Climatic Effects on the Nasal Complex: A CT Imaging, Comparative Anatomical, and Morphometric Investigation of Macaca mulatta and Macaca fascicularis



Previous studies exploring the effects of climate on the nasal region have largely focused on external craniofacial linear parameters, using dry crania of modern human populations. This investigation augments traditional craniofacial morphometrics with internal linear and volumetric measures of the anatomic units comprising the nasal complex (i.e., internal nasal cavity depth, maxillary sinus volumes). The study focuses on macaques (i.e., Macaca mulatta and Macaca fascicularis) living at high and low altitudes, rather than on humans, since the short residency of migratory human populations may preclude using them as reliable models to test the long-term relationship of climate to nasal morphology. It is hypothesized that there will be significant differences in nasal complex morphology among macaques inhabiting different climates. This study integrated three different approaches: CT imaging, comparative anatomy, and morphometrics—in an effort to better understand the morphological structure and adaptive nature of the nasal complex. Results showed statistically significant differences when subsets of splanchnocranial and neurocranial variables were regressed against total maxillary sinus volume for particular taxa. For example, basion–hormion was significant for M. fascicularis, whereas choanal dimensions were significant only for M. mulatta. Both taxa revealed strong correlation between sinus volume and prosthion to staphylion distance, which essentially represents the length of the nasal cavity floor—and is by extension an indicator of the air conditioning capacity of the nasal region. These results clearly show that climatic effects play a major role in shaping the anatomy of the nasal complex in closely related species. The major influence upon these differing structures appears to be related to respiratory-related adaptations subserving differing climatic factors. In addition, the interdependence of the paranasal sinuses with other parts of the complex strongly indicates a functional role for them in nasal complex/upper respiratory functions. Anat Rec, 291:1420–1445, 2008. © 2008 Wiley-Liss, Inc.

A continuing question in anthropology has been the adaptive significance of the human nasal complex, defined sensu latu as the nose and paranasal sinuses, and of the resultant possible climatic influences on its shape and form in present day populations. In addition, the extent that climate may have had on this seminal region in producing variation among populations, or in the path of evolution itself, has been at the core of many studies dating back to the eighteenth century. Indeed, although many central figures in anthropology and anatomy have produced detailed investigations in an attempt to assess the relationship of climate to nasal complex parameters, little basic biologic data has emerged that has allowed for the understanding of the functional anatomy of the region.

Previous investigations of the effects of climate on the nasal region have largely focused on external nasal measures, and have failed to consider internal structures (see historical background below). Such structures include: nasal cavity and paranasal sinus volumes, turbinal characteristics, choanal dimensions, and palatal length and width measures. Together, thepotency of this robust suite of measures may offer a more meaningful assessment of the region than previously available. This study integrates three distinct methodological approaches—comparative anatomy, morphometric assessment and CT imaging—in an effort to capture the basic biology underlining the morphological design and the anatomico–physiologic functions of this critical region. The examination employs a primate model using macaque groups living at high and low altitudes, thus providing a ‘natural experimental’ setting that allows evaluation of climatic effects on the nasal complex.

The primary hypothesis to be tested in this study is that differences in climate will show a significant correlation with differences among and between units of the nasal complex. The alternative, null hypothesis is that changes in anatomy will not reflect climatic differences. A second hypothesis to be tested is that shape and morphology of the nasal complex is determined to a considerable extent by respiratory related factors not by a predetermined genetic template. The third hypothesis to be tested is that the shape and morphology of the paranasal sinuses is significantly influenced by respiratory factors and not solely by dento-gnathic influences as has often been expressed in the literature (e.g., Keith, 1902).

Historical Background

Since the eighteenth century (e.g., Blumenbach, 1775; Broca, 1872, and Topinard, 1876; see Márquez et al., 2002 for review), indices of the nasal region (e.g., nasal index defined as maximum nasal breadth/nasal height * 100) have been widely used for distinguishing various geographic human population groups. These distinct differences were often related to physiologic adaptation in response to climate. For example, Holland (1902) reported on nasal index values of the Kanet from India living in two distinct habitats. The Kanets of Kullu, inhabiting the fertile valleys of the Punjab, presented a mean nasal index value of 74.1 whereas the Kanets of Lahoul, living along the barren sides of mountainous regions at an altitude of 10,000 feet, showed a nasal index value of 66.4. Although Holland did not report a link between climate and nasal form, Ales Hrdlicka was the first to call attention to an association of climate with bony nasal morphology (see also Hylander, 1972). In the case of Inuits (formerly referred to as “Eskimos”), he suggested that a typically narrow nasal aperture was directly related to the effects of the Arctic cold but did not address the functional significance of this narrowing (Hrdlicka, 1910). A number of studies focusing on the nasal index followed Hrdlicka's finding, and showed that populations from hot, humid climates generally have a larger nasal index value than populations from cold, dry climates. Studies by Thomson and Buxton (1923), and later by Davies (1932), showed Inuits presenting with a rather low nasal index reflecting a narrow but high nasal skeletal aperture whereas sub-Saharan populations displayed a high nasal index value exhibiting wide but short nasal apertures. In an effort to identify which climatic factors were influencing nasal form, Weiner (1954) suggested that because correlations between nasal index and temperature and relative humidity are not as high as the correlation between nasal index and absolute humidity, the key variable operating on nasal shape (i.e., the nasal index) is not temperature/relative humidity, but instead is absolute humidity. Wiener concluded, based on a 0.82 correlation found in his 1954 study, that absolute humidity was the operative critical factor in determining nose form and, indeed, some studies appear to support this finding (see for example Carey and Steegmann, 1981; Hall et al., 2004).

Studies began to question whether the nasal index alone can be a reliable measure of adaptation. Wolpoff (1968) reported that the nasal height measure does not correspond to any reliable aspect of internal respiratory passageway, and as such it may have no validity in such studies whereas nasal breadth has more of a biological value because it has a corresponding respiratory counterpart—namely the entry portal of the upper respiratory system. Wolpoff (1968) generated a model of nasal function on the effects of climate on the skeletal nasal aperture using Inuits and Australians as his sample and found that nasal breadth was the better indicator of climatic influences on the nasal region. Studies since Wolpoff began to explore other variables that centered on the nasal region in an effort to understand the associated selection pressures brought about by climatic conditions. Carey and Steegmann (1981) reported on the relationship between protrusion of nasal form in the sagittal plane and climate using the dataset of Woo and Morant (1934) in their global sample from 55 skeletal human populations. Findings from their study showed the strongest correlations are between absolute nasal projection and absolute humidity. Measures of non-nasal facial width and facial protrusion did not correlate strongly with environmental variables leading the authors to conclude that the nasal capsule operates somewhat independently from the rest of the face. Franscisus and Long (1991) exhaustive study on the components of the nasal index (using Howells' dataset of 2,048 crania representing 26 distinct regional populations) found that both nasal breadth and height exhibit equivalent intrinsic variation among populations, but intrinsic variation in nasal breadth was greater than that of nasal height among populations. The authors concluded that their findings support, but do not demonstrate, the variation of the human nasal index as an indicator of natural selection.

