An experimental analysis of temporomandibular joint reaction force in macaques.код для вставкиСкачать
An Experimental Analysis of Temporomandibular Joint Reaction Force in Macaques WILLIAM L. HYLANDER Departments of Anatomy and Anthropology, Duke University Medical Center, Durham, North Carolina 2771 0 K E Y WORDS Bone strain mandibular joint . Mandibular function . J a w mechanics - Temporo. ABSTRACT Mandibular bone strain in the region immediately below the temporomandibular ligament was analyzed in adult and sub-adult Macaca fascicularis and Macaca mulatta. Following recovery from the general anesthetic, the monkeys were presented food objects, a wooden rod, or a specially designed bite-force transducer. Bone strain was recorded during incisal biting and mastication of food, and also during isometric biting of the rod and/or the transducer. The bone strain data suggest the following: The macaque TMJ is loaded by a compressive reaction force during the power stroke of mastication and incision of food, and during isometric molar and incisor biting. TMJ reaction forces are larger on the contralateral side during both mastication and isometric molar biting. Patterns of ipsilateral TMJ reaction force in macaques during isometric biting vary markedly in response to the position of t h e bite point. During biting along the premolars or first two molars a compressive reaction force acts about the ipsilateral TMJ; however, when the bite point is positioned along the Mj, the ipsilateral TMJ has either very little compressive stress, no stress, or it is loaded in tension. An understanding of the biomechanics of mandibular function requires detailed information about (1) the gross and microscopic anatomy of the masticatory apparatus, (2) patterns of mandibular movement, and (3) the nature of the various external forces (muscular and reaction) that act upon the mandible. A good deal is known about the anatomy of the mammalian masticatory apparatus (Schumacher, '61; Turnbull, '70), and about in vivo patterns of mammalian mandibular movement (e.g., Crompton and Hiiemae, '70; Hiiemae and Kay, '73). However, with the exception of a number of studies on isometric biting forces in humans, there have been few experimental attempts to determine in vivo external forces which act upon the masticatory apparatus, although masticatory muscle forces have been analyzed indirectly by characterizing either patterns of muscle activity (e.g., Moller, '66; deVree and Gans, '76) or the histochemical and contractile properties of these muscles (e.g., Ringquist, '74; Thexton AM. J. PHYS. ANTHROP. (1979)51: 433-456. and Hiiemae, '75; and Maxwell et al., '79). Although there is some information about bite force magnitudes during mastication (Anderson, '561, there are no in vivo data on the direction of the bite force in mammals (except for some preliminary experimental work on humans by Graf "751 and Hylander "7811, or on the direction or magnitude of reaction forces that act upon the mammalian temporomandibular joint (TMJ). Since in vivo TMJ reaction forces have never been characterized experimentally, there is a wide range of opinion regarding t h e nature of these forces. Some workers categorically insist that the mammalian TMJ is not a load-bearing joint, and therefore reaction forces are absent during mandibular function (Wilson, '20, '21; Robinson, '46; Scott, '55; Steinhardt, '58; Tattersall, '73; Roberts and Tattersall, '74) ; others indicate 'Thm investigation was supported by the followng NIH Re search Career Development Award (DE 00027J, NIH Research Grant W E 4531) and a NSF Research Grant (BNS 76 11924) 433 434 WILLIAM L. HYLANDER t h a t the TMJ is frequently heavily loaded, but do not distinguish between ipsilateral and contralateral sides (e.g., Sicher, '51; Smith and Savage, '59; Davis, '64; Crompton and Hiiemae. '69; Turnbull, '70; Bramble, ' 7 8 ) . Those that have considered the ipsilateral and contralateral sides separately are divided in their opinion as to whether the two sides are loaded equally (Roydhouse, '55) or differentially. Page ('54) suggests that only the ipsilateral side is loaded during mastication, while Hekneby ('74) states that although both joints are loaded, the ipsilateral TMJ is loaded more than the contralateral TMJ; others suggest that the contralateral TMJ is loaded more than the ipsilateral TMJ (Gysi, '21; Hylander, '75; Smith, '781, and i t has even been suggested that only the contralateral TMJ is loaded (Walker, '78; Greaves, '78). Ideally, to resolve this disagreement it would be advantageous to measure directly the direction and magnitude of the reaction force between the mandibular condyle and the articular eminence. Unfortunately, this approach would necessitate surgical exposure of t h e TMJ and damage to the joint surfaces, which would interfere with normal masticatory movements. In addition, it would be technically difficult to build and implant a transducer that would sense in vivo external reaction force throughout the full range of condylar rotation and translation. Another approach, albeit indirect, is to characterize external reaction force through the joint by analyzing the pattern of in vivo bone deformation in the immediate vicinity of the TMJ, since the amount of bone deformation present presumably would be directly proportional to the magnitude of the external forces. Although this approach does not permit one to determine the absolute amount of reaction force along the TMJ, or its direction, it should allow us to tell whether the TMJ is loaded, which TMJ is loaded the most, and when the maximum and minimum loadings take place. The transducers necessary to measure bone deformation (strain gages) must also be implanted surgically, and therefore the subject will not be free of pain in the region overlying the joint. However, structures of the TMJ are not damaged surgically with this approach. In spite of the limitations of this indirect approach, it was decided to attempt to characterize in vivo reaction force along the TMJ by measuring bone strain immediately below the TMJ (the subcondylar region) during various phases of mandibular function in Macaca mulatta and Macaca fascicularis. This investigation was undertaken since an understanding of relative amounts of TMJ reaction force would be useful for testing various models of TMJ function. Expected patterns of stress and strain Let us consider what might be the expected patterns of stress and strain in the macaque subcondylar region during (1) the power stroke of mastication and incisal biting under conditions of no TMJ reaction force, and ( 2 ) under conditions of a reaction force which presses the mandibular condyle against the articular eminence. In addition, expected patterns of subcondylar stress and strain during the opening stroke will also be considered. The expected patterns of stress and strain were originally deduced from anatomical, electromyographical, and jaw movement data. They were also reproduced on a dry mandible which had a rosette bonded to the subcondylar region and upon which the various muscle and reaction forces were simulated. Fig. 1 Expected patterns of stress along the medial and lateral aspects of the subcondylar region during the opening stroke or during the power stroke, assuming TMJ reaction forces are absent. During the power stroke, F, is the medial component of force of the superior head of the lateral pterygoid muscle and the medial component of force of that portion of the temporalis which attaches to the coronoid process. During the opening stroke, F, is the medial component of force of the inferior head of the lateral pterygoid muscle. F, tends to bend the ramus on each side towards the midline. This causes the lateral aspect of the subcondylar region to experience tensile stress (t), while the medial aspect experiences compressive stress (c). Expected patterns of strain during such conditions are presented in figures 3A !power stroke) and 8 (opening stroke). This figure is not meant to imply that ipsilateral and contralateral muscle forces are of equal magnitude during the opening stroke or during the power stroke. TMJ REACTION FORCE Expected patterns of su bcondylar bone strain in absence of TMJreaction force (a) The effect of the temporalis and lateral pterygoid muscles Even if there is no compressive reaction force acting through the TMJ during the power stroke of mastication and incision, the subcondylar region will still experience stress and strain. For example, it has been shown that in humans (or macaques) i t is theoretically possible to have powerful contraction of the temporalis muscle during molar and incisor biting and a t the same time have no reaction force acting through the TMJ (Gingerich, '71). Under these conditions, a square finiteelement of bone in the subcondylar region will be deformed in various ways. First, since in macaques there is a medial component of force associated with the part of the temporalis muscle that attaches to the coronoid process, there will be a tendency for the mandibular ramus on each side to bend towards the midline; i.e., the tips of the coronoid processes will bend towards one another (fig. 1).This bending will result in tensile stress along the lateral surface of the subcondylar region and compressive stress along its medial surface. The resulting pattern of deformation along the lateral aspect of the subcondylar region will be like that in figures 2B or 3A; i.e., the maximum principal strain ( E , ) (tension) will be oriented upwards and backwards (or downwards and forwards), while the minimum Fig. 2 Expected patterns of subcondylar strain during the power stroke in the absence of TMJ reaction force. Flp and Ft represent the medial Component of force of the lateral pterygoid muscle and the medial component of force of that portion of the temporalis muscle which attaches to the coronoid process, respectively; the small arrows labeled c , and c 2 represent the direction of the maximum (el) and minimum ( c J principal strains. A As a first approximation, the macaque mandible is analogous to a curved bar that is bent sharply in the midline (dashed lines). B During the power stroke, the subcondylar region is bent by Fb and Flp. During the opening stroke, the subcondylar region is bent by Flp (temporalis inactive a t this time). This bending produces tensile stress along the lateral aspect of the subcondylar region. The resulting principal strains ( e l and g 2 , oriented as shown) would deform a square finite-element into a rectangle as indicated. C Since the twisting moment associated with Ft probably exceeds the one associated with Fl, during the power stroke, the subcondylar region is twisted as indicated by the curved arrows. The square finite-element is deformed into a skewed parallelogram by this torsional stress as shown by the directions of and ep. Flp has not been included in this figure. This figure is not meant to imply that ipsilateral and contralateral muscle forces are of equal magnitude during the power stroke. 