Pattern of collagen fiber orientation in the ovine calcaneal shaft and its relation to locomotor-induced strain.код для вставкиСкачать
THE ANATOMICAL RECORD 242147-158 (1995) Pattern of Collagen Fiber Orientation in the Ovine Calcaneal Shaft and Its Relation to Locomotor-Induced Strain JAMES M. MCMAKON, ALAN BOYDE, AND TIMOTHY G. BROMAGE Department of Anthropology (J.M.M.) and Hard Tissue Research Unit, Department of Anthropology, (T.G.B.), Hunter College, City University of New York, New York; Department of Anatomy and Developmental Biology, University College London, London, U.K. (A.B.) ABSTRACT Background: Gebhardt (1905. Arch. Entwickl. Org., 20:187322) originated the hypothesis that the direction of collagen fibers in bone is a structural response to the type of mechanical load to which the bone is subjected. He proposed that collagen fibers aligned parallel to the loading a x i s are best suited to withstand tensile strain, whereas fibers oriented perpendicular to the loading axis are best able to resist compressive strain. Research comparing load patterns with fiber alignment in bone have tended to support Gebhardt’s hypothesis. The aim of the present study is to further test this hypothesis by assessing the correspondence between the distribution of strain and the distribution of collagen fiber orientation in a bone that is subjected to compound loading (i.e., both tension and compression at different phases during the loading cycle). The ovine calcaneum was selected to meet this criterion. Methods: Calcaneum surface strain distributions were obtained from experimental results reported by Lanyon (1973. J. Biomech. 6:41-49). Histological sections of the calcaneal shaft were prepared and observed using circularly polarized light (CPL)microscopy to determine the distribution of collagen fiber alignment. The observed alignment pattern was then compared with the predicted pattern based on Gebhardt’s hypothesis. Results: Contrary to previous studies, our findings show no clear correspondence between the strain type of greatest magnitude and the direction of collagen fibers. Areas of bone characterized by high compression and low tension showed predominantly longitudinal collagen alignment (contra to Gebhardt). Conclusions: It is argued that even small magnitudes of tension operating on local areas of bone may be sufficient to induce collagen alignment favorable to this type of strain, even when greater magnitudes of compressive strain are acting on the same bone volume. 0 1995 Wiley-Liss, Inc. Key words: Collagen fiber alignment, Bone strain, Biomechanics, Calcaneum, Talocrural joint, Ouis aries, Circularly polarized light microscopy Calcified tissue research has shown unequivocally that mechanical function plays a crucial role in regulating the structure of bone. Studies of patients with chronic limb paralysis, or with experimental animals in which functional loading is arrested have allowed investigators to distinguish between structural properties of bone that do not depend upon external loading stimuli and structural features that are dependent upon external function (i.e., structural responses to strain). It has been shown in such cases that the normal size and shape of various processes and tuberosities, the girth, cortical thickness and cross-sectional geometry, and the extent of longitudinal curvature of the shaft of long bones all depend on the functional 0 1995 WILEY-LISS,INC. loading environment for their development (Howell, 1917; Stinchfield et al., 1949; Lanyon, 1980). Early recognition of these facts prompted Wolff (1892) to postulate that the external conformation of bone will develop so as to best withstand, and take advantage of, the functional strains to which it is subjected. This principle was soon extended to include microscopic features of bone structure, and research into ~~~~ Received March 21, 1994; accepted January 18, 1995. Address reprint requests to James McMahon Hard Tissue Research Unit, Department of Anthropology, Hunter College, C.U.N.Y., 695 Park Avenue, New York, NY 10021. J.M. McMAHON ET AL. 148 the mechanical properties of such features as bone density, distribution of mass, tissue type, and the alignment of various tissues has since been a major focus of research (see Evans, 1970; Currey, 1984a). Gebhardt (1905),in an early investigation of collagen fiber alignment, proposed that the ability to resist different types of forces covaried with the direction of collagen fibers in bone. More specifically, collagen fibers aligned parallel with the loading axis (i.e., longitudinal in a long bone) are best able to withstand tensile stress, whereas collagen fibers oriented perpendicular to the loading axis (i.e., transverse in a long bone) are best able to resist compressive stress. The purpose of the present study is to test Gebhardt’s hypothesis by assessing the degree of correspondence between the distribution of principal strains and the pattern of collagen fiber orientation over entire crosssectional areas of the ovine calcaneal midshaft. This study differs from previous related work in that the calcaneal diaphysis (unlike many other long bone diaphyses) is subjected to both tension and compression at different phases during the loading cycle, and it was therefore possible to examine the effects of this compound loading on the alignment of collagen fibers. RESEARCH BACKGROUND In lamellar bone, the arrangement of collagen fibers and mineral crystallites is highly ordered. Fibers within each lamella run parallel in their orientation and follow a helical path with respect to the axis of the canal. Hydroxyapatite crystallites are mostly aligned in nearly the same direction as the collagen fibers (Lees and Davidson, 1977; Ascenzi et al., 1983). Von Ebner (1875) and more recently, Cooper et al. (1966)have suggested that within osteons the direction of fibers changes dramatically in successive lamellae. Ascenzi and Bonucci (19681, however, contend that there are three distinct types of osteons with respect to fiber orientation. In the first type, all lamellae possess fiber bundles with a predominantly transverse helical course; in the second type, successive lamellae alternate between longitudinal and transverse helical orientation; and type three osteons have lamellae with nearly all longitudinally aligned collagen fibers. Gebhardt originally theorized that collagen fibers in Haversian systems serve to absorb forces applied to bone during functional loading and channel these forces (stress) along the plane of expansive strain. With this model it can be seen that compressive force acting on the long axis of an osteon, which results in expansive strain in the transverse plane, is best regulated by transversely oriented collagen fibers, and tensile force acting along the same axis, which results in expansive strain in the longitudinal axis, is best resisted by longitudinally directed fibers. Empirical support for Gebhardt’s hypothesis was forthcoming in the mid-l960s, when Ascenzi and Bonucci (1964, 1967, 1968) reported on a series of experiments in which they measured the ultimate cornpressive and tensile strengths of single isolated osteons of each of the three fiber orientation types. Table 1 lists the ultimate compressive strengths of isolated (wet) osteons taken from a 30-year-old human femur. These data show that osteons with longitudinal fibers (“longitudinal osteons,” for short) fracture under conditions TABLE 1. Ultimate compressive strength of single isolated osteons (kglrnm’)’ Orientation Transverse Alternate Longitudinal Full calcification 16.70 1.19 n = 13 13.66 k 0.95 n = 12 11.20 ? 