AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 80:173-185 (1989) Bone Strontium in Pregnant and Lactating Females From Archaeological Samples ROBERT L. B W L Y Department of Anthropology, Georgia State University, Atlanta, Georgia 30303 KEY WORDS Strontiudcalcium ratios, Diet, Prehistoric native Americans ABSTRACT Because plants and animals consume or absorb different amounts of strontium and calcium, anthropologists are able to use strontium/ calcium (Sr/Ca) ratios from archaeologically recovered human bone to estimate the relative contributions of meat and plants to paleodiets. Often females exhibit higher Sr/Ca ratios than males, a fact usually attributed to lower meat intake among women. However, in vivo and in vitro experiments with laboratory animals show that pregnancy and lactation elevate maternal bone strontium and depress maternal bone calcium because 1) strontium is discriminated against in favor of calcium in the transport of ions to the placenta and mammary glands and 2) pregnancy and lactation facilitate absorption of alkaline earth metals from the gut. In this study, bone Sr/Ca ratios and strontium concentrations were compared between reproductive-age females, postmenopausal females, and adult males from two late prehistoric Native American sites in Georgia: the King site (N = 43) and the Etowah site (N = 51). At the King site, the mean Sr/Ca ratio of females was over 14%greater than that of males. At Etowah, the mean strontium level of reproductive-age females exceeded that of postmenopausal females by almost 25%.Most of the difference, it is argued, is due to pregnancy and lactation. A dietary preference among pregnant and lactating women for foods high in alkaline earths, particularly nuts and corn, may also be partially responsible. Until we assess the influence variables other than nutrition exert on trace element concentrations, our reconstructions of paleodiets will be suspect. Among the outgrowths from above-ground tests of nuclear fission in the 1950s and 1960s was concern over the health hazards posed by radiostrontium ("Sr) in the ecosphere. Radioactive and stable strontium behave similarly in food chains (Comar, 1963). Most plants indiscriminantly absorb strontium and other alkaline earth metals -calcium, barium, and magnesium-from the soil (Hutchin and Vaughan, 1968; Sanzharova, 1978). Terrestrial vertebrates, on the other hand, discriminate against strontium in favor of calcium, since the proteins necessary for the transfer of ions across the intestinal mucosa have a lower affinity for strontium than for calcium (Ingersoll and Wasserman, 1971; Menczel and Mor, 1972). This may reflect the fact that calcium is a more essential nutrient in vertbrate metab- @ 1989 ALAN R. LISS, INC. olism than strontium (Comar and Wasserman, 1964; Schroeder et al., 1972; Spencer et al., 19731.l As a consequence, strontium! calcium (Sr/Ca) ratios decrease as one moves up the food chain. Herbivores absorb only a small amount of the strontium available to them in plants. Carnivores incorporate even less strontium than herbivores because of both continued discrimination and reduced amounts of strontium in their diets. OmniReceived May 4,1988; accepted December 3,1988. 'The essentiality of strontium in vertebrate metabolism is controversial. Although most investigators contend that strontium has no known independent metabolic function (Rosenthal, 1981; Sillen and Kavanagh,19821, some argue that, like calcium, it aids in the formation and maintenance of bones, enamel, and dentin. Strontium deficiencies are reported to result in depressed growth, impaired mineralization of bones and teeth, and high incidence of dental caries (Schroeder et al., 1972; Underwood, 1977). 174 R.L. B m L Y vores, such as humans, exhibit bone strontium levels intermediate to those of herbivores and carnivores, the amount dependent in part on the proportion of plants and animals consumed (Toots and Voorhies, 1965). Anthropologists have used this principle to infer the relative contributions of meat and plant foods t o paleodiets (see, for example, Brown, 1973; Gilbert, 1975; Schoeninger, 1979, 1982; Sillen, 1981; Price and Kavanagh, 1982; Katzenberg, 1984; for comprehensive overviews of the uses of strontium in paleodietary research, see Sillen and Kavanagh, 1982; Price et al., 1985). Some studies have been hampered by failure to take fully into account the influence of variables other than diet on Sr/Ca ratios in human bone. Diagenesis-postdepositional change in the chemical composition of buried bone-is one (see Lambert et al., 1979,1982, 1984,1985; Sillen, 1981; Vlasak, 1982; Pate and Brown, 1985; Klepinger et al., 1986; Nelson et al., 1986; Byrne and Parris, 1987; Price, 1988); disease is another (Sillen and Kavanagh, 1982; Price et al., 1985);the complexities of strontiudcalcium metabolism are a third (Comar, 1963; Lengemann, 1963; Thompson, 1963; Lenihan et al., 1967; Spencer et al., 1973; Elias et al., 1982; Sillen and Kavanagh, 1982; Price et al., 