435 principal strain (c2) (compression) will be oriented upwards and forwards (or downwards and backwards). Second, the lateral aspect of the subcondylar region will experience strain due to contraction of the superior head of the lateral pterygoid muscle, since in macaques this muscle also contracts during the power stroke of mastication (McNamara, '73). The force of this muscle will also bend the mandibular ramus towards the midline so that the subcondylar region experiences tensile stress along its lateral surface, while the medial surface experiences compressive stress (fig. 1). The deformation pattern in the lateral subcondylar region will also be similar to that seen in figures 2B or 3A. I--. F, 436 WILLIAM L. HYLANDER c2 A B u C Fig. 3 Expected patterns of strain of a square finiteelement along t h e lateral aspect of the subcondylar region during t h e power stroke when there is no reaction force across the TMJ. Ft is the resultant force from the temporalis muscle; Fb is the bite force; e l and c z are maximum and minimum principal strains. There are three possible deformation patterns t h a t are superposed on one another under the above conditions. These deformation patterns are figured separately in A, B, and C. A The subcondylar region experiences bending stress because the medial component of the temporalis muscle force bends the ramus towards the midline. B The above force also twists t h e ramus so t h a t the rostral border of the ramus inverts while the caudal border everts. C The ramus is bent in t h e sagittal plane by the sagittal component of the temporalis muscle force (see text for discussion). Contraction of the temporalis and the lateral pterygoid’s superior head will also cause the mandibular ramus to twist. Twisting which results from the medial component of the temporalis muscle will result in a medial deflection of the rostral surface of the ramus and a lateral deflection of its caudal surface, while twisting produced by the superior head of the lateral pterygoid muscle will have the opposite effect (fig. 4). Twisting from the temporalis muscle will normally exceed twisting from the lateral pterygoid muscles because the superior head of the lateral pterygoid muscle and its twisting moment arm are small relative to the size of the medial component of t h e temporalis muscle and its twisting moment arm. The net effect of this twisting will result in 6 , being directed vertically and c2 being directed horizontally (figs 2C or 3B). The combined effect of the above twisting and bending loads of the temporalis and lateral pterygoid muscles is that e l will now be aligned primarily upwards and slightly backwards, while will be aligned primarily backwards and slightly downwards, i.e., a combination between figures 3A and 3B. Finally, the upwards and backwards pull of the temporalis muscle will cause the ramus to be loaded axially and to be bent in the sagittal plane. Axial loading of the subcondylar region will be tensile and the direction of will be in line with the pull of the temporalis muscle, resulting in a pattern similar to the one seen in figure 3A. Since the subcondylar region is not in line with the resultant force of the temporalis muscle and bite force, the subcondylar region will also be bent slightly in the sagittal plane. The sagittal bending will result in tensile stress along the rostral portion of the ascending ramus and compressive stress along its most caudal portion; the resulting pattern of deformation along the caudal aspect of the subcondylar region for this type of loading will be like that in figure 3C. In the more rostral portion of the subcondylar region, which is closer to the bending axis of neutrality for sagittal bending, c 2 will be directed downwards and backwards and e l will be directed upwards and backwards. Since the combined strain pattern of bending towards the midline (fig. 3A), twisting (fig. 3B) and axial loading (fig. 3A) will be approximately opposite to the pattern for sagittal bending (fig. 3 0 , i t is theoretically possible to have little or no subcondylar strain with no TMJ reaction force if the amount of twisting, axial, and midline bend- TMJ REACTION FORCE 437 and strain for sagittal bending is located along the caudal edge of the mandibular ramus (i.e., as far away as possible from the bending axis of neutrality). (b) The effect of the medial pterygoid and masseter muscles Fig. 4 Expected patterns of stress during t h e power stroke in the absence of TMJ reaction forces. The arrows labeled Ft and F l p are described in figure 2. These forces tend to bend t h e ramus on each side towards the midline. In addition, Ft and F l p also have a tendency to twist the ramus on each side in opposite directions (curved arrows). However, since the twisting moment associated with Ft probably exceeds the twisting moment associated with Flp, twisting will result in inversion of the rostra1 border of the mandibular ramus (in t h e direction of the curved arrow along the rostra1 border of t h e mandibular ramus). See figures 2 and 3 for associated patterns of strain along the lateral aspect of the subcondylar region. This figure is not meant to imply t h a t ipsilateral and contralateral muscle forces are of equal magnitude during the power stroke. ing strain is equal to the amount of sagittal bending strain. More likely, however, the net deformation pattern will be similar to a combination of the patterns seen in figures 3A and 3B, since the amount of midline bending alone will presumably exceed the amount of sagittal bending. This is primarily because the section modulus of a section through t h e subcondylar region is much smaller for midline bending than it is for sagittal bending, thus making it much easier to bend the ramus towards the midline.2 In addition, whereas the finite-element in the subcondylar region or figure 3 is located in the region of maximum stress and strain for midline bending, this is not the case for sagittal bending, since maximum stress In addition to the temporalis and lateral pterygoid muscles, the masseter and medial pterygoid muscles are also active during the power stroke, and therefore they also have the potential to deform the subcondylar region. The medial pterygoid muscles will have a tendency to bend the angular processes of the mandibular rami towards one another while the masseter muscles will have a tendency to bend them away from one another. However, when acting together, they probably will have little tendency to bend the mandibular rami in either direction because even though t h e masseter muscle in macaques is about twice as large as the medial pterygoid muscle, t h e medial pterygoid has a more transverse orientation (Schumacher, '61).The net effect of muscle size and position probably will result in nearly equal transverse bending moments about each ramus. Even if transverse bending were to occur from these muscles, they probably would not bend the subcondylar region since only that portion of the mandible between the muscle attachment areas on the left and right sides would be bent. For example, if the left and right sides of the curved bar in figure 2B were pressed together and the point of applied force was below the square finiteelement, the finite-element would not be stressed or strained. Only that portion of the bar between the points of the applied forces would experience bending stress and strain. The medial pterygoid and masseter muscles on each side will also have a tendency to twist the mandibular ramus on each side. Whereas the masseter will have a tendency to evert the lower border of the mandible, the medial pterygoid will have the opposite effect. When acting together, however, they probably will have little tendency to twist each ramus for reasons similar to those outlined above in the discussion of bending. In this instance, how_____ 'The amount of bending stress in a member is inversely proportional to the section modulus. For midline hending of the ram-, t h e section modulus for a section passing through the subcondylar region is directly proportional t o a%, where a is t h e medio-lateral thickness of t h e ramus and h is the antero-posterior thickness nf t h e ramus. For sagittal hending, the section modulus for the aame a e c ~ tion is directly proportional to h'a. A first-level approximation indicates t h a t t h e section modulus is about seven t i m e larger for aagittal bending than for midline bending. 438 WILLIAM L. HYLANDER ever, i t is the twisting moments which will be nearly equal. Even if some twisting were to occur from these muscles, they probably would not twist the subcondylar region since only A Fig. 5 Expected patterns of strain of a square finiteelement along the lateral aspect of the ramus (deep to the superficial masseter muscle) during the power stroke. The same pattern will exist in the subcondylar region in the absence of the TMJ reaction force (see text for discussion). A F. A B Fig. 6 Expected patterns of strain of a square finiteelement along the lateral aspect of the subcondylar region during the power stroke when there is a compressive reaction force acting along the TMJ. The reaction force (Fr) is aligned vertically (not shown in B). In A the square finiteelement is positioned rostra1 to the reaction force and in B the square finite-element is immediately below the reaction force. Patterns of strain will vary between these two deformation patterns (see text for discussion). that portion of the mandible between the muscle attachment areas on t h e left and right sides would experience twisting stress. For example, if the left and right sides of the curved bar in figure 2C were twisted similarly (i.e., the lower border on each side is everted), and the point of application of the twisting force was below the square finite-element, the finite-element would not be stressed or strained. Only that portion of the bar between the points of the applied forces would experience twisting stress. In summary then, the action of these two muscles will probably not twist or bend the subcondylar region to any appreciable extent. However, if a square finite-element of bone is analyzed along a portion of the mandible immediately deep to the superficial masseter muscle, i.e., inferior to the subcondylar region, this element will be deformed by masseter and medial pterygoid muscle contraction, a s shown in figure 5. The subcondylar region will be deformed similarly except the magnitude of the strains will be less. Note that this deformation pattern is similar to the pattern of figure 3A. In conclusion, under conditions of no reaction force along the TMJ during biting, the net deformation pattern along the lateral aspect of the macaque subcondylar region probably