THE ANATOMICAL RECORD 217:223-228 (1987) Morphology of the Osteonal Cement Line in Human Bone MITCHELL B. SCHAFFLER, DAVID B. BURR, AND RICHARD G . FREDERICKSON Department of Anatomy, (M.B.S., D.B.B., R. G.F) and Orthopedic Research Laboratory (M.B.S., D.B.B.), West Virginia University Medical Center, Morgantown, WV 26506 ABSTRACT While current consensus suggests the absence of collagen in osteonal cement lines, the extent of cement line mineralization and the nature of the ground substance within the cement line are unclear. Samples of human radius were examined by using scanning electron microscopy, electron microprobe, and histochemical techniques. X-ray intensities were used to compare the amount of calcium, phosphorus, and sulfur in cement lines with amounts in surrounding lamellar bone. The results indicate that cement lines contain significantly less calcium and phosphorus, but significantly more sulfur, than surrounding bone matrix. The Ca/P ratio of cement lines was significantly greater than that of lamellar bone, suggesting that the mineral in cement lines may not be in the form of mature hydroxyapatite. No selective staining of the cement lines could be demonstrated by using periodic acid-Schiff, Sudan black B, or alcian blue critical electrolyte concentration techniques. Numerous functional and biomechanical theories and observations attach considerable functional importance to the cement lines which separate secondary osteons, or Haversian systems, from the surrounding bone matrix. They are variously implicated in fracture processes, energy absorption, viscous damping and elastic function, and fatigue processes in compact bone (Piekarski, 1970; Carter and Hayes, 1977; Lakes and Saha, 1979; Gottesman and Hashin, 1980; Katz, 1981; Martin and Burr, 1982; Burr et al., 1985). In short, there is no scarcity of ideas regarding cement line function. There are, however, inadequate morphological bases to either explain experimental observations of cement line action or to allow the evaluation of various theories of cement line function. While it has been suggested that cement lines are optical artefacts resulting from the juncture of the differently oriented components of osteonal and extraosteonal bone (Schmidt, 19591, these thin refractile structures are generally recognized as morphologically distinct entities since they were so described by von Ebner (1875; also see Sokoloff, 1973). However, the compositional correlates of this morphological distinctiveness, compared to the organic and mineral components of surrounding bone matrix, are unclear. Light microscopic methods, such as collagen specific stains and silver impregnation techniques, indicate that cement lines are collagen deficient (Weidenreich, 1930; Sokoloff, 1973); scanning electron microscopic study finds similarly (Frasca, 1981). Alternatively, collagen X-ray diffraction patterns have been shown in cement lines (Philipson, 1965). Investigations into, the nature of noncollagenous organic matrix, or ground substance, in cement lines yield no consensus as to which possible organic components are contained within (Sokoloff, 1973; Frasca, 1981). Similarly, it is unclear whether cement lines are more calcified than surrounding bone (Philipson, 1965; Frasca, 198l), less 0 1987 ALAN R. LISS, INC. calcified than surrounding bone (Fawns and Landells, 1953), or not different from surrounding bone in this regard (Mellors, 1964). It is also not known whether the mineral in cement lines differs qualitatively from the mineral contained in lamellar bone (i.e.,hydroxyapatite). The purpose of this investigation is to examine the osteon cement line by using scanning electron microscopy, X-ray microprobe, and histochemistry in order to address two principal areas of morphological ambiguity: 1)the mineralization of the cement line relative to surrounding bone matrix and 2) the presence and characteristics of cement line noncollagenous organic matrix components. MATERIALS AND METHODS Diaphyseal blocks were cut from six fresh human radii obtained at autopsy. Four bones were from males (ages 38, 57, 58, and 81 years) and the others were from females (ages 62 and 81 years); no history of metabolic or bone disease was noted. Undecalcified cross sections (150-200 pm thick) were cut with an Isomet metallurgical saw (Buehler Instruments, Evanston, E)and hand ground and polished, and each section was then cut into segeral smaller pieces. Specimens were then dried at 60 C, mounted on copper trays by using double-sided tape, sputter coated with a thin layer of carbon, and viewed in a JEOL 1OOCX TEMSCAN electron microscope. Specimens were imaged topographically by using the secondary electron image (SEMI and compositionally by using the back-scattered electron image @SE). Received July 28, 1986; accepted October 22,1986. M.B. Schaffler’s current address is the Division of Radiobiology, University of Utah School of Medicine, Salt Lake City, UT 84112. Richard G. Frederickson’s current address is the Veterans Administration Hospital, Seattle, WA 98108. 