Cortical Bone Loss with Age in Three Native American Populations MARY FRANCES ERICKSEN Department of Anatomy, George Washington University Medical Center, Washington, D.C. 20037 KEY WORDS Aging . Cortical bone . Eskimo . Pueblo . Arikara , Osteoporosis. ABSTRACT Age-related thinning of cortical bone was investigated in archaeological populations of Eskimos, Pueblos, and Arikaras. Medial-lateral cortical thickness was measured on radiographs of humerus and femur, and thickness of the anterior femoral cortex was measured directly on samples taken for histologic study. Maximum length of the bones was used to calculate indices of relative cortical thickness, in order to minimize differences due to body size and build. Bone loss in the humerus begins before middle age in all three populations and, except for Eskimo males, the same is true of the anterior femoral cortex. In general, overall female loss of cortical bone amounts to two or three times that of the males, and in the case of the humerus and the anterior cortex of the femur, this difference is evident by middle age. The weight-bearingfemoral medial-lateral cortex shows less sexual difference but has the greatest number of statistically significant differences between populations and the greatest contrast between populations in pattern of loss with age. It appears that of the cortical regions studied this is the area upon which environmental factors have the greatest effect, whereas areas more subject to tensile stress, the humerus and anterior femoral cortex. are less affected by these factors. Among the major findings of the recent maturity and cortical thickness at any upsurge of interest in bone biology are: given age level (Garn et al., '69; Garn, '70). bone involution is a normal concomitant of On the other hand, workers comparing two aging in all human populations; females or more ancient populations or comparing lose bone at a greater rate than males; ancient and modern populations found difand bone loss -thinning of cortex and ferences in the total amount of bone loss cancellous trabeculae -begins soon after over time and/or in the pattern (or timing) maturity is reached (Nordin, '66; Garn et of loss (Dewey, Bartley, and Armelagos, al., '67; Garn et al., '69; Garn, '70; De- '69; Perzigian, '73). queker, '72). Several recent studies have It is possible, of course, that these disinvestigated bone loss with age in archae- crepancies resulted from differences in ological populations (Dewey, Armelagos, technique, from differences between the and Bartley, '69; Dewey, Bartley, and Ar- areas of the skeleton studied, or from the melagos, '69; VanGerven et al., '69; Mazess limitations inherent to archaeological popand Jones, '72; Perzigian, '73). Garn and his ulations, especially the difficulty of detercolleagues, studying thousands of in- mining age at death. However, recently dividuals in several countries, at first found Garn ('72) reported differences between no differences in bone loss between popu- American white and black females in lations of diverse diet, environment, and amount of bone loss in the second metacargenetic background, although these factors pal. Similarly, Dequeker ('72) found that were found to affect overall skeletal size at once he had rigorously controlled for body -4~ J. .PHYS. ANTHROP., 45: 443-452. 443 444 MARY FRANCES ERICKSEN size, Nigerians differed from Europeans in rate of bone loss with age, although the differences were not statistically significant. Thus, it is possible that genetic and environmental factors do affect aging bone loss. The purpose of the present study is to add to the data on bone loss with age in archaeological populations and also to determine whether environmental and/or genetic factors may have affected rates or patterns of bone involution in these groups. MATERIAL AND METHODS Three native American groups differing greatly in lifestyle, diet, and physical environment were selected for study: Pueblo of the Southwest, Arikara of South Dakota, and coastal Alaskan Eskimo. Descriptions of the physical types and environmental and cultural settings of these groups are given elsewhere (Ericksen, '73; also Spencer et al., '65). Briefly, the Pueblo were sedentary town-dwelling farmers of the Colorado Plateau, whose diet