The cranium of the adult giraffe (Giraffa camelopardalis) presents with an unusual morphology in that it possesses an extensive and seemingly comparatively large frontal sinus (Fig. 1). Colbert (1938) noted that the giraffe frontal air sinuses (hereafter referred to as “frontal sinuses”) are “extremely large… [are] above the brain and [extend] into the occiput.” (Colbert, 1938). Churcher (1978) also noted that the giraffe has a “highly developed sinus.” Despite these prior observations, at present no quantification of this expansive sinus in the giraffe, either among adult or immature individuals, has been undertaken. Moreover, the extent to which the frontal sinus in the adult giraffe is “extremely large” has not been quantified in comparison to other artiodactyl species.
The lack of detailed anatomical descriptions or quantifications of the giraffe frontal sinus has not prevented researchers from suggesting several adaptive explanations for the structure. One suggestion is that the frontal sinus is a cooling mechanism for the brain (Ganey et al., 1990; Mitchell and Skinner, 2003), as the head is under constant, direct solar exposure given its position on an elongated neck relative to the height of vegetation in occupied ecozones. Mitchell and Skinner (2003) have further suggested that the giraffe frontal sinus may represent a biomechanical adaptation, where the large sinuses are “an important prerequisite for neck elongation” as they increase “head volume without increasing [the] weight [of the skull].” These authors have also implicated the frontal sinuses in both effectively increasing the surface area for attachment of masticatory muscles (e.g., temporalis), and improving olfaction (presumably by extending into the nasal region, increasing the overall size of the nasal conchae and the overlying olfactory epithelium).
Here, we report the results of the first quantification and comparative analysis of the giraffe frontal sinus using computerized tomography (CT). CT scans and three-dimensional image reconstruction programs have been employed to investigate the ontogeny of the frontal sinus in the giraffe from neonate to mature adult in both morphology and size. Our results are then compared to a sample of CT-reconstructed frontal sinuses from other artiodactyl species to evaluate the untested adaptive hypotheses of the giraffe frontal sinus that have been previously suggested.
MATERIALS AND METHODS
Crania of 14 individuals representing nine artiodactyl species were CT scanned for analysis (Table 1; giraffe ages following Hall-Martin, 1976). Specimens were obtained from the comparative skeletal collection of the School of Anatomical Sciences, University of the Witwatersrand (Johannesburg, South Africa), and the Department of Mammals, Ditsong National Museum of Natural History (Pretoria, South Africa). All specimens scanned, except for four of the six giraffe specimens, represent skeletally mature individuals with erupted third molars in occlusion.
Table 1. Raw data of cranial mass and frontal sinus volume of the artiodactyls studied
Approximate age (years)
Cranial mass (g)
Frontal sinus volume (cm3)
G. camelopardalis m
G. camelopardalis m
G. camelopardalis m
G. camelopardalis f
G. camelopardalis m
G. camelopardalis m
All specimens were scanned with a Phillips Brilliance6 180P3 CT Scanner at the University of the Witwatersrand Donald Gordon Medical Centre (Johannesburg, South Africa). Specimens were positioned with the frontal bone oriented superiorly, resting on the maxillary dentition, and scanned from the premaxilla to the occipital condyles (Fig. 2). Individual specimens were scanned at a slice-depth of 1.0, 1.5, 2, or 3 mm, depending on the overall size of the specimen (larger specimens were scanned at a 3 mm slice-depth).
All CT scans were evaluated using the 3D Slicer 3.5 program (www.slicer.org). Window and level values were manually adjusted for each scan before analysis. The frontal sinuses were identified in a coronal slice and then manually segmented as a label map using the “Editor” function in 3D Slicer. Manual segmentation of the sinus through coronal slices proceeded rostrally and caudally to include any spaces continuous with the sinuses of the frontal bones; therefore in this study, the “frontal sinus” was considered to be one continuous anatomical region and was not subdivided as done in other descriptions of large mammal cranial anatomy (Constantinescu and Constantinescu, 2004). Three-dimensional reconstructions for the crania, frontal sinuses, and endocranial cavities were then produced using the “ModelMaker” tool, which yielded the volumetric data.
