Presence of the Maxillary Sinus in Fossil Colobinae (Cercopithecoides williamsi) from South Africa



Extant cercopithecoid monkeys, except macaques, are distinguished among primates by their lack of paranasal pneumatization, including the maxillary sinus (MS). Analysis of this structure, widespread among Eutheria, suggests that its loss occurred in the cercopithecoid common ancestor; thus, the presence of the MS in macaques is not strictly homologous to that in other primates. CT analysis of the fossil species Victoriapithecus macinnesi supports this view, demonstrating the lack of the MS in this stem cercopithecoid. Recent evidence, however, has documented the presence of the MS in extinct cercopithecoids from the late Miocene and Pliocene. This study reports on CT examination of two fossil crania attributed to Cercopithecoides williamsi from South Africa, dated in the range, 3.0–1.5 Ma. BF 42a is a complete cranium from Bolts Farm; MP113 is an intact facial skeleton, including the anterior cranial vault, from the Makapansgat Limeworks. Both demonstrate MS presence, unknown in extant colobines and unexpected in most cercopithecoid monkeys. The relative size of the MS of BF 42a is similar to that of extant tropical and subtropical macaques. The presence of sinuses in several extinct colobines suggests that our understanding of the evolutionary history of these primates, and of the MS, is incomplete, and that other fossil cercopithecoids should be examined for this feature. The developmental plasticity exhibited in this feature, indicated by multiple loss and reemergence, provides further evidence that paranasal pneumatization has undergone a complex history of suppression and expression. Anat Rec, 291:1499–1505, 2008. © 2008 Wiley-Liss, Inc.

The presence of paranasal pneumatization, or sinuses, is a common morphological feature of the skull among eutherian mammals and is typical of many extant and extinct primate taxa (Lund, 1988; Rae and Koppe, 2004; Rossie, 2005). Not all primate taxa, however, possess each of the paranasal sinuses, and their size and configuration are variable (Cave and Haines, 1940; Koppe and Nagai, 1999; Koppe and Ohkawa, 1999). In particular, the distribution and variation of the maxillary sinus (MS) in primates has been a focus of much research (Cave and Haines, 1940; Lund, 1988; Koppe et al., 1999; Rae, 1999; Nishimura et al., 2005; Rossie, 2006). Despite a long history of research (e.g., Blanton and Biggs, 1969), the MS in general remains an enigmatic and little-understood anatomical structure.

The MS is one of four paranasal sinuses found in eutherian mammalian crania, which are variably present in many extant primates (Cave and Haines, 1940; Harrison, 1987; Rossie, 2005). Among anthropoid primates, the MS is present in most platyrrhine monkeys (Nishimura et al., 2005), absent in most cercopithecoid monkeys (Rae and Koppe, 2003), and present and frequently described as “enlarged” in hominoids (Cave and Haines, 1940; but see Rae and Koppe, 2000). The presence and morphological configuration of an MS among any specific primate group, however, is highly variable (Cave and Haines, 1940; Rae and Koppe, 2000; Rossi, 2005; Lund, 1988) and this poses many questions regarding the anatomy, development, and evolution of this structure among primates and mammals more generally (see Blanton and Biggs, 1969).

Morphologically, the MS is defined on the criterion that its ostium opens into the nasal cavity via the middle meatus (Cave and Haines, 1940; although other criteria have been proposed—see Rossie, 2006), resulting from its development as an epithelial diverticulum invading the maxillary bone from the nasal cavity. Thus, even with considerable diversity in craniofacial pneumatization, the MS is distinguished from other, presumably nonhomologous structures such as the “lateral recess” (Rae and Koppe, 2003) found in some cercopithecoid primates lacking the MS.

The MS is present in all extant hominoid primates and a few platyrrhine genera (Cave and Haines, 1940; Nishimura et al., 2005; Rossie, 2006). Among extant cercopithecoids, it is present only in the macaques (Ward and Brown, 1986; Koppe and Ohkawa, 1999; Rae and Koppe, 2003), and, until recently, the condition of extinct cercopithecoids with respect to pneumatization was unknown. The presence, absence, and size of the MS have been used as a diagnostic character in Primates (Cave and Haines, 1940; Harrison, 1987; Rae, 1997). All the same, the presence of the MS (or other paranasal sinuses) in the primate fossil record is known for only a few taxa (Andrews, 1978; Ward and Pilbeam, 1983; Ward and Brown, 1986; Rae, 1999). With the increasing use of CT scan imaging in the noninvasive analysis of fossil specimens, however, more attention is directed toward documenting and describing the paranasal sinuses and other aspects of the internal morphology of fossil primate crania.