Relatively few studies have examined internal measures that address the dimensions of the nasal cavity proper, internal naris, or paranasal sinuses, in relation to the possible interphase with climate. Among those that have, Wolpoff's (1968) study incorporated measures such as palatal width and length dimensions - surrogate measures for the nasal cavity floor that may be important indicators of the nasal complex because the inspiratory airflow passing through this region requires both temperature modification and complete humidification. Remarkably, there is almost no study that has addressed the possible role of internal naris dimensions in this regard. The paranasal sinuses have been used in studies investigating the effect of climatic conditions on them. Koertvelyessy (1972), for example, examined the relationship between frontal sinus and climatic conditions among Inuit populations. He found smaller frontal sinus areas associated with populations living in low wind chill equivalent temperatures. Shea (1977) followed with an investigation of 261 Inuit crania representing eight populations and reported smaller sinuses found in populations living in colder environments, whereas Rae et al. (2003) reported similar findings focusing on clinal variation of the maxillary sinus in M. fuscata, and from a more recent experimental study of cold stress on Rattus (Rae et al., 2006).

Aspects of nasal morphology—including the paranasal sinuses—have been important features in the description and analysis of both hominoid and homin fossil material and the evolution of human craniofacial form (e.g., Conroy et al., 1978; Ward and Pilbeam, 1983; Olson, 1985; Brown et al., 1985; Johanson et al., 1987; Koppe et al. 1997). For example, the great English otolaryngologist Sir Victor Negus reported paranasal sinuses to be a characteristic cranial feature of placental mammals, and in particular the maxillary sinus, which may be a primitive eutherian characteristic (Negus, 1957, 1958). Paranasal sinuses are such a prominent feature of the modern human face that their presence has been suggested as a paleontological marker for our own species, Homo sapiens (Stringer et al., 1979). Their potential significance as a phylogenetic marker may explain why workers have traditionally described these sinuses in fossil material, even though their taxonomic valence has not been fully documented (e.g., Weidenreich, 1941; Leakey and Tobias, 1965; Tillier, 1975; Ward and Brown, 1986; Wood, 1991; Márquez et al., 2001, but see Rossie, 2006; Kukendall and Rae, 2008; Rae, 2008). Begun (1992) utilized maxillary sinus morphology in conjunction with other facial features to assist in characterizing the hypothetical common ancestor of the living great apes. In a comprehensive study in which he details the morphology of sinuses in fossil catarrhines to reconstruct their evolutionary pathways, Rae (2008) concludes that maxillary sinus absence in cercopithecoids is because of the suppression rather than to complete phylogenetic loss. Because sinus components may act in parallel, better characterization of these structures would permit the clarification of the phylogentic implications of each component of the sinus complex. Given the importance that has been imputed to the paranasal sinuses in primate phylogeny, a better understanding of their functional underpinnings is crucial to revealing their adaptive role.


Species Studied

A mixed sex sample of adult wet cadaveric specimens of Macaca fascicularis (N = 31) and Macaca mulatta (N = 11) were used for gross dissection. Heads were frozen in a Fisher Scientific Isotemperature Freezer at –85°C before bisection with a butcher's band saw. The removal of turbinates was done with a no. 10 scalpel blade to expose the maxillary sinus whose medial wall required removal. Photodocumentation of the nasal cavity, particularly its lateral wall for turbinate and maxillary sinus assessment, was taken using a D100 Nikon digital camera with a Zoom Wide Angle AF 17-70 mm f/2.8-4.5 DC HSM Macro Autofocus Lens.

Dry crania of adult male M. fascicularis (N = 35) and M. mulatta (N = 34) were selected from osteological collections housed in different museums (Table 2). Specimens are from the departments of Anthropology and Mammalogy of the American Museum of Natural History, New York (AMNH); Museum of Comparative Zoology (Agassiz Museum), Harvard University (MCZ); Division of Primatology at the National Museum of Natural History, Washington, D C. (NMNH); and department of Mammalogy at the British Museum (Natural History), London (BMNH). The confounding factors of sexual dimorphism and age are controlled for by using adult males (aged determined by upper molar eruption patterns). All specimens were chosen based on the presence of the variables to be studied, the intact nature of all portals, and on the historical records of capture for each of the specimens that fulfills the provenience criteria for our study design (see below). Some skulls had damage because of the preservation of material, or injuries during capture of specimens (e.g., bullet wounds), but were still included in the sample. For example, damaged orbital floors were corrected by either covering them with tape to prevent seeds from escaping, or they were sealed by clay during CT scanning to enclose the sinus for volume determination. Broken skull bones precluded some measures to be taken while skulls with missing molar dentition were included provided that the remaining craniometric landmarks were sufficiently intact.

Macaques, rather than humans, have been selected as our model for studying the relationship between the nasal region and climate in primates. The use of primate models has been used extensively in the study of sinuses as these models allow for a natural experiment scenario to test the varied hypotheses of functional sinus roles that include biomechanical (Rae and Koppe, in press), physiological (e.g., Rae et al., 2003), and the utility of using sinus markers in primate systematics (Rossie, 2006).

The justification of using the macaque model here is that the recency of migratory human populations may preclude using them as reliable models to test the long-term relationship of climate to nasal morphology. Certain macaque groups, to the contrary, have been suggested to be inhabitants of different climates for extensive periods of time on the order of half a million years (Delson, 1980). Although macaques and humans have a number of morphological differences in their upper respiratory configuration—such as the macaque larynx being positioned higher in the neck (Laitman and Reidenberg, 1993), or that they exhibit only one of the four paranasal sinuses (i.e., maxillary sinus)—they do have many similarities including internal nasal turbinate structures and sinus construction. Additional similarities between macaques and modern humans is in their epithelial structure (Harkema, 1991), their similar patterns in the direction of their mucociliary movement within the nasal cavity (Lucas, 1933), and that they are considered microsmatic primates (but see Seibt and Laska, 2001 who strongly suggest that the concept of primates as primarily visual and microsmatic animals should be revised).

Macaca mulatta (the rhesus monkey) will serve as the primary focus of the study because they live at both low and high altitudes (Fooden, 2000). Although M. mulatta may lack the geographic distribution of humans, they can be found in quite diverse climatic habitats. For example, Macaca mulatta ranges from Paktia, Afganistan (Puget, 1971), through much of India (Frick, 1969), and into Indochina (Van Peenen et al., 1969). Of all the macaques, M. mulatta occupies the widest range of habitats: from sea level to 3,000 m in the mountainous areas of Afganistan (Kullmann, 1970), 3,050 m in Nepal (Richie et al., 1978), and even up to 4,300 m in the temperate forests of the Himalayas in northern Pakistan (Haltenorth, 1975). Seasonal variations may be extensive in these high altitude environments, with their 3 month wet seasons (averaging 38 cm of rain), warm dry springs, and freezing winters averaging 6.5 cm of snow (Roberts, 1977). Temperatures vary from below freezing to above 50°C.