will be similar to a combination of figures 3A and 3B, due to midline bending, twisting, and axial pull from the temporalis muscle, midline bending due to the superior head of the lateral pterygoid muscle, and compressive loading due to the masseter and medial pterygoid muscles. Expected patterns of subcondylar bone strain in the presence of TMJreaction force If there is a tendency for the bony articular surfaces of the TMJ to be pressed together during normal function, two additional patterns of subcondylar bone strain could conceivably result. First, a square finite-element of bone will be sheared as shown in figure 6A. Note that under these conditions, C , is directed upwards and forwards while c2 is directed upwards and backwards. Second, if t h e square finite-element is located immediately below the reaction force (as in fig. 6B), the finite-element will be simply compressed; will be aligned horizontally, and e 2 will be aligned vertically. In this instance the compressive strain is directly aligned with the compressive reaction force. Due to different TMJ REACTION FORCE possible positions of the finite-element within the subcondylar region, recorded strain patterns should vary between these two patterns, but in general, one would expect that the direction of the principal strains will be a compromise between the patterns seen in figures 6A and 6B. Thus, the net subcondylar deformation pattern during TMJ loading is almost opposite to the one hypothesized for no TMJ reaction force during biting (figs. 3A,B). The foregoing discussion of hypothetical deformation patterns is based upon the assumption that if a TMJ reaction force is present, its resultant acts along the middle of the articular surface of the mandible condyle. If this force acts along the lateral aspect of the condyle (fig. 71, the subcondylar region will also be bent so that its lateral surface experiences compressive stress, while its the medial surface experiences tensile stress, i.e., the two condyles will bend away from one another. Under these conditions, the resultant pattern of deformation will be similar to that seen in figure 6B; the direction of e 2 will be primarily vertical and E , will be primarily horizontal. If the resultant reaction force acts about the medial aspect of the articular surface of the condyle (fig. 71, rather than the middle or lateral aspect, strain patterns will be reversed since under these conditions the medial surface will experience compressive bending stress, while the lateral surface experiences tensile bending stress. The resultant pattern of deformation under these circumstances will be similar to the one in figure 3B. In summary, we might expect that patterns of deformation in the subcondylar region during TMJ reaction force will probably be a compromise between patterns seen in figures 6A and 6B if the articular surface of the condyle is loaded by a resultant force which acts along the middle or lateral aspect of the mandibular condyle. If the resultant reaction force acts along the medial articular surface of the condyle, the pattern of deformation will be similar to that seen in figure 3B, which is also one of the patterns hypothesized for no TMJ reaction force. Expected patterns of subcondylar bone strain duringjaw opening Presumably there would also be subcondylar bone strain during wide opening of the jaw and the opening stroke of mastication. Bone strain a t this time will be due to contraction of the inferior head of the lateral pterygoid mus- 439 cle which will cause the ramus t o bend towards the midline. This bending will result in compressive stress along the medial aspect of the subcondylar region and tensile stress along the lateral aspect (fig. 1).Therefore, the expected deformation pattern along the lateral aspect of the subcondylar region will be the Fig. 7 Expected patterns of stress during asymmetrical loading of t h e TMJ. The arrows indicate t h e point of application of a TMJ reaction force; t and c represent tensile and compressive stress, respectively. If the mandibular condyle is loaded i n the middle of its articular surface, t h e medial and lateral aspects of the subcondylar region will be loaded in compression in this projection. If the condyle is loaded along its lateral aspect (right), t h e subcondylar region along its lateral aspect will experience compressive bending stress, while its medial aspect will experience tensile bending stress. If t h e condyle is loaded along its medial aspect (left), the bending stress will be reversed. This figure is not meant to imply that the above loading patterns are occurring simultaneously on opposite sides of t h e mandible. Fig. 8 Expected patterns of strain of a square finiteelement along the lateral aspect of the subcondylar region during wide opening or during the opening stroke. The lateral pterygoid muscles bend t h e rami towards the midline and cause t h e above strain pattern (see fig. 2 and text for discussion). 440 WILLIAM L. HYLANDER following: E , is directed upwards and backwards while E ? is directed upwards and forwards. Figure 8 demonstrates the expected strain patterns for jaw opening. MATERIALS AND METHODS Subjects Four subadult male and one adult female Macaca mulatta and one subadult male and three adult female Macaca fascicularis served as subjects. The relative stage of dental development of the subadult males can be seen in the radiographic tracings of figures 10, 12, 13 and 15. Each Macaca mulatta was used for one experiment (Experiments 1-5). Each Macaca fascicularis was used for a t least two experiments (Experiments 6-13). Bite force measurement apparatus and procedure The apparatus and procedure for measuring bite force has been described elsewhere (Hylander, '77, '79a). Only the three adult female M. fascicularis were trained to bite the force transducer. Without training, macaques usually refuse to bite a force transducer more than a few times. Strain gage bondingprocedure and bone strain measurement apparatus Animals were food- and water-deprived for 24 hours and then general anesthesia was induced with fluothane. A preauricular approach (Morgan, '75) was used to expose subcondylar cortical bone or cortical bone immediately below the subcondylar region, deep t o the superficial masseter muscle. Extreme care was taken not to traumatize any of the neurovascular structures during the surgical procedure. Since the bonding site is quite small in area, particularly in Macaca fascicularis, the smallest available 120-ohm foil rectangular stacked rosettes (WA-06-030WR-120, MicroMeasurements) and 120-ohm single-element strain gages (BAE-06-015CC-120 LE, William T. Bean, Inc., Detroit, Michigan) were used. The bonding procedure was like that described previously (Hylander, '79a) except that the lead wires were not stabilized by cementation t o a screw embedded in mandibular bone. Instead, lead wires were stabilized by suturing the wires and plug t o the macaque's skin about 5 cm above its zygomatic arch. Either sham surgical operations were performed on the opposite side or strain gages were bonded Fig. 9 Radiograph of t h e left half of a Macaca mulatta mandible with a rosette bonded in the subcondylar region (Experiment 5). The long axis of the thick wire indicates t h e orientation of the middle strain gage element. to both left and right sides. Although gages were bonded bilaterally in some of the subjects, only one gage functioned properly in these instances. After all surgical wounds were sutured, a lateral radiograph of the head was taken in order to locate accurately the position of the strain gage. Figure 9 is a radiograph of an isolated mandible and rosette strain gage bonded along the subcondylar region. Following the radiographic procedures, the animals were given intramuscular administrations of benzathine penicillin G and procaine penicillin G and placed in a restraining apparatus previously described (Hylander, '79a). Recording procedure Details of the recording procedure are the same as those described previously (Hylander, "79a). Once the animal recovered from the general anesthetic, it was presented fresh apple, a small pine dowel (diameter = 5 mm) and/or the force transducer. When the animals masticated it was a simple matter to determine which side the animal chewed on. This was done by simply observing the subject's tongue position the apple bolus along its left or right postcanine teeth. Immediately before and after each bout of mastication or isometric biting the zero level of bone strain was determined. Due t o thermal stability, there was no detectable baseline drift during any bout of chewing or biting. Although untrained macaques will refuse to bite a force transducer, some macaques will repeatedly bite a small pine dowel. The animals were encouraged t o bite the dowel in order to characterize subcondylar bone strain during 44 1 TMJ REACTION FORCE I r W b ioY \\\ Exp. 5 Exp.12 Exp. 7 0 \ X \ Exp. 13 Fig. 10 Tracings of lateral radiographs of the macaque mandibles. The black dots indicate the position of the rosettes. The X axis is parallel to t h e A strain gage element (Hylander, '79a). The lines indicate only the mean value for the direction of the minimum principal strain (compression) ( € 3during apple mastication and incision; lines labeled b and w indicate contralateral (balancing) and ipsilateral (working) sides, respectively; lines labeled i indicate incision and lines labeled o indicate wide opening. Thus, a single rosette in each of t h e above experiments is used to characterize bone strain during t h e above behaviors. isometric biting. Mandibular bone strain was recorded during apple mastication or when the macaque bit the dowel or transducer. When animals bit on the dowel or the force transducer, bone strain could be easily correlated with bite force values. To assist in monitoring simultaneous bone strain and timing of portions of the masticatory cycle, the animal and the chart recorder pens and output were videotaped and analyzed as described elsewhere (Hylander, '79a). Recording sessions often lasted several hours with frequent intermittent rest periods. At the end of the recording period, the subjects were tranquilized with Ketamine hydrochloride and the strain gagek) were removed. All animals were returned to their cages within 24 hours of the original surgical procedure and all resumed eating after recovery from the tranquilizer. The post-operative periods were uneventful. Infections were never encountered in any of the subject^.^ Experimental Set A The specific aim of Experimental Set A was to analyze patterns of peak mandibular bone strain along the subcondylar region and deep to the superficial masseter muscle during incisal biting (incision) and mastication of food. One aspect of this analysis was t o compare the direction and magnitude of bone strain values 3All Macaca mulatta subjects were sacrificed a day or so following the hone-atrain experiments because they were experimental controls for a study on vascular plaque formation in macaques on a low and high cholesterol diet (Principal Investigator. Dr. Otto Hagen, Department of Surgery, Duke University, NIH grant HL. 15448). 