1986). Because the gastrointestinal tract discriminates against strontium in favor of calcium in the absorption of alkaline earths, only 2040% of ingested strontium passes through the intestinal wall, compared with 4040% of ingested calcium (Comar, 1963). The unabsorbed strontium is excreted in feces. Of the strontium absorbed through the intestinal wall, about 99% is stored in the bone mineral, or hydroxyapatite (Schroeder et al., 1972).The remainder is housed in soft tissue, excreted renally, and transferred to the placenta and mammary glands (Hartsook and Hershberger, 1973; Kostial et al., 1969; Spencer et al., 1973; Price et al., 1986). Sillen and Kavanagh (1982:75)note that A n important consequence of pregnancy and lactation may be an alteration in the strontium levels of maternal skeletons. The effect has been demonstrated in laboratory animals and occurs when there is (a) an increase in the absorption of all alkaline earths and (b) maintenance of discrimination against strontium in the transport of ions across the placenta and mammary glands. The result is a signifi- cant net increase in the Sr/Ca ratios of the maternal plasma, and, in turn, maternal skeletons (Kostial et al., 1969; Blanusa et al., 1970) [emphasis in the original].' The authors go on to point out that it is unclear whether this occurs in humans, since the rate of bone mineral turnover is slower than in laboratory animals (Bronner and Lemaire, 1969). Studies of modern industrial populations have not revealed significant differences in Sr/Ca ratios or in strontium levels of adult male and female skeletons (Turekian and Kulp, 1956; Tanaka et al., 19811, but these data may not be applicable to prehistoric populations, for whom it can be assumed that more women were either pregnant or lactating throughout most of the reproductive years. Price et al. (1986) estimate that 50% or more of the adult females in preagricultural samples were either pregnant or lactating. Studies of strontium in prehistoric human remains have almost routinely reported elevated strontium in adult females (Brown, 1973; Gilbert, 1975; Lambert et al., 1979, 1982; Sillen, 1981;Katzenberg, 1984).Usually this is attributed to dietary differences between the sexes, often to lower meat intake among women. (Katzenberg does relate the difference a t the Serpent Mounds site in Ontario to pregnancy and lactation.) SrKa ratios will go up when 1)the level of calcium is decreased, 2) the level of strontium is increased, or 3) both occur simultaneously. With absorption of more calcium into the placenta and mammary glands, pregnancy and lactation account for a reduction in the calcium available to maternal bone (Atkinson and West, 1970). According to Wasserman et al. (19581,about 85%of the calcium in maternal milk originates from the diet; the other 15% comes from the maternal skeleton. In vivo experiments by Price and coworkers (1986) show that strontium available to maternal bone is increased during preg'Because of discrimination against strontium in favor of calcium in the transuort of ions to the mammarv glands. and the fact that lactose facilitates the absorption of cal"cikn (Lengemann et al.. 1959: Fournier et al.. 1971). maternal milk has exceDtionallv low Sr/Ca ratios-n the order of 0.14 to 0.3 mg'Sr/g Ca (Rosenthal, 1981). Therefore, even though infants discriminate against strontium less efficiently than adults, and alkaline earths are more completely absorbed from liquids than from solid foods (Marcus and Lengemann, 19621,the skeletons of unweaned infants also have low SriCa ratios (Lough et al., 1963;Twardock, 1963). At weaning, Sr/Ca ratios rise sharply before declining to adult values (Sillen and Smith, 1984). 175 STRONTIUM IN FEMALES IN PREHISTORIC AMERICA nancy and lactation. In vitro studies by Kostial and colleagues (1969) demonstrate that this is due to the fact that pregnancy and lactation produce morphological changes in the intestinal mucosa and serosa that facilitate the absorption of alkaline earths from the gut. Their results indicate that the absorption of strontium and calcium in lactating rats was two to three times greater than in controls. Whether this is sufficient to account for the elevated bone strontium observed in female skeletons from archaeological sites is not known. One test for the association between Sr/Ca ratios and pregnancy and lactation would be the demonstration that the highest Sr/Ca ratios occur in females of reproductive age. In the present study, archaeological samples from the southeastern United States showed that reproductive-age women had both elevated strontium and depressed calcium compared to adult males and postmenopausal females. Lowered calcium is attributed to discrimination against strontium in favor of calcium in the transport of ions to the mammary glands. Increased strontium is associated with both enhanced intestinal absorp- Tennessee : N o r t h Carolina tion of alkaline earth metals and a dietary preference among reproductively active women for foods high in