A possible complication of relying solely on M. mulatta, however, is that a lowland population may migrate from a highland area, thereby confounding the analysis. As a result, M. fascicularis (also known as the cynomolgus, crab-eating, longtail and kra macaque), a natural lowland dweller has been chosen as an additional comparitor. M. fascicularis has a wide range of distribution in Southeast Asia across 30-degree of latitude and 35-degree of longitude (Fooden and Albrecht, 1993; Fooden, 1995). M. fascicularis has been reported to live in primary forest, disturbed forest, and secondary forest, as well as in riverine and lowland coastal forest (Medway, 1970).

M. fascicularis was selected for comparison as it is usually considered to be closely related to M. mulatta. Indeed, there previously was controversy as to whether M. fascicularis is a subspecies of M. mulatta or a distinct species. At one time, M. fascicularis was considered conspecific with M. mulatta based on a phenotypically intermediate M. fascicularis population from Indochina (Fooden, 1964). However, it is now generally held that these animals do not represent a single intergrading species, but rather two separate species whose ranges came into secondary contact, thereby resulting in their hybridization (Fooden, 1971). Fooden (1980) has placed M. fascicularis and M. mulatta into a single species group that also includes M. fuscata, and M. cyclopis. Unfortunately, morphological, chromosomal, allozyme, and mtDNA-based studies are in disagreement about the relationships among these four taxa: M. fascicularis may well be a conservative Asian macaque rather than a member of the same species group/clade as the other three (e.g., Cronin et al., 1980; Hayasaka et al., 1988; Melnick and Hoelzer, 1993; Morales and Melnick, 1998; Fooden, 2006; Smith et al., 2007; Chu et al., 2007). For example, Chiarelli (1962) reports that M. mulatta and M. fascicularis exhibit similar amounts of lengthening of the X chromosome, whereas Pasztor (1976) states that they cannot be differentiated chromosomally by the G (Giemsa) banding technique. Other studies have shown a high degree of similarity in the proteins between these two species (Weiss et al., 1973). In addition, M. mulatta and M. fascicularis share similar facial dimensions, which make these two species prime candidates for a comparative morphological study. Because of their great craniofacial similarity, differences between aspects of their nasal complex should reflect climatic and not species differences. However, the confounding effects of masticatory adaptations to dietary regimes may contribute to differences in craniofacial morphology between these two taxa and thus may affect nasal complex variation in an incidental rather than a causal manner. So in effect, masticatory compensatory response to dietary staples could potentially indirectly affect nasal complex morphology with differences reflecting an architectural instead of a thermoregulatory basis. To ensure that the interpretation for the differences in nasal complex components will be of a respiratory rather than a biomechanical adaptation, multiple regression statistical treatments will be implemented. This statistical technique allows one to select an independent variable (i.e., a nasal complex component) and regress a set of dependent variables simultaneously. The groups of dependent variables to be studied will include the most likely masticatory dependent variables (i.e., zygoma and maxilla height), respiratory variables (i.e., choanal and nasal breadth), and neutral variables (i.e., external auditory meatus and foramen magnum length). This line of investigation will identify intimate relationships among variables, potentially eliminate the possible explanation of nasal complex differences attributed to biomechanical rather than climatic influences, and support the bivariate linear regression treatment of the sample data.

Quantitative and Qualitative Assessments

Linear measurements of craniofacial and dental parameters and volume determinations of the maxillary sinus and endocranial capacity of dry crania were obtained by using the following methods: (1) standard craniometric techniques, (2) a seed-filling technique, (3) plain film radiography and (4) CT-imaging with two- and three-dimensional (3D) reconstruction and reformatting techniques.

Linear Measurements

Craniodental linear measurements were taken with digital sliding calipers to the nearest 0.1 mm from which three indices were derived yielding a total of 66 measures, including maxillary sinus and endocranial volumes (Tables 1, 3–11). Linear measuresments were obtained directly from the dry crania. The selection criteria for these measurements were based on the following: (a) they have previously been shown to reflect patterns of craniofacial morphology (see McCollum, 2000; Franciscus, 2003), and (b) they are traditional measures and thus allow for direct comparison to previous studies (Buiskstra and Ubelaker, 1994).

Table 1. Linear measures associated with height and width of nasal complex parameters for Macaca fascicularis (in millimeters)
MuseumCat. No.Species12345678910111213
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 1, Superior bimaxillary breadth; 2, inferior bimaxillary breadth; 3, prosthion to left inferior maxillare pt; 4, prosthion to right inferior maxillare pt; 5, prosthion to nasion; 6, intercanine breadth; 7, ectomolar breadth; 8, endomolar breadth; 9, nasal height; 10, nasal breadth; 11, aperture height; 12, nasion to alveolare; 13, basion to nasion.

Table 2. Sample size and source of primate material
  1. Note: Species sample grouped by collection source. AMNH, American Museum of Natural History, New York; MCZ, Museum of Comparative Zoology, Harvard University; NMNH, National Museum of Natural History, Washington, DC; BMNH, British Museum of Natural History, London.

Macaca mulatta11561234
Macaca fascicularis12176035
Table 3. Linear measures associated with height and width of nasal complex parameters for Macaca mulatta (in millimeters)
MuseumCat. No.Species12345678910111213
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • BMNH, British Museum of Natural History; NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 1, Superior bimaxillary breadth; 2, inferior bimaxillary breadth; 3, prosthion to left inferior maxillare pt; 4, prosthion to right inferior maxillare pt; 5, prosthion to nasion; 6, intercanine breadth; 7, ectomolar breadth; 8, endomolar breadth 9, nasal height; 10, nasal breadth; 11, aperture height; 12, nasion to alveolare; 13, basion to nasion.

Table 4. Linear measures associated with depth of nasal complex parameters for Macaca fascicularis (in millimeters)
MuseumCat. No.Species14151617181920212223242526
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 14, prosthion to hormion; 15, prosthion to staphylion; 16, prosthion to alveolar spine; 17, prosthion to sphenobasion; 18, prosthion to basion; 19, staphylion to hormion; 20, choanal height left; 21, choanal height right; 22, choanal width left; 23, choanal width right; 24, prosthion to palatalmaxillary suture; 25, internal palatal length; 26, basion to hormion.