To compare both the relative and absolute size of the giraffe frontal sinuses to those of other artiodactyls, cranial mass (minus any associated mandible) was measured. The log-transformed values of cranial mass were then regressed against log-transformed frontal sinus volumes using ordinary least squares regression (OLS). Although we acknowledge the presence of measurement error in cranial mass (X), the use of OLS rather than other bivariate line-fitting methods (e.g., reduced major axis) is justified because the relationship between cranial mass and frontal sinus volume (Y) is asymmetrical (see Smith, 2009).
Initial observations of a sagittally sectioned adult giraffe cranium revealed a large frontal sinus that extends to the caudal end of the cranium, into the occipital bone, completely overlying the endocranial cavity (Fig. 1). CT scanning and reconstruction of the giraffe age series confirmed this caudal sinus extension, and demonstrates how the sinus becomes more prominent with age, hence increasing volume as the animal matures (Figs. 2, 3, 4, 2–4; Table 1). The frontal sinus of adult giraffe is far larger than the adult okapi (Okapia johnstoni; Fig. 4) and is far larger than several other artiodactyls studied (Figs. 4, 5, 6, 4–6). There is also a relationship between the cranial mass and frontal sinus volume for all the artiodactyls studied, but the adult giraffes exhibit a different relationship to that seen in other species.
Anatomy of the Frontal Sinus in the Adult Giraffe
Several previous authors, from as early as Colbert (1938), and in more recent studies (Churcher, 1978; Ganey et al., 1990; Mitchell and Skinner, 2003), have all recorded that the frontal sinus of the giraffe is large and extensive. Our observations show that the adult giraffe frontal sinus is very large compared to the other ungulate species studied, including the okapi, the giraffe's closest living relative. The frontal sinus of the giraffe begins in the frontal bone, just anterior to the dorsal chonchal sinus. The sinus extends caudally into the parietal and interparietal bones and laterally into the temporal bones (Constantinescu and Constantinescu, 2004). Within the parietal bones, the cranial vault expands dorsally, resulting in a dome-shaped region at the bases of the ossicones. The frontal sinus extends into that dome, but as shown in the CT reconstruction (Fig. 2), does not extend into the ossicones. Caudal to the parietal bones, the frontal sinus extends over the endocranial cavity and into the occipital bone. The sinus has numerous plate-like trabeculae (Fig. 1), reflecting the progressive separation of the inner and outer bony mantles as the sinus invaginates into the diploë during frontal sinus development (see below). A bony septum extends from the nasal bones to the occiput along the midsagittal plane that completely divides the frontal sinus into left and right halves (Figs. 1, 2).
Development of the Giraffe Frontal Sinus
The ontogenetic sequence of male giraffe crania (Fig. 3) illustrates the progressive expansion of the frontal sinus caudally and laterally during development as it attains the adult morphology. In the neonate individual (ZA 1265), the frontal sinus extends only partly into the parietal bone and does not extend caudally over the braincase. At this stage of development, there is partial extension of the sinus into the temporal bone (Fig. 3). CT scans of the newborn crania did not reveal the large separation of the inner and outer bony mantles seen in the adult cranium (Figs. 1, 2), rather the region overlying the endocranial cavity in the neonate is composed of a thick layer of diploë.
CT scans of the somewhat older juvenile giraffe cranium (ZA 1253) revealed a greater separation of the bony mantles compared to the newborn. These scans also showed a more caudal extension of the sinus over the brain case as well as a more lateral expansion into the temporal bone compared to the neonate (Fig. 3). In the sub-adult (TM 16429), the frontal sinus was seen to extend over the endocranial cavity in the occipital bone and into the temporal bone to a greater degree compared with the younger juvenile and neonate animals (Fig. 3). Our observations from this developmental series add support to the observations made by Ganey et al. (1990) that the enlarged frontal sinus in the giraffe results from the “frontal, parietal, and supraoccipital bones [growing at] a much faster rate than the rest of the cranium.”