224 M.B. SCHAFFLER, D.B. BURR, AND R.G. FREDERICKSON X-ray microanalysis was undertaken on the same sections employed for electron microscopy; the analyses were performed by using a Tracor-Northern X-ray spectrophotometer with a 30-mm2energy-dispersive detector (model TN-2000, Middleton, WI) installed on the JEOL lOOCX unit. Continuum suppression was achieved by digital filtering through the analytic system. Sampling for the microprobe study consisted of three sites for each of 32 randomly selected complete secondary osteons; the sample sites were the cement line and areas of interstitial and osteonal lamellae 25 pm on either side of the cement line sample site. The electron beam was positioned over a sample by viewing the BSE image at ~ 4 , 0 0 0magnification. A 1-minute sampling period was used to obtain X-ray spectra; the accelerating voltage was 60 kV. Sampling was confined to either the cement line or the sample lamellae, as these structures are a t least 1 pm or greater in width. Relative amounts of calcium and phosphorus among three sample sites were compared from X-ray intensities by using correlated ttests. Quantification of these intensities was achieved by using a variation of the internal standard ratio technique (Dorge et al., 1978) as a ZAF correction was not available for data reduction. Characteristic X-ray intensities obtained for calcium and phosphorus a t cement lines and osteonal lamellae were compared directly to the intensities for interstitial bone, defined as the internal standard for a sample group, from the corresponding sample area. The internal standard method of normalization assumes that interstitial bone is fully mineralized, mature hydroxyapatite [3Ca3 (PO& * Ca(OH)2]; this assumption is well supported in the literature (Amprino and Engstrom, 1952, Wergedal and Baylink, 1974; Landis and Glimcher, 1978; Glimcher, 1984). Comparison of Ca/P molar ratios was performed by using data normalized by the internal standard; correlated t-tests were used. Specimen portions not used for electron microscopyKray omicroprobe analysis were frozen and stored at -80 C . Specimens were subsequently thawed and fixed in formol-cetylpyridinium chloride (4-8 hours). Undecalcified specimens were cut as described earlier. Another set of fixed specimens was completely decalcified in 5% EDTA and paraffin embedded. Sudan black B was used to determine whether lipid is present in cement lines (Pearse, 1980). Periodic acid-Schiff reaction was used to assess the presence of glycoproteins and covalently bound sugars in cement lines (Pearse, 1980). The alcian blue critical electrolyte concentration (CEC) technique, which uses different molar concentrations of MgC12 to differentially stain glycosaminoglycans, was employed to examine cement lines as well (Scott et al., 1964; Scott and Dorling, 1965). RESULTS Scanning Electron Microscopy (SEM) a dark band between osteonal and interstitial lamellae (Fig. 2). Because backscattered electrons reflect the mass density of a region, the dark-band appearance of cement lines indicates that they have less mass and are compositionally different from surrounding bone. This difference may indicate that cement lines contain less calcified material than the surrounding bone matrix, that the materials contained within are composed of elements of lower atomic number, or some combination of both possibilities. Electron Microprobe Analysis Relative intensities for calcium and phosphorus (Fig. 3) indicate that cement lines are less mineralized than surrounding bone matrix. Cement line calcium and phosphorus intensities were significantly lower than those in either interstitial lamellae (mean differences: Ca= 10%, P= 18%; P < .01) or osteonal lamellae (mean differences Ca = 6%,P= 12%;P < .01).While the amount of sulfur in bone matrix is typically small and the current data support this, elevated sulfur peaks (12-15% greater intensity than surrounding bone) were found in cement lines. It should be noted that sulfur, as examined in cells with the electron microprobe, tends to evaporate or burn off rapidly (Andrews et al., 1983). As no attempt was made to prevent this, such as using a cooled specimen stage, the actual sulfur differences observed may underestimate systematically the amount of sulfur present. Because no standard was used for sulfur, higher sulfur intensities at cement lines can only be interpreted qualitatively relative to surrounding bone at a sample site to indicate that sulfur-containing moieties are contained within cement lines. Osteonal lamellae were found to be less highly mineralized than extraosteonal bone, which is consistent with results of previous studies (Amprino and Engstrom, 1952; Bonucci et al., 1970). Osteonal and interstitial bone, however, did not differ with regard to sulfur. Quantitative comparison is based on molar proportions of calcium and phosphorus in hydroxyapatite; mean normalized number of moles for each of these elements at cement lines and osteonal and interstitial lamellae are shown in Table 1. Calcium-phosphorus molar ratios (CdP), calculated from these normalized data, are shown in Figure 4. Ca/P molar ratio in cement lines (mean= 1.84) is significantly greater than in interstitial or osteonal lamellae (P<.OOl). Ca/P molar ratios in osteonal and interstitial bone do not differ significantly (means= 1.69 vs. 1.67, respectively). Histochemistry Examination of bone sections with periodic acid-Schiff (PAS) reaction did not reveal any staining in cement lines, which suggests that little, if any, glycoprotein is contained within cement lines. Similarly, Sudan black B did not reveal any lipid in cement lines. Studies of cement lines with the alcian blue CEC technique did not show selective intensification of staining at cement lines, although some diffuse staining of the cement lines and the entire bone matrix was observed in the chondroitin sulfate and keratan sulfate ranges. SEM indicated that cement lines are topographically indistinct from surrounding bone (Fig. 1).These topographic views reveal no separation of osteons from the surrounding bone matrix, a potential artefact of drying the specimens for electron microscopic examination. Images generated by using BSE, which are produced DISCUSSION by interactions of incident electrons with the various elemental masses within the specimen, revealed a dis- When intracortical bone resorption and refilling occur, tinct morphology for cement lines; these interfwes of the resulting secondary osteon, or Haversian system, is osteons with surrounding bone appeared consistently as separated from the the surrounding bone matrix by an OSTEONAL CEMENT LINE COMPOSITION 225 Fig. 1. The secondary electron image revealed that the cement line (CL) is topographically indistinct from surrounding lamellar bone. X 3,000 original magnification. interface structure known as the “cement line.” While standard histological descriptions of osteonal bone indicate that cement lines are characteristic features of this type of tissue, the question of whether they are optical artefacts resulting from the juncture of osteonal and extraosteonal lamellae has been raised (Schmidt, 1959). Nevertheless, the prevailing opinion in the literature is that these thin, refractile bands, several micrometers wide, are real and unique structural entitites (von Ebner, 1875; Weidenreich, 1930;Fawns and Landells, 1953; Philipson, 1965; Sokoloff, 1973; Frasca, 1981). However, the nature of cement line composition is controversial. Bone is composed principally of collagen fibers, calcium phosphate salts primarily in the form of hydroxyapatite, and small amounts of glycoprotein and proteoglycan collectively termed ground substance and water. The extent to which each of these components appears in lamellar bone is well documented (Howland et al., 1926; Easthoe and Easthoe, 1954; Woodward, 1964; Robinson, 1975). The extent to which these components appear in cement lines is unclear. Weidenreich (1930) and Sokoloff (1973) indicate that stains for collagen fail to reveal its presence in cement lines; silver impregnation techniques, which are specificfor collagen and reticular fibers, are selectively rejected by cement lines but not by the surrounding bone matrix. Scanning electron microscopy studies of mechanically disrupted cement lines similarly do not show collagen fibers in cement lines (Frasca, 1981). One study stands in contradistinction to these other studies; Philipson (1965) used X-ray diffraction t o study orangutan and whale bone samples and obtained characteristic collagen diffraction patterns in cement lines. However, inherent methodologicalfaults and