and that African ground squirrels could be derived from arboreal squirrels contrary to Moore’s (1959) hypothesis. SYSTEMATICS AND EVOLUTION OF AFRICAN PLIO-PLEISTOCENE XERINI The systematics and the evolution of the African members of the Xerini remains very unclear without further analyses of other anatomical and molecular characters. The clade consisting of all Xerus species shows a similar general trend toward the development of bunodonty as reported by Cuenca (1986) for species of Heteroxerus in Europe. The North African terrestrial squirrel Atlantoxerus is basalmost among the Xerini (Fig. 9) and may have evolved independently of those of tropical Africa. Its divergence would date to about 7 to 8 Ma in North Africa according to the fossil record and this may correlate with a major modification of climate with the onset of the Asian monsoon and the dominance of C4 plants (grass). Then, by vicariance or by dispersal, the isolation of tropical Africa led to speciation among Xerus, first in West Africa (X. erythropus) and later in East and South Africa. The development of bunodonty as well as the crestiform arrangement of the back of the M3 may be an adaptation to a more granivorous diet. Although X. daamsi shares morphological features with both X. rutilus and X. erythropus, both cladistic analyses determined that X. daamsi has more synapomorphies with X. rutilus. PALEOECOLOGICAL AND PALEOBIOLOGICAL IMPLICATIONS Hubert (1978) reported modern Xerus rutilus in the Omo Valley area of Ethiopia on plateau covered by open shrub or Acacia bush savanna well away from the river. Similarly, Coe (1972) reported Xerus in the transition zone between thicket and grassland. Generally, Xerus is found in dry savannas (300 to 900 mm annual rainfall; Kingdon, 1974), and is essentially a bushland, burrowing animal. It is omnivorous and diurnal. X. erythropus has been recorded in Bandia (Senegal) in the Sudanian zone in a woodland with 570 mm isohyet (Hubert, 1977) and in southeastern Chad in bushes in a Sudanian savanna with 700 mm rainfall (pers. obs.). Despite a few similarities between the fauna of KB and those from North (Sahabi), East (Lothagham, Aramis), and South (Langebaanweg) Africa, this site displays rather marked provincialism (Brunet and MPFT, 2000), and Xerus daamsi could be a new indicator of its endemism.

DENYS ET AL.—NEW XERUS FROM CHAD

685

FIGURE 9. Phylogenetic affinities of Xerus daamsi, sp. nov. with other African sciurids. A. Xerini phylogeny, showing X. daamsi close to X. rutilus. Most parsimonious tree for the partial matrix including African Xerini, with Sciurus as outgroup (32 steps, CI 0.625, RI 0.520). Only synapomorphies are shown here together with the state of the character in parentheses on the cladogram (1: C6(1), 2: C13(0), 4: C7(1), 5: C14(1), 6: C17(0), 7: C2(2), 8: C12(2), 9: C15(2), 11: C5(1), 12: C10(1)). X. rutilus and X. princeps are respectively characterized by one autapomorphy (3: C17(1), 10: C10(0)) but not X. daamsi. B, African sciurid phylogeny (total matrix) with Sciurus as outgroup. Strict consensus of 6 most parsimonious trees (71 steps, CI 0.394, RI 0.566, Rohlf index 0.808). Note that Xerus daamsi and X. rutilus belong to the same clade but its position is not well defined among Xerus.

Today, Xerus belongs to the domain of the Sudano-Guinean, Somali Masai, and Zambezian savannas (Denys, 1999). X. daamsi is clearly different from the Tanzanian lineage of X. inauris, and seems closer to Xerus rutilus presently found only today in the Horn of Africa in the Somali-Masai zone. However, X. rutilus is currently not found in Chad, instead X. erythropus is present. It is primarily an animal of open woodland and sudanic savannas, living in more arid regions than X. rutilus. However, the two species are sympatric over parts of Uganda, Sudan, and Kenya. Their natural foods are roots, grass seeds, storage leaf bases of some grass species, green leaves, fallen fruits, nuts, Acacia pods, and some insects, especially termites (Kingdon, 1997).