442 WILLIAM L. HYLANDER A B C 11 J a ul W I- 4 Q u I2 , 0 u TENSION COMPRESSION 0 W rI- F Fig. 11 Chart recording of subcondylar bone strain recorded from the A, B, and C elements of a rosette during apple mastication (Experiment 5). Bone strain scale is in microstain units; time is in 1-second intervals. The zero level of bone strain is indicated for each channel. In channel B, peaks t o the right correlate with the power stroke and peaks to the left correlate with opening. on the working or ipsilateral side of the mandible with bone strain values on the balancing or contralateral side during mastication. Bone strain values during incisal biting of whole apples were also compared t o the mastication data. In addition, bone strain patterns during jaw opening were also analyzed. The above experimental set consisted of seven subcondylar bone strain experiments and two experiments on bone strain deep to the superficial masseter. Experimental Set B The specific aim of Experimental Set B was t o analyze patterns of peak mandibular bone strain along the subcondylar region and deep to the superficial masseter muscle during isometric biting. An important part of this analysis was to compare the direction and magnitude of bone strain values on the ipsilatera1 side of the mandible with the bone strain values on the contralateral side. This set of experiments is made up of two subsets. Subset 1 involved isometric biting on a wooden dowel while subset 2 consists of isometric biting on a force transducer. Subset 1consists of four subcondylar bone strain experiments and one experiment on bone strain below the superficial masseter; subset 2 consists of one experiment on subcondylar bone strain and one on bone strain below the superficial masseter muscle. The direction and magnitude of the maximum and minimum principal strain were calculated following Dally and Riley ('65). Descriptive statistics for bone strain and bite force values were generated. Bone strain (the dependent variable) and bite force (the independent variable) data were also analyzed utilizing linear regression techniques (regression line fit by method of least squares). Strain (€1, a dimensionless unit, equals the change in length (AL) of an object divided by the original length (L) of the object, i.e., E = &. The unit of strain is the microstrain ( ~ € 1Lwhich , . equals 1 X inchedinch or millimeters/millimeter, etc. By convention, tensile strain is a positive value and compressive strain is a negative value. In this study the maximum principal strain (E ,) is the largest tensile strain value whereas the minimum principal strain ( E J is the largest compressive strain value. The direction of e 2 is given in degrees, and its direction is relative to the long axis (the X axis) of one of the three strain gage elements within a rectangular rosette. Adding or subtracting 90" t o the direction of yields the direction of e l . RESULTS AND DISCUSSION In this section, emphasis is placed on both the magnitude and direction of the maximum amount of recorded compressive strain ( € 3 . The values of the maximum amount of recorded tensile strain (E ,) are included in tables 1, 3 and 4 but will not be dealt with here. Experimental Set A 1. Rosette analysis of bone strain during the power stroke of mastication and incision In these experiments, the rosette was posi- 443 TMJ REACTION FORCE TABLE 1 Descriptive statistics of subcondylar bone strain values during incision and mastication of apples (rosette analysis) Principal strains Experiment Maximum Minimum (6,) Angle of minimum principal strain ( c J (6 S.D. Mean 23 66 17 21.8 4.0 33.7 110.8 - 341 - 353 - 531 - 664 43 66 177 - 183 - 286 - 306 - 633 -544 - 981 84 - 130 N Mean Largest S.D. Mean Largest 80 102 137 237 27 106 27 50 - 105 - 155 - 180 - 280 33 - 359 - 108 - 359 - 145 - 14 88 149 27 69 47 40 13 136 191 36 323 254 186 59 33 84 - 202 140 84 8 45 123 254 90 293 431 24 61 131 26 109 28 14 60 36 65 52 114 60 117 82 27 13 24 20 Range S.D. M. mulatta (3-82) Experiment 1 Ipsilateral Contralateral Incision Opening (threat) M. mulatta (6-20) Experiment 2 Ipsilateral Contralateral Incision M. mulatta (6-41) Experiment 5 Ipsilateral Contralateral Incision M.fascicularis (1) Experiment 7 Ipsilateral Contralateral Incision Opening (threat) M. fascicularis (4) Experiment 12 Ipsilateral Contralateral Incision M.fascicularis (4) Experiment 13 Ipsilateral Contralateral Incision Opening (threat) 1 - - 344 - - 49 - 77 - 55 - 95 - 132 - 82 6.2to 34.0 2.7to 22.5 6.6 4.1 103.3to 124.3 5.9 67.5 81.2 58.2 45.8to 88.0 74.6to 90.3 37.7to 73.4 6.8 3.6 12.0 56 113 284 56.1 46.2 50.2 48.8to 80.8 35.3to 60.8 48.9to 51.8 4.2 4.6 1.1 24 19 26 20 89.4 76.8 101.5 103.4 77.7 to 100.2 51.8 to 104.5 86.0 to 111.7 91.6 to 111.5 7.1 - - - 5.8 6.0 6.6 - - - - - - - - - - 11 4 36 146 85 203 28 66 - 85 - 226 - 153 - 347 43 123 61.6 68.3 49.lto 68.6 67.5to 69.3 5.3 1.0 134 97 35 16 60 77 46 128 189 186 145 234 31 45 45 50 - 74 - 99 - 283 - 213 - 207 - 753 - 47 34 44 167 14 3.6 8.5 41.8 109.4 -27.2to 27.7 6.4to 28.5 9.2to 81.0 100.8 to 115.4 7.8 5.6 18.9 4.1 - 22 - Note: Bone strain values are in microstrain units. N = number of power strokes or the number of times mouth was opened wide. Angle values are in degrees. Animal identification number in parentheses. See table 2 for additional data from Experiment 12. tioned in the subcondylar region (Experiments 1,2, 5, 12, 13) or deep to the superficial masseter muscle (Experiment 7). The location of the rosette in each experiment is shown in figure 10. A representative tracing of subcondylar bone strain during mastication is illustrated in figure 11. (a) Bone strain magnitudes When the rosette was positioned in the subcondylar region, the mean values of c 2 on the contralateral side are larger than and differ significantly from the corresponding ipsilateral-side values (p < 0.001) (table 1). In con- trast to the above pattern, the mean value of c 2 deep to the superficial masseter muscle (Experiment 7) on the ipsilateral side is larger than and significantly different from the mean value on the ipsilateral side (p < 0.001). When the animals bit apples with their incisors, mean values of c 2 for the subcondylar region generally exceed those values for mastication. In all but one instance these mean values are significantly different from the mastication mean values (p < 0.001). During incisal biting the mean and maximum values of bone strain below the superficial masseter muscle are larger than and significantly dif- 444 WILLIAM L. HYLANDER ferent from contralateral values (p < 0.001), and virtually identical to ipsilateral values during mastication (Experiment 7). (b) Bone strain direction In every instance during the power stroke of mastication and incision, E ? in the subcondylar region is directed upwards and backwards (or downwards and forwards). In contrast, the direction of c 2 below the superficial masseter muscle (Experiment 7) is directed upwards and forwards. The mean values of the direction of e ? during incision and mastication (ipsilateral and contralateral sides) are all significantly different from one another within each experiment (p < 0.051, but in the subcondylar region these values show no consistent pattern from one experiment t o another. Unlike patterns of bone strain in the corpus of the macaque mandible (Hylander, '79a), in some instances the ipsilateral values are directed more vertically than the contralateral values, while in other instances the reverse pattern is found. 2. Single-element strain gage analysis of bone strain during the power stroke of mastication and incision A single-element strain gage was positioned either in the subcondylar region (Experiments 8, 9, 12) or deep to the superficial masseter muscle (Experiment 4). The strain gage position and alignment are illustrated in figure 12. In the subcondylar experiments, mean values of compressive subcondylar bone strain during mastication are larger on the contralateral side than on the ipsilateral side (Experiments 8, 9,121, and these mean values are significantly different from one another (p < 0.001) (table 2). Mean values of compressive bone strain during incision exceed corresponding values during mastication (table 21, although these values are not always significantly different from one another (contralatera1 and incision mean values are not significantly different from one another in Experiments 8 and 9). Similar t o the rosette analysis of bone strain deep to the superficial masseter muscle, a very different pattern of bone strain occurs deep to the superficial masseter. Whereas compression was sensed by the single-element gages in the subcondylar region (Experiments 8,9,12), the single-element gage deep t o the superficial masseter muscle sensed tensile strain (Experiment 4) (table 2 and fig. 12). Exp. 4 A Exp. 9 Fig. 12 Tracings of lateral radiographs of the macaque mandibles. The black dots indicate the position of the single-element gages and the lines indicate the orientation of the long axis of each gage. In Experiment 12 the lines labeled B and C represent the B and C elements of a rosette. 