consisted almost entirely of vegetable foods. The Arikara were village-dwelling horticulturists, but they also spent a large part of each year roaming the Plains hunting bison. Their diet was well balanced b y modern standards, between animal and vegetable food. The Eskimo lifestyle was dominated by the seasonal movements of the birds, fish, and sea mammals they hunted, and their diet consisted almost entirely of animal products. Although all three groups lived in harsh environments, that of the Eskimo was especially rigorous. Genetically, the Arikara and Pueblo are considered to be more closely related to each other than to the Eskimo. The skeletons studied are in the Anthropology Department of the Smithsonian Institution. The 142 Pueblo skeletons are from three archeological sites in New Mexico: Pueblo Bonito, Puye, and Hawikuh. Pueblo Bonito was abandoned around 1130 A.D., and Hewett ('09) found no sign of European influence at Puye. Smith et al., ('66)consider the Hawikuh skeletons in the Smithsonian collection to date mainly from the late pre-Spanish and early Spanish pe- riods, that is, the sixteenth and seventeenth centuries. The 134 Arikara skeletons were excavated from 11 archaeological sites in northern South Dakota along the upper Missouri River and its tributaries. Principal sources for site identification and dating were: Wedel, '55; Lehmer, '54; Lehmer and Jones, '68; Jantz, '70; Bass et al., '71. A few of the skeletons are from sites dated t o t h e Extended Coalescent Horizon, 1550-1675 A.D., but most are from the Post-Contact Coalescent Horizon, 16751845 A.D. (Lehmer and Jones, '68). The Alaskan Eskimo sample consists of 123 skeletons found from 14 mainland coastal settlements between Point Hope on the Arctic Ocean and Pastolik on the south shore of Norton Sound. According to the Anthropology Department catalog, the material from Point Hope dates from the late eighteenth and early nineteenth centuries, and the other skeletons are from the late nineteenth and early twentieth centuries. These samples obviously are not true populations, especially the Eskimo and Arikara. However, when juvenile, fragmentary, and pathological skeletons, and those whose age or sex is doubtful are eliminated, no collection from a single site is large enough to provide a reasonable sample for statistical analysis. The three groups are sufficiently separated geographically so that there is no question of intermixture, and within -group differences are likely to be less than between-group differences. Corruccini ('72) demonstrates this for the Pueblo groups used in the present study. Skeletons showing pathological change other than simple osteoarthritis or trauma were excluded from the study. Age determination was based on assessment of the whole skeleton, mainly using the method of McKern and Stewart ('57).Although their method may be unreliable for groups like those of the present study, especially for females over the age of 40 (Gilbert, '73; Gilbert and McKern, '731, they should provide reasonably accurate estimates for dividing the samples into three gross age 44s CORTICAL BONE LOSS WITH AGE groups: young (18-251, middle-aged (30under 501, and old (50 and above). Individuals between the ages of 25 and 30 were excluded, in the hope that any earlydeveloping age changes might thereby become more evident. Except among the Pueblo and the Arikara males, the collections include very few individuals over 40. The humerus and femur were chosen for study because they are two of the bones which Trotter et al. ('59) report to be unlike in the amount of sexual differences in density loss with age. These authors consider that differences in functional stress -weight bearing and tension -are related to their finding less difference between the sexes in density of the femur than in the humerus. For each skeleton, the maximum lengths of one humerus and one femur were measured on an osteometric board, whenever possible, the bones used were those of the left side. In 23 cases maximum Length of one of the bones was estimated by the regression formulae of Steele and McKern ('691. The midpoint of the shaft was marked by a lead shot, so that it could be located on a radiograph of the bone. Radiographs were taken at a tube-film distance of 60 inches, with