CT analysis has highlighted three regions where enlargements of the frontal sinus occur during the ontogeny of this structure in the giraffe that add considerably to the resulting total volume in the adult. The first enlargement is located at the crest just rostral to, and at the base of, the ossicones (see sub-adult reconstruction in Fig. 3) resulting in nasal and parietal “bulges” in the outer bony mantle. The second enlargement is the lateral expansion of the sinus into the temporal bone (Figs. 2, 3), which while not as extensive as the first is a significant departure from the morphology typically seen in other artiodactyls (see below). The third significant expansion is the caudal extension of the frontal sinus over the endocranial cavity into the occipital bone. These extensions, while seen in the crania of some other artiodactyls examined (see below), differ morphologically in the extensive dorso-ventral separation of the inner and outer bone mantles that accommodate both the caudal expansion of the frontal sinus into the occipital region and the enlargement of the naso-parietal portion of the sinus in the adult giraffe.
The Frontal Sinus in the Okapi
CT scans and reconstructions of the okapi cranium reveal a large frontal sinus relative to most of the other ungulates studied (Table 1; Fig. 4), confirming Colbert's (1938) observation that the okapi frontal sinus is “large [and] anterior to the brain.” The relative size and morphology of the sinus, however, is not as large or extensive as that of the adult giraffe. As in the adult giraffe, the okapi frontal sinus extends superiorly into the dome-shaped part of the parietal bone at the bases of the ossicones (Fig. 4). The frontal sinus does not extend into either the temporal bones, the occipital region, or result in the notable nasal and parietal swellings that occur in adult giraffes.
The Frontal Sinus in Other Artiodactyls
Among the other scanned artiodactyls, frontal sinuses are both present and variable in size and morphology (Table 1; Fig. 5). Generally among the sampled species, as in the giraffe, the sinus is present in the frontal bones just caudal to the dorsal conchal sinus. The frontal sinus typically extends through the parietal bones, and at least partially overlies the endocranial cavity; however, it does not extend laterally into the temporal bones as seen in the giraffe cranium (Fig. 5). In the warthog (Phacochoerus aethiopicus) the frontal sinus extends through the parietal bones and into the occipital bones, effectively pneumatizing the nuchal crest (Fig. 5). In the two sampled bovids, wildebeest (Connochaetes taurinus) and domestic cow (Bos taurus), the frontal sinus extends into the horn core pedicles, and in the latter species extends into the body of the horn core itself (Fig. 5).
The Relationship of Cranial Mass to Frontal Sinus Volume
A statistically significant isometric correlation exists between log-transformed frontal sinuses and cranial mass in the non-giraffe artiodactyls studied (including the okapi) (r2 = 0.693, P = 0.010, slope = 1.08; Fig. 6; Table 1); however there is considerable spread of the artiodactyl values around the best-fit line (standard error of the estimate = 0.28). In the male giraffe ontogenetic sequence, however, very small increases in cranial mass were significantly correlated with very large increases in the frontal sinus volume (r2 = 0.961, P = 0.003, slope = 2.04; Fig. 6; Table 1), resulting in a radically different regression slope indicating strong positive allometry between the two variables. There is little scatter of points around the best-fit line (standard error of the estimate = 0.15). Although the single neonate giraffe specimen (ZA 1265) falls near the general artiodactyl regression line, the next oldest immature individual (ZA 1253) deviates from the artiodactyl line along the same trajectory as the other, more mature giraffe specimens. Interestingly, the absolute, but not relative, size of the frontal sinus of the single adult female giraffe specimen included in this analysis is smaller than that seen for the adult males (Fig. 6; Table 1). While only one specimen, this may reflect sexual dimorphism in the absolute size of the frontal sinus within giraffes, likely related to overall differences in body size (Skinner and Chimimba, 2005). The strong, positive correlation within the male giraffe subgroup, and the relatively large size of the single female specimen, indicates that the large relative size of the frontal sinus in giraffe is a species-specific phenomenon at nearly every period in male ontogeny and for adults of both sexes.