inconsistencies within that study (see Sokoloff, 1973 for discussion) have led to the belief that Philipson’s data for cement line collagen are incorrect. Consequently, consensus is that cement lines are collagen deficient. The current histochemical studies reveal no PAS staining at cement lines, suggesting that little or no glycoprotein is contained within. Because glycoproteins in connective tissue are typically intimately associated with collagen, the lack of PAS staining also implies that cement lines are collagen deficient. In addition, no lipid was found in cement lines. These data are consistent with earlier characterizations of cement line constituents (Weidenreich, 1930; Sokoloff, 1973; Frasca, 1981). Previous studies have not examined whether glycosaminoglycans are present in cement lines. The current studies with alcian blue failed to reveal any differential staining between the cement lines and the bone matrix, suggesting that sulfated glycosaminoglycans are no 226 M.B. SCHAFFLER, D.B. BURR, AND R.G. FREDERICKSON Fig. 2. The backscattered electron image indicated that the cement line (CL) is a region of low mass density which is compositionally distinct from adjacent lamellar bone. x 1,000 original magnification. I_ Sullur C a l ciuiii 400.000 inlerstitial Lamellae 1,600 70.000 1.400 65,000 Lo m r c * = 13 L 350.001 60,000 55.000 300.001 50.00 Fig. 3.The relative X-ray intensities for calcium and phosphorus indicate that cement lines contain 510% less calcium and 10-16% less phosphorus than adjacent bone matrix ( P < .001). Cement lines contain about 20% more sulfur than adjacent bone ( P < .001). 227 OSTEONAL CEMENT LINE COMPOSITION more prevalent in cement lines than in the lamellar bone matrix. However, the microprobe data consistently indicated that higher sulfur levels are present in cement lines, and the most likely location for large amounts of sulfur in bone is in sulfated glycosaminoglycan. Baylink et al. (1972), using electron microprobe techniques, demonstrated high sulfur levels in osteoid and also showed that the amount of sulfur present varied directly with the amount of chondroitin sulfate present in the tissue. Accordingly, while the histochemical analyses were unable to demonstrate glycosaminoglycans, the sulfur found in cement lines suggests that these complexes may be present. Most equivocal is whether cement lines are more, less, or similarly mineralized than surrounding bone. Fawns and Landells (1953)were unable to demonstrate calcific materials in cement lines with the aid of either alizarin red or von Kossa stain. Philipson (1965) used X-ray absorption and found that mineralization of cement lines was comparable to heavily mineralized osteons. Similarly, Frasca (1981) indicated that cement lines are more highly mineralized than surrounding bone. Mellors (1964) found that mineralization of cement lines does not differ from surrounding bone. Cement lines are characteristically radiolucent when seen in microradiographs of cortical bone; this has been generally a.. a :.:.:.: I .90 Interstitial lamellae (Internal Standard) Cement Line Osteonal Lamellae I .8 Ca/P I .7 I .6 Fig. 4. The C a R ratio in cement lines was significantly greater than that in surrounding lamellar bone (P < .001),suggesting that cement lines do not contain mature hydroxyapatite. interpreted as indicating a lower relative mineralization for cement lines. However, Sokoloff (1973) pointed out that such radiolucency could also be a n artefact due to specimen shrinkage resulting in cement line separation during the preparation of undecalcified bone sections. In the current study, SEM indicated that shrinkage separation a t cement lines did not occur. BSE images, like typical microradiographs, showed that cement lines are regions of lower mass density than either osteonal or interstitial bone. The X-ray microprobe data indicated that this relative lower mass density corresponds to both less calcium and less phosphorus in cement lines than in surrounding bone. From the empirical formula for hydroxyapatite, the Ca/P molar ratio can be calculated. Experimentally derived Ca/P molar ratios for bone approximate this theoretical value well. Strandh (1960) used microchemical analyses and found a Ca/P molar ratio of 1.71for human bone. Mellors (1964) determined Ca/P molar ratio in human cortical bone by using electron microprobe; his data show a mean ratio of 1.70. Wergedal and Baylink (1974) employed electron microprobe techniques to examine mineralization in developing chick bones; a t the completion of mineralization