the bones laid directly on the film holder. An exposurt. of 70 milliampere-seconds at 70 kilovolts was found to be best for distinguishing trabecular from cortical bone on the endosteal surfaces. The humerus was oriented with its posterior surface on the film holder, but the femur was positioned with its anterior surface on the film holder, to avoid distortion of the image due to bowing of the shaft. Medial-lateral diameter and medullary cavity diameter were measured on the radiographs, to the nearest 0.5 mm, using a Helios dial caliper and excluding trabecularized endosteal bone. The femur was measured at midshaft, but preliminary examination of the radiographs indicated that it would be impractical to measure the humerus at midshaft. The lateral cortex in the area of the deltoid tuberosity tends to become divided into distinct inner and outer "tables" with increasing age and does not give a realistic picture of simple cortical involution. Therefore, the measurements were taken at a level immediately below the tuberosity, where the shaft becomes relatively parallel-sided. In addition to t h e data from the radiographs, it was possible to measure thickness of the anterior femoral cortex directly in a few individuals of each age and sex group. For purposes of future histologic study, I used a 5/8-inch core drill to take samples of the anterior cortex just below the midshaft (to preserve this anthropometric landmark). The midline cortical thickness of these samples was measured to the nearest 0.5 mm with the same caliper used for the radiographs. Three indices of relative cortical thickness were calculated from the measurements, to minimize differences due to size and body build. Kerley ('61) notes, for example, that gross cortical thickness is not consistently associated with age, and Dewey, Bartley, and Armelagos 1'69) found correction for body size extremely important in comparing populations. Dequeker ('72) also strongly emphasizes the importance of correcting for body size. The three indices are: (1) Femur medial-lateral index (FM- LI): shaft diameter - medullary diameter 2 maximum length (21 Femur anterior index (FAI): midline anterior cortical thickness maximum length (3) Humerus medial-lateral index (HMLI): same as (11,above. RESULTS The data were processed by the Computer Center of the Smithsonian Institution; differences between sample means were tested by Student's t and the variances b y the Fratio. Differences significant according to the t test but whose F ratios were also significant were discarded from the analysis. Test results were analyzed according to differences between age levels 446 MARY FRANCES ERICKSEN (sex and population held constant), sex (age and population constant), and population (age and sex constant). Table 1 gives sample sizes, means, and standard errors of the three indices and also indicates statistically significant sex differences within each agepopulation group. Figures 1-3present the indices graphically. Cortical loss (or gain) between age levels is presented in table 2, and the statistical ly significant differences are indicated. Change from one age level to the next is expressed as a percentage of the mean of the younger age group: younger mean - older mean younger mean All but three values in table 2 represent cortical loss, but the most striking features of the table are the differences between the sexes in amount of loss and the contrast between femur and humerus in regularity of pattern. Medial-lateral femoral cortex Ten of the 12 statistically significant population differences occur in the index measuring relative thickness of the mediallateral femoral cortex, FM-LI. T h e greatest contrast in cortical thickness is between Eskimos and Pueblos, with the Arikara tending to be intermediate, except in old age, when Eskimo females and both sexes of Arikara show pronounced loss of cortex. Eskimo males have a statistically significant advantage over other males in relative amount of cortex at each age level. Female Eskimos have greater cortical T4RLE 1 Indices: cortical thickness as a percentage of hone length. Sample size, means, standard errors, and statistically significant male-femule differences Young Mean Mean Mean (W) S.E. No. (%I S.E. 28 10 27 1.67 0.04 1.11 0.05 1.43 0.05 26 