3. A comparison of expected and recorded patterns of strain during the power stroke of mastication and incision Patterns of recorded subcondylar bone strain support the hypothesis that a compressive reaction force is present along the TMJ during the power stroke of mastication and incision. Thus, the direction of the recorded principal strains (fig. 10) range between the theoretical patterns seen in figures 6A and 6B. Since eubcondylar strain levels are always larger on the contralateral side, these data further suggest that among macaques (and by inference humans and all other anthropoids) the contralateral TMJ is loaded more than the ipsilateral TMJ during mastication. This confirms the patterns predicted TMJ REACTION FORCE 445 'CABLE 2 Descriptive statistics ofsubcondylar bone strain values during incision and mastication ofapples (single-elementgage analysis) Experiment N Mean Largest S.D M. mulatta (6-17) Experiment 4 Ipsilateral Contralateral Opening (contralateralside) - - - 132 200 27 29 - 258 - 420 42 103 87 - 43 38 63 162 222 33 - 146 102 83 16 - 105 100 26 - 65 - 175 - 237 90 M. fascicularis (2) Experiment 8 Ipsilateral Contralateral Incision M. fascicularis (2) Experiment 9 Ipsilateral Contralateral Incision M . fascicularis (4) Experiment 12 Ipsilateral Contralateral Incision - 181 - 204 - 420 42 - 83 - 190 - 63 - 68 - 113 19 37 24 Element B - 134 - 380 - 514 23 81 145 Element C Ipsilateral Contralateral Incision 105 100 26 - 70 - 182 - 256 - 146 434 - 560 - 28 86 166 Xote. Bone strain values are in microstrain units N = number of power strokes or openingstrokes. Animal identification num ber in parentheses. In Experiment 12, one element of three-element rosette malfunctioned (A element) after a limited amount of data were recorded (table 1). for apes, macaques, and humans (Gysi, '21; Hylander, '75, '79b; Hylander and Sicher, '79; Smith, "78). These data make it unlikely that in humans both joints ordinarily are loaded equally during mastication (Roydhouse, '551, or that in humans the ipsilateral side is loaded more than the contralateral side (Hekneby, '74). Since both sides are apparently loaded, these data do not support the hypothesis that TMJ reaction forces are only present on the contralateral side (Walker, '78; Greaves, '78) or ipsilateral side (Page, '54). Moreover, these data are incompatible with the hypothesis that TMJ reaction forces are absent during mastication (Wilson, '20, '21; Robinson, '46; Scott, '55; Steinhardt, '58; Tattersall, '73; Roberts and Tattersall, '74). Patterns of recorded bone strain deep to the superficial masseter muscle are very similar t o the expected patterns seen in figure 5 . These data support the hypothesis that bone strain in this region is due primarily to muscle force from the medial pterygoid and masseter muscles. In Experiment 7 bone strain along the ipsilateral side exceeded contralateral- side values during mastication, suggesting that the ipsilateral-side muscle force exceeds the contralateral-side muscle force during unilateral mastication of apples. Further support for this possibility comes from EMG studies by Luschei and Goodwin ('74) which show that the peak electrical activity of the masseter muscle in macaques is greater on the ipsilateral side than on the contralateral side during mastication. The strain data also are consistent with EMG studies in humans which demonstrate that masseter and medial pterygoid muscles are more active on the ipsilateral side during mastication of soft food objects (Moller, '66; Ahlgren, '66). The findings here are also consistent with the lever model of mandibular function. The lever hypothesis predicts that reaction forces are present along the TMJ during the power stroke. An alternative model suggests the mandible functions as a beam, rather than a lever, during the power stroke (Smith, '78; Walker, '78). This suggestion is apt to confuse rather than enlighten, since every loaded lever functions as a beam, i.e., a beam is a 446 WILLIAM L. HYLANDER member subjected to transverse loads (loads perpendicular to its long axis) (Arges and Palmer, '63). When analyzing external forces on the mandible (muscular, bite, and condylar reaction forces) i t is appropriate to consider whether or not the mandible functions as a lever. When analyzing internal forces of the mandible (i.e., those forces tending to resist bone deformation), i t is appropriate to consider the mandible as a beam and to analyze bending stress within it utilizing beam theory if the mandible is being loaded transversely. Stating that i t is more appropriate to consider the mandible as a beam, rather than a lever, is much like saying it is more appropriate to consider a macaque a mammal, rather than a primate. It of course all depends on the problem. 4. Rosette analysis of bone strain during jaw opening Bone strain data during wide opening are presented in table 1.The magnitude of strains recorded from the subcondylar region and deep to the superficial masseter muscle during wide jaw opening is usually less than strain magnitudes during the power stroke or during isometric biting (Experiments 1, 13, 7). Moreover, strain magnitudes are greater during wide opening than during the opening stroke of mastication. The direction of E~ during wide opening is upwards and forwards in both the subcondylar region and deep to the superficial masseter muscle (fig. 10).The direction of t qduring the opening stroke of mastication is positioned similarly, although not figured here. 5. A comparison of expected and recorded patterns of bone strain during jaw opening The direction of the principal strains during jaw opening are similar to the direction of principal strains predicted if the subcondylar region and the region deep to the superficial masseter muscle were experiencing tensile bending stress due to midline bending of the mandibular ramus (fig. 8). This bending is presumably due to the force of the inferior head of the lateral pterygoid muscle. Experimental Set B 1. Rosette analysis of bone strain during isometric biting on the wooden dowel In these experiments the rosette was positioned in the subcondylar region (Experiments 1, 5, 11, 13) and deep to the superficial Exp. 1 wx w1 E x p . 11 n E x p . 13 Fig. 13 Tracings of lateral radiographs of the macaque mandibles. The black dots indicate the position of the rosettes. The X axis is parallel to the A strain gage element. The lines indicate the mean value for the direction of c 2 during isometric biting of the wooden dowel; lines labeled b and w indicate contralateral and ipsilateral aides, respectively; lines labeled i indicate incisal biting. In Experiment 13, 1 and 3 indicate biting along the M1 and M3, respectively. masseter muscle (Experiment 7). The location of the rosette in each experiment is shown in figure 13. (a) Bone strain magnitudes When the rosette was positioned in the subcondylar region the mean value of e 2 on the 447 TMJ REACTION FORCE TABLE 3 Descriptive statistics of subcondylar bone strain values during mnmetric bzting (rosette analysB) Principal strains Exprriment Angle of minimum principal strain Minimum Ic,) Maximum ( c j N Mean Largest SD Mean 1,argest SU Mean 37 28 61 115 261 204 236 338 320 59 55 75 -109 -373 -243 -215 -472 -372 55 71 89 11.6 6.4 6.7 23 31 96 144 200 404 53 108 -224 -331 -360 -972 84 234 32 57 71 59 111 121 23 37 91 -105 -131 -183 7 23 42 5 109 200 43 115 - 48 -305 - 69 -857 16 105 247 72 20 22 7 99 146 92 168 509 166 68 121 64 Range SD M mulatta (3-82) Experiment 1 Ipsilateral Contralateral Incisor -14.5to 39.5 12.9 22.5 15.0 3.4 5.9 56.9 55.3 51.lto 63.8 48.7to 66.2 3.6 4.8 24 39 96.2 89.6 71.9 to 124.4 67.5 to 105.6 13.5 7.0 17 260 37.2 -13.3to -263 72 14.3 - -183 -768 -220 58 184 25 100.2 39.9 25.8 - 1.1to - 1.7to M. mulatta (6-41) Experiment 5 Ipsilateral Contralateral M. fascicularis (1) Experiment 7 Ipsilateral Contralateral ~ M. fascicularis (3) Experiment 11 Ipsilateral' Contralateral M. fascicularis (4) Experiment 13 Ipsilateral (M1) Ipsilateral (M3) (strain reversal) Contralateral Incisor - 86 -101 -410 -183 - - 72.7 17.4 5.1to 47.0 16.7 77.8 to 129.3 28.9to 57.0 5.2to 55.8 111.9 7.8 23.6 Note: Bone strain values are In microstrain units. N = number of isometric bites. Angle values are in degrees. Animal identification number in parentheses *Accurate determination of minimum principal strain angles not possible due to low strain levels. The hite point is unilateral along the M1-MZ region unless otherwise noted in parentheses. contralateral side during molar biting exceeds the corresponding ipsilateral-side value by a considerable amount within each experiment (table 3). In each instance, these mean values are significantly different from one another (p < 0.05). In contrast, although the mean value of c 2 deep to the superficial masseter muscle is larger on the contralateral side during molar biting, these values are not significantly different from one another. In each experiment the mean value of e q during incisor biting is larger than the mean value on the ipsilateral side and smaller than the mean value on the contralateral side during molar biting (Experiments 1, 13). These values are significantly different from one another (p < 0.01). (b) Bone strain direction Generally the direction of e 2 in the subcondylar region during molar biting is upwards and backwards (Experiments 1, 5, 11, but see Experiment 13 discussed below). In contrast, the direction of c 2 deep t o the superficial masseter muscle is upwards and forwards, rather than upwards and backwards. All of the mean values for the direction of e 2 within each experiment are significantly different from one another (p < 0.05) except as indicated in figure 13 (table 3). In Experiment 1 3 the monkey was allowed t o bite the dowel along its third molars (M3) as well as the first molars ( M l ) , whereas only M1-M2 biting data were obtained from the other experiments. Whenever biting took place along the M3 and when ipsilateral subcondylar bone strain was being monitored, bone strain patterns suddenly were reversed from the patterns which occurred during biting on the M1 (Experiment 13). This reversal did not occur during M3 biting when subcondylar bone strain was being monitored along the contralateral side. It also did not occur in any of the experiments in which biting was restricted solely to the M1-M2. Since the two strain patterns for the ipsilateral side in Ex- 448 WILLIAM L. HYLANDER periment 13 are quite different, they are analyzed and presented separately. The mean values for the magnitude of e 2 for these two patterns are not significantly different from one another, but the mean values for the direction of are dramatically different. The direction of e 2 during biting along the M1 is upwards and backwards, as it is in the other subcondylar bone strain experiments. In contrast, when biting occurred along the M3 the direction of e 2 was shifted almost 90" (fig. 13). 