Papio Cranium from the Hominin-Bearing Site of Malapa: Implications for the Evolution of Modern Baboon Cranial Morphology and South African Plio-Pleistocene Biochronology

A new partial cranium (UW 88-886) of the Plio-Pleistocene baboon Papio angusticeps from Malapa is identified, described and discussed. UW 88-886 represents the only non-hominin primate yet recovered from Malapa and is important both in the context of baboon evolution as well as South African hominin site biochronology. The new specimen may represent the first appearance of modern baboon anatomy and coincides almost perfectly with molecular divergence date estimates for the origin of the modern P. hamadryas radiation. The fact that the Malapa specimen is dated between ~2.026–2.36 million years ago (Ma) also has implications for the biochronology of other South African Plio-Pleistocene sites where P. angusticeps is found.


Introduction
The recent discovery of fossiliferous deposits at Malapa, South Africa, has had a major impact on the study of human evolution [1][2][3]. Over the past few years, three Australopithecus sediba partial skeletons have been described [1,[4][5][6][7][8][9][10][11][12][13], and the preparation of additional specimens (including other partial skeletons) is currently underway. Along with the hominin material, numerous faunal specimens from Malapa have also been recovered [2,[14][15], and most are still awaiting formal analysis and description. The description and analysis of the fauna are crucial to understanding aspects of paleoecology, biochronology, and mammalian evolution during the African Plio-Pleistocene. In this paper, we describe the only known non-hominin primate specimen from Malapa, UW 88-886, a specimen of the fossil baboon Papio angusticeps. Furthermore, we evaluate Papio angusticeps in the broader context of modern baboon craniodental evolution and South African Plio-Pleistocene biochronology.

Geological context
Malapa cave is hosted in the stromatolite-rich, chert-free dolomite of the Lyttelton Formation of the late Archaean (2.64-2.5 Ga) Malmani Subgroup [2], and presently lies near the base of a cave system that was originally >30-m deep prior to erosion by valley incision. Minimal limestone mining in the early 20th century exposed in situ cave deposits [2]. Two shallow pits have been recognized at Malapa, of which the main (Pit 1) contains the well-preserved and welldated (~1.977 Ma) Au. sediba fossil remains [1][2]. The cave deposits comprise six distinct stratigraphic lithofacies assigned, A to F, from base to top [3]. Facies A and B represent the oldest deposits that were partially eroded before they were covered by flowstone (Flowstone 1, Fig  1). Facies D overlies Flowstone 1, followed by Facies E and topped by Facies F, while Facies C occurs as a remnant between the dolomite on the western sidewall of Pit 1 and Facies B and E (and possibly D) [3]. Flowstone 1 is dated to 2.026 ± 0.021 Ma, suggesting that Facies A and B are older than 2 million years [2]. The P. angusticeps partial cranium was found in a block of calcified clastic sediments within two metres of Pit 1 and was almost certainly moved to this position by early mining activities. The cranium is embedded within fine-grained clastic dark brown sediment with clastic sandstone. It is unlikely that the specimen comes from Facies A, D and E for a couple of reasons. First, Facies A and D are comprised of course grain sediments, both abundant in fossil remains, and second, Facies E sediments are best characterized as coarse to fine upwards and also contain abundant fossil remains. It is also unlikely that the specimen derives from Facies C and F, since Facies C lacks clastic intercalations while Facies F is devoid of sandstone intercalations. We also considered the possibility that the specimen came from a nearby cave. Indeed, there are two caves in close proximity to Malapa, both less than 150 metres south of the Malapa pits. Currently there are no indications of fossil remains in situ or in the mining dumps near these caves. Therefore, it is most parsimonious to presume that the block containing P. angusticeps was removed from Pit 1, from Facies B, and the resulting age of this specimen is thus > 2.026 ± 0.021 Ma based on the U-Pb dates of the overlying Flowstone 1 (Fig 1). Geological studies indicate that the different Facies were deposited in quick succession of each other with limited transport distance [2], and the discovery of Equus sp., which has its earliest recorded appearance in Africa at~2.36 Ma [16][17][18], in Facies D, provides a probable maximum age for the site more generally. Therefore, the probable maximum age for this specimen is 2.36 million years ago (Ma) based on the appearance of Equus sp. in Facies D [2]. Preservation. The specimen is sheared nearly in half along an approximate sagittal plane thus preserving most of the right part of the cranium. There is little postdepositional compaction with minor broken areas in the maxillary and infratemporal fossae. The preserved portion of the cranium is approximately half-filled with fine sediment capped by a thin layer of flowstone; stage 1 weathering is apparent [19].

