Phenotypic plasticity, or environmental modulation of the phenotype, is a significant source of morphological variation in populations. Such postnatal responses are generally thought to reflect the functional or adaptive nature of a structure or system of interest (Gotthard and Nylin,1995; Agrawal,2001). Among mammals, the plasticity of masticatory elements, especially the mandibular corpus and temporomandibular joint, to differential loading is well documented in lagomorphs, murids, carnivorans, and primates (Beecher and Corruccini,1981; Bouvier and Hylander,1981,1982,1984,1996a,b; Beecher et al.,1983; Kiliardis et al.,1985; Bouvier,1987,1988; Bouvier and Zimny,1987; Block et al.,1988; He and Kiliardis,2003; Nicholson et al.,2006; Ravosa et al.,2007a,b,2008a,b ). Such studies demonstrate that the external dimensions, cortical bone thickness, trabecular density, and cartilage composition of nondental mandibular elements are influenced significantly by ontogenetic variation in masticatory peak and/or cyclical loads during an organism's lifetime. In this regard, masticatory elements appear to show similar patterns of plasticity as observed in postcranial tissues and structures (Bouvier and Hylander,1981,1982; Lanyon and Rubin,1985; Biewener,1993; Ravosa et al.,2007b,2008b).
Although there is reason to believe that other craniofacial elements are likewise affected, less is known about the functional bases of masticatory plasticity in other regions of the skull. For instance, among Plio-Pleistocene hominins, Paranthropus is distinguished by a heavily developed masticatory complex widely considered as adapted for producing and countering increased masticatory stresses associated with hard and/or tough food items (Tobias,1967; du Brul,1977; Walker,1981; Rak,1983; Demes and Creel,1988; Hylander,1988; Daegling,1989; Constantino and Wood,2007). Current understanding of the primate face from experimental data supports assertions that this region is under strong selective pressures related to masticatory activity (Hylander,1979a,b,c,1988,1992; Hylander and Johnson,1997; Ravosa et al.,2000,2007a,b,2008a,b). The hard palate of Paranthropus, which is relatively “deep” or superoinferiorly tall, further characterizes the genus when compared with its sister taxon Australopithecus (Ward and Kimbel,1983; McCollum et al.,1993).
Several experiments have described diet-induced plasticity in the hard palate, revealing that hard/tough diets tend to result in wider maxillary arches (Beecher and Corruccini,1981; Beecher et al.,1983) and descended orientation of the palate (He and Kiliardis,2003). Nevertheless, there are no experimental data regarding the dynamic influence of mechanical forces on the internal anatomy of the hard palate. With this in mind, this study provides unique experimental evidence on long-term, naturalistic responses to dietary manipulation in an animal model to address the functional role of masticatory loading on palatal anatomy and feeding behavior in living and fossil mammals. In doing so, this study also evaluates suggestions that a deep palate may counter elevated stresses during unilateral postcanine biting and chewing (Rak,1983). Additionally, these analyses have potential implications for research on the functional consequences of the evolution of the secondary palate in mammals (Thomason and Russell,1986; Russell and Thomason,1993) as well as for surgical approaches to pathologies of the human hard palate (Gedrange et al.,2001).
Experimental Model of Craniofacial Plasticity
As palate anatomy largely precludes an in vivo strain analysis, this research examines the long-term, diet-induced plasticity responses of cortical and trabecular bone in the hard palate of growing rabbits during postweaning ontogeny. The domestic white rabbit (Oryctolagus cuniculus) provides a unique opportunity to understand norms of reaction in the mammalian masticatory apparatus, particularly anthropoid primates and other mammals that use significant transverse jaw movements during unilateral mastication. A major benefit of rabbits is that considerable in vivo data exist regarding jaw-adductor muscle activity, jaw-kinematic and jaw-loading patterns, masticatory function during ontogeny, intracortical remodeling, and the relationship between masticatory behaviors and diet (Weijs and de Jongh,1977; Weijs and Dantuma,1981; Weijs et al.,1987,1989; Langenbach et al.,1991,1992,2001; Hirano et al.,2000; Langenbach and van Eijden,2001). Additionally, a number of similarities exist in the masticatory apparatus of rabbits and anthropoids, including a vertically deep face, position, and movements of the temporomandibular joint (Weijs and Dantuma,1981; Crompton et al.,2006) as well as transverse jaw movements and jaw-muscle activity patterns (Weijs and Dantuma,1981; Weijs et al.,1989; Hylander et al.,2000; Langenbach et al.2001; Vinyard et al.,2008). Similar to many mammals, rabbit jaw-adductor activity patterns vary with dietary properties (Herring and Scapino,1973; Luschei and Goodwin,1974; Gorniak and Gans,1980; Thexton et al.,1980; Weijs et al.,1987,1989; Gans et al.,1990; Dessem and Druzinsky,1992; Hylander et al.,1992,2000,2005), such that increased jaw-adductor recruitment results in elevated peak strains along the mandible and higher TMJ reaction forces (Weijs and de Jongh,1977; Hylander,1979a,b,c,1992; Hylander et al.,1998; Ravosa et al.,2000). Lastly, previous work on rabbit mandibular plasticity responses to postweaning alteration of dietary properties and masticatory stresses is consistent with similar experiments in a variety of other mammals (Beecher and Corruccini,1981; Bouvier and Hylander,1981,1982,1984,1996a,b; Beecher et al.,1983; Kiliardis et al.,1985; Bouvier,1987,1988; Bouvier and Zimny,1987; Block et al.,1988; Yamada and Kimmel,1991; Ravosa et al.,2007b,2008a,b).