speciation in the different savannas of Africa from a common ancestor in North Africa in late Miocene times. Xerus daamsi shares some dental and cranial characters with X. janenschi from Laetolil and X. erythropus from Omo, indicating some relationships from a common ancestor with an early divergence. However, little is known of the variability of fossil species because of the paucity of the fossil record. Due to the absence of fossils of this genus at other sites of the same age, there are no possibilities of biochronologic correlations and thus no new information about the age of the KB site. Further discoveries are required to improve our understanding of the evolution of this group of rodents in the Pliocene of Africa. ACKNOWLEDGMENTS

CONCLUSIONS The close phylogenetic affinities of Xerus daamsi, sp. nov. and extant Xerus rutilus suggests that Somali Masai elements of vegetation were present at the KB site about 5 Ma. Moreover, differences between Xerus daamsi and the two modern species from South Africa supports the hypothesis that this area of Chad was biogeographically separated from southern savannas by 5 Ma. But connections with Ethiopia are confirmed and more comparisons need to be made with rodent fossils from Hadar and Aramis. Unfortunately, Xerus is currently not yet known for these sites. The African modern Xerini, which form a monophyletic clade, but with low support, may have evolved by allopatric

We are indebted to the Chad authorities (Ministe`re de l’Education Nationale, de l’Enseignement Supe´rieur et de la Recherche), to the French Ministe`re de l’Enseignement Supe´rieur, de la Recherche et de la Technologie, to the French Ministe`re des Affaires Etrange`res (Coope´ration: MCAC N’Djame´na), to Re´gion Poitou-Charente, De´partement de la Vienne, Groupe Elf, Association pour le Prix scientifique Philip Morris, to the French armed forces, and to all the troops of MAM and Epervier, who, through their logistical support, contributed along with members of MPFT to the success of field programs. We are extremely grateful to Paula Jenckins (Natural History Museum, London) and Ducan MacFadyan (Transvaal Museum, Pretoria) for access to comparative material of mod-