2. A comparison of expected and recorded patterns of strain during isometric biting With the exception of recorded subcondylar bone strain patterns along the ipsilateral side during biting on the M3 (Experiment 131, subcondylar bone strain patterns during isometric biting suggest that the TMJ is loaded by a compressive reaction force (compare fig. 13 with figs. 6A,B). Furthermore, since subcondylar strain levels are higher on the contralateral side, these data suggest that the contralateral TMJ is loaded more than the ipsilateral TMJ. Strain patterns deep to the superficial masseter muscle indicate that this region is deformed primarily by the medial pterygoid and masseter muscles (note the similarity between fig. 13 and fig. 5). The data suggest that the muscle forces on the ipsilateral and contralateral sides may not differ significantly from one another during isometric biting. These data are in accord with Blair's data (cited in Luschei and Goodwin, '74) which show that during isometric molar biting, EMG activity patterns in the masseter muscle on both ipsilateral and contralateral sides in macaques are very similar. Similar observations on human subjects have been reported (Finn et al., '79). This EMG and bone-strain pattern is in contrast t o the pattern found during mastication of soft foods which suggest that the masticatory muscles on the ipsilateral side generate much more force than the muscles on the contralateral side. Perhaps biting hard objects or masticating tough foods requires relatively more contralateral jaw muscle activity than chewing soft foods. Moller ('74) notes that whereas the masseter and medial pterygoid muscles are much more active on the ipsilateral side than on the contralateral side during gum chewing in humans, levels of activity for both sides are nearly equal during carrot mastication. Furthermore, Luschei and Goodwin ('74) note that differences in peak EMG amplitudes between masticatory muscles on each side usually are very small when macaques chew monkey biscuits. In summary, i t appears that among macaques (and perhaps humans also?) a greater relative amount of contralateral muscle force is generated during isometric biting or powerful mastication, than during the mastication of soft foods such as apples. Another notable feature of these data is the fact that subcondylar bone strain during isometric biting along the M3 suddenly reversed itself from the pattern seen during biting along the M1 (Experiment 13). Moreover, whereas strain patterns during biting along the M1 are very similar to expected patterns associated with a compressive reaction force along the TMJ (compare fig. 13 with fig. 6A), the recorded strain pattern during biting along the M3 is very similar to the pattern seen in figure 3A. The strain reversal data suggest a number of possible interpretations: (11 Compressive reaction forces are very small or absent along the ipsilateral TMJ during biting along the' M3; (2) Compressive reaction forces are not only absent, but there is a tendency for the mandibular condyle t o lift off the articular eminence on the ipsilateral side; furthermore, this movement is checked by the temporomandibular ligament (sphenomandibular and stylomandibular ligaments also?), resulting in tensile stress across the TMJ which contributes to the recorded strain patterns. (3) The articular surface of the mandibular condyle is asymmetrically loaded, i.e., although there is still a compressive reaction force across the TMJ, the resultant reaction force acts along the medial aspect of the mandibular condyle rather than along the middle or lateral portion. This latter type of loading causes the subcondylar region t o be bent so that the medial portion experiences compressive stress while the lateral portion experiences tensile stress, thereby causing the recorded deformation pattern (fig. 7). The available data do not allow any of the above possibilities t o be confidently eliminated, but a mechanical analysis of the ipsilateral side of the mandible increases the plausibility of the first two possibilities, i.e., either the existence of no compressive stress or possibly tensile stress along the ipsilateral TMJ during M3 biting. As seen in figure 14, there are a t least three forces acting on the ipsilateral side of the ma- TMJ REACTION FORCE caque mandible during unilateral isometric biting. These are bite force (Fb), an ipsilateral resultant muscle force (F,) and a force being transmitted through the symphysis from the contralateral t o the ipsilateral side (F,). The last force is associated with the contralateral muscle force. In order to simplify this analysis, only the vertical components of these forces are considered here and no consideration is given to the effects they have on mandibular twisting. If moments are taken about the bite point in figure 14,then in order for there not to be a reaction force through the TMJ during conditions of equilibrium, the moment associated with F, must equal the moment associated with F,, i.e., (FJ(y1 = (F,J(x). If (F,)(y) is greater than (Fml(x),a tensile reaction force will be present along the ipsilateral TMJ. If (F,)(y) is less than (Fm)(x),a compressive reaction force must be present along the ipsilateral TMJ. From the above analysis, it can be seen that altering the bite point (i.e., shifting it from MI to M J should markedly change the reaction force patterns along the ipsilatera1 TMJ since moving the bite point posteriorly decreases x and increases y while moving the bite point anteriorly has the opposite effect. This analysis also demonstrates that it is theoretically possible t o have either no stress or tensile stress acting across the ipsilateral TMJ during unilateral molar biting. This possibility supports models of TMJ function advanced by Walker ('78) and Greaves ('78). In macaques, however, these models are only supported by data during biting on the MB. Models advanced by Gysi ('211, Hylander ('75, '79b), Hylander and Sicher ('791, and Smith ('78) are also supported by these data since these models predict either no stress or tensile stress along the ipsilateral TMJ during biting along the more posterior aspect of the tooth row. Models advanced by Page ('54) (only stress on ipsilateral TMJ), Hekneby ('74) (more stress on ipsilateral TMJ), and Roydhouse ('55) (equal stress on ipsilateral and contralateral TMJ), however, are not supported by these data. The possibility of tensile stress across the ipsilateral TMJ correlates with cinefluorographic observations by Koivumaa ('61). Koivumaa reported that the mandible rotated about the bite point during unilateral molar biting of a hard caramel in several human subjects. At this time the mandibular incisors moved upwards while the ipsi- 449 Fig. 14 Mechanics of the ipsilateral side of t h e macaque mandible during unilateral isometric molar biting. Condylar reactions are not included here. This is a firstlevel approximation since only the vertical components of the (1) bite force (Fb), (2) ipsilateral muscle force (Fm), and (3) force transmitted from contralateral to ipsilateral sides (F,), are considered. If moments are taken about the bite point, Y and X are t h e distances between the bite point and F, and F,, respectively. In order to satisfy conditions of equilibrium and not have reaction force along t h e TMJ. F, X must equal F, . Y. If F,, X is larger or smaller than F, Y, there will be a compressive or tensile reaction force along the TMJ, respectively. Moving the bite point caudally increases Y and decreases X. Moving t h e bite m i n t rostrallv has the omosite effect (see text for discussion). . . . I. lateral mandibular condyle lifted off the articular eminence several millimeters. Perhaps at this time the temporomandibular ligament checked further condylar movement, resulting in tensile stress across the ipsilateral TMJ. 3. Comparison between isometric biting data and mastication data In general, the isometric biting data are very similar t o the mastication and incision data in that (1) mean compressive bone strain values in the subcondylar region on the contralateral side exceed ipsilateral-side values; (2) mean values for the direction of c 2 in the subcondylar region are generally upwards and backwards; and (3) mean values for the direction of e* deep to the superficial masseter muscle are upwards and forwards. Mastication and incision data differ from isometric biting data in the following respects: (1) bone strain values tend to be larger during isometric molar biting; (2) in contrast t o mastication the difference between bone strain values on the ipsilateral and contralateral sides during isometric molar biting is larger; (3) whereas subcondylar bone strain values 450 WILLIAM L. HYLANDER during apple incision generally exceed subcondylar bone strain values during apple mastication, bone strain values during isometric incisor biting do not generally exceed bone strain values during isometric molar biting; and (4) subcondylar bone strain reversals on the ipsilateral side were not encountered during mastication, as they were during isometric molar biting. It is not surprising that bone strain values during isometric biting exceed bone strain values during mastication. A soft object like an apple does not require a powerful bite, and therefore is not associated with powerful muscular and reaction forces. There are two possible reasons why isometric molar biting results in larger bone strain differences between ipsilateral and contralateral sides. First, in contrast to apple mastication, powerful isometric molar biting might be associated with larger relative amounts of muscle force on the contralateral side. Second, the bite point during isometric molar biting (usually along the M I M,region) might be slightly caudal to the resultant bite point during mastication (position unknown). Both of these possibilities have the effect of increasing F,. y relative to F,.x (fig. 141, and therefore they will both reduce the amount of compressive reaction force along the ipsilateral TMJ, which will result in larger differences in reaction force between ipsilateral and contralateral sides. Furthermore, increasing the amount of muscle force along the contralateral side will directly increase the compressive reaction force along the contralateral TMJ, contributing to a larger difference between ipsilateral and contralateral sides. As noted above, in contrast to apple mastication, incisal biting of apples, involves larger amounts of subcondylar bone strain. In addition to obvious mechanical factors which should favor larger condylar reaction forces during incisal biting, this observation probably also reflects the relative ease of apple mastication as opposed to incisal biting of apples. A similar pattern was found while recording bone strain in the mandibular corpus during apple incision and mastication (Hylander, '79a). Isometric biting on the incisors, however, usually results in lower subcondylar strain values than during isometric molar biting. This observation probably indicates relatively low incisor bite force values. In macaques, a single incisor tooth and its periodontium is not nearly as well adapted as a molar tooth and its periodontium to withstand large X Exp. 7 Fig. 15 Tracings of lateral radiographs of two macaque mandibles. The black dots indicate the position of the rosettes. The X axis is parallel to the A strain gage element. The lines indicate only the mean value for the direction of c 2 during transducer biting; lines labeled b and w indicate contralateral and ipsilateral sides, respectively; p indicates premolar biting and 1 and 3 indicate biting along the M1 and M3, respectively. biting forces. Perhaps isometric incisor bite force values are low in order to avoid pain in the periodontium of the one or two incisors involved in biting a dowel. In contrast, when several incisors are loaded simultaneously, as in apple incision, larger muscle, bite, and TMJ reaction forces can be generated without discomfort. Finally, it is not surprising that strain reversals were only encountered during isometric molar biting; during mastication, the moment (F,) (x) probably always exceeds (F,) (y) because the resultant bite force is located in the M,-Mz region (fig. 14).On the other hand, altering the position of the bite point to the M 3during isometric biting results in large values of y and small values of x, causing the moment (F,) (XIto be exceeded by (or equal to) the moment (FJ (y). If this occurs, tensile stress (or no stress), rather than compressive stress, will act across the joint. 