Table 5. Linear measures associated with depth of nasal complex parameters for Macaca mulatta (in millimeters)
MuseumCat. No.Species14151617181920212223242526
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • BMNH, British Museum of Natural History; NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 14, prosthion to hormion; 15, prosthion to staphylion; 16, prosthion to alveolar spine; 17, prosthion to sphenobasion; 18, prostion to basion; 19, staphylion to hormion; 20, choanal height left; 21, choanal height right; 22, choanal width left; 23, choanal width right; 24, prosthion to palatalmaxillary suture; 25, internal palatal length; 26, basion to hormion.

Table 6. Cranial linear measures not directly associated with nasal complex parameters for Macaca fascicularis (in millimeters)
MuseumCat. No.Species2728293031323334353637383940
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • NMNH, National Museum Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; Cat. No., Catalogue number; F, Macaca fascicularis; 27, Bizygomatic breadth; 28, inferior maxillary midpoint to glenoid tubercle left; 29, inferior maxillary midpoint to glenoid tubercle right; 30, inferior maxillary midpoint to inion left; 31, inferior maxillary midpoint to inion right; 32, staphylion to sphenobasion; 33, staphylion to basion; 34, orbital height left; 35, orbital height right; 36, orbital width left; 37, orbital width right; 38, glabella to opisthocranion; 39, alveolon to opisthocranion; 40, prosthion to opisthocranion.

Table 7. Cranial linear measures not directly associated with nasal complex parameters for Macaca mulatta (in millimeters)
MuseumCat. No.Species2728293031323334353637383940
  • a

    Measurement could not be taken because of missing craniometric landmark, or landmarks.

  • BMNH, British Museum of Natural History; NMNH, National Museum Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; Cat. No., Catalogue number; M, Macaca mulatta; 27, Bizygomatic breadth; 28, inferior maxillary midpoint to glenoid tubercle left; 29, inferior maxillary midpoint to glenoid tubercle right; 30, inferior maxillary midpoint to inion left; 31, inferior maxillary midpoint to inion right; 32, staphylion to sphenobasion; 33, staphylion to basion; 34, orbital height left; 35, orbital height right; 36, orbital width left; 37, orbital width right; 38, glabella to opisthocranion; 39, alveolon to opisthocranion; 40, prosthion to opisthocranion.

Table 8. Volumes measures (in cubic centimeters) & derived indices for Macaca fascicularis
MuseumCat. No.Species414243444546
  1. NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 41, endocranial volume; 42, left maxillary sinus volume; 43, right maxillary sinus volume; 44, nasal index; 45, maxilloalveolar index; 46, palatal index.

Table 9. Volumes measures (in cubic centimeters) & derived indices for Macaca mulatta
MuseumCat. No.Species414243444546
  • a

    Index could not be computed because of missing measure.

  • BMNH, British Museum of Natural History; NMNH, National Museum of Natural History; MCZ, Museum of Comparative Zoology; AMNH, American Museum of Natural History; 41, endocranial volume; 42, left maxillary sinus volume; 43, right maxillary sinus volume; 44, nasal index; 45, maxilloalveolar index; 46, palatal index.

Table 10. Dental measures for Macaca fascicularis and Macaca mulatta
MuseumCat. No.Species47484950515253545556575859
  1. MCZ, Museum of Comparative Zoology; BMNH, British Museum of Natural History; 47, mesial distal length M3 left; 48, buccal-lingual length M3 left; 49, inferior superior length M3 left; 50, mesial distal length right; 51, buccal lingual length M3 right; 52, inferior superior length M3 right; 53, mesial distal length M1_M2_M3 left; 54, mesial distal length M1_M2_M3 right; 55, buccal lingual length M2 left; 56, buccal lingual length M1 left; 57, inferior superior length M2 left; 58, inferior superior length M1; 59, buccal lingual length M2 right; *, measurement could not be taken due to missing dentition.

Table 11. Dental measures for Macaca fascicularis and Macaca mulatta
MuseumCat. No.Species60616263646566
  1. MCZ, Museum of Comparative Zoology; BMNH, British Museum of Natural History; 60, buccal lingual length M1 right; 61, inferior superior length M2 right; 62, inferior superior length M1 right; 63, mesial distal length M2 left; 64, mesial distal length M1 left; 65, mesial distal length M2 right; 66, mesial distal length M1 right; *, measurement could not be taken due to missing dentition.


Volume Measures

1) The following protocol for establishing maxillary sinus and nasal cavity volumes for M. mulatta and M. fascicularis has been developed as a way to test some of the current approaches of craniometric analysis. A shortcoming of these prior approaches relates to the difficulty in obtaining accurate volumetric measurements. Volume measurements on these macaque groups were obtained by the use of oil rapeseeds. These seeds were chosen because of their small size and superior packing qualities. Filling the sinuses required that all communicating portals to the nasal cavity be clearly identified and subsequently sealed with cotton or modeling clay (inert materials which do not damage the specimens upon removal). With each cranium positioned on its side, the seeds were then introduced into the maxillary sinus via a funnel placed either through the choana or piriform aperture. Continual gentle tapping of the cranium was done throughout the filling process to ensure that the seeds packed and settled. Any seeds projecting through the opening of the medial wall of the maxillary sinus were leveled off and removed through the piriform aperture. Transillumination was required to confirm the complete filling of the sinus. Crania were weighed before and after seed-filling so the difference obtained represented the weight of the seeds alone. This procedure was repeated three times for each cranium. Data for weight (in grams) was then transformed to a volumetric measurement (in milliliters) by linear regression. The reference regression line was created by use of weights for nine known volumes of seeds (5–45 mL at 5mL intervals), whereby each observation was measured and recorded twice by using two different shaped graduated cylinders. Endocranial volumes were similarly obtained by filling the cranium with seeds through the foramen magnum. Endocranial volumes were taken to standardize measures to body/skull size differences. The linear regression equations used for weight to volume transformations were generated in each of the collection sites where the seeds were utilized to reduce measurement error from the effects of humidity (see Albrecht, 1983). As the outer coverings of the rapeseeds are of an organic nature, humidity factors would effect expansion parameters of the seeds. Generating the linear regression equation on the day, and in the environment, in which volume determinations were conducted ensures reduction in the source of measurement error. The oil rapeseed filling method used here requires inspection of the material to ensure that all foramina, fissures, and internal nasal morphologies are intact. Once approved, the skull is catalogued and weighed on a scale with an accuracy to 0.1 gm.

2) An additional method of determining volumes was employed using plain film radiography. The X-ray method requires two separate plain radiographs per individual: a frontal view to assess the width of the maxillary sinus, and a lateral view to assess the anteroposterior length and height of the maxillary sinus (Fig. 1). Treating the maxillary sinus as a prism shaped object and then following the calculation of Aust and Helmius (1974) volumes are estimated.

Figure 1.

Lateral and frontal plain film radiographs of Macaca fascicularis. Figures a and b shows an adult male (above) and adult female (below). Schematic drawing represents the lateral and frontal view of X-ray used to derive sinus volumes. Note from schematic drawing from lateral view how measures of height (1) and length (2) of the maxillary sinus are derived. Schematic drawing from frontal view shows how to obtain the two dimensions of height (1) and width (3) of the maxillary sinus.