Material and Methods
Description. U.W. 88-886 is a partial cranium preserving most of the right facial skeleton and the right half of the neurocranium (Figs 2 and 3). The overall size, strong maxillary ridges, and strong temporalis muscle markings suggest the specimen is most likely a male. In addition, the overall size of the specimen, apparent anteorbital drop, long and narrow muzzle, strong maxillary ridges, deep maxillary fossae, and tall malar region make the specimen referable to Papio among Plio-Pleistocene cercopithecoid taxa, generally, and P. angusticeps among Plio-Pleistocene Papio taxa, more specifically (Table 1). Unfortunately, the dentition is not preserved, with the exception of remnants of the M 2 and M 3 roots. The facial skeleton includes  portions of the right orbit, right zygomatic, right nasal, and right half of the muzzle including most of the maxilla and a small remnant of the premaxilla (Figs 2 and 3). In addition to the right side of the face, fragments of right frontal, parietal, temporal and zygomatic arch are also preserved. The occipital, basicranium, most of the dorsal surface of the neurocranium, and left side of the cranium have not been recovered.
The upper part of the face is distinguished by tall, oval-shaped orbits. Because only half of the orbital rim is preserved (the inferior, lateral, and superolateral aspects), it is impossible to measure the exact size of the orbit, but the orbital height and width can be estimated (Tables 2  and 3; Figs 2-4). Thus, UW 88-886 appears to have large orbits compared to other specimens of P. angusticeps (a feature more typical of Papio izodi among Plio-Pleistocene Papio taxa), but additional specimens preserving the entire orbit are needed to confirm this morphology as typical of the Malapa population. Given the other features present in the specimen (see below) and the low sample size of known P. angusticeps orbits, it seems most reasonable to accept that UW 88-886 simply adds to the known variation of P. angusticeps at this time. In fact, including UW 88-886 in the existing P. angusticeps sample of orbit height measurements results in a range of variation (20-27 mm) similar to that seen in a small sample of extant P. h. cynocephalus (18-27 mm, variances statistically equivalent, Levene statistic p = 0.344; see Tables 2 and 3; Fig 4). From the preserved anatomy of the upper face and orbital region, particularly the nasals  [39]. P. izodi is estimated at~20 kg for males and~15 for females [39]. The most similar extant taxon in terms of body mass is P. h.
cynocephalus,~23 kg for males and 12.5 kg for females, which is why P. h. cynocephalus is used in the above comparisons. All specimens were pooled regardless of sex in order to increase sample size. For sex-specific values, see

Notes:
Orbit height defined as the maximum distance between the inferior and superior orbit borders. Orbit width defined as the maximum distance between the lateral and medial orbit borders. Orbit area is defined as orbit width x orbit height and is measured in mm 2 and maxilla just medial to the infraorbital border, it appears that a pronounced anteorbital drop would have been present. The zygomatic bone bulges outward slightly lateral to a series of infraorbital/maxillary foramina, then pinches inferiorly, forming sharp cheekbones bordering the pronounced infraorbital/maxillary fossa. The right lateral border of the orbit is intact from the frontal-zygomatic suture to the beginning of the temporal process of the zygomatic. The zygomatic arch is not preserved (Figs 2 and 3). Malar height, as characterized by the minimum distance from inferior orbital margin to the inferior border of zygomatic process of the maxilla, is relatively tall in UW 88-886, a feature it shares with other P. angusticeps specimens and similarly-sized modern P. hamadrayas baboons (e.g., P. h. cynocephalus) to the exclusion of P. izodi (Tables 1-3; Figs 4 and 5). On the lateral orbit border, there are two zygomatico-facial foramina, the larger in the central body of the zygoma and the smaller is placed superiorly near the frontal-zygomatic suture. The zygomatic root begins above the maxillary alveoli at the level of the M 2 -M 3 contact. The lateral margin of the zygomatic appears to flare slightly outward (laterally). than P. angusticeps, but a non-significant difference compared to P. h. cynocephalus. The difference between P. angusticeps and P. h. cynocephalus is also non-significant. UW 88-886 appears to have large orbits compared to other P. angusticeps specimens, but again within a reasonable range of expected variation for a species. Bottow Row: Boxplots of relative malar height considering UW 88-886 separately (left) and within P. angusticeps (right). P. izodi has a significantly shorter malar height compared to P. angusticeps and P. h. cynocephalus, but the difference between P. angusticeps and P. h. cynocephalus is non-significant. UW 88-886 is closest to the average of other P. angusticeps specimens. See also Table 2   The preserved portion of the muzzle dorsum is flat to slightly rounded moving medially, and there is strong development of a prominent maxillary ridge bordering the lateral aspect of the muzzle dorsum. The superior rim of the nasal aperture is broken in this specimen; however, what is preserved of the superior portion of the rim appears to have been rounded. The premaxillae do not extend up along the nasals as far as in some other specimens of P. angusticeps (e.g., female CO 101). There is a slight groove medial to the canine juga extending to a large infraorbital foramen. Lateral and inferior to this foramen, there are several smaller maxillary foramina. The lateral portion of the muzzle is dominated by a deep and pronounced maxillary fossa extending into the zygomatic and maxillary bone inferior to the orbital margin (i.e., the infraorbital plate).
In dorsal view, there is evidence for a strong supraorbital region based on the preserved morphology of the right superolateral orbit border. The superolateral portion of the orbital rim is relatively flat and the supraorbital torus has a sharp border demarcated posteriorly by the temporal line with a distinct lip above the post-orbitally constricted frontal bone. There is marked post-orbital constriction with evidence of a probable post-orbital sulcus, not as much constriction as in Theropithecus, but a similar degree to that observed in typical male P. angusticeps and Papio hamadryas specimens (Figs 2-5). Also similar to P. angusticeps and modern Papio hamadryas males, in particular, UW 88-886 clearly displays pinched temporal lines as far medially as they are preserved, although it is unclear whether they would have met posteriorly to form a slight sagittal crest. The temporal lines emanate from the supero-lateral corner above the orbits medial to the post-orbital constriction of the frontals.

