Pneumatization, in the form of paranasal sinuses, is a feature of the cranium in most eutherian mammals (Novacek, 1993). The sinuses are bony cul-de-sacs connected to the nasal cavity (Fig. 1); there are four such sinuses found in the facial skeleton of primates (Cave and Haines, 1940). The names of the maxillary, sphenoidal, ethmoidal, and frontal sinuses are taken from the (principal) bones they occupy. These hollow spaces develop as outgrowths of the mucosa that line the nasal capsule during ontogeny, and traditional anatomical definitions (e.g., Negus, 1958) of the sinuses are derived from the distinctive placement of their openings, or ostia, into the nasal cavity, although other classificatory systems have been proposed (e.g., Rossie, 2006).
The presence of sinuses has intrigued scholars for millennia, including such luminaries as Hippocrates (Stierna and Westrin, 1999), Vesalius (Blanton and Biggs, 1969), and da Vinci (Clark, 1968). It is perhaps surprising, then, that a concerted effort to unravel the mysteries of sinus growth, distribution, and possible function did not emerge until the advent of modern imaging techniques (Rae and Koppe, 2002), although a handful of previous studies developed significant insights into paranasal pneumatization in primates (Paulli, 1900; Cave and Haines, 1940; Ward and Pilbeam, 1983; Lund, 1988). Since computed tomography (CT and microCT) imaging have become commonplace, however, there has been a sharp increase in our knowledge of both the growth and development (Koppe and Nagai, 1997; Koppe et al., 1995, 1999; Rossie, 2006) and size and distribution (Koppe and Ohkawa, 1999; Rae and Koppe, 2000; Rae et al., 2003; Rossie, 2003; Nishimura et al., 2005) of sinuses in both extant and extinct (Rae, 1999; Spoor and Zonneveld, 1999; Rossie et al., 2002; Rossie, 2005; Nishimura et al., 2007; Rae et al., 2007) primates.
One of the most contentious issues in the study of craniofacial pneumatization is the possible function of the paranasal sinuses; on this topic, there is little consensus (Rae and Koppe, 2004). At present, there are nearly as many hypotheses as to the biological role of the sinuses as there are investigators. Table 1 lists a few of the mooted functional explanations for the presence of sinuses. Although few of these “functions” have been tested critically, obvious counter-examples exist for the majority of the proposed explanations of sinus function (Blanton and Biggs, 1969). The two most favored explanations for sinus presence and/or variation are that they contribute in some way to the conditioning of inspired air, or that they are the result of biomechanical forces acting upon the cranium.
Table 1. Probable functions of paranasal sinuses
Modified after Blanton and Biggs (1969) and Witmer (1997).
1. Olfactory function (in certain mammals)
2. Respiratory function (heating/humidifying air)
3. Thermoregulatory function
4. Resonance to the voice
5. Floatation devices
6. Role in facial ontogeny
7. Balancing head on neck
8. Trauma protection
9. Buttressing of the skull
10. Providing maximal strength with minimal material
11. Reduction of skull weight
12. Increasing facial dimensions for the origin of cranial muscles
13. Evolutionary remnants (vestigial)
The idea that paranasal sinuses contribute to the warming and humidification of air before it reaches the lungs (e.g., Coon, 1962) has been dealt severe blows by comparative work on humans and monkeys. Humans living in cold environments tend to have smaller maxillary sinuses the further north they live (Shea, 1977), and the same is true for Japanese macaques (Rae et al., 2003). Also, physiological work has shown that the transfer of air from the sinuses to the nasal cavity happens at a rate that is too low to contribute substantially to conditioning air (Proetz, 1941).
The other most frequently cited explanations for paranasal sinuses is that their presence and/or size is related (in some way) to biomechanical stresses and strains, usually associated with masticatory behavior (Weidenreich, 1924; Hofer, 1965; Prossinger et al., 2000). Although rarely stated explicitly, the implication is that the size and/or presence of the sinuses are linked to forces present in the cranium during chewing; the size and direction of these forces is directly related to the degree of paranasal pneumatization (e.g., Preuschoft et al., 2002).