3) Computer tomography (CT) imaging has provided a new and robust means of visualizing sinusal anatomy, which was not available previously (Noyak, 1997; see Fig. 2) and permitted visualization of internal nasal structures in a non-destructive manner (Kelves, 1997; see Fig. 3). To analyze the morphology of the sinuses from the CT images, segmentation of the structures from the volume data set was undertaken (Udupa and Herman, 1991). Segmentation refers to the extraction of the voxels comprising the individual structures (see Spoor et al., 2000). The nasal cavity and maxillary sinus are defined traditionally based on anatomical landmarks (see Cave, 1967, 1973). Instead of developing an analogous strategy for identifying the landmarks in the CT image, physical isolation of the structures was required before scanning. The nasal cavity and maxillary sinus region was enclosed by sealing the choanae, incisive foramen, the medial wall of the maxillary sinus, the pterygomaxillary fissure, and the piriform aperture with clay. The enclosed cavity was then segmented by thresholding the voxels of low intensity (air) from the high intensity (bone and clay) enclosing structures. The burden of the segmentation process is on physically sealing the openings with clay and not on the image processing strategies. To scale and orient the CT image to the physical specimen, small capsules filled with copper sulphate (a highly detectable marker for CT scans) was placed at distinct, non-coplanar, osteometric points on the cranium.

Figure 2.

Frontal view of a 3D CT reconstruction of adult male human (author SM) showing the topographical relationship between frontal sinus (seen in green) and maxillary sinus (seen in purple) to the nasal cavity proper (seen in red). Maxillary sinus is the largest of the four paranasal sinuses exhibited by humans and dominates the midfacial architectural space. Sphenoid sinuses are not visible in this coronal plane.

Figure 3.

Three-dimensional CT reconstruction of the same individual in Fig. 2 shown in oblique parasagittal view where ethmoid (es) and sphenoid air sinuses (ss) can be viewed. Black asterisk indicating frontal sinus and black arrow is pointing to piriform aperture rim where just posterior to it is the site of attachment of inferior turbinate.

Coronal CT scans (1.0-mm slice thickness) of M. fascicularis and M. mulatta crania (samples from AMNH only) were taken using a 9800 HiSpeed Advantage General Electric CT-scanning machine in the Department of Radiology at The Mount Sinai Hospital. The computerized tomograms were digitized and subsequently transferred to a Silicon Graphics Indigo 2-XZ Workstation and then analyzed with VoxelView software that allows for 3D image analyses. VoxelView is a comprehensive volume rendering application program that permits 3D reconstruction of CT scan data. In addition, it allows for specified measures to be made of discrete components thereby isolating internal anatomical structures outside the surrounding area (e.g., turbinates, nasal cavity, maxillary sinus). The volumes calculated from the CT image were compared to the volumes calculated from the seed-filling method. This component of the study provided assessments of maxillary sinus and nasal cavity volumes and thus can provide a quantitative baseline understanding of the contributing role of sinus to nasal cavity volume.

Statistics Employed

There are a variety of methods regarding size standardization and each has been debated (e.g., Martin, 1993). Size standardization utilizing endocranial volume has been historically used and evaluated by a number of investigators (e.g., Bauchot and Stephan, 1969; Martin, 1993). Accordingly, endocranial volume was used to standardize size in this study. Size standardization for volumetric measurements to be used in this study consists of dividing maxillary sinus volumes by endocranial volume for each specimen. Taking the cube root of each of these ratios converts the volumetric data to linear data. In addition, standardization of linear measurements was undertaken by dividing each measurement by the cube root of endocranial volume for each specimen. This allows each raw linear measurement to be standardized as a function of endocranial volume per individual. By standardizing linear and volumetric measurements in this way, both sets of measurements will have uniform units and thus can be compared statistically.

In measuring two samples for a number of variables, one can use the Hotelling T2 statistic to determine whether the set of means for one group differs significantly from that of the other group (Tabachnick and Fidell, 2006). If they do differ significantly, it would reflect the differential effects of the alternative treatments that the two independent samples received namely different climatic conditions. The T2 test is a special case of multivariate analysis of variance just as the Student t-test is a special case of univariate analysis of variance in which two groups comprise the independent variable. Because the T2 test assesses the dependent variables simultaneously, it may be difficult to discern which variables contribute most to a significant difference between the two groups. As the number of variables increases, the explanatory power diminishes for any given variable. For this reason a bivariate analysis was also be conducted to determine whether relationships are present.

The Student t-test provides pair wise contrasts, and discerns which dependent variable(s) contribute most to a significant difference between the two groups for the particular morphological feature. Test for significance will be at the 0.05 level. However, a Bonferoni inequality calculation was required, because this procedure is an adjustment of the probability level for the evaluation of most simultaneous tests. The adjustment takes into account the number of contrast made, usually estimated as level of significance 0.05 divided by number of contrasts, to protect against falsely accepting a significant statistical difference between samples when none exist (Sokal and Rolf, 1981). However, it is possible to have a significant multivariate test despite insignificant univariate tests. This is attributed to the accumulation of the evidence from the individual variables on the overall test. Alternatively, an insignificant multivariate test can occur when some univariate tests are significant because the evidence of a difference provided by the significant variables is swamped by the evidence of no difference provided by the other variables (Manly, 1994). Consequently, a principal component analysis (PCA) was also applied to the data in the form of a correlation matrix. The importance of applying a principal component analysis is in the identification of factors, underlying a set of variables, which finds linear combinations accounting for a relatively large amount of the total variation. PCA, in effect, reduces the number of variables to principal components of a descending order of importance with regard to explaining the variation. If the data contains such information, then each variable, by looking at its loading or eigenvalue, can be studied for its relative contribution to each component (Neff and Marcus, 1980). Hotelling T2 test, Student t-tests, and PCA analysis of the data was performed using the Statistical Analysis System (SAS) for personal computer.


Gross Dissection

Wet specimen gross dissections of M. fascicularis and M. mulatta revealed similar internal nasal morphology in their turbinate system and maxillary sinus configuration (see Fig. 4). Both taxa exhibited similar single scroll inferior and middle turbinate construction. A distinct superior turbinate was lacking but a somewhat small mass of bone appears to meld with the middle turbinate. Each species presented with only one paranasal sinuses, the maxillary sinus. Pneumatization in the frontal, sphenoid, or ethmoid bony regions was not observed. The maxillary sinus was confined within the maxilla above the molar dentition and comprised most of the orbital floor. No specimen exhibited pneumatization of the hard palate. No differences were shown between sexes.

Figure 4.