1.03 n = 7 * Initial calcification 10.02 1.05n=8 7.99 5 1.37n=7 8.96 + 0.94n = 7 * ‘From Ascenzi and Bonucci, 1968. of relatively low compressive loading (11.20 kgimm’) compared to transverse osteons (16.70 kg/mm 1, with alternate osteons intermediate between the two (13.66 kg/mm’). Conversely, as shown in Table 2, transverse osteons subjected to tensile strain appear to fracture at a much lower loading threshold (3.49 kg/mm2)than do longitudinal osteons (11.07-12.64 kg/mm’), with alternate osteons intermediate but trending toward the higher values (9.59 kg/mm2). The degree of calcification also appears to have an effect on the ability of osteons to withstand compressive and tensile strain (Evans and Vincentelli, 1974; Riggs et al., 199313). This fact was not emphasized in these studies, but perhaps for good reason. The lag between the onset and initial completion of osteonic calcification is relatively short (days). Thus even formative bone segments will consist predominantly of fully calcified osteons. Apart from these experimental studies, another approach to test Gebhardt’s hypothesis is to assess the correspondence between strain type and collagen fiber orientation over entire cross sections of bone. Two types of data are thus required: (1)the distribution pattern of transverse and longitudinal lamellae, and (2) the distribution of compressive and tensile strain, over corresponding cross-sectional areas of bone. Fewer than a dozen such correlative studies have appeared in the literature, the vast majority of which, especially early on, have focused on the human femur. This is not surprising since most investigators have relied on published strain data for their assessments, and the human femur is one of the most thoroughly studied of all human long bones. A typical example of this type of study is provided by Portigliatti Barbos et al. (1984), who documented the distribution of transverse and longitudinal lamellae in transverse femoral sections (nearly the entire diaphysis of the human femur was represented). These authors found that “. . . the orientation of transverse lamellae undergoes a rotation from the medial toward the posterior wall in going from the proximal to the distal end of the shaft” (p. 313). In order to assess the correspondence between fiber orientation and functional strain, they reviewed the available literature on patterns of strain operating on the human femur. They conclude: “The recent literature reports data that consistently suggest a distribution of bending forces in the femur that could account for the rotational distribution of the transverse and longitudinal osteonic and interstitial lamellae along the femoral shaft . . .” (p. 314). Boyde and Riggs (1990)recently conducted a similar study on the equine radius. The radius is the major load bearing long bone of the horse forelimb. Its in vivo strain pattern is highly predictable due to the marked CALCANEAL STRAIN AND COLLAGEN ALIGNMENT TABLE 2. Ultimate tensile strength of single isolated osteons (kg/mm2) Orientation Transverse Alternate Longitudinal Full calcification 3.49 0.79n= 4l 9.59 I+_ 1.55n= g2 12.64 _t 1.51n=12’ 11.07 rt 0.68n= 4l * Initial calcification 3.05 k 0.94 n = 3l 9.06 * 1.48 n = 9’ 10.95 _t 1.88 n 9.41 2 1.71 n = = 12’ 4l ‘From Simkin and Robin, 1974. 2From Ascenzi and Bonucci. 1968. longitudinal concave curvature of its caudal aspect. This pattern-tensile strain on the cranial, medial, and lateral sides, and compressive strain posteriorlyhas been confirmed experimentally using electric strain gauge technology (e.g., Rubin and Lanyon, 1982). The correspondence between published strain data and the distribution of transverse and longitudinal lamellae over cross sections of the horse radius is unambiguous: “A striking pattern for the distribution of TS [transverse spiral] and LS [longitudinal spiral] collagen is apparent. The bulk of the cranial, medial and lateral cortices has a very high proportion of LS collagen-the caudal cortex contains predominantly TS collagen” (Boyde and Riggs, 1990, p. 37). Results of these and other correlative studies (e.g., Portigliatti Barbos et al., 1983, 1984, 1987, 1995; Boyde et al., 1984; Ascenzi et al., 1987a, b; Carando et al., 1989; Boyde and Riggs, 1990; Riggs et al., 1993a; Bromage, 1992) demonstrate support, at least in part, for Gebhardt’s hypothesis of the relationship between functional loading and preferential collagen alignment. 149 bers appear bright. The advantage of using circularily polarized light, as compared to linearly polarized light, is that sections can be viewed in all 360” rotational positions without affecting the pattern or magnitude of bright- and darkfields, which is thus a more accurate representation of the orientations of collagen arrays (Boyde et al., 1984). More recently, Boyde and Riggs (1990) elaborated this method to include both simple qualitative and quantitative imaging techniques. Qualitative observations are easily made as a result of displaying whole ground sections of even very large bones by placing the section between two large circular polars. One can then readily comprehend the macroscopic patterns of preferentially oriented collagen. Quantitative procedures for producing a digitized, pseudo-color coded image of the signal intensity returned from whole ground sections examined by a macro lens between crossed circular polars has been described in detail by Boyde and Riggs (1990). Preparation of Materials Five ovine calcanei were prepared for analysis in the present study (Table 3). The diaphysis was removed from each calcaneum with a high-speed diamond saw. Specimens were then cleaned in a 1%Tergazyme (Alconox, New York City) solution a t 40”C, with occasional ultrasonication. A series of plane parallel 100-km sections was made with a water-cooled, slowspeed diamond saw transverse (perpendicular) to the longitudinal axes of the shafts. Sections were retreated with 1%Tergazyme and ultrasonicated for further cleaning. Each section was then treated with chlorofordmethanol (1:l ratio), rinsed in methanol, and MATERIALS AND METHODS mounted in DPX. Sections were then observed and phoThe calcaneum was selected for study here because it tomicrographed in CPL. shares several properties with long bones-its function is primarily mechanical, it acts as a lever, and it has a RESULTS shaft and medullary cavity-and results are therefore CPL Results comparable with those of previous studies involving long bones. Yet, unlike previously studied long bones, Photomicrographs of the ovine calcaneal sections all elements of the calcaneum experience compound load- revealed a consistent pattern of bright and dark distriing (i.e., both principal tension and compression) dur- bution under circularly polarized light. Unlike the reing normal locomotion. For example, in the sheep, the sults obtained for the human femoral shaft, in which caudal aspect of the calcaneal shaft undergoes both the pattern rotates from medial to posterior moving compressive and tensile strain a t different phases of distally, the pattern of bright and dark lamellae in the the locomotor cycle. This allows study of microstruc- sheep calcaneum remains constant along the length of tural responses to more complex loading situations the shaft. The following patterns in the distribution of Erigkit than have previously been reported. The sheep calcaneum was chosen for study because reliable strain data and dark lamellae of a typical section were observed for this bone is available in the literature (Lanyon, (see Fig. 1). First, a progressive increase in the density of bright (transverse) lamellae was observed proceed1973). ing from the exterior surface to the interior