686

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 23, NO. 3, 2003

ern Xerini. We would also like to thank Alain Beauvillain for field help, Guy Mouchelin for skeleton preparation, Sabine Riffaut for drawings and photographic plate mounting, Ghislaine Florent for administrative guidance of MPFT, and Christopher Sutcliffe for critical reading of this manuscript. LITERATURE CITED Behrensmeyer, A. K. 1975. The taphonomy and paleoecology of PioPleistocene vertebrate assemblages east of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology 146:473–578. Black, C. C. 1965. New species of Heteroxerus (Rodentia, Sciuridae) in the French Tertiary. Verhandlungen der Naturforschenden Gesellschaft, Basel 76:185–196. Bremer, K. 1994. Branch support and tree stability. Cladistics 10:295– 304. Brunet, M., A. Beauvilain, Y. Coppens, E. Heintz, A. H. E. Moutaye, and D. Pilbeam. 1995. The first australopithecine 2,500 kilometers west of the Rift Valley (Chad). Nature 378:273–275. ———, ———, ———, ———, ———, and ——— 1996. Australopithecus bahrelghazali, une nouvelle espe`ce d’Hominide´ ancien de la re´gion de Koro Toro (Tchad). Comptes Rendus de l’Acade´mie des Sciences, Paris 322:907–913. ———, ———, D. Geraads, F. Guy, M. Kasser, H. T. Mackaye, L. MacLatchy, G. Mouchelin, J. Sudre, and P. Vignaud. 1997. Tchad: un nouveau site a` Hominide´s Plioce`ne. Comptes Rendus de l’Acade´mie des Sciences, Paris 324:341–345. ———, ———, ———, ———, ———, ———, ———, ———, ———, and ——— 1998. Tchad: de´couverte d’une faune de Mammife`res du Plioce`ne infe´rieur. Comptes Rendus de l’Acade´mie des Sciences, Paris 326:153–158. ———, and MPFT (Mission Pale´oanthropologique Franco-Tchadienne). 2000. Chad: discovery of a vertebrate fauna close to the Mio-Pliocene boundary. Journal of Vertebrate Paleontology 20: 205–209. Coe, M. J. 1972. The South Turkana expedition IX. Ecological studies of the small mammals of South Turkana. Geographical Journal 138: 316–338. Cuenca, G. 1986. Heteroxerus insignis, n. sp. (Sciuridae, Rodentia, Mammalia) from the lower Miocene of Spain. Casopis pro Mineralogii a Geologii 31:131–142. Cuenca Bescos, G. 1988. Revision de los Sciuridae del Aragoniense y del Rambliense en la fosa de Catalayud-Montalban. Scripta Geologica 87:1–116. Denys, C. 1987. Fossil rodents (other than Pedetidae) from Laetolil; pp. 118–170 in M. D. Leakey and J. M. Harris (eds.), Laetolil, a Pliocene Site in Tanzania. Oxford University Press, London. ——— 1990. First occurrence of Xerus cf. inauris (Rodentia, Sciuridae) at Olduvai Bed I (lower Pleistocene, Tanzania). Pala¨ontologische Zeitschrift 64:359–365. ——— 1999. Of mice and men. Evolution in East and South Africa during Plio-Pleistocene times; pp. 226–252 in T. Bromage and F. Schrenk (eds.), African Biogeography, Climate Change and Human Evolution. The Human Evolution series, Oxford University Press. ———, and J. J. Jaeger. 1992. Rodents of the Miocene site of Fort Ternan (Kenya) first part: phiomyids, bathyergids, sciurids and anomalurids. Neues Jahrbuch fu¨r Geologie und Pala¨ontologie Abhandlungen 185:63–84. Hendey, Q. B. 1981. Paleoecology of the late Tertiary fossil occurrences in ‘E’ Quarry, Langebaanweg, and a reinterpretation of their geological context. Annals of the South African Museum 1:1–104. Hoffmann, R. S., C. G. Anderson, R. W. Thorington, and L. R. Heaney. 1993. Family Sciuridae; pp. 419–465 in D. E. Wilson and D. A. Reeder (eds.), Mammal Species of the World. A Taxonomic and Geographic Reference. Smithsonian Institution Press, Washington and London. Hubert, B. 1977. Ecologie des populations de rongeurs de Bandia (Se´ne´gal), en zone sahe´lo-soudanienne. La Terre et la Vie 31:33–100. ——— 1978. Modern rodent fauna of the Lower Omo Valley, Ethiopia. Bulletin of the Carnegie Museum of Natural History 6:109–112. Jaeger, J. J. 1977. Les rongeurs du Mioce`ne moyen et supe´rieur du Maghreb. Paleovertebrata 8:1–166. Kingdon, J. 1974. East African Mammals. An Atlas of Evolution in Africa, Vol. II, part B: Hares and Rodents. Academic Press, London and New York, pp. 343–703.

——— 1997. The Kingdon Field Guide to African Mammals. Academic Press, London and New York, 464 pp. Lavocat, R., and P. Mein. 1973. Sous-Ordre Sciuromorpha; pp. 239– 243 in Les Rongeurs du Mioce`ne d’Afrique Orientale. I. Mioce`ne infe´rieur. Me´moires et Travaux de l’Ecole Pratique des Hautes Etudes, Institut de Montpellier 1:1–284. Moore, J. C. 1959. Relationships among the living squirrels of the Sciurinae. Bulletin of the American Museum of Natural History 118: 153–206. Munthe, J. 1987. Small-mammal fossils from the Pliocene Sahabi formation of Libya; pp. 135–144 in N. T. Boaz, A. El-Arnauti, A. W. Gaziry, J. de Heinzelin, and D. Dechant Boaz (eds.), Neogene Paleontology and Geology of Sahabi. A. R. Liss, Inc., New York. Pickford, M., P. Mein, and B. Senut. 1992. Primate bearing Plio-Pleistocene cave deposits of Humpata, Southern Angola. Human Evolution 7:17–33. Senut, B., M. Pickford, P. Mein, G. Conroy, and J. Van Couvering. 1992. Discovery of 12 new Late Cenozoic fossiliferous sites in palaeokarsts of the Otavi Mountains, Namibia. Comptes Rendus de l’ Acade´mie des Sciences, Paris 314:727–733. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bulletin of the American Museum of Natural History 85:1–350. Wesselman, H. B. 1984. The Omo micromammals; pp. 1–219 in M. K. Hecht and F. S. Szalay (eds.), Contributions to Vertebrate Evolution. S. Karger, Basel. Received 7 September 2000; accepted 24 June 2002.