Based on these important similarities in the form and function of the mammalian feeding complex, the research presented herein utilizes a rabbit sample to examine plasticity of the mammalian hard palate. This study tests the hypothesis that, much as demonstrated with the mandible, variation in the morphology of the hard palate is linked to postweaning variation in masticatory stress. Thus, significant differences should exist in palate thickness, cortical bone thickness, and/or overall palatal bone area between dietary treatment groups. A consideration of the relative degree of palatal plasticity vis-à-vis dietary properties will be facilitated via comparisons with similar data for jaw joints from the same rabbit sample, where increased masticatory loading results in larger joint proportions, greater cortical bone thickness, and elevated hard-tissue biomineralization (Ravosa et al.,2007b,2008a,b).
MATERIALS AND METHODS
To evaluate plasticity of masticatory elements vis-à-vis altered loading levels, 20 New Zealand white rabbits (Oryctolagus cuniculus) were obtained as weanlings (4 weeks old) from an approved domestic commercial source and housed in the AALAC-accredited Center for Comparative Medicine (Northwestern University) for 15 weeks until attaining subadult status at 19 weeks old (Sorensen et al.,1968; Yardin,1974). To control for variation in genetics and thus ensure that observed responses are linked to postnatal loading modification, only siblings were used. Two dietary cohorts of 10 rabbits each were established to induce postweaning variation in jaw-adductor muscle activity and masticatory loads. Weaning was chosen as the starting point for dietary manipulation because plasticity can decrease with age (Hinton and McNamara,1984; Meyer,1987; Bouvier,1988; Rubin et al.,1992) and to minimize the confounding influence of postweaning diets other than those used herein. Weanlings were fed ad-lib comparable amounts of either a “soft” diet of ground pellets to model under-use of the chewing complex or a “tough/hard” diet of Harlan TekLad rabbit pellets supplemented daily with two 2.5-cm hay blocks to model over-use. As the between-cohort comparisons largely accentuate the duration of oral processing (i.e., crushed pellets exhibit similar properties to whole pellets), under-use diet rabbits are posited to more closely resemble normal/non-pathological loading conditions. Indeed, unlike the jaw joints of over-use diet rabbits, the anatomy of under-use diet rabbits is similar to that for a limited number of 6-month old adult rabbits raised on a “normal/control” diet of intact pellets (Ravosa et al.,2008b).
The inclusion of pellets in the diet of all weanling rabbits ensured adequate nutrition for normal growth. In this regard, behavioral analyses and observations indicate that under-use diet rabbits neither exhibit failure to thrive nor did they develop incisor malocclusions; 90% of the under-use diet sample is within the skull-length range for the 10 over-use diet rabbits (Ravosa et al.,2007b). This suggests the absence of a nutritional discrepancy between dietary cohorts that could have deleteriously affected skeletal growth. Procedures for dietary manipulation, animal monitoring, and euthanasia under heavy sedation were conducted in accordance with an ACUC-approved protocol.