Broad comparisons of sinus presence and/or size and masticatory stress between taxa are contradictory. For example, some pitheciines lack a maxillary sinus (Nishimura et al., 2005; Rossie, 2006), which might suggest that high masticatory stresses associated with hard-object feeding (Martin et al., 2003) may be associated with a reduction or loss of paranasal pneumatization. Pongo, on the other hand, has a maxillary sinus that is indistinguishable in size from other hominoids (Rae and Koppe, 2000), despite a higher reliance on mastication of hard objects (Ungar, 1994).
A number of recent studies using finite element analysis have shed new light on the distribution of masticatory stress throughout the craniofacial skeleton (e.g., Witzel and Preuschoft 1999, 2002; Ross et al. 2005). While these studies principally support the existence of the well-known bony pillars of the facial skeleton, they clearly identify areas in the facial skeleton, such as the maxillary sinuses, that are free from stress (Witzel and Preuschoft, 2002). Indeed, modeling studies have shown that there is little biomechanical difference between primate facial skeletons with sinuses and those without (Smith et al., 2007). The conclusion from these studies is that the thin-walled shells of the maxillary sinuses are the necessary part of the facial skeleton. The corollary to this inference is that changes to these biomechanically important pillars, which impart resistance to high masticatory loads, are responsible for any observable differences in pneumatization (Preuschoft et al., 2002). Thus, mechanical forces do not act directly on the maxillary sinus, but the sinus may be affected by changes in the bony pillars that surround them. Indeed, the individual bony trabeculae observed in human upper jaws are orientated in such a way as to shift the strain caused by mechanical loads away from the maxillary sinus (Wetzel, 1925; Wetzel and Schröder, 1925). This system of trabelular orientation seems to work with a high degree of safety, as crash-tests of the human maxillary sinus floor suggest (Wetzel and Schröder, 1925).
To test the hypothesis that masticatory stress has an effect on the paranasal sinuses, we adopt the narrow allometry approach (Conroy, 1987) of comparing closely related species, to eliminate the confounding factor of phylogeny. Here, two species of the genus Cebus (C. albifrons and C. apella) are compared; the former is a primary frugivore (Masterson, 1997), and the latter is classified as a hard-object feeder (Dumont, 1995; Scott et al., 2005), and the form of their crania and mandibles reflect this difference in diet (Cole, 1992; Daegling, 1992).
MATERIALS AND METHODS
A mixed sex sample (Table 2) of adult dry crania comprising 16 individuals (eight C. albifrons, eight C. apella) from the collections of the Primate Research Center (Kyoto University, Japan) and the Japan Monkey Center (Inuyama, Japan) was subjected to computed tomography (CT) scanning. Serial coronal scans were taken at 1-mm intervals on a HiSpeed Advantage RP CT scanner (General Electric Medical Systems, Waukesha, WI) with intensity settings of 120 kV and 150 mA (Fig. 2). Sections were digitized on an ALLEGRO graphics workstation (ISG Technologies, Mississauga, Ontario, Canada) using the ALLEGRO software package to create virtual three-dimensional reconstructions, from which sinus volumes can be obtained directly. Visualization was performed with Amira 4.0 (Mercury Computer Systems, Chelmsford, MA); see Fig. 3.
All specimens are adult, as judged by dental eruption.
JMC, Japan Monkey Center, Inuyama; PRC, Primate Research Center; Kyoto University.
?, indeterminate sex.
To compare the relative size of the maxillary sinuses, sinus volume was scaled by two measures of cranial size. In the first instance, volumes of the right maxillary sinus (in cc) are divided by basicranial length (basion-nasion), measured with sliding calipers. This is a standard measure of cranial size, although it may not be completely appropriate for scaling paranasal pneumatization (Rae and Koppe, 2000). The second analysis scales sinus size by an approximation of facial volume; raw sinus volumes were divided by the product of palatal length (orale-staphylion), facial height (nasion-prosthion), and bimaxillary width (zygomaxillare-zygomaxillare; Rae and Koppe, 2000). Both relative sinus size data sets were analyzed by t-tests with a significance level of P < 0.05 using SPSS for Windows 12.0.1 (SPSS Inc., Chicago, IL).