The current series of analyses and comparisons indicate that frontal sinus of the giraffe is both absolutely and relatively large in comparison to a sampling of eight other artiodactyls, and morphologically unique in the expansion of the sinuses through the temporal, parietal, and occipital bones. In the giraffe there are dramatic increases in frontal sinus volume with small increases in cranial mass during ontogeny. The findings and comparisons made in this study indicate that the size of the frontal sinus in the giraffe may have evolved to serve an alternative function to that of the frontal sinuses found among the other sampled artiodactyls.
Interpreting the observed variation in frontal sinus dimension and morphology across the sampled artiodactyls, and the interrelationship of this structure to cranial mass, is not a straightforward task. The enlargement of the frontal sinus into the parietals and nuchal region of the occiput in the warthog (Fig. 5) may be an economical means for enlarging the nuchal attachment area for their species-specific biomechanical needs without substantially increasing cranial mass through the addition of solid bone. The warthog possesses an elongated cranium and enlarged canines near the rostrum which shifts the weight of the head anteriorly. Furthermore, warthogs engage in several behaviors, from pushing duels to using the rostrum as a lever in substrate during feeding and burrow digging, which regularly engage the nuchal musculature beyond simple postural support (Cumming, 1975; Kingdon, 1979; Estes, 1991).
Although the occurrence of frontal sinuses in the other sampled artiodactyls may provide some biomechanical benefit by reducing overall cranial mass, they do not appear to directly influence the size and shape of the nuchal region as occurs in the warthog. Given the modest and isometric expansion of frontal sinuses with increasing cranial mass observed among our artiodactyl sample, the variation in frontal sinus morphology between these species likely reflects specific cranial adaptations, such as increasing cranial vault size for muscles of mastication, horn core support, or the development of tribe- or species-specific secondary sexual characteristics.
Integrating Frontal Sinus Volume and Cranial Mass in Giraffe
The morphology of the frontal sinus in both male and female adult giraffe, with extensive intrusion into both the occipital and temporal bones, is unparalleled among the other artiodactyls considered here. Both Ganey et al. (1990) and Mitchell and Skinner (2003) have suggested that the extensive giraffe frontal sinus may assist in thermoregulation of the brain in a head that is constantly under solar exposure. Thermoregulation as a possible explanation for an extensive sinus is problematic due to a lack of adequate airflow within the sinus, as no significant, direct openings between the frontal sinuses and the nasal cavities were observed that would sustain air exchange to support this conclusion. Moreover, as in most artiodactyls, there is a rete mirable vasculature within the base of the skull (Mitchell and Skinner, 1993) and this is more likely to be an efficient regulatory of brain temperature than air circulating through the sinuses (see also discussion in Mitchell and Skinner, 1993).
Mitchell and Skinner (2003) have also suggested that the expansive giraffe frontal sinus effectively reduces the mass of the cranium while maintaining an overall “large head volume” necessary for the support of masticatory musculature (e.g., the temporalis) and function. By itself, the frontal sinus does reduce overall cranial mass than if the region were solid bone, although some mass is retained through the presence of extensive trabeculae in the sinus. It appears unlikely, however, that the increased cranial volume due to the enlarged giraffe frontal sinus is an adaptive requirement for, or directly supports, masticatory musculature. First, the temporalis muscles do not extend over the area most affected by the enlarged frontal sinuses, which as noted above increase the surface area of the dorsal-most portions of the cranium. Second, the most common adaptation among mammals requiring greater temporalis muscle attachment area than afforded by the neurocranium is the development of a sagittal crest, which forms in response to the two muscle bellies approaching the midsagittal plane as the individual matures (Washburn, 1947; Ashton and Zuckerman, 1956). While less common in artiodactyls than in carnivores and some primates, sagittal crests do occur on hippopotamus (Hippopotamus amphibious), camel (Camelus dromedarius) and chevrotain (e.g., Hyemoschus aquaticus) crania, which suggests that sagittal crest development would be a more predictable adaptation for supporting the temporalis muscles than hypertrophied frontal sinuses. Third, giraffes browse on quality foliage, and have evolved a unique and highly effective stomach and digestive tract, even relative to the other ruminants, to process nutrients from their food (Kingdon, 1979, 1997). Giraffes have been noted as relying less on mastication than other ruminants during browsing (Kingdon, 1979, 1997), which further suggests that support for the muscles of mastication were not a significant selective pressure for the evolution of such enlarged frontal sinuses.