(a) Right lateral view of nasal cavity wall of adult male M. fascicularis showing hard palate (HP), inferior turbinate (IT), and middle turbinate (MT). (b). Right lateral view of nasal cavity wall of M. fascicularis in a. The middle and inferior turbinates have been removed revealing the internal morphology of the maxillary sinus (black asterisk is within the sinus). Note the margin of the sinus cavity has been cut away. (c). Right lateral view of nasal cavity wall of adult male Macaca mulatta showing hard palate (HP), inferior turbinate (IT), and middle turbinate (MT). (d) Right lateral view of nasal cavity wall of M. mulatta in c showing hard palate (HP) and maxillary sinus (black asterisk within sinus).

Linear Measures

The Shapiro-Wilk statistic (i.e., W-statistic) was used to test for normality since it is appropriate for samples under 2000 (Cody and Smith, 2005). Normality tests were performed on various forms of the data including raw and log transformed. Results showed that no p-values of less than 0.01 were found thus demonstrating that all measures fell within the range of normal distribution.

Reliability Analysis

Results from the traditional approach of conventional X-rays show that it was not reliable in obtaining maxillary sinus volumes because of the inability to define with precision its border (particularly in lateral view), and thus unable to ensure confidence in either accuracy or reproducibility. Both seed-filling and CT imaging were shown to be more reliable methods for obtaining maxillary sinus volumes. For the seed-filling method, a one-way random effects analysis of variance (ANOVA) was conducted for left and right maxillary sinus volume measurements separately. The root mean squared error (RMSE)—a measure of typical intra-object disagreement, and intraclass correlation coefficient (ICC)—an index of reliability associated with this model are reported in Table 12). RMSE values are below 0.1, which is the nominal level of precision with which individual measurements are reported. This indicates that individual measurements are satisfactorily made to this level of precision. The high level of reproducibility is based upon small standard deviations, low variances, and high intraclass correlations between repeated measures for the same individual. The use of oil rapeseeds is preferable because of the small size, spherical shape, and packing qualities of the seed, thereby allowing the mass of the seeds to follow the contours of the irregular 3D shape of the maxillary sinus. Results from CT to derive maxillary sinus volume determinations are listed in Table 13. Volume differences derived from CT and seed-filling methods were evaluated using Student t-tests and were found not to be statistically significant.

Table 12. Seed filling method reliability analysis for sinus volume determination
 Left maxillary sinus volumeRight maxillary sinus volume
  1. Note: RMSE, root mean squared error; ICC, intraclass coefficient.

95% confidence interval(0.89, 0.99)(0.90, 0.99)
Table 13. Maxillary sinus volumes of Macaca fascicularis derived from CT
S: No.SpeciesAMNH Cat. #Left MS (cc)Right MS (cc)

CT Imaging

CT scans analyzed with the use of VoxelView provided a comprehensive image of the morphology of the maxillary sinus by generating 2- and 3D reconstructions. Figure 5a depicts the reconstruction of a 2D coronal slice image of M. mulatta showing maxillary sinus (ms), nasal cavity (nc), endocranial cavity (ec), and clay (cl). Figure 5b shows a 3D reconstruction of Figure 5a structures yielding volumes of 1.4 and 1.5 mL for left and right maxillary sinus volumes, respectively. Although a total reconstruction of the cranium can be generated and viewed at any orientation, the advantage of using the volume rendering program of VoxelView is that it can digitize the space where the maxillary sinus resides permitting its graphical examination in situ and by extracting the sinus from the same cranium (Fig. 6). Visualization of the bony contours of the right maxillary sinus shows that the most posterior excavation occurs in the coronal plane of the maxillary tuberosity whereas the most anterior limit occurs in the coronal plane at the juncture of the first molar tooth closely abutting with the large canine root (Fig. 7). Sinuses appeared mostly spherically-shaped taking up most of the orbital floor region with the average greatest height of the sinus occurring at the midpoint of the third molar dentition whereas the greatest height occurs in the coronal plane at the distal border of M3 for both taxa (Figs. 8 and 9). Sinuses for both primate groups were confined to the maxilla but above the roots of the molar dentition where the lateral extension was confined to the body of the maxilla and the medial extension was restricted to the lateral nasal cavity wall. Although the sinus floor incorporated the molar dentition for both primate groups they also occupied most of the orbital floor. No medial extension into the hard palate was observed.

Figure 5.

(a) A two-dimensional transverse slice showing internal structures of nasal cavity (nc), maxillary sinus (ms), endocranial cavity (ec), and the high density detection of the clay (cl) used to seal the pterygomaxillary fissure. (b) shows a 3D CT reconstruction of figure a.

Figure 6.

(a) Three-dimensional CT reconstruction of an entire M. mulatta cranium. The clay (cl) used to seal the nasal aperture can be seen in this figure. (b) and (c) show the left maxillary sinus of M.mulatta highlighted in purple in situ (b) and extracted from the cranium (c).

Figure 7.

(a) A parasagittal view of a 3D CT reconstruction of M.mulatta showing the relationship of the right maxillary sinus and its contiguous structures. White arrow is pointing to the outer border of the sinus. The floor of the sinus in (a) covers most of the molar dentition. (b) A 3D CT reconstruction of the same specimen from another parasagittal perspective showing the left maxillary sinus. Note the difference in the relative position of the anterior portion of the sinus and the canine seen in b (indicated by white arrow) as compared with (a).

Figure 8.

The three dimensional space of the maxillary sinus of M.fascicularis has been digitized and graphically represented both in situ and extracted in parasagittal view revealing its spherical construction seen in red. Identification of the maxillary sinus in this manner both affords the opportunity to visualize the detailed morphology of the sinus to its contiguous craniodental structures in situ and in isolated form.

Figure 9.

A composite plate showing: (a) a 3D CT reconstructed skull of an adult male M. fascicularis viewed anteriorly and (b) a reference coronal slice transection line (seen in yellow) viewed superiorly. The coronal slice can be reconstructed 3D or presented in 2D (d). Such reconstructions allow quantitative and qualitative sinus assessments.

Statistical Analysis

Principal component analysis (PCA) reveals that the first three components accounted for 64% of total variance for M. fascicularis and 77% for M. mulatta. The loadings of the variables for M. fascicularis and M. mulatta were of homogeneous contribution with the maxillary sinus volume being one of the lower loadings. Results from the bivariate linear regression analysis between the dependent variable of maxillary sinus volume and 30 linear measurements show six measures to be statistically significantly correlated for M. fascicularis, and two for M. mulatta; whereas only one of these measures are common to both taxa. Selected bivariate linear regression plots including r2 and P values are shown in Figures 10–17.

Figure 10.

Bivariate regression plot of total maxillary sinus volume regressed against nasal breadth in M. fascicularis showing a r2 value of 0.20, P < 0.007.

Figure 11.

Bivariate regression plot of total maxillary sinus volume regressed against choanal width (left) in M. fascicularis showing a r2 value of 0.19, P < 0.04.