Discussion
Based on preserved morphology, UW 88-886, a presumed male, is most similar to the South African taxon Papio angusticeps among known Plio-Pleistocene baboon populations. The combination of an anteorbital drop, long and narrow muzzle, well-developed maxillary ridges, welldeveloped and deep maxillary fossae, and a relatively tall malar region make the specimen statistically distinct from other Plio-Pleistocene Papio species and most similar to P. angusticeps as well as modern P. hamadryas subspecies (Tables 1-3; Figs 4 and 5). The identification of P. angusticeps at Malapa has broader implications for baboon evolution and hominin biochronology as well.
Modern baboons (extant P. hamadryas ssp.) represent one of the most successful living primate radiations, with populations distributed all across sub-Saharan Africa and into the Arabian Peninsula. Despite their evolutionary success, modern baboon origins in the fossil record are not well understood or agreed upon. In East Africa prior to the Middle Pleistocene, there are craniodental fragments from various localities that possibly represent the genus Papio, but none of these fossils are definitively diagnostic (e.g., [20][21][22][23][24][25]). Fossils more clearly attributable to the modern Papio hamadryas population appear in the Middle Pleistocene at Olduvai Gorge, Asbole, and Bodo [26][27][28][29].
During the South African Plio-Pleistocene, at least two species of "small-bodied" papionins exist that have been typically placed into the genus Papio: P. izodi and P. angusticeps. P. izodi is currently recognized at Taung as well as Members 2 and 4 at Sterkfontein [30][31][32][33], suggesting an age range anywhere from~3.7-2.0 Ma depending on the estimate (e.g., see [30,34] for Taung; see [34][35] for most recent estimates of Members 2 and 4 at Sterkfontein). P. angusticeps on the other hand, is securely documented at sites generally suggested to be less than~2.0 Ma, such as Kromdraai A and Cooper's A. One exception might be the P. angusticeps population at Haasgat, which was recently argued to be slightly older, between~2.58-1.95 Ma, with a more likely range for most fossils between~2.3-2.0 Ma [36][37]. The age of P. angusticeps at Gladysvale is difficult to assess at this time as much of the faunal material was initially collected from ex situ breccia blocks of unknown provenience [34].
While some authors lump both P. izodi and P. angusticeps into a single taxon (P. izodi sensu lato; e.g., [22,38]), other experts recognize them as different taxa (e.g., [30][31]39]). Some authors even suggest that P. angusticeps is most likely a subspecies of the modern baboon radiation (i.e., P. hamadryas angusticeps) and note many similarities with modern P. h. kindae and P. h. cynocephalus populations, in particular [30,33,40]. Similar to East Africa, specimens definitively attributable to P. hamadryas appear in southern Africa by the Middle Pleistocene, best represented by the distinct subspecies P. h. botswanae in Botswana [41] and the South African Pleistocene/Holocene P. "spelaeus" specimen of unknown provenience that appears to represent a large P. h. ursinus male [22,[41][42].
P. angusticeps shares features with modern P. hamadryas such as a relatively tall malar region, relatively small orbits, a definitive anteorbital drop, deep maxillary/suborbital fossae, pronounced maxillary ridges in males, a definitive anteorbital drop, a relatively narrow muzzle, a relatively long muzzle, and relatively small teeth (Table 1; Fig 5). When Freedman [42] transferred the species to Papio from Parapapio (originally Parapapio angusticeps Broom, 1940), he also noted that P. angusticeps cranial morphology closely resembles that of P. h. ursinus, except that the maxillary ridges are slightly weaker and the maxillary fossae are slightly deeper in P. angusticeps. P. izodi, on the other hand, is much more primitive in that it displays a relatively short malar region, relatively large orbits, a variable anteorbital drop, shallower maxillary/suborbital fossae (particularly in males), little to no development of the maxillary ridges in males, a relatively broad muzzle, a relatively short muzzle, and relatively large teeth (Tables 1 and 3; Fig  5; see also [30][31]33]). In fact, one recent phylogenetic analysis places P. izodi outside of the crown African papionins and questions its status as a member of the modern genus Papio at all [33]. In any case, it is clear that the morphology of P. izodi is distinct from that of P. angusticeps as well as the modern baboon populations, P. hamadryas ssp. (Tables 1-3