In terms of the more behavior-based hypotheses, Von Muggenthaler and Baes (2001) have presented data indicating that giraffe produce infrasonic vocalizations, and suggested that the extensive frontal sinus may act as a resonance chamber for the production of such low frequency sounds. The infrasonic vocalizations in the giraffe were recorded during “neck stretch” and “head throw behavior” (von Muggenthaler and Baes, 2001). While it is possible that the enlarged giraffe frontal sinus is related to vocal resonance, von Muggenthaler and Baes (2001) have also reported infrasonic vocalizations in the okapi which lacks a giraffe-like frontal sinus (in either size or morphology). It is possible that the variation in sinus volume is explained by species-specific differences in the infrasonic frequencies or sound levels produced during vocalization, with the giraffe simply being capable of producing vocalizations of lower frequency or greater sound levels than the okapi. Such a hypothesis linking vocalizations with the frontal sinus requires further independent testing in both species, but raises an interesting possibility for explaining the giraffe frontal sinus morphology.
The enlarged dorsal prominences, produced as a by-product of the enlarged frontal sinus near the ossicones, are important secondary sexual characteristics among giraffes that feature prominently in several giraffe behaviors. The external boney table over the frontals and parietals, and the integument overlying it, progressively develop rugosities and thickenings during adulthood that signal sexual maturation (and come to dwarf underlying cranial structures among older bull giraffes, Kingdon, 1979, 1997; Simmons and Scheepers, 1996). The dorsal surface of the cranium is aggressively employed during neck-sparring, clubbing and other dominance contests used in competition for mates and the establishment of hierarchies among individuals in bachelor herds (Kingdon, 1979; Simmons and Scheepers, 1996). Such secondary sexual characteristics and functions may explain the tentative sexual dimorphism observed in the current study, where the adult female giraffe had a smaller absolute frontal sinus volume than the adult male giraffe. This conclusion may be offest by the fact that adult male giraffes are generally larger than adult females giraffes (Skinner and Chimimba, 2005), thus the size difference of the frontal sinus may be a purely isometric effect, this being supported by the observation that the female giraffe does not differ significantly from the allometry displayed by the developing males. Despite this, the size and shape of the frontal sinus in the giraffe may represent an adaptation for increasing the effective surface area for cranial contact during such aggressive behaviors without increasing cranial mass. Further work determining the size of the frontal sinus in more specimens of giraffe, and specifically an ontogenetic sequence of females, would determine the extent of sexual dimorphism and the period during development when this potential dimorphism may arise.
Despite this, it is difficult to use intrasexual selection as an explanation for the overall large size of the giraffe frontal sinus in both sexes. While intrasexual competition can lead to the exaggeration in the size of a particular structure in the selected sex, and through genetic pleiotropy lead to an increased size of the homologous structure in the non-selected sex (Lande, 1980), there is no established criterion for determining the level of dimorphism that must be present to explain the evolution of a feature purely on the basis of male-male competition. In other words, is the potential dimorphism in frontal sinus volume between male and female giraffes of such a size that intrasexual selection fully explains the evolution of the overall large size of the sinus for the species as a whole? While intra-sexual selection may be adequate to explain size dimorphism in extant giraffe, further assessment of frontal sinus morphology among the artiodactyls and extinct giraffe, with an emphasis on sex-related differences and functional interrelationships, is required before other potential selective pressures (such as the development of infrasonic communication or other yet unproposed or unknown adaptive functions) can be ruled out in favor of sexual selection.
The authors thank the School of Anatomical Sciences, University of the Witwatersrand, Johannesburg, for allowing us the access to the specimens and for providing transportation to and from the WITS-Donald Gordon Medical Centre for CT scanning. They also thank Teresa Kearney at the Department of Mammals, Ditsong National Museum of Natural History, Pretoria, South Africa, for allowing access to the giraffe and okapi crania. They thank the MRI staff at the WITS-Donald Gordon Medical Centre, Mark Haagensen, and Claire Gibbs, for their invaluable assistance in scanning the crania.