Figure 12.

Bivariate regression plot of total maxillary sinus volume regressed against basion-hormion in M. fascicularis showing a r2 value of 0.55, P < 0.0001.

Figure 13.

Bivariate regression plot of total maxillary sinus volume regressed against prosthion-staphylion in M. mulatta showing a r2 value of 0.13, P < 0.007.

Figure 14.

Bivariate regression plot of total maxillary sinus volume regressed against prosthion-staphylion in M. fascicularis showing a r2 value of 0.11, P < 0.05.

Figure 15.

Bivariate regression plot of total maxillary sinus volume regressed against prosthion-hormion in M. mulatta showing a r2 value of 0.12, P < 0.05.

Figure 16.

Bivariate regression plot of total maxillary sinus volume regressed against superior bimaxillare breadth in M. fascicularis showing a r2 value of 0.18, P < 0.01.

Figure 17.

Bivariate regression plot of total maxillary sinus volume regressed against inferior maxillare–glenoid tubercle in M. fascicularis showing a r2 value of 0.25, P < 0.002.

To explain statistically, the greatest amount of variance in maxillary sinus morphology, several of the most likely sets of variables were entered into tests for multiple regressions. A statistical significant r2 value was revealed when the two traditional external nasal measures of nasal breadth and nasal height were regressed against total maxillary sinus volume (Table 14). When selected, splanchnocranial and neurocranial variables were added to the above subset and regressed against total maxillary sinus volume, prosthion to basion is shown to be statistically significant only for M. fascicularis (Table 15). Another set included aperture height, staphylion to hormion, and basion to nasion and when regressed against maxillary sinus volume it was shown that M. fascicularis was the only taxa that showed a significant r2 value (Table 16). When a combination of choanal dimensions were entered and regressed against maxillary sinus volume it was M.mulatta that resulted with a statistical significant r2 value. Finally, regressing a subset of only splanchnocranial variables that reflect nasal cavity floor, sinus and internal naris characteristics (i.e., staphylion to hormion, endomolar breadth, maxillary sinus volume, and choanal width) against nasal breadth revealed statistical significance for both taxa (Table 17, 18).

Table 14. Multiple regression result for Macaca fascicularis
Maxillary sinus volume = nasal breadth and nasal height
Summary of Fit
 Mean of Response0.4673R-Square 0.1987
 Root MSE0.0366Adj R-Sq 0.1487
Analysis of Variance
 SourceDFSum of squaresMean squareF StatPr > F
Table 15. Multiple regression result for Macaca fascicularis
Maxillary sinus volume = nasal_breadth, nasal height, prosthion to staphylion, prosthion to alveolon, prosthion to basion, prosthion to nasion
Summary of Fit
 Mean of response0.4673R-Square 0.5042
 Root MSE0.0308Adj R-Sq 0.3979
Analysis of Variance
 SourceDFSum of squaresMean squareF StatPr > F
Table 16. Multiple regression result for Macaca fascicularis
Maxillary sinus volume = aperture height, basion nasion, staphylion to hormion
Summary of Fit
 Mean of response0.4662R-Square 0.5662
 Root MSE0.0274Adj R-Sq 0.5228
Analysis of Variance
 SourceDFSum of squaresMean squareF StatPr > F
Table 17. Multiple regression result for Macaca mulatta
Nasal breadth = maxillary sinus volume, choanal width, staphylion to hormion, endomolar breadth
Summary of Fit
 Mean of Response7.6342R-Square 0.3507
 Root MSE0.9687Adj R-Sq 0.2579
Analysis of variance
 SourceDFSum of squaresMean squareF StatPr > F
Table 18. Multiple regression result for Macaca fascicularis
Nasal breadth = maxillary sinus volume, choanal width, staphylion to hormion, endomolar breadth
Summary of Fit
 Mean of Response6.0858R-Square 0.4884
 Root MSE0.7448Adj R-Sq 0.4202
Analysis of Variance
 SourceDFSum of squaresMean squareF StatPr > F


Methodological Approaches

The traditional approach to investigating sinus morphology has involved linear measures derived from plain radiographic films. In turn, volumes have been mathematically derived from such linear measures. However, reliance upon plain film radiography has its weaknesses as we found in our study. For example, the margins of the maxillary sinus cannot be identified clearly, thus precluding precise determination of landmarks. Dimensional measures are thus by necessity estimates, and volumetric measures based upon these are estimates as well. An additional problem is that use of conventional radiographs requires two separate views per individual, (an A-P for width, and lateral for length and height). This procedure thus compounds the error potential.

Data from this study shows that the seed filling method, although labor intensive, provides a more reliable and reproducible means for obtaining volumetric data. This is partly because of the size and packing qualities of the seeds, which allow them to follow the contours of the 3D shape of the maxillary sinus. Another advantage of the seed-filling method is that it has the ability to provide volumes for very irregular spaces, such as the nasal cavity proper with its turbinates.

CT-imaging markedly increases visualization of maxillary sinuses and internal nasal structures as compared to conventional X-ray. CT-imaging, assisted by VoxelView, also provides multiplanar two-dimensional visualizations, which can be generated and viewed at any orientation by taking the points of intersection of a particular plane with the full series of coronal CT-scans. A 3D reconstruction of the cranium can also be generated, and viewed in any orientation, by taking a variable number of points of intersection of the transverse and sagittal planes with the full series of coronal CT-scans. Both seed filling and reconstructed CT-imaging provide reliable and reproducible volumetric measures, with CT-imaging affording views and details of morphology previously unavailable.

Analysis of Results

Bivariate plots, using maxillary sinus volume (MSV) as the dependent variable for all regressions, offered differing results. The bivariate regression analysis indicated only one statistically significant measure for both M. mulatta and M. fascicularis. This was somewhat surprising, especially in light of their close craniofacial similarity and, as suggested by molecular evidence, their close phylogenetic relationship. Although the regression formula between MSV and nasal breadth for M. fascicularis is significant, the r2 value accounts for about 20% of the variance, leaving the other 80% of the variance to unknown factors. Similarly, MSV regressed against choanal width reveals a significant relationship but again with a low r2 value of 0.11.

One of the more interesting findings was a significant r2 regression score for the measure of prosthion to staphylion. This measure essentially represents the length of the nasal cavity floor, and is by extension an indicator of the air conditioning process of the nasal region. It was significant for both taxa, with r2 values of 0.13 for M. mulatta and 0.12 for M. fascicularis. The finding of common significance of this measure for both primate groups may be the result of allometric effects. However, when prosthion to hormion, essentially the previous measure extended from staphylion to hormion (this measure thus incorporates nasal cavity floor and choanal region) is considered, M. mulatta continues to show a significant r2 value of 0.12, whereas M. fascicularis does not demonstrate a significant association in this measure. This may reflect that choanal dimensions may not be as sensitive to respiratory influences in M. fascicularis as in M. mulatta.