4. Rosette and single-elementstrain gage analysis of bone strain during isometric transducer biting In these experiments, a rosette or single-element strain gage was positioned in the sub- 451 TMJ REACTION FORCE Y t contralateral Confralotcml r=-077 Slops = - 3 0 . 5 : # . .* .. r : -0.96 (1 G -400- a E g ... . - ...- . . . .... . . . . . ...-..*:-. . ..... - -200- - . . . . . *. . x : * ..... . - . . *. . ~ slope: -15.0 -100- * I -200- & i a 0- .. . . . -3 - f In s a.. -300- -4001 Exp.il -600- * I Ex0.9 0 0 4 12 8 Bite Farce (kgl . -5 J 8 24 16 16 32 Bite Force lkgl OC. r--060 Slope--l5.4 . . * * lpsllateral r: -0.53 s1o.e: -1.7 . t 5 .. : .. -300- E *- -400 0 4 8 Bite Force 1kq) 12 I 16 -looj 0 Elp9 8 * 16 24 32 B i l e Force I k g l Fig. 16 Scattergram of bite force and subcondylar bone strain values in Experiment 9; r is Pearson's product-moment correlation coefficient. Fig. 17 Scattergram of bite force and subcondylar values of c 2 in Experiment 11 during transducer biting; r is Pearson's product-moment correlation coefficient. condylar region (Experiments 8, 9, 11) and deep to the superficial masseter muscle (Experiment 7 ) .The location of the rosette for Experiments 7 and 11 is shown in figure 15, and the location and alignment of the single-element gage for Experiments 8 and 9 is shown in figure 12. side than on the ipsilateral side during molar biting. Ipsilateral bone strain in the subcondylar region was also recorded for premolar biting (Experiment 111, and it was found that bone strain on the ipsilateral side was relatively greater during premolar biting than during molar biting since the slope value for premolar biting is twice as large as, and significantly different from, the slope value for molar biting (p < 0.01) (table 4). In the one experiment measuring bone strain deep to the superficial masseter muscle during isometric transducer biting (Experiment 7), the slope values for the ipsilateral and contralateral sides are not significantly different from one another, and are much smaller than the slope values associated with the subcondylar rosette values. (a) Bone strain magnitude Both descriptive statistics and the regression analysis of bone strain (dependent variable) and bite force (independent variable) values are presented in tables 4 and 5. Plots of bite force versus bone strain are presented in figures 16 and 17 for Experiments 9 and 11. In the three subcondylar bone strain experiments the slope values for the regression analysis of molar bite force and bone strain on the contralateral side always exceed the slope values for the regression of molar bite force and bone strain on the ipsilateral side, and these values are significantly different from one another (p < 0.01) (Experiments 8, 9, 11). These data clearly demonstrate that for a given amount of molar bite force, there is usually more bone strain on the contralateral (b) Bone strain direction The direction of the strains in the transducer biting experiments exhibit certain similarities encountered in previous experiments. For example, the gages in Experiments 8 and 9 both sensed compressive subcondylar bone strain during transducer biting, as they did 452 WILLIAM L. HYLANDER during apple mastication (tables 2,5). In addition, the direction of t 2deep to the superficial masseter muscle during transducer biting is similar to its direction during both apple mastication and isometric biting on the dowel (ExIn one experiment, periment 7) (tables 1,2,4). however, the direction of E during transducer biting was unlike anything previously encountered for the analysis of bone strain in the subcondylar region (Experiment 11).Whereas c 2 in all previous subcondylar experiments is directed upwards and backwards (except for strain reversals in Experiment 13 during isometric dowel-biting along the M3), the direction of t 2 i n Experiment l l during transducer biting is upwards and forwards (fig. 15). In contrast, the direction of c 2 in Experiment 11 during isometric biting on the dowel is quite different (compare fig. 13 with fig. 15); the comparison between the mean values of the direction of t 2 on the contralatera1 side for transducer biting and isometric dowel biting reveals a difference of 33" ( < 0.001). Similar to isometric dowel biting along the MB,strain reversals were encountered whenever the transducer was bitten along the M 3 (Experiment 11).The two strain patterns are presented separately in table 4 and figure 15. Unlike Experiment 13, however, the direction of c 2during the strain reversal in Experiment 11 is upwards and backwards, rather than upwards and forwards. 5. A comparison of expected and recorded patterns of strain during transducer biting One of the isometric transducer-biting experiments is somewhat a t variance with other transducer-biting experiments and with the previously presented mastication and isometric biting data. That is, in Experiments 8 and 9 the long axis of each single-element gage is aligned upwards and backwards (fig. 12) and both of these gages sensed compressive bone strain during transducer biting. These data support the hypothesis that the mandibular condyle is loaded by a compressive reaction force and furthermore are very similar to the mastication data. In one experiment, however, the direction of t 2 during transducer biting is upwards and forwards, rather than upwards and backwards as it was during isometric dowel biting (Experiment 11). The directional data in the transducerbiting experiment support the existence of one TMJ REACTION FORCE 453 TABLE 5 Descriptiue statistics and regression analysis ofsubcondylar bone strain duringtransducer biting (single-elementgage analysis) Bite values (kg) Experiment Compressive strain N Mean Largest S.D. Mean Largest S.D. Slope 175 233 15 11.6 8.0 2.9 34.4 26.4 4.8 8.8 6.0 1.5 -145 -249 -118 -320 -980 -238 65 233 76 - 1.5 -0.20 -34.8 -36.4 99 160 10.8 12.0 32.5 31.0 8.0 9.1 - 63 -161 -114 -474 26 143 -15.0 r M. fascicularis (2) Experiment 8 Ipsilateral Contralateral Incisor M fascicularis (2) Experiment 9 Ipsilateral Contralateral - -0.90 -0.69 1.7 -0.53 -0.96 Note: Bone strain values are in microstrain units. N = number of bites un the transducer; r = Pearson's product-moment correlation coefficient; all values of r are significantly different from zero (p < 0.05).The bite point is unilateral along the M1-MZ repion. of two possible situations. In some subjects variability may be partly related to object size during transducer biting either there is (1)no since the height of the transducer block is 10 TMJ reaction force on either ipsilateral or mm while the dowel is only 2-5 mm thick contralateral sides or (2) the mandibular con- (depending on how much the animal crushed dyle is loaded along its medial aspect, rather the dowel during biting). Perhaps the wider than along its middle or lateral aspect, result- gape during transducer biting which is correing in compressive bending stress and strain lated with greater rotation and translation of along the medial aspect of the subcondylar re- the mandibular condyle somehow resulted in gion and tensile stress and strain along the altered TMJ loading patterns. lateral aspect (fig. 7). The regression analysis Recorded bone strain patterns deep t o the of bite force and bone strain magnitudes tends superficial masseter muscle during transto support the hypothesis that a TMJ reaction ducer biting are quite similar to expected patforce is present in Experiment 11. For exam- terns (compare fig. 5 with fig. 15) (Experiple, there is much more subcondylar bone ment 7). The regression analysis of bite force strain (per unit bite force) along the contra- and bone strain values in this experiment suglateral side than along the ipsilateral side dur- gest that peak magnitude of the ipsilateral ing molar biting. This is similar to the data for and contralateral muscle force of the masseter mastication and isometric dowel biting. In and medial pterygoid muscles are not signifiaddition, there is more subcondylar bone cantly different from one another during strain on the ipsilateral side during premolar isometric transducer biting. biting than during molar biting. These findMorphological and experimental evidence ings correlate with what would be expected if for differential TMJloading reaction forces are both present and vary with bite point position (fig. 14). Many workers have noted that although the Similar to isometric biting of the dowel, a TMJ differs from a typical synovial joint, its subcondylar bone strain reversal was encoun- articular surfaces are nonetheless well tered along the ipsilateral side when the sub- adapted to bear stress (see Hylander, '75, for a jects were allowed to bite a transducer along review). In most primates, the presumed their M,. There is no apparent explanation as stress-bearing articular surfaces of the joint to why the direction of t 2is upwards and back- are confined primarily t o the articular emiwards at this time. Unfortunately, there is nence of the temporal bone and the mandibuonly one experiment on the direction of E~ lar condyle. There is evidence that the reacalong the subcondylar region during trans- tion forces within the stress-bearing region of ducer biting. Until more data are collected, the primate TMJ during the power stroke of the only conclusion to be drawn is that among mastication and incision are not always equalsome subjects the direction of e 2during trans- ly distributed. This evidence comes from exducer biting varies from its direction during perimental studies of jaw movement and manmastication and isometric dowel biting. This dibular bone strain in primates, patterns of re- 454 WILLIAM L. HYLANDER modeling in the human TMJ, and pathological changes in the human TMJ. I t has been demonstrated cinefluorographically that during the power stroke of mastication the mandibular corpus of galagos rotates about its long axis (Hiiemae and Kay, '73). This rotation or twisting which is mediated by an unfused mandibular symphysis, causes the lower border of the galago mandible to evert and the alveolar and coronoid processes to invert. Since the galago mandibular condyle maintains contact with the temporal bone during the power stroke of mastication (via the interposed articular disc), the lateral surface of the galago condyle must press against the intervening disc and temporal bone, while the medial surface of the mandibular condyle, disc, and temporal bone are slightly separated. Presumably then, among galagos (and all other mammals whose dentaries rotate as described above) only the lateral aspect of the TMJ is loaded during certain portions of the masticatory