Sinuses in Cebus
The paranasal cavity system of Cebus consists of two sets of paired sinuses: maxillary and frontal sinuses. The maxillary sinus pneumatizes almost the entire maxilla beyond the maxillary canine. Several recesses, including a palatal recess, a frontal recess and a zygomatic recess, enlarge the size of the maxillary sinus. The sinus is in close proximity to the maxillary molars. In adults, the roots of the maxillary molars regularly project into the maxillary sinus floor. The paired frontal sinus pneumatizes part of the interorbital septum and enlarges into the supraorbital torus; large parts of the orbital roof are also pneumatized. Whether ethmoidal air cells are also developed in Cebus is unclear from the available CT scans. As the homology of the frontal sinuses in platyrrhines and catarrhines is currently subject to debate (Rossie, 2003), only the maxillary sinuses are considered here.
Sinus Size in Cebus
Statistical comparisons between the relative volumes of the maxillary sinuses of the two species are summarized in Table 3. Independent sample t-tests for both scaled variables show no significant differences between species at P < 0.05. Figure 4 shows box-and-whisker plots for each analysis. Thus, it can be concluded that the presence of higher masticatory stresses and strains in the facial skeleton do not have a significant effect on the degree of pneumatization in cebid primates. It is worth noting that the size of the samples is somewhat modest, increasing the chance of Type II error. Even given this, however, the degree of overlap between the ranges is compelling.
Table 3. Results of the statistical analyses
Std. error difference
C. apella v
C. albifrons (basicranial length)
C. apella v
C. albifrons (facial volume)
The above demonstration that masticatory stress models fail to explain the size and/or presence of paranasal sinuses in anthropoid primates leave a lacuna in our understanding of the architecture of the anthropoid craniofacial skeleton. One alternative explanation may be consideration of space-related trade-offs in cranial form, which have been advocated previously as explanations for the existence/function of craniofacial pneumatization. This type of explanation treats paranasal sinuses essentially as “spandrels” (Gould and Lewontin, 1979), either dead-end spaces created by the structural incongruence between various cranial capsules (Weidenreich, 1924) or interspaces between cranial organs such as the brain and the orbits (Hofer, 1965; Hershkovitz, 1977). For example, Cacajao and Saimiri seem to lack a true maxillary sinus (Rossie, 2006), unlike other platyrrhines (Hershkovitz, 1977; Nishimura et al., 2005, Rossie, 2006); this may be due to spatial limitations caused by large orbits and a distinct inferior nasal meatus in these genera, which could hinder secondary pneumatization. Using the “trade-off” approach, the well-developed maxillary sinus of Cebus could be interpreted as a consequence of the combination of a small nasal cavity breadth and relatively wide palate (Rossie, 2006).
There are obviously ways to counteract the high forces generated by the masticatory muscles during feeding of hard objects (Antón, 1996) other than changing sinus volume. For example, C. apella is known to possess thick enamel caps on its molar teeth, a condition which has been linked to hard object feeding (Dumont, 1995). Unfortunately, the other platyrrhine hard object specialists, the pitheciines, posses thin enamel; indeed, processing a mechanically tough diet does not require enamel of any particular thickness (Martin et al., 2003) and other aspects of C. apella dental morphology are not specifically associated with its tough diet (Wright, 2005). More convincing are the morphological differences of the craniofacial skeleton observed in C. apella and C. albifrons, which seem to be evident relatively early in ontogeny (Cole, 1992).
Paranasal sinuses grow within the framework of the craniofacial skeleton. During ontogeny they seem to pneumatize bone as much as is possible (Witmer, 1997). Although there is ample evidence that the distribution of mechanical loads within the facial skeleton has a major impact on craniofacial morphology (e.g., Antón, 1996), it is questionable whether the same holds true for paranasal sinus volume. In this context it is interesting to note that even humans with cleft lips and palates possess a maxillary sinus that can be predicted from the size of the facial skeleton (Koppe et al., 2006). These observations, together with the results of this study, support the idea (Hylander and Johnson, 1997) that the bony structures of the facial skeleton cannot be understood solely as a means of counteracting masticatory stress. As a result, the biomechanical hypothesis of sinus function can be considered falsified. It is only with continued explicit testing of proposed sinus functions that we can hope to understand these fascinating “empty spaces”.
We thank Y. Hamada (Inuyama, Japan) for providing the material used in this study. We are grateful to Y. Ohkawa (Okayama, Japan) and S. Hadlich (Greifswald, Germany) for the acquisition of the CT data, and to B. Demes for comments on an earlier version. The final version was vastly improved by comments from Eric Delson and an anonymous reviewer. Finally, we thank S. Marquez for inviting us for this special issue on the paranasal sinuses.