One measure that has been traditionally linked to biomechanical or masticatory regimes is the measure of inferior bimaxillary suture to midpoint of glenoid tubercle. M. fascicularis showed a r2 value of 0.25. On the other hand, a measure that may show mixed signal for thermoregulation or biomechanical adaptation is the superior bimaxillary suture. This measure was significant for M. fascicularis, but had a somewhat low r2 value of 0.18.

Student t-tests analysis of the study revealed many differences between the two groups, especially in nasal complex components. Indices of upper face (i.e., maxilloalveolar index), nasal index, and palatal index (a measure that incorporates the floor of nasal cavity) were found to be statistically significant between the two groups. Aside from nasal complex differences, the two groups of macaques also differed in a number of other craniofacial parameters. Such measures include the cranial length, both superior and inferior bimaxillary breadths, and bizygomatic breadth. Two interesting findings are that nasal breadths were found to be smaller in the lowland group when compared to the highland M. mulatta group, and that M. fascicularis exhibited larger paranasal sinuses than the highland macaque group. The latter result corroborates the finding of smaller sinuses in colder environments from studies of human groups (Shea, 1977), of primate groups (looking at clinal variation of the maxillary sinus in M. fuscata: Rae et al., 2003), and from a more recent experimental study of cold stress on Rattus (Rae et al., 2006). The finding of wider nasal breadths in M. mulatta is contrary to the morphology seen in diverse modern human populations living in colder localities.

Another focus of the study was on the relationship between dentition and paranasal sinus dimension because the floor of the maxillary sinus is related, at least anatomically, to the adult upper and posterior dentition. Although not all dental measures were taken for each specimen because of missing tooth structures, more than 600 dental measures were obtained and analyzed with respect to maxillary sinus size. Bivariate plots for all dental linear measures against MSV were not statistically significant. Moreover, when height, width, and length were multiplied to derive a volumetric surrogate value for the molar tooth, no significant linear regression score was revealed. When the volumetric estimate for all three molars was regressed against MSV, no significant r2 value appeared. Thus, no single dental measure, or combination of measures, could be reliably used to predict MSV for any of the taxa in this study. The fact that no combination of molar volumes showed a significant relationship with MSV was surprising, given that some researchers have proposed a strong relationship between fully erupted molar teeth and adult maxillary sinus form (Underwood, 1910). This finding demonstrated that although dentition has a robust functional relationship with dietary constraints, it does not have such a relationship with respiration and/or the maxillary sinuses.

Perhaps the most important observations arise from multiple regression analyses, which are useful in indicating specific relationships. When a number of splanchnocranial variable sets were regressed against maxillary sinus volume, measures representing the respiratory portion of the nasal region (i.e., aperture height and choanal region) were revealed to be statistically significant for M. fascicularis with a relatively high r2 value of 0.57, whereas no significance was found for M. mulatta. When a subset of variables that included the nasal cavity floor was entered in the multiple regression analysis, and combined with nasal breadth, M. fascicularis was significant with an r2 value of 0.50. What these significant r2 scores may be indicating is that M. fascicularis is more dependent upon the nasal cavity proper with respect to either air conditioning or heat dissipation. M. mulatta had a significant r2 score of 0.35 when nasal cavity floor width and length (endomolar breadth and staphylion to hormion, respectively) were entered in the multiple regression analysis. Here, M. mulatta put more of an emphasis on choanal dimensions, in conjunction with nasal cavity proportions. It appears that nasal breadth and prosthion to staphylion are covarying in the same direction as maxillary sinus size, as reflected by the negative slope for each taxon. This suggests that for M. mulatta nasal breadth, and for M. fascicularis prosthion to staphylion, are the strongest predictors of maxillary sinus size.

A major observation from multiple regression results is that both nasal breadth (a part of the entry portal of the upper respiratory tract) and prosthion to staphylion (the length of the hard palate) relate to aspects of nasal complex structures rather than to other regions of the cranium. This suggests a tight functional coupling of one moiety of the nasal complex (e.g., nasal breadth) with another (e.g., maxillary sinus volume). Indeed, what these covarying relationships may give insight into are functional units within the nasal complex proper that involve both splanchnocranial bones derived from membraneous development and the more conservative endochondral bones. Change in these “functional units” may, in turn, relate to particular climatic or environmental influences. Although further studies are needed to decipher the specific effects of differing environments upon the nasal complex, these results strongly suggest the interdependence of the maxillary sinus with other parts of the complex. Such ties strongly support a functional role for the sinuses in nasal complex/upper respiratory functions.


This study tested three hypotheses integral to our understanding of the basic biology of the effect of climate on craniofacial form. The first, tested whether or not a relationship exists between climate and nasal complex parameters; the second, that respiratory related factors play a significant role in shaping structure above a predetermined genetic template; and the third, that the morphology of the sinuses is influenced by respiratory related factors and not governed solely by dentognathic influences. Our findings in which nasal complex differences were found to be associated with climatic conditions supports the first hypothesis. Given the close genetic affinity between the two primate groups, and yet, presenting with distinctly different nasal complex dimensions supports rather than refutes the second hypothesis. The third hypothesis is supported by the lack of any statistical morphometric linkage between sinus anatomy and cheek dentition. Taken together, this study revealed the biologic emergent properties of the region in which nasal complex components are not static structures but are individual units that are highly sensitive to the forces of selection initiated by climatic conditions. It is important to treat the nasal complex as a fully well integrated functional unit, and not as a cluster of distinct anatomic entities. The study demonstrated the importance of examining all components of the nasal complex simultaneously, even though they may individually be subject to selection at different rates and magnitudes resulting in the distinct distribution of nasal form seen today. Finally, findings from this study strongly support paranasal sinuses having an active functional role rather than serving solely a biomechanical role, or existing as an incidental structure in craniofacial construction.


We thank the following people who granted permission to study the primate cranial material at their institutions: Ross MacPhee of Mammalogy and Ian Tattersall of Anthropology at the American Museum of Natural History; Maria Rutzmoser of the Museum of Comparative Zoology, Harvard University; Linda Gordon at the National Museum of Natural History, Smithsonian Institution; and Paula Jenkins, Natural History Museum, London. Patrick Gannon of Touro University College of Medicine provided study advisory consultation and primate cadaveric material. We also thank Eric Delson of Lehman College, Timothy Bromage of New York University College of Dentistry, Todd Disotel of New York University, and Joy Reidenberg of Mount Sinai School of Medicine for many suggestions and thoughtful conversations regarding aspects of our work. We are most grateful to Daniel Sklare of the National Institutes of Health and Mark Weiss of the National Science Foundation for their consistently positive encouragement, advice and guidance through the years. Finally, we would like to acknowledge the late Les Marcus who advised the various statistical techniques used in this study; his warmth and wisdom is acutely missed.