cycle. The mandible of macaques and other anthropoids does not rotate visibly during the power stroke because the mandibular symphysis is fused; therefore, the medial aspect of the mandibular condyle does not necessarily lose contact with the articular eminence and intervening disc during the power stroke. Nevertheless, because of the similar organization of the galago and macaque masticatory apparatus, differential loading within the TMJ probably also takes place among macaques. A stress analysis of the mandible supports this assumption by demonstrating that the mandibular corpus of macaques is twisted about its long axis during the power stroke of mastication (as among galagos) (Hylander, '79a). If the macaque TMJ is loaded by a compressive reaction force during the power stroke (as the data in this study suggest), and if the macaque mandible is simultaneously twisted (as the data presented by Hylander, '79a indicate), then the macaque TMJ should also be loaded differentially so long as the twisting moment associated with the bite force does not completely counter the twisting moment associated with the ipsilateral resultant muscle force. Moreover, differential joint loading will be more pronounced within the ipsilateral TMJ among both galagos and macaques, since twisting is much more pronounced on the ipsilateral side than on the contralateral side (Hylander, '79a). Therefore, specific portions of the ipsilateral TMJ are perhaps stressed more (i.e., transmit more force per unit area) than the contralateral joint, although the data in this study suggest that among macaques the largest overall net reaction forces ordinarily will exist on the contralateral side. This latter finding is in contrast t o experimental evidence which suggests that in galagos the ipsilateral TMJ is loaded with a larger net reaction force than is the contralateral TMJ during isometric molar biting (Hylander, '79b). Patterns of TMJ remodelling also suggest differential loading within the joint. For example, Moffett et al. ('64) have shown that the lateral aspect of the adult human TMJ remodels differently from the medial aspect, and although these remodelling patterns do not directly indicate the nature of the loading pattern, they do suggest differences in mechanical loading patterns within the TMJ. In a recent study of the human TMJ (Oberg et al., '71) a number of cases were found with degenerative TMJ changes, most of which were associated with the lateral aspect of the joint. For example, in 13 of 14 individuals with perforations of the articular disc, the perforations were restricted t o the lateral aspect of the joint. (In the fourteenth case, the perforation was restricted t o the middle aspect.) Kopp et al. ('76) also noted degenerative changes along the lateral aspect of the human TMJ. Since pathologic changes in joints are often related t o local mechanical factors (i.e., excessive or highly repetitive stresses), these data suggest that the lateral aspect of the human TMJ experiences more stress, and therefore more "wear-and-tear,'' than the medial aspect. CONCLUSIONS In vivo bone strain data from the subcondylar region of macaques suggest the following: The macaque TMJ is loaded by a compressive reaction force during the power stroke of mastication and incision, and during isometric molar and incisor biting. TMJ reaction forces are larger on the contralateral side during both mastication and isometric molar biting. Patterns of ipsilateral TMJ reaction force in macaques during isometric biting vary markedly in response to the position of the bite point. During biting along the premolars or first two molars a compressive reaction force acts about the ipsilateral TMJ; however, when the bite point is positioned along the M3, the ipsilateral TMJ has either very little com- TMJ REACTION FORCE pressive stress, no stress, or it is loaded in tension. ACKNOWLEDGMENTS I am grateful to Dr. R. Bays, for without his surgical expertise this study would not have been possible. I would also like to thank Marianne Bouvier, Matt Cartmill, Karen Hiiemae, Richard F. Kay, George Lauder, and Jack Stern for reading and commenting on various drafts of this manuscripts, and helping to prepare the figures. Finally, I am grateful to Dr. Otto Hagen, Department of Surgery, Duke University, for allowing me to use his experimental animals (NIH grant HL-15448). LITERATURE CITED Ahlgren, J. 1966 Mechanism of mastication: a quantitative cinematographic and electromyographic study of m s t i c a t o r y movements in children, with special reference to occlusion of the teeth. Acta Odontol. Scand., 24 (Suppl. 44): 1.109. Anderson, D. J. 1956 Measurement of stress in mastication. I. J. Dent. Res., 35: 664-670. Arges, K. P., and A. E. Palmer 1963 Mechanics of Materials. McGraw-Hill Book Co., New York. Bramble, D. M. 1978 Origin of the mammalian feeding complex; models and mechanisms. Paleobio., 4: 271-301. Crompton, A. W., andK. M. Hiiemae 1969 How mammalian molar teeth work. Discover, Yale Peabody Museum, 5: 23-34. 1970 Molar occlusion and mandibular movements during occlusion i n t h e American opossum, Didelphis marsupialis L. Zool. J. Linn. SOC.,49: 21-47. Dally, J. W., and W. F. Riley 1965 Experimental Stress Analysis. McGraw-Hill Book Co., New York. Davis, D. D. 1964 The Giant Panda. Fieldiana: Zoology Memoirs, 3: 1-337. de Vree, F.. and C. Gans 1976 Mastication in pygmy goats (Capra hircusi. Ann. Soc. Roy. 2001. Bel., 105: 255-306. Finn, R., G. Throckmorton, W. Gonyea, D. Barker and W. Bell 1979 Neuromuscular characteristics of vertical maxillary dysplasias. In: Surgical Correction of Dentofacial Deformities. W. Bell, W. Proffit, and R. White, eds. W.B. Saunders Go., Philadelphia, in press. Gingerich, P. D. 1971 Functional significance of mandibular translation in vertebrate jaw mechanics. Postilla, 152: 1-10. Graf, H. 1975 Occlusal forces duringfunction. In: Occlusion: Research in Form and Function. N. H. Rowe, ed. The University of Michigan School of Dentistry and the Dental Research Institute, Ann Arbor, pp. 90-109. Greaves, W. S. 1978 The jaw lever system in ungulates: a new model. J. Zool. (London), 184: 271-285. Gysi, A. 1921 Studies on the leverage problem of t h e mandible. Dent. Digest, 27: 74-84, 144-150, 203-208. Hekneby, M. 1974 The load of t h e temporamandibular joint: physical calculations and analyses. J. Prosthet. Dent., 31: 303-312. Hiiemae, K. M., and R. F. Kay 1973 Evolutionary trends in t h e dynamics of primate mastication. In: Symp. Fourth Int. Cong. Primatology, 3. M. R. Zingeser, ed. Karger, Basel, pp. 28-64. Hylander, W. L. 1975 The human mandible: lever or link? Am. J. Phys. Anthrop., 43: 227-242. 1977 In uiuo bone strain in the mandible of 455 Galago crassicuudatus. Am. J. Phys. Anthrop., 46: 309-326. 1978 Incisal bite force direction in humans and t h e functional significance of mammalian mandibular translation. Am. J. Phys. Anthrop., 48: 1-7. 1979a Mandibular function in Galago c r a s sicaudatus and Macaca fascicularis: an i n vivo approach to stress analysis. J. Morph., 159: 253-296. 1979b Functional significance of primate mandibular form. J. Morph., 160: 223-240. Hylander, W. L., and H. Sicher 1979 Functional anatomy of the temporomandibular joint. In: The Temporomandibular Joint: A Biologic Basis for Clinical Practice. Third ed. B. G. Sernat and D. Laskin, eds. Charles C Thomas, SprinRfield, in press. Koivumaa, K. 1961 Cineflurographic analysis of t h e masticatory movements of the mandible. Suom. Hammaslaak. Toim., 57: 306-336. Kopp, S.,G. E. Carlsson, T. Hansson and T. Oberg 1976 Degenerative disease in the temporomandibular, metatarsophalangeal and sterno-clavicular joints. Acta Odontol. Scand., 34: 23-32. Luschei, E. S., and G . M. Goodwin 1974 Patterns of mandibular movement and jaw muxcle activity during mastication in the monkey. J. Neurophysiol., 37: 954-966. Maxwell, L. C., D. S. Carlson, J . A. McNamara and J. A. Faulkner 1979 Histochemical characteristics of the masseter and temporalie muscles of t h e rhesus monkey ~Mucacamulattd. Anat. Rec., 193: 389-401. McNamara, J. A,, Jr. 1973 The independent functions of the two heads of the lateral pterygoid muscle. Am. J. Anat., 138: 197-204. Moffett, B. D., L. C. Johnson, J. B. Micabe and H. C. Askew 1964 Articular remodeling i n the adult human temporamandibular joint. Am. J. Anat., 215: 119-142. Moller, E. 1966 The chewing apparatus: an electromyographic study of the action of t h e muscles of mastication and its correlation to facial morphology. Acta Physiol. Scand., 69 (Suppl. 280): 1-129. 1974 Action of the muscles of mastication. In: Frontiers of Oral Physiology. Y. Kawamura, ed. Basel, Karger, pp. 121-158. Morgan, D. H. 1975 Surgical correction of temporomandibular joint arthritis. J. Oral Surg., 23: 766-773. Oberg, T., G. E. Carlsson and C. Fajers 1971 The temporomandibular joint. A morphologic study on human autopsy material. Acta odontol. Scand., 29: 349-384. Page, H. L. 1954 Temporomandibular joint physiology and jaw synergy. Dent. Digest, 60: 54-59. Ringquist, M. 1974 Fiber types i n human masticatory muscles. Relation to function. Scand. J. Dent. Res., 82: 333-355. Roberts, D., and I. Tattersall 1974 Skull form and the mechanics of mandibular elevation in mammals. Amer. Mus. Novitates, No. 2536, 1-9. Robinson, M . 1946 T h e temporomandibular j o i n t : theory of reflex controlled nonlever action of the mandible. J. Am. Dent. Assoc., 33: 1260-1271. Roydhouse, R. 1955 Upward force of the condyles on t h e cranium. J. Am. Dent. Assoc., 50: 166-172. Schumacher, G. H. 1961 Funktionelle Morphologie der Kaumuskulatur. Gustav Fisher Verlag, Jena. Scott, J. H. 1955 A contribution to the study of mandibular joint function. Brit. Dent. J., 94: 345-349. Sicher, H. 1951 Oral Anatomy. C. V. Mosby and Co., St. Louis. Smith, J. M., and R. J . G. Savage 1959 The mechanics of mammalian jaws. School Science Review, 40: 289-301. Smith, R. J . 1978 Mandibular biomechanics and tempo- 456 WILLIAM L. HYLANDER romandibular joint function in primates. Am. J. Phys. Anthrop., 49: 341-349. Steinhardt, G. 1958 Anatomy and function of the ternporomandibular joint. Inter. Dent. J., 8: 155-156. Tattersall, I. 1973 Cranial anatomy of Archeolemurinae (Lemuroidea, Primates). Anthropol. Pap. Am. Mus. Natl. Hist., 52: 1.110. Thexton, A. J., and K.M. Hiiemae 1975 The twitch-contraction characteristics of opossum jaw musculature. Arch. oral Biol., 20: 734-748. Turnbull, W. D. 1970 Mammalian Masticatory Apparatus. Fieldiana: Geology, 18: 1-356. Walker, A. 1978 Functional anatomy of oral tissues. In: Textbook of Oral Biology. E. A. Sweeney, C. C. Capuccino and S. M. Melfer, eds. W.B. Saunders Co.,Philadelphia, pp. 277-296. Wilson, G . H. 1920 The anatomy and physics of the temporomandibular joint. J . Natl. Dent. Assoc., 7: 414-420. 1921 The anatomy and physics of the temporomandibular joint. J. Natl. Dent. ASSOC., 8: 236-241.