The first explicit use of CT to attempt to detect the presence of the MS in fossil cercopithecoids was an examination (Rae et al., 2002) of a complete cranium of the stem cercopithecoid (Benefit and McCrossin, 1993) Victoriapithecus macinnesi. The maxilla of this extinct primate, however, is most similar to non-macaque extant cercopithecine taxa, such as Allenopithecus, which do not possess the MS. The lack of the MS in this proposed stem cercopithecoid supports the “early loss” hypothesis (Rae, 1999), which accounts for the widespread lack of this structure among cercopithecoids. In this scenario, the MS observed in extant macaques is explained as an independent evolutionary acquisition, which is nonhomologous to the MS in other primates.

Recent investigations of other fossil crania, however, indicate that at least some extinct cercopithecoid species of the Neogene possessed the MS (Nishimura et al., 2007; Rae et al., 2007; Rae, 2008). These reports have implications for reconstructing the evolutionary history of the MS in cercopithecoid primates, and suggest that its complexity is not fully reflected in the distribution of this structure among modern cercopithecoid taxa. These results highlight the need for more data on pneumatization in extinct cercopithecoid taxa, so that a complete picture of its evolution in Old World monkeys can be achieved. To that end, we report here on CT investigation of craniofacial skeletons attributed to the Plio-Pleistocene colobine Cercopithecoides williamsi (Mollett, 1947) from South Africa, and discuss its implications in an adaptive and evolutionary context among primates.


C. williamsi is known from multiple fossil localities in East and South Africa (Freedman, 1960, 1965; Freedman and Brain, 1972; Szalay and Delson, 1979; Leakey, 1982; Delson, 1984; Keyser, 1991; Jablonski, 2002). The deposits that these fossil specimens derive from are dated in the range, 1.5–3.0 Ma. In East Africa, additional fossil material has been attributed to the species C. kimeui (Leakey, 1982), C. meaveae (Frost and Delson, 2002), and C. kerioensis (Leakey et al., 2003).

Species of Cercopithecoides are described as “medium to very large colobines” (Frost and Delson, 2002: 725), varying in size; C. kimeui is larger than C. williamsi (Leakey, 1982; Jablonski, 2002; Frost et al., 2003), and the two newer species are smaller still (Frost and Delson, 2002; Leakey et al., 2003). Unlike extant arboreal colobines, these extinct colobines are postcranially similar (where known) to more terrestrial cercopithecoids (Leakey, 1982; Frost and Delson, 2002; Jablonski, 2002; Leakey et al., 2003). Evidence for sexual dimorphism consists of the presence of a longer, narrower cranium, a larger canine, and a larger P3 in presumed males. The cranium of C. williamsi is similar to that of modern colobines in possessing the wide face, wide interorbital space, and large orbits, but the mandible differs in being shallow with an obliquely oriented ramus.

Two specimens of C. williamsi housed at the University of the Witwatersrand (Fig. 1) were examined for this study: BF42 from Bolt's Farm and MP113 from the Makapansgat Limeworks (Freedman, 1960, 1965). BF42 is a nearly complete female cranium including the maxillary dentition. MP113 includes an intact male facial skeleton encompassing the anterior cranial vault and complete maxillary alveoli.

Figure 1.

Frontal view of C. williamsi crania; MP113 from Makapansgat Limeworks (left) and BF42 from Bolt's Farm (right). Scale bar is 33 mm.

CT scans of these specimens were obtained using a Phillips Brilliance6 CT Scanning System at the Radiology Department of the Donald Gordon Hospital in Johannesburg. Scans were taken at 1.0-mm slice thickness in coronal orientation, at 120 kVP/120 mA. Serial CT scan images were viewed using the AccuLite DICOM Viewer version 3.108 (AccuImage Diagnostics Corporation, Three-dimensional reconstructions of crania and sinuses were obtained using Amira 4.0 (Mercury Computer Systems, San Diego, CA). All statistical analyses conducted with SPSS for Windows 12.0.1 (SPSS, Chicago, IL).