Material Properties of Experimental Foods
Using a portable food tester (Darvell et al.,1996; Lucas et al.,2001), the material properties of pellets and hay were assessed (Table 1) and monitored to assure consistency (Wainright et al.,1976; Vincent,1992; Lucas,1994; Currey,2002). The elastic, or Young's, modulus (E) is the stress/strain ratio at small deformations, characterizing the stiffness or resistance to elastic deformation. Toughness (R) is an energetic property describing the work performed propagating a crack through an item. Hardness (H) is used to quantify indentation. Although it is not feasible to directly measure the material properties of crushed pellets because of the specifications of the above food tester, crushed pellets differ little from intact pellets. This is because the difference exists in the scale of the food particles, not in their physical properties. The notable distinction is that whole pellets entail greater repetitive loading because of a longer processing time. Thus, the sequence from crushed pellets to whole pellets with hay tracks a diet with longer preparation time and greater elastic moduli, hardness and toughness, all of which are well known to result in increasingly elevated masticatory peak loads and cyclical loading.
Table 1. Dietary mechanical properties of rabbit experimental foods measured with a portable food tester
Food items (sample size)
Young's modulus: E in MPa (mean, range)
Toughness: R in Jm−2 (mean, range)
Hardness: H in MPa (mean, range)
Ground pellets require minimal oral preparation, which reduces the amount of cyclical loading during unilateral mastication (Under-use). Hay requires greater forces to process that, in addition to the greater processing required for intact pellets, results in increased peak loads as well as increased cyclical loading during molar biting and chewing (Over-use). The mechanical properties of wet hay model the exposure of hay to saliva.
Pellets (n = 10)
Wet hay (n = 15)
Dry hay (n = 15)
MicroCT Analysis and Measures
Intra- and between-group variation in palatal form was assessed via microCT (Nicholson et al.,2006; Ravosa et al.,2007a,b,c,2008a,b). The Siemens microCAT II X-ray tube was operated at 80 kV and 500 μA, with reconstruction using 0.103 mm3 voxels (volume elements). Amira 4 software (Mercury Computer Systems, Inc., Germany) was used to obtain linear and area measurements from the microCT dataset (Fig. 1). Hard palate dimensions were obtained from three representative coronal slices: anterior, at the level of the mesial premolars; middle, at the intermediate premolars; and posterior, at the distal premolars (Fig. 1). Palate width is the transverse width of the palatal arch between the interior alveolar processes of the maxillae (and used to control for subtle variation in palate shape between loading cohorts). The following measures were sampled at four locations (10%, 35%, 65%, and 90% of total palate width): palate depth, or the superoinferior height of the palate from the nasal to the oral cavity; and cortical thickness, or the height of the cortical bone along the oral lamina of the palate as measured perpendicular to the curvature of the palatal arch. Palatal bone area is a measure of the amount of palatal bone minus the sinuses/air spaces in a given microCT slice (Fig. 1). Segmenting tools in Amira were used to calculate the area of palatal bone, which is equal to the number of voxels defined as bone multiplied by 2D voxel dimensions (0.103 mm2). In addition to microCT, a series of standard linear external dimensions were taken on each rabbit cranium (Ravosa et al.,2007b,2008a,b). Some external and microCT data were used for comparing levels of plasticity among cranial regions as well as to control for subtle size-related variation in the skull and masticatory apparatus in comparisons of loading cohorts (sensu Bouvier and Hylander,1981,1982,1984; Nicholson et al.,2006; Ravosa et al.,2007a,b,c,2008a,b). Gross and microCT inspection confirmed early fusion of the midpalatal suture in rabbits (Persson and Roy,1979), with extensive obliteration at 4.5 months old prohibiting a functional analysis of sutural interdigitation in this sample.
Nonparametric ANOVA (Mann Whitney U-test, P < 0.05) was used to investigate variation between treatment groups in raw and size-adjusted data. Linear measures such as palate depth and cortical thickness were sampled at four locations per slice and then averaged per slice in order to facilitate comparisons of between-group variation at different locations along the palate. To account for variation in skull size/shape between loading cohorts, averaged linear measures (palate depth, cortical thickness) or square roots of area measures (palate bone area) were scaled against palate width in single-slice (anterior, middle, or posterior) comparisons.