Ca/P molar ratio closely approximated the expected values for hydroxyapatite. Accordingly, the high degree of concordance between theoretically and experimentally determined calciumphosphorus molar ratios supports the use of the internal standard (interstitial bone) normalization used here. Because interstitial bone was assumed for this study to have a C a / p molar ratio of 1.67, comparison among interstitial bone sample sites cannot be performed. However, comparison among interstitial bone, osteonal bone, and cement line ratios are possible. The mean Ca/p molar ratio of osteonal lamellae in this study was 1.69+0.02, which is essentially identical to that expected for hydroxyapatite and to previous determinations in mature bone (Strandh, 1960; Woodward, 1964; Mellors, 1964; Wergedal and Baylink, 1974; Glimcher, 1984). In the current study, cement line Ca/P molar ratio was significantly higher (1.84 Ifr 0.03) than in any surrounding area of bone. This higher ratio primarily represents the influence of the relatively decreased amounts of phosphorus observed in cement lines. Phosphorus emits “soft”or low-energy X-rays under X-ray excitation; these X-rays are likely to be attenuated by the matrix through which they must pass. It is possible that the different composition of the cement line matrix, evidenced by the relatively increased presence of sulfur, absorbs more TABLE 1. Elemental analysis of cement line composition in bone Cement line Ca P Intensity (cpm) Relative intensity 361,048 (+20,532) 90.2 (k0.8) 52,431 (+3,411) 80.2 (k1.7) Osteonal lamellae Ca P 405,896 (+20,413) 95.2 (kl.1) 59,217 (+3,763) 95.0 (k1.7) Interstitial lamellae Ca P 416,264 (f21,505) 100 62,851 (+3,738) 100 (-1 (-1 (%) Relative no. moles CafP molar ratio 9.02 (k0.08) 4.81 (kO.10) 1.84 (k0.03) 9.52 (kO.11) 5.70 (+0.10) 1.69 (k0.02) 10 6 (-1 (-1 1.67 (-1 228 M.B. SCHAFFLER. D.B. BURR, AND R.G. FREDERICKSON phosphorus-emitted X-rays than the surrounding bone, and consequently the intensity of phosphorus peaks detected for cement lines would appear relatively decreased. Alternatively, the relatively increased CaiP molar ratio in cement lines may suggest a real difference in their mineralization relative to lamellar bone. Pellegrino and Biltz (1968) showed a linear relationship between increasing Ca/P molar ratio and increasing amounts of carbonate in bone mineral. Hohling et al. (1970), using electron microprobe, demonstrated higher CaiP molar ratios than expected for hydroxyapatite in initial mineralization deposits in embryonic chick bones. In a similar developmental model, an inverse relationship was shown between acid phosphate and carbonate concentrations (Pellegrino and Biltz, 1972). The presence of carbonate rather than phosphate in bone mineral would result in an increased calcium-phosphorus molar ratio and is a likely explanation for the high Ca/P molar ratio seen in cement lines. This cannot be assessed, however, within the current detection limits of electron microprobe analysis. Similarly, in order to determine the influence of cement line matrix absorption on X-rays emitted from phosphorus, a ZAF correction would have to be used. Data from the current investigation indicate that cement lines are compositionally distinct from lamellar bone. They are both less highly mineralized and differently mineralized than the surrounding bone matrix. Taken together with the higher sulfur levels present in cement lines, these data suggest that cement lines may represent a residuum of mineralized “ground substance” from the initial reversal phase of formation of a secondary osteon. These data also suggest that cement lines may be more mechanically compliant than surrounding bone and may also help explain some of the biomechanical functions attributed to cement lines. ACKNOWLEDGMENTS This work was supported by NM Biomedical Research Grant SO7 RR0443-18 and NM grant AM27127. The authors wish to thank Dr. Gary Kirk for his helpful discussion, comments, and criticism on the manuscript, and Professor John Scott for supplying the alcian blue. LITERATURE CITED Amprino, R., and A. Engstrom (1952) Studies on X-ray absorption and diffraction of bone tissue. Acta Anat., 15:l-22. Andrews, S.B., J.E. Mazurkiewicz, and R.G. Kirk (1983)The distribution of intracellular ions in the avian salt gland. J. Cell Biol., 96:1389-1399. Bavlink, D.. J. Wergedal, and E. Thompson (1972) Loss of protein polysaccharides sites where bone mineralization is initiated. J. Histochem. Cytochem. 