APPENDIX 1 List of dental (C1–13) and skull (C14–22) characters used in this analysis: C1. P3: present (0), absent (1). C2. Metaloph connection and development on M1–M2 (after morphotypes defined by Black, 1965). Long metaloph faintly connected to the protocone (0), connected to the hypocone (1), short or reduced metaloph connected with the posteroloph (2), short unattached metaloph (3). C3. Metaconule on M1–M2 absent (0), or well individualized (1). C4. Anterior cingulum connected to the protocone on M1 (0), or ending very low to the cusp and separated from it by a transverse valley (1). C5. Mesostyle on M1–M2 present (0), absent (1), or poorly differentiated (2). C6. Hypocone and protocone connected on the M3 (0), or not connected (1). C7. Roots visible on upper molars in lateral view (0), or not visible (1). C8. Anterolophid on m1 present (0), or absent (1). C9. Ectolophid on m1–m2 present (0), or absent (1). C10. Entoconid on m1–m2 oblique to longitudinal (0), transverse (1), or bunodont and median on labial wall (2). C11. Entolophid on m1–m2 not visible (entoconid included in the posterolophid) (0), or visible (1). C12. Protoconid and metaconid on m1–m2 at same level unfused (0), fused by the anterior side (1), or fused by posterior part of cusps (protolophid) (2). C13. Hypoconulid differentiated on m3 (0), or barely visible or absent (1). C14. Nasal extended at front of the skull (0), or not extended (1). C15. Tympanic bullae septa: 0 to 2 (0), or 3 (1). (Moore, 1959 reported that most Sciurinae have 1 or 2 tympanic septae, so this is taken to be the primitive condition). C16. Interparietal present (0), or absent (1). C17. Mastoid process marked (0), or absent (1). C18. Bony palate extended beyond the end of maxillary tooth rows (1), not extended (0). C19. Squamosal extended to the base of postorbital process of the frontal (1), not extended (0). C20. Length of lacrimomaxillary contact: lower or identical to lacrimojugal one (0), larger (1). C21. Masseteric tubercle of the maxilla: prominent (0), normal (1), absent (2). C22. In the pterygoid fossa close to the tympanic bullae: large round opening of alisphenoid canal (0), small round opening (1).

DENYS ET AL.—NEW XERUS FROM CHAD

687

APPENDIX 2 Matrix of character states used in the cladistic analyses of relationships. C2, C12, C15 and C21 are multistate characters. ? 5 unknown state for the species. Taxon Xerus daamsi Xerus inauris Xerus princeps Xerus rutilus Xerus erythropus Atlantoxerus getulus Heliosciurus Funisciurus Protoxerus Funambulus Epixerus Paraxerus Myosciurus Sciurus

C1

C2

C3

C4

C5

C6

C7

C8

0 1 1 1 0 0 1 0 1 1 1 0 1 0

3 2 2 3 3 3 0 2 0 0 1 0 ? 3

1 1 1 1 1 1 0 1 1 1 1 0 ? 0

1 1 0 1 0 0 0 0 0 1 0 0 1 0

1 1 0 1 1 1 0 1 0 1 1 1 1 0

1 0 0 1 0 0 0 0 1 1 0 0 ? 1

1 1 1 1 1 0 0 1 0 0 0 0 0 0

1 1 1 0 0 1 0 1 0 0 0 1 1 0

C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 1 1 0 1 1 0 1 0 0 0 0 0 1 1

0 1 0 0 1 0 0 1 1 0 0 1 0 2

1 1 1 0 0 0 0 1 0 0 0 1 0 0

1 2 2 0 1 1 2 2 1 0 2 1 0 0

0 1 1 0 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 0 0 0 0 0 0 0 0 0

1 2 2 1 1 1 0 0 0 0 0 0 0 0

1 0 1 0 1 0 1 1 0 1 0 1 1 1

0 0 0 1 0 1 0 1 1 0 0 1 1 1

1 1 1 1 1 1 0 0 0 0 0 0 0 1

1 1 1 1 1 1 0 0 0 0 0 0 1 0

? 1 1 1 1 1 0 0 0 1 0 0 0 0

1 0 0 0 0 0 0 1 2 0 2 0 2 1

? 1 1 1 1 0 0 0 0 0 0 1 0 0