Comparisons of size-adjusted measures of hard palate structures between dietary cohorts indicate that over-use rabbits develop relatively larger masticatory proportions (Table 2; Fig. 2). In particular, cortical bone thickness at all three palate sites (mean differences of 15.4%–42.6%), bone area at all three palate sites (mean differences of 10.2%–17.8%), and anterior palate thickness (9.5% difference) are significantly larger in growing rabbits routinely subjected to elevated peak loads and greater cyclical loading during postcanine biting and chewing. Such norms of reaction are comparable to those for masticatory structures in the same rabbit sample: TMJ condyle AP and ML width (7.3% and 18.7%); corpus height and width (2.1% and 6.7%); symphysis length and width (12.8% and 14.7%) (Ravosa et al.,2007b).
Table 2. Comparisons between dietary cohorts of mean ± standard deviation (mm, except where noted) of palate dimensions
Variable and slice
Over-use (n = 10)
Under-use (n = 10)
Asterisks denote significant differences,
P < 0.05,
P < 0.01, Mann Whitney U-test. % Difference values without asterisks are non-significant (P > 0.05).
Values are size-adjusted.
Values represent averages of four sample sites within the slice.
In terms of the adaptive plasticity of the hard palate, growing rabbits subjected to elevated masticatory loads developed significantly greater trabecular and cortical bone area, and thus decreased sinus area, as well as significant increases in cortical bone thickness along the oral lamina. Interestingly, while anterior palate thickness differed significantly between cohorts, the external architecture of the remainder of the palate showed more subtle changes related to dietary properties. Nonetheless, levels of palatal plasticity are comparable to those observed for more well-studied masticatory elements such as the mandible (Ravosa et al.,2007b). Therefore, findings from this study indicate that morphological differences between dietary cohorts reflect adaptive osteogenic responses for maintaining the structural integrity of the hard palate vis-à-vis altered masticatory stresses. The differentially greater variation in palate cortical thickness and bone area further reinforce the significance of such parameters of internal geometry for interpreting the variation and plasticity of skeletal tissues in extant and extinct taxa (Daegling,1989; Nicholson et al.,2006; Ravosa et al.,2007a,b,c,2008a,b).
Furthermore, the rabbits examined herein were exposed to altered loads from weaning (at 4 weeks old) to subadulthood (19 weeks old), which is a long duration of life history for an experimental study. Such long-term studies regarding the impact of loading regimes upon the skull more closely approximate what is experienced by organisms in the wild and are thus fundamental for inferring behavior and performance in the fossil record and in the field (Ravosa et al.,2007b,2008a,b).
Implications for Evolutionary and Translational Issues
A biomechanical consideration of the thickened palate of Paranthropus proposes that it counters elevated stresses during unilateral postcanine biting and chewing, particularly midline shearing and sagittal bending (Rak,1983). As a discussion of the principal-strain directions experienced along the hard palate is best informed by an in vivo stress analysis, data of which are currently unavailable for any alert mammal, Rak's model is considered in terms of its functional link between a thick hard palate and increased loading of the masticatory complex. In this regard, it mirrors comparative studies of the external and internal geometry of the mandible in Paranthropus, which suggest that these extinct hominins experienced greater masticatory stresses—higher peak and/or cyclical loads—associated with a differentially harder and/or tougher diet (Hylander,1988; Daegling,1989). This study provides further support for such a functional interpretation of variation in the hard palate (cf., Rak,1983).
A criticism of this biomechanical model invokes the high metabolic costs of osteogenesis and suggests that, rather than cortical and structural thickening, increased interdigitation of the midpalatal suture is the primary mechanism by which a bony palate counters elevated stresses during unilateral mastication (McCollum,1997,1999). The midpalatal suture may indeed reflect the mechanical demands placed upon the palate, as in vitro data indicate that palatal strains are highest at the suture (Hotzman et al.,2007). However, although bone is a metabolically expensive tissue, remodeling and modeling are not an uncommon means by which cortical bone in the cranium and postcranium adapts to postnatal alterations in dynamic loads (e.g., Bouvier and Hylander,1981; Lanyon and Rubin,1985; Biewener,1993). It is also worth noting that the complexity of the midpalatal suture in living and extinct (e.g., Paranthropus) mammals has begun to be quantified only recently (Hotzman,2008) and this suture has yet to be considered from a broadly comparative perspective.