of the secAssessing Distribution of Collagen Fiber Alignment in Bone tion, culminating with endosteal and trabecular tissue Boyde et al. (1984) used circularly polarized light that was nearly all transverse. This pattern has also (CPL) microscopy to detect the pattern of collagen fiber been observed in shafts of the human femur (Portigliorientation in bone sections. This technique involves atti Barbos et al., 1984; Ascenzi et al., 1987a) and tibia the use of two polarizing filters and two quarter wave (Portigliatti Barbos et al., 19951, and the horse radius plates, each set a t a specified orientation. Thin sections (Riggs et al., 1993a). There were two notable exceptions t o this pattern: of lamellar bone observed between crossed circular polars appear either bright or dark depending on the ori- (1) in some sections, areas of outer circumferential entation of their collagen fiber: lamellae composed of lamellae appeared bright, and (2) the occurrence of a longitudinally aligned collagen fibers appear dark, few bright lamellae (bright “rings”) around each of othwhereas lamellae composed of transversely aligned fi- erwise dark osteons was observed in both the caudal J.M. McMAHON ET AL. 150 TABLE 3. Specimen data for calcanei of Ovis aries used in this study Specimen # OVAl OVA2 OVA4 OVA5 OVA6 Side right left right left left Age adult subadult adult adult subadult Sex unknown unknown female female female Condition dry dry fresh fresh fresh Fig. 1, Photomicrograph of the diaphysis of a left subadult sheep calcaneum (OVA6 in Table 3) under circularly polarized light (CPL), showing the pattern of collagen fiber orientation. The cranial aspect is at the top and the medial aspect is to the left. Cranial (A),lateral (B),medial (0,and caudal (D) aspects are shown to the right at higher resolution. The scale bars shown on the main figure and insert D represent 1 mm (A-C scale to D). and, especially, cranial aspects of nearly all specimens. In some sections, particularly in the cranial cortex of OVA4, the greater density of these bright-ringed osteons resulted in a relatively higher concentration of bright lamellae (see Riggs et al., 1993a). Second, bright (transverse) lamellae were observed predominantly in the medial and (to a lesser extent in some sections) lateral aspects of the shaft, as exemplified by OVA6 (see Fig. 1).However, there are several important differences between the bright lamellae in the medial and lateral segments: (1) in OVAl and OVA2, bright lamellae of the medial aspect extends over the entire cortical thickness (with the exception in some instances of the external circumferential lamellae), whereas the bright lamellae in the lateral aspect are concentrated along the endosteal surface of the cortex, and ( 2 )bright lamellae in the medial aspect consist mostly of interstitial and compacted coarse cancellous bone with some bright secondary osteons, whereas bright lamellae of the lateral aspect consist almost exclusively of secondary osteons (with the exception that in several sections of OVA2 resorption of the medial side has left only bright cancellous bone). Finally, the cranial and caudal aspects of the shaft were composed predominantly of dark lamellae. Along with the bright “rings,” the only other source of bright lamellae in the cranial and caudal aspects of the shaft is from interstitial tissue. Previously Published Strain Data Data used in this study to reconstruct the strain cycles experienced by the sheep calcaneal shaft during normal locomotion comes from two sources. Badoux (1987) has proposed a biomechanical model for the equine talocrural joint that, with modifications, can be applied to the sheep talocrural joint and used to predict the distribution of strain along the ovine calcaneal shaft. Lanyon (1973) has published in vivo strain data recorded from rosette strain gauges adherent to the sheep calcaneal shaft during locomotor activity. Lanyon (1973)implanted rosette foil strain gauges in various positions on the calcaneum of live sheep (For a discussion of the in vivo strain gauge technique, see Cochran, 1972, and Caler et al., 1981). In each case, at least one of the gauge elements was aligned parallel to the long axis of the shaft. The changing direction and 151 CALCANEAL STRAIN AND COLLAGEN ALIGNMENT I I I I + Caudolateral Aspect 0 + Craniolateral & Lateral Aspects 0 Event # 1 2 3 1 Fig. 2. Cycle of strain on caudolateral (top), craniolateral, and lateral (bottom) aspects of the sheep calcaneum during normal locomotion (after Lanyon, 1973). The horizontal lines running through each of the two graphs indicate the point of zero strain: tension ( + ) is above the line; compression (-1 is below it. The horizontal axis represents time, and the numbers signify important locomotor events: (1) midpoint of swing phase, (2) midpoint of weight-bearingphase, and (3)just after hindlimb lift, before swing phase. magnitude of principal compressive and tensile strain were then recorded during walking a t various speeds. Figure 2 shows several cycles of strain (corresponding with locomotor cycles) for three strain gauges adherent to the lateral, craniolateral, and caudolateral aspects of the shaft (the medial surface was not measured). Each is aligned parallel to the long axis of the shaft. These data show that the lateral and craniolateral (dorsal) aspects underwent a cycle of relatively high compression followed by low tension and that the occurrence of maximum compressive strain coincided with full weight bearing of the hindlimb (event 2 of Fig. 2). During this phase of the cycle, the craniolateral aspect of the shaft reached a greater magnitude of compression than did the lateral aspect (not shown in Fig. 2). The strain cycle on the caudolateral (plantar) side was somewhat different. Here, the shaft underwent a cycle of tension and compression of nearly equal magnitude. Moreover, the pattern of correspondence to locomotion is nearly opposite to that of the other two surfaces measured. Whereas the lateral and craniolatera1 surfaces tended toward maximum compression during maximum weight-bearing, the caudolateral surface experienced maximum tension. Once the hindlimb passed the point of maximum weight-bearing, the lateral and craniolateral regions met with decreased compression (and moved toward tension), whereas the caudolateral aspect decreased in tension (and moved toward compression). Once the hindlimb was lifted (event 3 of Fig. 2), the lateral and craniolateral sides experienced relatively small magnitudes of tension, whereas the caudolateral aspect underwent maximum compression. Subsequently, all regions of the shaft tested met with neutral strain during the swing phase before the cycle was repeated. Lanyon’s (1973) own interpretation of his strain data is somewhat different from that proposed here. He did not view the calcaneum as a lever with alternating bending cycles. He surmised that “If the calcaneum acted as a lever, then pull by the gastrocnemius muscle . . . would have bent the bone, causing compression on one side [cranial] and tension on the other [caudal].” But he insisted, “This did not happen, indeed in the plantar [caudal] region there was either little strain change, or a degree of compression, during the dominant compression period (Lanyon 1973, p.47). In his experiments, there were only two test limbs for which strain data on the craniolateral and caudolateral surfaces of the calcaneum were measured parallel to the long axis of the shaft (denoted as S34 and S35). Lanyon’s (1973) interpretation is based primarily on one of these (S35) in which peak compression on the craniolateral side during maximum weight bearing is