; Figs 4 and 5).
Excluding P. izodi, there are a few additional South African fossil Papio specimens that are relevant to the origins of the modern baboon. Papio robinsoni is a large form of Papio (larger on average than P. izodi, P. angusticeps, and most modern subspecies) securely documented at sites estimated to be anywhere between~2.6-1.5 Ma: Swartkrans Member 1, Skurweburg, Bolt's Farm Pit 23, and Drimolen; it is also less securely identified at Sterkfontein Member 4 [30,[42][43][44][45][46]. While P. robinsoni has also been argued to be a member of the modern Papio hamadryas radiation by some (e.g., [30,33,38,44]), there are possible reasons to suspect it is distinct. P. robinsoni clearly displays synapomorphies with modern P. hamadryas such as a distinct anteorbital drop, large overall size, definitive maxillary ridges, a relatively long and narrow muzzle, relatively small orbits, relatively small teeth, and definitive maxillary fossae in males (Tables 1 and 3). However, there are also consistent differences between P. robinsoni and modern P. hamadryas ssp. P. robinsoni displays weaker development of the maxillary and mandibular corpus fossae than modern baboons (Table 1), maxillae that occasionally meet in the midline (covering the nasals near nasion), and nasals that lie below the level of the maxillary ridges on the muzzle dorsum [40,42]. Our interpretation is that these features are primitive relative to modern baboons, particularly the slight development of the facial fossae. P. angusticeps, on the other hand, displays no features that are obviously distinct from the modern P. hamadryas populations (see Tables 1 and 3), suggesting that it is possibly the first well-documented modern P. hamadryas subspecies in the fossil record (i.e., P. h. angusticeps). While P. angusticeps displays slightly deeper fossae and less-developed maxillary ridges on average when compared to P. h. ursinus [42], similar morphology to P. angusticeps can be found among the other P. hamadryas susbspecies. A more complete and formal taxonomic reassessment of P. angusticeps and P. robinsoni is beyond the scope of this paper, but future studies may further support the sinking of P. angusticeps into P. hamadrayas (as originally suggested by Delson [30]), depending on the species concept being used. In particular, given the quantitative morphometric differences documented among modern P. hamadryas populations [41,47], future 3D morphometric studies including both P. angusticeps and P. robinsoni may help illuminate the affinities of these taxa relative to modern P. hamadryas populations and relative to the recently described P. h. botswanae.
Interestingly, recent molecular dates place the origin of the modern P. hamadryas radiation between~1.8 and 2.2 Ma in South Africa [48][49][50][51], which closely brackets the estimated age of the Malapa specimen at~2.026-2.36 Ma. Thus, the P. angusticeps specimen described here may not only represent the earliest appearance of P. angusticeps, but potentially the earliest appearance of the modern baboon population in the fossil record, in almost perfect agreement with molecular estimates. While it is true that the molecular clock must always be calibrated through the fossil record and, therefore, there is always a bit of circular reasoning involved, it is still noteworthy and perhaps reassuring whenever new fossil discoveries seem to confirm molecular estimates, thereby lending support to hypothesized divergence dates. Because the estimated age of the Malapa fauna seems well-constrained around the dated~2.026 Ma flowstone (see Geological Context above), the presence of P. angusticeps also assists in the biochronological assessment of other sites where P. angusticeps is documented, including Gladysvale, Haasgat, Kromdraai A, and Cooper's A [30,33,42,44,[52][53]. In particular, the Malapa specimen supports recent revised estimates for the age of the Haasgat fauna (and P. angusticeps) between~2.3-2.0 Ma, and suggests that Haasgat and Malapa together document a more secure first appearance date (FAD) for P. angusticeps between~2.3-2.0 Ma. Cercopithecoid taxa such as P. angusticeps have long been used as important biochronological markers in age assessments of the South African Plio-Pleistocene hominin sites (e.g., [30-31, 44, 32, 54]), thus a more certain FAD for P. angusticeps may consequently lead to a slight adjustment in the estimated chronology of other hominin sites pending future faunal correlation analyses.
Supporting Information S1 Dataset. Raw measurements used in this study (Table A). (XLS)