Observations from serial CT scans through relevant parts of the maxilla of BF42 (Fig. 2) and MP113 (Fig. 3) were compared with similar scans of an extant colobine skull (Fig. 4) and a macaque (Fig. 5). The extant colobine clearly demonstrates cancellous bone throughout the body of the maxilla without any pneumatization. In contrast, CT scan images of both fossil specimens demonstrate the presence of an MS as indicated in Figures 2 and 3. Although similar in general configuration to the MS demonstrated in the macaque specimen, the MS in C. williamsi differs in that its floor is positioned well above the roots of the molar teeth.

Figure 2.

CT scan slices of BF42 (C. williamsi) in the region of the M2 (a) and M3 (b) demonstrating the presence of the maxillary sinus. The arrow in (a) is positioned approximately at the level of the middle meatus in the nasal cavity, into which the ostium (not clearly visible) opens. The line in (b) indicates the reconstructed position of the bony partition separating the MS from the nasal cavity, which was partially damaged in this specimen.

Figure 3.

CT scan slices of MP113 (C. williamsi) through the mesial (a) and distal (b) molar regions. In (a) the anterior end of the right MS is visible as a darkened oval shape lateral to the nasal cavity. In (b) the MS appears larger, but the bony partition between the MS and nasal cavity is not fully preserved on either side; the irregular black space within the left maxilla is an air pocket communicating through the ostium between the nasal cavity and the MS.

Figure 4.

CT scans through the premolar (a) and molar (b) regions of an extant colobine (Colobus) demonstrating the presence of cancellous bone and the lack of the maxillary sinus in the maxilla.

Figure 5.

CT scan through the distal molar region in an extant macaque (M. cyclopis) demonstrating the presence of a maxillary sinus (arrow indicates the ostium opening through the lateral wall of the nasal cavity into the middle meatus).

Three-dimensional virtual reconstructions were obtained from the CT scans using Amira 4.0 (Fig. 6). Both sinus volume and univariate morphometrics were obtained from the reconstructions. To evaluate the relative size of the sinus, only BF42 was used, as MP113 is too damaged externally for sufficient craniometrics to be obtained. Even in BF42, sinus volume is still an approximation, as the small amount of damage to the lateral wall of the nasal cavity that is present required the position of this bony septum to be estimated in places.

Figure 6.

Semitransparent 3D CT reconstruction of the BF42 cranium (C. williamsi) demonstrating the size and position of the maxillary sinus. Note the elevated position of the floor of the maxillary sinus.

To evaluate the sinus volume obtained from C. williamsi relative to that of other primates, it was compared with a small sample of similarly sized warm-climate extant cercopithecoids, Macaca fascicularis and M. cyclopis (data from T. Koppe); the remaining members of the M. fascicularis species group, M. mulatta and M. fuscata, were excluded, as sinus volume in the latter is significantly correlated with its cold environment (Rae et al., 2003). The volume of the sinus was regressed against a measure of facial volume, the product of bimaxillary width, facial height, and palatal length (see Rae and Koppe, 2000).

Table 1 provides a summary of the data for this comparison. The size of the sinus found in C. williamsi is large compared with those obtained from the macaques, but it also possesses a larger face overall. Figure 7 displays the results of the regression analysis in log–log space; C. williamsi falls well within the 95% individual confidence interval for the sample. It is worth noting that this relationship between sinus volume and facial size is isometric (slope = 1.04, 95% confidence interval = 0.41–1.68), similar to that reported for hominoids (Rae and Koppe, 2000); the difference between this result and that reported previously for a larger sample of macaques (Koppe et al., 1999) presumably is due to the inclusion of the cold-climate macaques in that study.

Figure 7.

Log–log regression plot of maxillary sinus volume relative to face size in C. williamsi and extant macaques. Pearson's r = 0.681, P< 0.01. Straight line is the least squares regression (y = 1.044029x–4.330215); upper and lower curves are 95% individual confidence intervals.

Table 1. Summary data for sinus and facial volume in Cercopithecoides (this study) and Macaca (from T. Koppe)
TaxonNMean maxillary sinus volumeStd. deviationMean facial volumeStd. deviation
Macaca cyclopis82.640.89147.9839.51
M. fascicularis81.510.7792.0719.78
Cercopithecoides williamsi14.69 159.45 


We report on the presence of the MS in a fossil colobine monkey, suggesting that examination of additional fossil cercopithecoid crania for this structure is warranted. Detailed comparison of the MS is hampered by fragmentation and the quality of CT scans in such fossil specimens, but in these C. williamsi specimens, it can be characterized generally as a moderately sized sinus positioned well above the molar tooth roots. The CT scans of C. williamsi presented earlier, however, demonstrate clearly all the anatomical features of a “true” MS and there is much less damage to interpret than in previous studies (Nishimura et al., 2007; Rae et al., 2007; Rae, 2008). This may assist in clarifying the presence of the MS in other extinct primate taxa.