separated from peak tension on the caudolateral side by -.06 seconds. On S34, these opposing peaks occur virtually simultaneously. Data presented in Table 4 are derived from the original strain cycle graphs in Figure 4 of Lanyon (1973) for S34 and S35. These data support the interpretation proposed here that the sheep calcaneum undergoes alternating bending forces during each locomotor cycle. First, bending in a cranial direction causes compression cranially and tension caudally (0-.36 seconds in S34), then, slight bending in a caudal direction results in slight tension cranially and compression caudally (.36-.5 seconds in ,534). Although the timing of opposing peak strains and instant of crossover are not precisely simultaneous for the cranial and caudal aspects of 5335, a similar trend is evident. This interpretation is consistent with the biomechanical model proposed for the equine talocrural joint by Badoux (1987) and reproduced (with modifications reflecting ovine anatomy) in Figure 3. When the hindlimb comes into substrate contact, the resultant forces 152 J.M. McMAHON ET AL. TABLE 4. Comparison of strain values for various gauge elements read at .l-sec intervals through a typical loading cycle’ Time (sec) .o .1 .2 .3 .4 .5 Specimen S34 Caudolateral Craniolateral (gauge 1) (gauge 2 ) .oo + 83.50 + 225.45 + 117.00 -10.10 .oo .oo - 43.60 -50.10 -25.10 + 58.50 .oo Specimen S35 Caudolateral Cranialateral (gauge 1) (gauge 2) .oo +41.50 + 190.90 + 24.90 -6.70 .oo .oo -74.70 -16.60 +20.80 + 43.20 .oo ‘Source: Lanyon (1973),Fig. 4. F1 and F2-from the loads applied by the superficial digital flexor (SDF) and gastrocnemius tendons (GAS), respectively-cause the calcaneal shaft to bend, resulting in compressive strain cranially and tensile strain caudally (see Fig. 3a). That the shaft is subjected to bending in the craniocaudal (parasagittal) plane is further indicated by its cross-sectional geometry in which the maximum diameter from centroid is also along the craniocaudal plane. Once the point of maximum weight bearing has been attained, there follows a gradual decrease in the amount of weight supported by the hindlimb, resulting in a progressive decrease of compression cranially and of tension caudally. This condition is further advanced by the fact that, in the sheep, the gastrocnemius muscle actively contracts only during the phase leading up t o maximum weight-bearing (Lanyon, 1973j. Subsequently, the plantar ligament, which functions to stabilize the calcaneum against the metatarsus, can more easily neutralize the bending effects of the superficial digital flexor tendon. Moreover, as the angle at the talocrural joint widens during the phase leading up to hindlimb lift, the bending force of the superficial digital flexor diminishes, as more of its applied load is directed axially along the calcaneum (see Fig. 3b). The tensile strain recorded on the craniolateral and lateral surfaces of the calcaneum during maximum plantar flexion a t hindlimb lift can be explained biomechanically by observing how the angle between the calcaneum and the gastrocnemius tendon changes during plantar flexion. The gastrocnemius muscle originates from the distal femur and inserts on the cranial side of the proximal end of the calcaneum via the tendo calcaneum. In a dorsiflexed position (see Fig. 3a), the gastrocnemius tendon pulls on the calcaneal tuber at an oblique angle (resulting in cranially oriented bending of the shaft and compression on the cranial side). During plantar flexion (see Fig. 3b), this angle is increased until the tendon is pulling nearly parallel to the calcaneal axis (resulting in tension on the cranial side). It is much more difficult to account for the compressive strain recorded on the caudolateral side during this phase. Lanyon (1973)postulated that pull from the superficial digital flexor tendon could cause overall compression of the shaft. But this seems unlikely, since it could occur only if both the distal (plantar) and proximal portions of the SDF pulled downward from the tip of the calcaneum, a condition requiring an exceedingly small talocrural angle, in contrast to the large talocru- ral angle observed during caudolateral compression. More likely is Lanyon’s (1973) second suggestion, that plantar flexion results in stretching the plantar ligament: “This effect, from fibres inserting above the gauge site, could explain the continued compression [of the caudolateral aspect] during the time when the heel was lifting from the ground” (p. 47). DISCUSSION Assumptions The four principal underlying assumptions upon which our interpretations are based are made explicit here. All deal with the degree of confidence with which the strain histories of the calcanei used in our histological analysis can be inferred. Since collagen fiber orientation data were derived from calcaneal specimens different from those used to generate strain cycle data, it is first assumed (assumption 1)that the strain cycles experienced by the sheep calcanei in Lanyon’s (1973) experiments are the same as those experienced by the calcanei used in our CPL microscopic analysis. But even if both structural and functional analyses were performed on the same specimens, three further suppositions are required: assumption 2, Lanyon’s in vivo strain gauge data for sheep locomotion are accurate; assumption 3, the strain cycles recorded are representative of the total loading repertoire influencing microstructural response in the sheep calcaneum; and assumption 4, strain gauge data read on a limited surface area of bone can be extrapolated to include bone deep t o the surface. Assumption 1 : lntraspecific invariability of strain cycle Lanyon (1973) found little interindividual variation in strain cycles among the seven sheep that were studied. Of the four specimens (S34, S35, S47, and S48) to which gauges were attached to the cranial aspect and aligned parallel to the long axis of their calcanei, all showed the same wave form and relative compression to tension magnitudes (although the absolute magnitudes of each varied somewhat). As previously discussed, the wave forms indicated by gauges adherent to the caudolateral aspects of specimens S34 and S35 were slightly different, but their relative compression and tension magnitudes and correspondence with strain cycles on the cranial side were similar. Although not explicitly stated by Lanyon, it is clear from his diagrams (e.g., Fig. 2 in Lanyon, 1973) that specimens of Ovis aries (domestic sheep) were used as experimental subjects. There are many breeds, but this CALCANEAL STRAIN AND COLLAGEN ALIGNMENT f- Cranial Dorsifexion 153 Caudal + Plantar flexion Fig. 3.Ovine talocrural joint in (a)dorsiflexion and (b) plantar flexion (see text for details). species exhibits remarkable morphological homogeneity (prompting Hafez et al., 1962, to argue for a single origin of domestic sheep). Prummel and Frisch (1986), in a morphological comparison of domestic sheep and goats, have drawn attention to numerous invariabilities in the postcranium (including the calcaneum) of European domestic breeds. Indeed, our own dissections of several domestic ewe hindlimbs revealed a configuration of skeletal, muscular, tendinous, and ligamentous relationships identical to that shown in Lanyon’s (1973) diagram (his Fig. 1)of an ovine talocrural joint. This evidence favors the supposition of high interindividual invariability of calcaneal strain cycles among members of Ovis aries. Assumption 2: Accuracy of strain gauge data Although encouraged by the consistency of his strain gauge results, Lanyon (1973) cautioned that the accuracy with which the gauge technique can effectively