20:279-292. Bonucci, E., A. Ascenzi, F. Vittur, M.C. Pugiarello, and B. DeBernard (1970) Density of osteoid tissue and osteons at different degrees of calcification. Calcif. Tissue Res., 5:lOO-107. Burr, D.B., R.B. Martin, M.B. Schaffler, and E.L. Radin (1985) Bone remodeling in response to in uiuo fatigue microdamage. J. Biomech., 18:189-200. Carter, D.R., and W.C. Hayes (1977) Compact bone fatigue damage: A microscopic examination. Clin. Orthop. Rel. Res., 127t265-274. Dorge, A., R. Rick, K. Gehring, and K. Thurau (1978) Preparation of ~ at freeze-dried cryosections for quantitative X-ray microanalysis of electrolytes in biological soft tissues. Mugers Arch., 373:85-97. Easthoe, J.E., and B. Easthoe (1954)The organic constituents of mammalian compact bone. Biochem. J., 57:453-459. Ebner, V. von (1875) Ueber den feineren bau der knochensubstanz. Bandes der Sitzgsber. Akad. Wiss., 111, 72t49-138. Fawns, H.T., and J.W. Landells (1953) Histochemical studies of rheumatic conditions. I. Observations on the fine structures of the matrix of normal bone and cartilage. Ann. Rheumat. Dis., 12:105113. Frasca, P. (1981) Scanning electron microscopy studies of ‘ground substance’ in the cement lines, resting lines, hypercalcified rings and reversal lines of human cortical bone. Acta Anat., 109t115-121. Glimcher, M.J. (1984)Recent studies of the mineral phase in bone and its possible linkage to the organic matrix by protein-bound phosphate bonds. Philos. Trans. R. Soc. Lond. [Biol.], 304:479-508. Gottesman, T., and Z. Hashin (1980) Analysis of viscoelastic behavior of bones on the basis of microstructure. J. Biomech., 13t89-96. Hohling, H.J., T.A. Hall, R. Boothroyd, C.J. Cooke, P. Duncomb, R.W. Fearnhead, and S. Fitton-Jackson (1970) Electron probe studies of prestages of apatite formation in bone and aorta. In: Les Tissues Calcifies. G. Milhoud, M. Owen, and H.J.J. Blackwood, eds. Soc. DEdition D’ Enseignement, Paris, pp. 323-328. Howland, J., W.M. Marriott, and B. Kramer (1926) Studies upon the inorganic composition of bones. J. Biol. Chem., 68t721-728. Katz, J.L. (1981) Composite material models for cortical bone. In: Mechanical Properties of Bone. S.C. Cowin, ed. ASME, New York, 45~171-184. Lakes, R.S., and S. Saha (1979) Cement line motion. in bone. Science, 204.501-503. Landis, W.J., and M.J. Glimcher (1978) Electron diffraction and electron probe microanalysis of the mineral phase of bone tissue prepared by anhydrous techniques. J. Ultrastruct. Res., 63:188-223. Martin, R.B., and D.B. Burr (1982) A hypothetical mechanism for the stimulation of osteonal remodeling by fatique damage. J. Biomech., 15:137-139. Mellors, R.C. (1964) Electron probe microanalysis. 1. Calcium and phosphorus in normal human cortical bone. Lab Invest. 13:183195. Pearse, A.G.E. (1980) Histochemistry. Williams and Wilkins, Baltimore. Pellegrino, E.D., and R.M. Biltz (1968)Bone carbonate and the Ca and P molar ratio. Nature, 219t1261. Pellegrino, E.D., and R.M. Biltz (1972) Mineralization in the chick embryo. I. Monohydrogen phosphate and carbonate relationships during maturation of the bone crystal complex. Calcif. Tissue Res., 10:128-135. Philipson, B. (1965)Composition of cement lines in bone. J. Histochem. Cytochem., 13:270-281. Piekarski, K. (1970) Fracture of bone. J. Appl. Phys. 41:215-223. Robinson, R.A. (1975)Phvsicochemical structure of bone. Clin. Orthop. Rel. Res., 112:261. Scott, J.E., and J. Dorling (1965) Differential staining of acid glycosaminoglycans (mucopolysaccharides) by alcian blue in salt solutions. Histochemie, 5:221-233. Scott, J.E., J. Dorling, and G. Quintarelli (1964) Differential staining of acid glycosaminoglycans by alcian blue in salt solutions. - . Biochem.-J, 90:4-5. Schmidt. W.J. (1959) Grenzscheiden der lakunen und kittlinien des knochengewebes. Polarisationoptische analyse kollagen freier kongorotgefarbten schliffe. Z. Zellforsch., 50:275-296. Sokoloff, L. (1973)A note on the histology of cement lines. In: Perspectives in Biomedical Engineering. R.M. Kenedi, ed. University Park Press, Baltimore, pp. 135-138. Strandh, J. (1960)Microchemical studies on single Haversian systems. I. Methodological considerations with special reference to variations in mineral content. Exp. Cell. Res., 19:515-530. Weidenreich, F. (1930) Das knochengewebe. In: Handbuch der Mikroskopischen Anatomie des Menschen. W. von Mollendorff, ed. Springer, Berlin, Vol. 2., pp. 391-520. Wergedal, J.E., and D.J. Baylink (1974)Electron microprobe measurements of bone mineralization rate in uiuo. Am. J. Physiol., 226t345352. Woodward, H.Q. (1964)The composition of human cortical bone. Clin. Orthop. Rel. Res., 37:187-193.