10 24 1.71 1.17 1.37 0.03 0.05 0.05 7 7 7 1.64 1.07 1.20 0.08 0.08 0.12 27 8 26 1.65 0.05 1.24 0.05 1.39 0.04 27 12 25 1.63 0.05 1.12 0.05 1.16' 0.06 6 6 6 1.29 0.70 0.80 2 0.06 0.07 0.09 17 10 16 1.48 0.06 1.18 0.05 1.38 0.06 26 12 26 1.41 1.14 1.34 0.04 0.06 0.04 20 11 22 1.37 1.08 1.16 0.04 0.05 22 13 24 1.41 0.04 1.10 0.03 1.30 0.05 27 10 27 1.36 0.05 0.91 8 0.08 1.09 0.06 23 13 24 1.26 0.82 1 0.80 0.05 0.05 0.05 20 10 18 1.50 0.04 1.24 0.04 1.46 0.05 49 24 49 1.59 1.17 1.36 0.03 0.03 0.03 10 10 10 1.39 1.02 1.16 0.06 0.07 0.05 25 12 26 1.53 0.03 1.14 0.04 1.37 0.03 25 1.52 0.04 1.022 0.04 1.15' 0.05 2 2 2 1.18 0.73 9 0.60 1 0.42 0.10 0.06 No. Eskimo males Femur medial-lateral index Femur anterior index Humerus medial-lateral index Eskimo females Femur medial-lateral index Femur anterior index Humerus medial-lateral index Puehlo males Femur medial-lateral index Femur anterior index Humerus medial-lateral index Pueblofemales Femur medial-lateral index Femur anterior index Humerus medial-lateral index Arikara males Femur medial-lateral index Femur anterior index Humerus medial-lateral index Arikarafemales Femur medial-lateral index Femur anterior index Humerus medial-lateral index (%I Old Middle-aged 10 25 Female mean significantly below male mean of the same age-population group, P Female mean significantly below male mean of the same age-population group, P < < 0.01. 0.05. ?Go. S.E. 0.05 447 CORTICAL BONE LOSS WITH ACE I .5 f \ \ ‘1 \\ I .3- Y M O &3 \ \ I Y ESK I MO M O PUEBLO \ Y M O AR IKARA Fig. 1 Femur medial-lateral index means (Y axis) plotted against age (X axis) for three native American populations. Index expresses average medial-lateral cortical thickness as a percentage of maximum femur length. Y, young (18-25);M, middle-aged (30-under 50);0, old (50 and above). 1.2 1 t \ I .o / \ \ \ \ L 0.8 0.6 Y M ESKIMO O Y M O PUEBLO Y M O AR I K A R A Fig. 2 Femur anterior index means (Y axis) plotted against age (X axis) for three native American populations. Index expresses anterior cortical thickness as a percentage of maximum femur length. Y, young (18-25);M,middle-aged (30-under 50); 0, old (50 and above). thickness than other females in youth and middle age, and the differences from Pueblo females are significant. Pueblo female bone loss between youth and middle age is enough so that they also have a significantly lower FM-LI than Arikara females in middle age. Although all groups show a drop in FMLI between youth and old age, the curves or “patterns” of loss are different (fig. 1). 448 MARY FRANCES EHICKSEN $ \ \ \ \ \ \ L 0.5 Y M ESKIMO O Y M O PUEBLO Y M O AR I KARA Fig. 3 Humerus media1-lateral index means (Y axis) plotted against age (Xaxis) for three native American populations. Index expresses average medial-lateral cortical thickness as a percentage of maximum humerus length. Y,young (18-25), M, middle-aged (30-under 501, 0, old (50 and above). The decline is not statistically significant in either Eskimo or Pueblo males, but the Pueblo males show a steady loss, amounting to 7.43%between youth and old age. Eskimo males gain in this index into middle age, and their total loss is a mere 1.80%. Arikara males also gain between youth and middle age, so that in middle age they have a significantly higher FM-LI than Pueblo males, but they lose much more than the other males between middle and old age. Pueblo females show the same sort of steady reduction in FM-LI as Pueblo males, but Eskimo and Arikara females show little loss until after middle age. Loss in FM-LI is greater between youth and middle age in Pueblos than in the other groups. On the other hand, Pueblo cortical loss between middle and old age is considerably less than in the other groups of the same sex. Table 2 indicates that the female rate of cortical loss is appreciably greater than that of the males. However, it is uncertain how much of the difference among old Eskimos and Arikaras is due to small sample size in these groups. The sexual difference between old Pueblos is considerably less. Anterior femoral cortex In contrast to the medial-lateral cortex, sex differences in loss of anterior femoral cortex are evident by middle age. In all three female groups loss in FA1 between youth and old age is statistically significant, but among males only