measure in vivo bone strain has never been adequately determined. There are, in fact, several characteristics of the strain cycle that require confirmation. Of particular importance to our own analysis are the peak compressive and tensile strain magnitudes attained during loading. Cochran (1972) obtained consistent results when he compared strain magnitudes from gauges that had been implanted in vivo on dog tibiae with in vitro gauge measurements on the same bones removed postmortem and applied with a known load. Lanyon (1973) employed a different postmortem calibration technique on three of the gauges used in his study. He attempted to leave the gauges attached only to a small pieces of the original bone that he then applied to tensiontesting machine. The three gauges tested (S34 and S35 not among them) demonstrated ‘l. . . a linear response t o applied strain over the range encountered in vivo, and a faithful indication of strain direction . . .” (p. 46). The peak in vivo microstrain’ ranges (200-300 compression, and 60-75 tension, Fig. 4 in Lanyon, 1973) recorded for the axially aligned gauges applied to the caudolateral and craniolateral aspects of the sheep calcanei (S34 and S35) are low when compared to maximum strain magnitudes recorded on the caudal and cranial aspects of other long bones (at similar stride durations), e.g., - 1205 and 760 for sheep radii (Lanyon and Baggott, 1976), -947 and 1122 for tibiae (Lanyon and Bourn, 1979). But given that the calcaneal shaft has no longitudinal curvature and is primarily designed to resist cranial bending in the parasagittal plane (having a high craniocaudal t o mediolateral dimension), it is perhaps not surprising to find such low deformation values in the craniocaudal plane. Another important characteristic of the strain cycle is the wave form. Interpretation of the wave forms reported for gauges S34 and S35 has already been given. That the wave shapes produced by the two caudolateral gauges of these specimens are different suggests either that strain is variable a t this position or, more likely, that either or both of them are inaccurate. However, the overall pattern observed in both is consistent: as the craniolateral gauge records increases toward maximum compression, the direction of strain indicated by the caudolateral gauge is toward maximum tension. This “mirror image” wave form is typical for gauges ‘Strain is the ratio of the change of length of a material after loading to the original length before loading. One unit of microstrain is equal to units of strain. All measures of strain are given in microstrain units. 154 J.M. McMAHON ET AL. TABLE 5. Correspondence of predicted and observed collagen fiber orientation Calcaneal aspect Craniolateral Lateral Medial Caudolateral Recorded strain High compression, low tension High compression, low tension Not measured Equal compression and tension Predicted collagen orientation Mostly transverse Observed collagen orientation Mostly longitudinal Observed CPL image Dark (some bright) Mostly transverse Transverse and longitudinal Bright and dark - Mostly transverse Mostly longitudinal Bright (some dark) Dark (some bright) Mixed longitudinal and transverse attached to opposite sides of long bones loaded in bending (e.g., Lanyon and Baggott, 1976, Lanyon and Bourn, 1979). The alternating bending strain cycles (described previously) characteristic of the calcaneal shaft have not received independent confirmation. flexion of one or both hindlimbs. However, if collagen fiber ultrastructure responds to the same strain properties as do other structural features (e.g., redistribution of mass), then strain intermittency, as well as magnitude, is an essential condition for structural response. Low stretching involves dynamic rather than intermittent strain. This contrasts sharply with sexual mounting, which is accompanied by pelvic and hindlimb oscillations. Dominant rams may engage in mounting activity for extensive periods of time, especially during the beginning of estrous in females (Hulet, 1966).The degree of talocrural plantar flexion during sexual mounting by the male has not been measured, but in such activity the hindlimbs function to balance the bulk of the ram’s weight, thus actually limiting the extent of plantar flexion. Conversely, during male-male head butting, the hindlimbs act to thrust the body forward and slightly upward resulting in pronounced talocrural plantar flexion and, presumably, high magnitudes of stress on the calcaneum. This could help explain the predominantly dark (“tension resistant”) lamellae found in the cranial segments of the sheep calcaneum, but this explanation, as with the behavior, refers only to males. Of these behaviors only pseudosexual mounting is performed by females, but such occurrences are rarely observed (Hafez et al., 1962). It is therefore difficult to account for the dark (longitudinal) lamellae found on the cranial aspect of OVA4, a confirmed female. Assumption 3: Experimentally measured strain represents natural strain repertoire To assess the correspondence properly between strain distribution and pattern of collagen fiber orientation, a complete history of the total strain regimen influencing collagen alignment must be attained. Incomplete strain information could lead to inaccurate assessments. For example, the discordance between predicted and observed collagen fiber orientation for the craniolateral and lateral aspects of the calcaneal shaft, as presented in Table 5, could be due to a flaw in Gebhardt’s (1905)original hypothesis, or alternatively, to incomplete data on the pattern of applied strain. Lanyon (1973) recorded calcaneal strain data for locomoting sheep at four different speeds: slow walk, medium walk, fast walMslow trot, and medium trot. His results show that peak strain magnitudes increase slightly at each faster pace, but that the wave form generally remains unchanged. However, members of the genus Ovis (including domestic sheep) exhibit many more locomotor and positional behaviors than simply walking and trotting at variable speeds. Are there any other kinesiological activities that have an effect on microstructural response to functional loading? If so, and the cranial and lateral aspects of the Assumption 4: Correspondence of deep strain with ovine test calcaneum had actually experienced high surface strain tension during some locomotor behavior not included However accurate, an electrical resistance strain in Lanyon’s (1973) study, then the predicted collagen gauge measures strain only on the limited surface of orientation might more closely match that observed. bone directly covered by the gauge apparatus. AccordLanyon’s (1973) strain data show that the instances ing to Huiskes (1982) and Huiskes and Chao (19831, at which the craniolateral and lateral sides of the cal- surface strains recorded using strain gauges are concaneum undergo tension correspond to maximum sistent with those predicted by finite element analysis, talocrural flexion (both at the beginning and a t the end a computer model used to approximate the distribution of the weight-bearing phase). Therefore, any positional of stresses and strains in structures of complex shape behavior in sheep in which the hindlimb is loaded dur- and material characteristics, such as bone. The finite ing maximal plantar flexion would tend to induce element method predicts that, in the case of long bones, great tensile strain on the cranial and lateral sides. At the direction of principal strain will be similar on the least three such motor activities have been described bone surface and deep to the surface. This finding also for the male