As a result of the CT examinations of fossil primate taxa reported earlier and elsewhere (Nishimura et al., 2007; Rae et al., 2007; Rae, 2008), it is now clear that the distribution of the MS in extinct cercopithecoids is nonintuitive based on its expression in extant taxa. To date, more fossil colobines (Libypithecus, two species of Cercopithecoides) than cercopithecines (M. majori) possess the MS, and nearly as many fossil taxa contradict the extant pattern (Libypithecus, two species of Cercopithecoides) as support it (Mesopithecus, Theropithecus, Paradolichopithecus, M. majori; Victoriapithecus is a special case—see above). This implies that the presence of pneumatisation per se may not be a reliable taxonomic or phylogenetic character among this group of primates (see discussion in Rae, 2008). The analysis reported earlier shows that the inferred independent acquisition of the MS in various clades is not indicated by obvious differences in size, either; the fossil colobine possesses a MS indistinguishable from a (non-temperate climate) macaque of a similar size, and these sinus volumes are comparable to those seen in hominoids. In this case, the degree of convergence is very high.

Overall, the observed pattern of MS distribution suggests that the development and presence of the MS as a structure involves a considerable amount of evolutionary plasticity in primates as well as in other taxa (Witmer, 1999). This is also generally supported by the observed variation in paranasal sinus morphology among primates (Cave and Haines, 1940; Nishimura et al., 2005; Rossie, 2005).

Given the absence of the MS in the Miocene stem cercopithecoid Victoriapithecus, the “early loss” hypothesis (Rae, 1999; Rae et al., 2002) remains the most useful framework for reconstructing the evolutionary history of the MS in cercopithecoid monkeys. Thus, the presence of maxillary pneumatization in some or all of the extant cercopithecoid primates must be nonhomologous with the MS sensu stricto in other primates and in eutherian mammals generally (Rae et al., 2002). This requires consideration of more complicated evolutionary scenarios for MS presence and distribution in this group of primates. For example, the nonintuitive distribution (“loss” and “reappearance”) of the MS among extinct and extant cercopithecoids, including colobines, may reflect the effect of unknown selective factors suppressing the developmental mechanism controlling the epithelial diverticulum from the nasal capsule into the body of the maxilla, rather than complete loss of the mechanism of pneumatization itself (Rae et al., 2007; Rae, 2008). It may be that the MS itself is not a “structure” or “trait” upon which selection acts, but that expression of pneumatization is contingent on another factor (or factors) that determines whether air spaces in the cranium are expressed or suppressed.

Clarifying the pattern of distribution of the MS, and more complete documentation of factors involved in the variation of the MS with factors, such as overall facial configuration, dental size, and other aspects of craniofacial morphology, may help to understand the nature of the selection pressures responsible in the development and evolution of the MS in this group of primates (Rae and Koppe, 2004) as well as the actual “target” of selection. Virtually, all explanations regarding the presence and distribution of the MS presented to date, however, are hampered by obvious exceptions among other taxa (Blanton and Biggs, 1969).

As more information about the distribution of the MS is gleaned from the fossil record, it is becoming more apparent that the evolutionary history of this structure in primates is poorly understood. Although the early loss model remains supported by available character-state analyses of cercopithecoid primates, there remains a lengthy temporal gap in the fossil record between Victoriapithecus (the proposed stem cercopithecoid at 15 Ma) and the later Neogene cercopithecoids such as Paradolicopithecus, Libypithecus, and Cercopitheoides, which currently leaves much to be clarified about the evolutionary history and actual distribution of the MS in extinct cercopithecoid primates.


The authors thank Mike Raath for assistance with the fossil collections at Wits University, and Claire Gibbs and Justin Leonard at the Donald Gordon Medical Centre, Johannesburg for assistance obtaining the CT data of the fossil crania used in this research. They thank Thomas Koppe for providing the comparative volumetric data presented on extant macaques and the comments of Eric Delson and an anonymous reviewer. Finally, they extend their gratitude to S. Marquez for the invitation to contribute to this special issue.