the Arikara show a significant difference between young and old. By middle age Arikara and Pueblo female means are significantly below male means, and in all three groups FA1 is significantly lower in old females than in old males. In all groups, except the Eskimo males, loss begins between youth and middle age and is steady thereafter (fig. 2). Pueblo 449 CORTICAL BONE LOSS WITH AGE TABLE 2 Cortical hone loss with age: dijferences between means of successive age levels, expressed as a percentage of the younger group mean Population group Eskimo males Young vs. middle-aged Middle-aged vs. old Young vs. old Puehlo males Young vs. middle-aged Middle-aged vs. old Young vs. old Arikara males Young vs. middle-aged Middle-aged vs. old Young vs. old Eskimofemales Young vs. middle-aged Middle-agrd vs. old Young vs. old Pueblofemales Young vs. middle-aged Middle-aged vs. old Young vs. old Arikara females Young vs. middle-aged Middle-aged vs. old Young vs. old I Fernur medial-lateral index Femur anterior index Humerus medial-lateral index % % % difference difference difference + + - 4.73 - 2.84 - 7.43 - 2.40 - 4.09 - 1.80 3.40 - 8.55 - 3.60 - 4.19 - 12.41 - 16.08 3.39 - 5.26 - 8.47 - 5.64 - 12.82 - 17.74 2 - 2.90 13.43 ' - 15.94 1 - + 6.002 - 12.58 1 - 7.33 - 1.21 - 20.86 I - 21.82 - 9.68 - 37.50 - 43.55 ' - 31.03 - 3.55 - 7.35 - 10.64 - 17.27 - 9.89 - 25.45 1 - 16.15 2 - 26.60 1 - 38.46 - J ' - 0.65 - 22.37 - 22.88 Difference between age levels statistically significant, P Difference between age levels statistically significant, P - 10.53 - 28.43 - 35.96 < < ' 6.85 - 14.70 1 - 20.55 1 - 16.55 1 - 42.45 16.06 - 47.83 - 56.20 ' ' - J 0.01. 0.05. females, but not the males, differ from the other groups in that the greatest percentage of loss occurs between youth and middle age, but as in FM-LI, they appear to lose appreciably less anterior cortex than the other females over a total lifespan (table 2 ) . ences in HM-LI are evident by middle age. From the beginning females show higher percentages of loss, and their total loss is at least twice that of the males. The sexual difference in HM-LI is statistically significant in middle-aged Eskimos and Arikaras and in all groups in old age. Medial -lateral humeral cortex The pattern of involution of the mediallateral humeral cortex is similar to that of the anterior femoral cortex and cven more regular (fig. 3 ) . All groups show a steady decline in HM-LI from youth onward, statistically significant between all age levels in Eskimo and Pueblo females and from middle age onward in all but Eskimo males (table 2). As in FAI, sexual differ- DISCUSSION The present investigation agrees with many previous studies in finding that sex is consistently related to rates of cortical bone involution. However, it is evident that another factor must be taken into account when interpreting the findings of a given study -the effect of function upon the patterns of aging in different bones and in different areas within the same bone. What 450 MARY FRANCES ERICKSEN we perceive depends upon where we look. As mentioned above, Trotter and her coworkers ('59) noted that sex differences in density were less in weight -bearing bones than in bones more subject to tensile stress. Similarly, Garn et al. ('69) found that the tibia and second metacarpal differed in both onset (earlier in the tibia) and rate (less in the tibia) of cortical involution. VanGerven et al. ('69) presented evidence that different areas around the circumference of the mid-femoral cortex change differently with aging, just as they do during growth (Johnson, '66). The present study confirms these findings. The humerus shows the most consistent and steady change in all three groups, but by middle age females have lost at least twice as much cortex as males (table 2). No differences appear which could be attributed to race, and such environmental factors as physical activity and nutrition, which differed a great deal in these groups, apparently had little or no effect, as the curves of loss are parallel and even. These findings for the humerus are similar to those reported by Garn et al. ('70) for cortical loss in the second metacarpal in several thousand subjects in North and Central America. They found that bone loss began in both sexes by age 40 and