domestic sheep; these include low stretch- gains support from data obtained using the photoelasing, sexual mounting, and butting (Hafez et al., 1962; tic method2 of stress analysis (see Bianchi et al., 1983) Banks, 1964; Grubb, 1974), all of which occur during the breeding season (generally during the autumn and ‘Certain plastic materials become birefringent when loaded and early winter months in temperate climates) and relate observed in polarized light. By constructing and loading a plastic to agonistic or sexual activities. Low stretching by the model of a bone, it is possible to see lines (or fringes) of differing ram occurs during positional behavior associated with optical intensity which reveal the trajectories of stress that occurred pre-mating displays and involves talocrural plantar in the element during loading. CALCANEAL STRAIN AND COLLAGEN ALIGNMENT 155 TABLE 6. Correspondence of predicted and observed collagen fiber orientation, assuming Lanyon’s “dorsal” and lateral gauges were both cranially placed, and his “plantar” gauges were laterally placed Calcaneal aspect Craniolateral Lateral (Lanyon’s “plantar”) Medial Caudolateral Recorded strain High compression, low tension Equal compression and tension Not measured Not measured Predicted collagen orientation Mostly transverse Observed collagen orientation Mostly longitudinal Observed CPL image Dark (some bright) Equal transverse and longitudinal Transverse and longitudinal Bright and dark - Transverse Longitudinal Bright (some dark) Dark (some bright) - for various bone elements, including the calcaneal shaft (Preuschoft, 1970). Correspondence between Strain Data and Collagen Fiber Orientation and Interpretation of Results Potential Effects of Tissue Type on Collagen Fiber Alignment As noted, most correlation studies involve two components: distribution of strain and pattern of collagen fiber orientation. Bromage (n.d.) introduced a third variable, namely, tissue type, which is the product of a number of factors, including origin, location, and rate of osteogenesis. He argued that the direction of collagen fibers in bone is subject to developmental constraints and may therefore be influenced by remodeling processes. For example, results obtained by Bromage (n.d.) for the macaque mandibular corpus showed that bone of periosteal origin was composed predominantly of longitudinal lamellae, whereas lamellae of endosteal origin was primarily transverse. This led him to suggest that tissue of endosteal origin (under certain conditions) may be physically and developmentally constrained by the remodeling process to lay down collagen fibers with a predominantly transverse orientation. Thus the observation made in this and other correlation studies that there is a progressive increase in the amount of transverse lamellae toward the endosteal surface of the bone does not necessarily require a functional explanation (e.g., Carando et al., 1989). Moreover, it would mean that both tissue type and bone growth history need to be considered in any evaluation of the relationship between functional strain and collagen alignment. The predominantly bright (transverse) area of bone in the caudomedial cortex of the sheep calcaneal shaft, for instance, may be due not to strain stimulus but t o a concentration of compacted coarse cancellous bone (of endosteal origin) relocated to this position by the craniolateral drift apparent in the diaphysis of this bone. Riggs et al. (1993a) found that primary lamellar bone in the horse radius was composed predominantly of longitudinal collagen fibers, independent of the strain regime. Even if collagen orientation in primary cortical bone in the sheep calcaneum is influenced by functional strain, such an effect must have occurred during collagen deposition, on either a periosteal or endosteal surface. Thus collagen alignment of primary cortical bone would reflect structural response to past, rather than present, functional strains. Secondarily remodeled structures, however, have the potential to respond intracortically to current loading regimes (Lanyon et al., 1982). Yet, Bromage (n.d.) further cautioned that even secondarily remodeled osteons may in some instances be influenced by various growth factors and the tissue type within which they develop. The effects that developmental constraints have on the microstructural remodeling of localized regions of bone are largely unknown, and clearly much work is needed in this area. Correspondence between Strain and Collagen Fiber Orientation in Secondary Osteons If only secondarily remodeled structures are considered, it can be seen that only the medial and lateral aspects of the sheep calcaneum contain a significant proportion of bright secondary osteons, whereas these lamellae are primarily dark in all other areas of the diaphysis. Table 5 summarizes the correspondence between the distribution of collagen fiber orientation and strain in four sectors of the ovine calcaneal shaft. It can be seen that in contrast to previous studies of long bones, there is no simple correspondence between the predicted (based on Gebhardt’s hypothesis) and the observed collagen fiber orientation. For instance, although the strain signatures for the craniolateral and lateral segments of the shaft are (apparently) virtually identical (see Fig. 2), CPL results demonstrate a greater concentration of bright (transverse) lamellae on the lateral side. One way in which this particular inconsistency can be eliminated is by postulating that Lanyon’s (1973) dorsal (craniolateral) and lateral strain gauges were actually both positioned craniolaterally and that his “plantar” (caudal) gauges were positioned laterally. This speculation is prompted by the fact that in all of the sheep calcanei observed in dissections, the plantar ligament completely covered (was attached to) the caudal side of the bone and, in fact, extended considerably up on the lateral side. Since Lanyon (1973, p. 47) described the position of his “plantar” strain gauges as “adjacent to the [fibers of the1 plantar ligament,” it is not unreasonable to assume that they were actually laterally, rather than caudally, positioned. Table 6 shows a revised comparison of predicted and observed collagen orientation based on this interpretation. However, although this eliminates certain inconsistencies (e.g., the difficulty in accounting for the occurrence of compressive strain on the caudal surface, and the greater than predicted amount of longitudinal lamellae on the lateral side) it raises new ones: how to explain compression operating on the lateral aspect during a phase of tension directed cranially. Considering only the craniolateral aspect, it can be 156 J.M. McMAHON ET AL. seen with reference to both Tables 5 and 6 that the predicted and observed collagen fiber orientations are inconsistent. Gebhardt’s (1905) hypothesis predicts that the high degree of compression recorded by Lanyon (1973) on the cranial surface should result in the formation of bright (transverse) lamellae, rather than the dark (longitudinal) lamellae actually observed. Proposed Modification of Gebhardt’s Hypothesis The fact that bone, as a tissue, generally appears to be less resistant to tensile strain than compressive strain (Evans, 1957; Currey, 1959, 1984a,b; Yamada, 1970; also see Tables 1 and 2) may provide at least a partial (albeit tentative) explanation for the inconsistencies between the predicted and observed results, summarized in Tables 5 and 6. Since the ultimate strength (and by implication, the yield point) is lower under tension than it is under compression, the occurrence of even a small magnitude of tensile strain acting