Furthermore, given the influence of masticatory loading on cortical bone thickness along the lateral and medial portions of the hard palate in growing rabbits, the level of midpalatal suture interdigitation in extinct and living mammals likely was and continues to be affected by variation in such forces. Given the differentially closer proximity of the hard palate to the application of postcanine bite forces during unilateral mastication, such loads must be transmitted medially to the midpalate via the hard palate. It stands to reason that if diet-induced increases in masticatory stresses do affect the development of midpalatal suture interdigitation, such forces are likewise of sufficient magnitude to influence variation in cortical bone thickness along the hard palate. The converse may not be true if there is a steep lateral-to-medial gradient in the influence of masticatory stress on palate thickness and sutural interdigitation. Alternatively, if cortical bone thickness along the hard palate were unaffected by variation in masticatory loads, this would indicate that the hard palate is overbuilt to counter routine stresses during postcanine biting and chewing (presumably because of some other factor). In this scenario, masticatory loads transmitted medially to the midpalate would be significantly dissipated by an overbuilt hard palate, with it being unlikely that stresses along the midpalatal suture are sufficiently large to result in diet-induced variation in the degree of interdigitation.
Results from this experimental model support an evolutionary paradigm in which the variation in hominin palatal morphology is considered adaptive to the mechanical properties of disparate diets. This is in accord with indirect evidence from finite element analysis that a vertically thickened hard palate is part of a suite of cranial features that may act to resist strains resulting from elevated masticatory stresses (Strait et al.,2007). Indeed, paleobiological arguments that do not consider the role of diet-induced plasticity in adaptive explanations for all aspects of hominin feeding anatomy (e.g., Ungar and Scott,2008, and a related article by Gibbons,2008) overlook the significant influence of long-term mechanical loading on the patterning of craniomandibular form throughout an organism's lifetime (Ravosa et al.,2008b). From a paleontological and ecomorphological perspective, experimental analyses that strive to more closely model masticatory forces experienced throughout an organism's lifetime will be critical for more accurately characterizing regional and temporal effects on craniomandibular form and function. In addition, the ability to extrapolate conclusions on functional morphology from a laboratory model species (i.e., the rabbit) to wild or extinct species (i.e. Paranthropus) is based upon demonstrated commonalities in the masticatory apparatus (see above). However, such comparisons may be mitigated by differences in dental formulae and eruption, presence of a dental diastema, daily feeding frequencies, and tongue activity. For this reason, it is important to employ caution when extrapolating from within-species analyses and, when doing so, to account for the extent and limitations of known general mammalian trends of functional adaptation (Lanyon and Rubin,1985; Biewener,1993; Ravosa et al.,2007b,2008a,b).
This research is also relevant for models regarding the function and evolution of the secondary palate in mammals (Thomason and Russell,1986; Russell and Thomason,1993). The secondary palate in mammals, by separating the oral and nasal cavities, is theorized to have initially facilitated simultaneous feeding and breathing (Brink,1956). However, reinforcement of the palate has also been modeled as a primary adaptation to counter rostral torsion during unilateral mastication (Thomason and Russell,1986; Russell and Thomason,1993; Strait et al.,2007). The effect of loading on palatal morphology documented herein provides the most direct support to date for suggestions that an osseous palate functions to resist forces during postcanine biting and chewing. Given such findings, the biomechanical significance of palate morphology may represent an exaptation from the adaptive pressures experienced by stem mammals. In other words, once the hard palate was interposed between the maxillary tooth rows to fully separate oral and nasal cavities (sensu Brink,1956), it became subjected to masticatory stresses experienced during biting and chewing. Therefore, as a product of its location, palatal morphology should reflect the functional demands of resisting mechanical forces.
In clinical practice, the anchorage of teeth or dental implants and manipulation of the dental arcade may require the use of orthodontic implants fixed to the hard palate. Deformation of the palate via loads induced by these implants is more severe given thinner, less biomineralized bone, with the efficacy of implant procedures diminished if palatal bone is not able to counter forces generated by such devices (Gedrange et al.,2001). Disruption and damage of palatal tissues by implants has long-term consequences for craniofacial elasticity and growth (Hansson et al.,1983). Results from this study suggest that the growing palate subjected to surgical intervention and implantation may be modulated via altered mechanical loads.
The authors thank Kris Aldridge for the use of computers and imaging software, Ashley Gosselin-Ildari for segmentation instructions, Barth Wright for performing the analyses of rabbit food properties, and Scott Miller and anonymous reviewers for comments on the manuscript. Authors gratefully acknowledge the support provided by the VA Biomolecular Imaging Center at the Harry S. Truman VA Hospital and the University of Missouri–Columbia. A version of this manuscript was awarded the 2008 Mildred Trotter student prize at the 77th annual meeting of the American Association of Physical Anthropologists.