averaged 30%in females between ages 30 and 80, whereas males lost only 12%in the same time span. The sexual difference was evident by age 40, and the average percentage of overall loss was no greater in Central American women, who tended to nurse babies from one to three years (Frisancho et al., '71) than in Ohio white females, among whom breast feeding is rare and tends to be of short duration. For this reason and because Garn and his colleagues (Garn et al., '70) have presented evidence that high parity is associated with decreased cortical thinning, I am reluctant to ascribe the sex differences found in the present study to the effects of pregnancy and lactation. It appears that a greater rate of aging bone loss in females is a universal fact of life and not dependent upon such factors . The anterior femoral cortex, which is away from the main axis of weight bearing (Evans, '571, follows a pattern of involution much like that of the humerus, but the medial-lateral cortex presents a more complex picture of bone loss with age. Most notable is the late appearance of appreciable male-female differences in amount of cortical loss. For instance, Eskimo females show a 1.21%loss in FM-LI between youth and middle age, and the males show a 2.40%gain in this period, but this difference is much less than that seen in HM-LI, in which Eskimo females lose almost four times as much cortical thickness as the males between youth and middle age. Similarly, in the anterior femoral cortex Eskimo males gain between youth and middle age, while the females lose almost 10%.The midshaft medial-lateral cortex of the femur is closer to the main axis of weight bearing than is the anterior cortex (Johnson, '661, and it is possible that this has a "maintenance" effect, somewhat suppressing cortical loss in the female during the early decades of maturity. Pueblos show earlier and more regular loss of femoral medial-lateral cortex than the other groups. Eskimo and Arikara males have a rise in FM-LI between youth and middle age and a drop thereafter, minimal in Eskimos. Eskimo and Arikara females show a slight decline into middle age and a steep one thereafter, with the result that their lifetime loss of cortex is greater than that of the Pueblo females. Apparently in this area environmental or genetic factors are superimposed upon the "normal" pattern of aging, so that population differences appear in both onset and rate of involution. Although, as described above, Garn ('72) and Dequeker ('72) found racial differences in aging bone loss in the second metacarpal, especially among females, it is difficult to ascribe a major role in the differences between the present groups to genetic factors for two reasons. Principally, if this were the main cause of the differences, the Arikara and Pueblo should most closely resemble each other and the Eskimo stand apart. Instead, CORTICAL BONE LOSS WITH AGE 45 1 it is the Pueblo who differ from the other Frisancho, A. R., S. M. Garn and W. Ascoli 1971 Unaltered cortical area of pregnant and lactating two groups. Second, genetically deterwomen: studies of the second metacarpal bone in mined differences, if they are important in North and Central American populations. Invest. . . this case, should appear in the humerus as Radlol., 6: 119-121. well as the femur. Therefore, I feel that Garn. S. M. 1970 The Earlier Gain and the Later h s s o f Cortical Bone. Chas. C Thomas, Springfield, some environmental factor or factors -at Illinois. a guess, both nutrition and level of physical 1972 The course of bone gain and the activity -must play a part in determining phases of bone loss. Orthop. Clin. N. America, 3: the pattern and amount of age-related loss 503-520. in the medial-lateral cortex of the femur. Garn, S. M., C. G. Rohmann and B. Wagner 1967 ACKNOWLEDGMENTS I am grateful to Dr. J. Lawrence Angel, Curator of Physical Anthropology, and to the other members of the Department of Anthropology of the Smithsonian Institution for their hospitality and assistance during the gathering of the data for this study. I am especially indebted to Dr. Donald J. Ortner for arranging the computer processing of the data and to Dr. T. Dale Stewart for his guidance and support throughout the project. The material presented here is taken