on a localized bone volume may be sufficient to stimulate (by as yet unknown processes) the formation of longitudinal collagen fibers (which are purportedly better suited to resist tension), even when the same bone region is undergoing greater compressive strain during another phase of the locomotor cycle. This “tension-resistance priority’’ (TRP) hypothesis gains some support from several studies involving compound strain. Bromage (n.d.) found that collagen fibers along the inferior border of the macaque mandibular corpus were aligned predominantly in a longitudinal (tension-resistant) direction even though this region is subject to both tension and compression during different phases of the various masticatory cycles. Hobby et aI. (in prep.) found that sectors of the macaque femoral diaphysis that appear to have been loaded in both compression and tension are also composed of longitudinal collagen fibers. Findings from other studies, however, contradict the TRP hypothesis. Portigliatti Barbos et al. (1995) have reported that the anteromedial surface of the distal portion of the human tibia1 shaft, which was made up primarily of transverse lamellae, is subject “mainly” to compression (which implies at least a small magnitude in the direction of tensile strain). Bromage (n.d.) has also reported findings of this type with respect to several localized areas of the macaque mandible. Indeed, our own results demonstrate the occurrence of both tensile strain and a high proportion of bright (compressive resistant) lamellae on the lateral side of the ovine calcaneum (this holds true for both Tables 5 and 6 results). Thus by proposing that collagen fibers will align preferentially to resist tensile strain, when both tension and compression are alternately acting on bone tissue, we can account for some but not all of the observed inconsistencies in Gebhardt’s hypothesis. The particular ultrastructural response (i.e., the direction in which collagen fibers are laid down) to varying magnitudes of tensile strain, acting as part of a compound loading regimen, may be dependent on other mechanical and physiological properties of bone in a localized area. This “holistic response” is consistent with recent findings that indicate that both macrostructural and microstructural features of bone function together to maintain a constant strain environment in the face of changing load regimes, both throughout ontogeny (Biewener et al., 1986) and phylogeny (Rubin and Lanyon, 1985). Viewed in this way, the longitudinal alignment of collagen fibers observed in the craniolateral aspect of the ovine calcaneum may function to preserve Young’s modulus of elasticity in a direction optimally suited to the local loading environment. In other words, collagen orientation (and other microstructural features) in bone may be developmentally constrained to contribute to the maintenance of local strain regimes in combination with macrostructural properties. More work on the role of bone growth and development, and on the interactions among the mechanical properties of various bone structures (both at the micro and macro level) is needed before Gebhardt’s (1905) hypothesis can be adequately modified and placed within the wider context of the relationship between bone structure and function. CONCLUSIONS Following is a summary of the conclusions reached in this study. 1. Collagen fiber alignment in the diaphysis of the sheep calcaneum is characterized by the following properties. There is a progressive increase in the density of bright (transverse) lamellae proceeding from the exterior surface to the interior of each section, with the exception that some outer circumferential lamellae are bright. Bright lamellar “rings” around the perimeter of otherwise dark osteons were a common occurrence. There was little interindividual variation in the regional pattern of bright and dark secondary osteons among the cross-sectional specimens observed. The general pattern (as shown in Table 5) was characterized by the presence of mostly longitudinal (dark) osteons in the craniolateral and caudolateral aspects of the shaft, mostly transverse (bright) osteons on the medial side, and a combination of transverse and longitudinal osteons laterally. 2. Reinterpretation of Lanyon’s (1973) published strain data for the ovine calcaneum, recorded in vivo during normal locomotion, indicates that this bone acts as a lever with alternating bending cycles. 3. Correlative studies of the type presented here rely on a number of assumptions that are rarely, if ever, stated or explored. Many of these involve the assumed accuracy with which we know (or think we know) the strain histories of the skeletal elements on which collagen fiber orientation analysis is undertaken. The following questions need to be addressed in correlative studies: What is the effect of interindividual variation of strain pattern on the external validity of the study results? How can we determine the accuracy of strain measurement or inference techniques? Is experimentally measured strain representative of the naturally experienced strain (should the ethological literature pertaining to mechanical movement be consulted)? 4.Both tissue type and bone growth history need to be considered in any evaluation of the relationship between functional strain and collagen alignment: (1) tissue type (which reflects the mechanisms of bone remodeling) is a potential variable affecting the orientation of collagen fibers, and (2) since bone tissue is deposited only on surfaces, collagen fibers of primary bone deep to the surface may have been laid down under different strain conditions from that experienced on CALCANEAL STRAIN AND COLLAGEN ALIGNMENT the existing surface of the bone. Thus not only must one consider the remodeling history but also the strain history acting on the bone. 5. If our assumptions and interpretations are correct, then at least some of our results are inconsistent with Gebhardt’s (1905) original hypothesis. For example, the craniolateral side of the sheep calcaneum, which clearly exhibits much greater magnitudes of compression than tension, should, according to Gebhardt, exhibit collagen fibers aligned predominantly transversely (bright under CPL). Yet this sector of the calcaneum consistently exhibits predominantly longitudinal (dark under CPL) collagen fibers. 6.The results reported here indicate that Gebhardt’s hypothesis (1905) may require modification. Two hypotheses are proposed to account for the observed inconsistencies between reported and predicted results. The “tension-resistance priority” hypothesis states that the occurrence of even a small magnitude of tensile strain acting on a localized bone volume may be sufficient to stimulate the formation of longitudinal collagen fibers, even when the same bone region is undergoing greater compressive strain during another phase of the locomotor cycle. The “holistic response” hypothesis holds that collagen fiber orientation functions to contribute to the overall maintenance of Young’s modulus in combination with other microstructural and macrostructural properties. These hypotheses are tentative and require further systematic testing. ACKNOWLEDGEMENTS This research was supported, in part, by the Horserace Betting Levy Board (A.B.), and by a PSC-CUNY Faculty Award of the City University of New York (T.G.B.). Research facilities used in this study are courtesy of an NSF grant to the New York Consortium in Evolutionary Primatology (NYCEP). The authors wish to thank Professor F.S. Szalay for contributing the dry bone specimens. The original manuscript was improved by the helpful comments of three anonymous