from a dissertation presented to the George Washington University in partial fulfillment of the requirements for the Doctor of Philosophy degree. LITERATURE CITED Bass, W. M., D. R. Evans and R. L. Jantz 1971 The Leavenworth Site Cemetery: Archaeology and Physical Anthropology. Univ. Kansas Publ. in Anthrop., 2. University of Kansas, Lawrence. Corruccini, R. S. 1972 The biologic relationships of some prehistoric and historic Pueblo populations. Am. J. Phys. Anthrop., 37: 373-388. Dequeker, J. 1972 Bone Loss in Normal and Pathological Conditions. Leuven Univ. Press, Leuven, Belgium. Dewey, J. R., G. J. Amelagos and M.H. Bartley 1969 Femoral cortical involution in three Nubian archaeological populations. Hum. Biol., 41: 13-28. Dewey, J. H., M. H. Bartley and G. J. Amelagos 1969 Rates of femoral cortical bone loss in two Nubian populations: utilizing normalized and non-normalized data. Clin. Orthop., 65: 61-66. Ericksen, M. F. 1973 Age-related bone remodeling in three aboriginal American populations. Doctoral dissertation. George Washington University, Washington. Evans, F. G. 1957 Stress and Strain in Bones, their Relation to Fractures and Osteogenesis. Chas. C Thomas, Springfiield, Illinois. Bone loss as a general phenomenon in man. Fed. Proc., 26: 1729-1736. Cam, S. M., C. G. Rohmann, B. Wagner and G. H. D a d a 1970 Dynamics of change at the endosteal surface of tubular bones. In: Progress in Methods of Bone Mineral Measurements. Whedon, G. D., and J. R. Cameron, eds. Nat. Inst. Health, Washington, pp. 430-453. Garn, S. M., C. G. Rohmann, B. Wagner, G. H. D a d a and W. Ascoli 1969 Population similarities in the onset and rate of adult endosteal bone loss. Clin. Orthop., 6551-60. Gilbert, B. M. 1973 Misapplication to females of the standard for aging the male os pubis. Am. J. Phys. Anthrop., 38: 39-40. Gilhert, €3. M., and T. W. McKem 1973 A method for aging the female 0s pubis. Am. J. Phys. Anthrop., 38: 31-38. Hewett, E. L. 1909 The Pajaritan Culture. Pap. Sch. Am. Arch., 3. Archeological Institute of America. Jantz, R. L. 1970 Change and variation in skeletal populations of Arikara Indians. Doctoral dissertation, University of Kansas, Lawrence. Johnson, L. C. 1966 The kinetics of skeletal remodeling. Birth Defects, Orig. Art., Ser. 2: 66-142. Kerley, E. R. 1961 The microscopic determination of age in human bone. Doctoral dissertation, University of Michigan, Ann Arbor. Lehmer, D. J. 1954 Archeological Investigations in the Oahe Dam Area, South Dakota, 1950-51. Bur. Am. Ethnol. Bull. 158. Government Printing Offce, Washington. Lehmer, D. J., and D. T. Jones 1968 Arikara Archeology: the Bad River Phase. Publ. in Salvage Arch. 7. Smithsonian Inst., River Basin Surveys, Lincoln, Nebraska. McKem, T. W., and T. D. Stewart 1957 Skeletal Age Changes in Young American Males. Quartermaster Res. and Dev. Command, Tech. Rep. EP-45, Natick, Massachusetts. Mazess, R. B., and K.Jones 1972 Weight and density of Sadlermiut Eskimo long bones. Hum. Biol., 44: 537-548. Nordin, B. E. C. 1966 International patterns of osteoporosis. Clin. Orthop., 45: 17-30. Perzigian, A. J. 1973 Osteoporotic bone loss in two prehistoric Indian populations. Am. J. Phys. Anthrop., 39: 87-95. Smith, W., R. B. Woodbury and N. F. S. Woodbury 1966 The Excdvdtion of Hawikuh by Federick Webb Hodge. Contrib. Mus. Am. Indian 20. Heye Foundation, New York. Spencer, R. F.,J. 1).Jennings, C. E. Dibble, E. Johnson, A. R. King, T. C . Stern, K. M. Stewart, D. C. Stewart and W. J. Wallace 1965 The Native Americans. Harper and How, New York, Evanston, and London. Steele, D. G., and T. W. McKern 1969 A method for assessment of maximum long bone length arid living stature from fragmentary long hones. Am. J. Phys. Anthrop., 31: 215-227. Trotter, M., G. E. Broman and H. H. Peterson 19.59 Density of cervical vertebrae and comparison with densities of other bones. Am. J. Phys. Anthrop., 17: 19-25. VanGcrven, D. P., G. J. Armelagos and M.H. Bartley 1969 Roentgenographic and direct nieasrirernent of femoral cortical involution in a prehistoric Mississippian population. Am. J. Phys. Anthrop., 31; 23-38. Wedel, 1%'. R. 1955 Archeological materials from t h e vicinity of Mobridge, South Dakota. Anth. Pap. 45, Bur. Am. Ethnol. Bull. 157, Govtwunent Printing Office, Washington, pp. 69-188.