reviewers. LITERATURE CITED Ascenzi, A,, A. Bigi, A. Ripamonti and N. Roveri 1983 X-ray diffraction analysis of transversal osteonic lamellae. Cal. Tiss. Inter., 35:279-283. Ascenzi, A,, and E. Bonucci 1964 The ultimate tensile strength of single osteons. Acta Anat., 58:160-183. Ascenzi, A., and E. Bonucci 1967 The Tensile properties of single osteons. Anat. Rec., 158:375-386. Ascenzi, A,, and E. Bonucci 1968 The compressive properties of single osteons. Anat. Rec., 161:377-392. Ascenzi, A,, A. Boyde, M. Portigliatti Barbos, and S. Carando 1987a Micro-biomechanics vs. macro-biomechanics in cortical bone: A micro-mechanical investigation on femurs deformed by bending. J . Biomech., 20:1045-1053. Ascenzi, A,, S. Improta, M. Portigliatti Barbos, S. Carando, and A. Boyde 19871, Distribution of lamellae in human femoral shafts deformed by bending with inferences on mechanical properties. Bone, 8t319-325. Badoux, D.M. 1987 Some biomechanical aspects of the structure of the equine tarsus. Anat. Anz., Jena., 16453-61. Banks, E.M. 1964 Some aspects of sexual behavior in domestic sheep, Ovis aries. Behaviour., 23:249-279. Bianchi, R., P. Clerici, L. Vizzotto, and A. Miani 1983 Development of a method for quantitative stress analysis in bones by three-dimensional photoelasticity. Anat. Rec., 206:227-237. Biewener, A.A., S.M. Swartz, and J.E.A. Bertram 1986 Bone modeling during growth: Dynamic strain equilibrium in the chick tibiotarsus. Calcif. Tiss. Int., 39:390-395. 157 Boyde, A., P. Bianco, M. Portigliatti Barbos, and A. Ascenzi 1984 Collagen orientation in compact bone: A new method for the determination of the proportion of collagen parallel to the plane of compact bone sections. Metab. Bone Dis. Rel. Res., 5:299-308. Boyde, A., and C.M. Riggs 1990 The quantitative study of the orientation of collagen in compact bone slices. Bone., 11:35-39. Bromage, T.G. 1992 Microstructural organization and biomechanics of the macaque circumorbital region. In: Structure, Function and Evolution of Teeth. P. Smith and E. Tchernov, eds., Freund, London, pp. 257-272. Bromage, T.G. n.d. Preferential intracortical remodeling and growth of the macaque mandible: Developmental constraints on growth and adaptation of bone. Caler, W.E., D.R. Carter, and W.H. Harris 1981 Technique for implementing an in vivo bone strain gage system. J . Biomech., 14(7): 503-507. Carando, S., M. Portigliatti Barbos, A. Ascenzi, and A. Boyde 1989 Orientation of collagen in human tibia1 and fibular shaft and possible correlation with mechanical properties. Bone, 1Ot139142. Cochran, G.V.B. 1972 Implantation of strain gages on bone in vivo. J . Biomech., 5:119-123. Cooper, R.R., J.W. Milgram, and R.A. Robinson 1966 Morphology of the osteon. J . Bone Joint Surg., 48A:1239-1271. Currey, J.D. 1959 Differences in the tensile strength of bone of different histological types. J . Anat., 93:87-95. Currey, J.D. 1984a The Mechanical Adaptations of Bone. Princeton University Press, Princeton, NJ. Currey, J.D. 1984b What should bone be designed to do? Calcif. Tiss. Inter., 36:S7-S10. Evans, F.G. 1957 Stress and Strain in Bones. Charles C. Thomas, SprinMield, IL. Evans, F.G. 1970 The Mechanical Properties of Bone. Charles C. Thomas, Springfield, IL. Evans, F.G., and R. Vincentelli 1974 Relations of the compressive properties of human cortical bone to histological structure and calcification. J . Biomech., 7:l-10. Gebhardt, W. 1905 Uber funktionell wichtige Anordnungsweisen der feineren und groberen Bauelemente des Wirbeltierknochens. Arch. Entwickl. Org., 20:187-322. Grubb, P. 1974 The rut and behaviour of Soay rams. In: Island Survivors. P.A. Jewel1 and J.M. Boyd, eds. Athlone Press, London. pp. 195-223. Hafez E.S.E, R.B. Cairns, C.V. Hulet, and J.P. Scott 1962 The behaviour of sheep and goats. In: Behaviour of Domestic Animals. E.S.E. Hafez, ed. Balliere Tindal and Cox, London., pp. 297-333. Howell, J.A. 1917 An experimental study of the effect of stress and strain on bone development. Anat. Rec., 13:233-252. Huiskes, R. 1982 On the modelling of long bones in structural analyses. J . Biomech., 15(1):65-69. Huiskes, R., and E.Y.S Chao 1983 A survey of finite element analysis in orthopedic biomechanics: the first decade. J . Biomech., 16: 385-409. Hulet, C.V. 1966 Behavioral, social and psychological factors affecting mating time and breeding efficiency in sheep. J . Anim. Sci., 25: 5-20. Lanyon, L.E. 1973 Analysis of surface bone strain in the calcaneum of sheep during normal locomotion. J . Biomech., 6:41-49. Lanyon, L.E. 1980 The influence of function on the development of bone curvature: a n experimental study on the rat tibia. J . Zool., London., 192:457-466. Lanyon, L.E., and D.G. Baggott 1976 Mechanical function as an influence on the structure and form of bone. J . Bone Joint Surg., 58B:436-443. Lanyon, L.E., and S. Bourn 1979 The influence of mechanical function on the development and remodeling of the tibia. J . Bone Joint Surz 61A:263-273. Lanyon, L.E., A.E. Goodship, C.J. Pye, and J.H. MacFie 1982 Mechanically adaptive bone remodeling. J . Biomech., 15:141-154. Lees, S., and C.L. Davidson 1977 The role of collagen in the elastic properties of calcified tissues. J. Biomech., 10:473-486. Portigliatti Barbos, M., P. Bianco, and A. Ascenzi 1983 Distribution of osteonic and interstitial components in the human femoral shaft with reference to structure, calcification and mechanical properties. Acta Anat., 115:178-186. Portigliatti Barbos, M., P. Bianco, A. Ascenzi, and A. Boyde 1984 Collagen orientation in compact bone: 11. Distribution of lamellae in the whole of the human femoral shaft with reference to its mechanical properties. Metab. Bone Dis. Rel. Res., 5:309-315. Portigliatti Barbos, M., S. Carando, A. Ascenzi, and A. Boyde 1987 On the structural symmetry of human femurs. Bone, 8:165-169. ,. 158 J.M. McMAHON ET AL. Portigliatti Barbos, M., S. Carando, A. Boyde, and A. Ascenzi 1995 Collagen orientation in compact bone: Ill. Distribution of lamellae over the whole of the human tibia1 and fibular shaft, and its correlation with mechanical properties (in press). Preuschoft, H. 1970 Statische untersuchungen am fub der primaten. 11. Statik des gazen fubes. Z. Anat. Entwick1.-Gesch., 131:156192. Prummel, W., and H.J. Frisch 1986 A guide for the distinction of species, sex and body side in bones of sheep and goat. J . Archaeol. Sci., 13:567-577. Riggs, C.M., L.E. Lanyon, and A. Boyde 1993a Functional associations between collagen fibre orientation and locomotor strain direction in cortical bone of the equine radius. Anat. Embryol., 187:231-238. Riggs, C.M., L.C. Vaughan, G.P. Evans, L.E. Lanyon, and A. Boyde 1993b Mechanical implications of collagen fibre orientation in cortical bone of the equine radius. Anat. Embryol., 187:239-248. Rubin, C.T., and L.E. Lanyon 1982 Limb mechanics as a function of speed and gait: A study of functional strain in the radius and tibia of horse and dog. J . Exp. Biol., 101:187-211. Rubin, C.T., and L.E. Lanyon 1985 Regulation of bone mass by mechanical strain magnitude. Calcif. Tiss. Int., 37:411-417. Simkin, A,, and G. Robin 1974 Fracture formation in differing collagen fiber pattern of compact bone. J. Biomech., 7:183-188. Stinchfield, A.J., J.A. Reidy, and J.S. Barr 1949 Prediction of unequal growth of the lower extremities in anterior poliomyelitis. J. Bone Joint Surg., 31A:478-486. Von Ebner, V. 1875 Ueber den feineren Bau der Knochensubstanz. Sber. Akad. Wiss. Wien III., 72:49-138 (cited in Ascenzi and Benvenuit, 1986). Wolff, J. 1892 Das Gesetz der Transformation der Knochen. A Hirschwald, Berlin. Yamada, H. 1970 Strength of Biological Materials. Williams & Wilkins, Baltimore.