these elements. (The data do not reveal what portion, if any, of the difference might be due to dietary choice or necessity among reproductive-age females, but it is not unreasonable t o think that dietary supplements would have been beneficial. The Creek Indians, who archaeological and ethnohistoric accounts show were direct descendants of the peoples represented by the skeletal series [Hally, 19881,had taboos against eating meat at times during pregnancy and throughout menstruation [Swanton, 19221.) We include a cautionary note: Until we know the effects that normal physiological processes and attendant behaviors have on trace element concentrations, we will continue t o reconstruct paleodiets incorrectly. MATERIALS AND METHODS The samples used in this study were recovered from the Etowah and King sites in northwestern Georgia (Fig. 1). The Etowah site is situated on the flood plain of the Etowah River as it winds westward through A.D. 1600 -, Etowah 1400- 1350- 1300 - 1250- 1200- Fig. 1. Location and chronology of the Etowah and King sites. Etowah King 176 R.L. BIAKELY the Piedmont Province. The site’s location allowed its inhabitants to exploit three distinct ecological zones-Piedmont, Great Valley, and Appalachian Plateau-as well as to utilize the rich alluvial valley soil for agriculture (Larson, 1971). Etowah is best known for the elaborate ceremonial paraphernalia and design motifs associated with burials in Mound C and ascribed to the “Southeastern Ceremonial Complex,” or “Southern Cult” (Waring and Holder, 1945).These Wilbanksphase interments, numbering 171 individuals, date to the thirteenth and fourteenth centuries m (Larson, 1971). In a village area at Etowah, another segment of the population is interred. These Lamar-phase burials, numbering 76 individuals, date principally to the fifteenth and sixteenth centuries m (Kelly and Larson, 1957). In sharp contrast to the Mound C interments, those in the village generally contained fewer, less exotic, and more utilitarian grave accoutrements. Because the two burial grounds were not contemporaneous, they do not represent different ranks within the same social system; but Larson (1971) is convinced on the basis of grave goods and demographic composition that Mound C constituted the superordinate dimension of its society. The social ranks of the village interments remain unclear. Thirty-five kilometers west of the Etowah site, the Etowah and Oostanaula Rivers join to form the Coosa River. On the alluvial flood plain of the Coosa, some 50 km west of Etowah, the King site is located. As at Etowah, residents at the King site exploited the diverse biotic communities of the Piedmont, Great Valley, and Appalachian Plateau. Hally (1988) notes that the approximately 3,200 km2 of fertile bottomland surrounding the site forms one of the largest tracts of alluvial land in the northwestern part ofthe state. Cross-dating shows that the site dates to the sixteenth centurym (Smith, 1987). Because the stockade received only limited repair and there is little crowding of architectural features, the site seems to have been occupied briefly-probably for less than 50 years (Hally, 1988). One hundred eighty-nine skeletons were recovered. Artifactually defined low- and high-ranking individuals exhibit no appreciable differences in trace element concentrations (Brown and Blakely, 1985). For the trace element analyses, individuals under 14 years of age were excluded, since strontium and calcium levels fluctuate dramatically during growth and development (Sillen and Kavanagh, 1982; Price et al., 1986). Specimens with little intact cortical bone were also excluded, reducing the samples to 26 individuals (13 females, 13 males) from Etowah Mound C, 25 individuals (12 females, 13 males) from the Etowah village, and 43 individuals (19 females, 24 males) from the King site. The Smirnov test, a two-sample, nonparametric test of ordinally scaled variables, indicated that only the age structure of Mound C males differed significantly from the others, having an overrepresentation of older individuals. To compare strontium and calcium levels between reproductively active and reproductively inactive females, the samples were arbitrarily divided into subsamples consisting of individuals under and over age 40 years.3 Diagenesis Diagenesis is one of the most nettlesome problems facing those who use trace element analysis as a paleodietary technique (Price et al., 1985; Price, 1988). Lambert et al. (1985:481) argue that “under any circumstance, the potential effects of diagenesis are so pervasive that no study of ancient diet based on inorganic composition [of bone] should be carried out without proper diagenetic controls.” A growing body of literature offers contradictory information on the extent to which strontium and calcium are subject to diagenetic change. Discrepancies arise in part from different methods of analysis (Lee, 19831,different bones-or parts of bones-tested (Tanaka et al., 1981; Lambert et al., 1982; Price et al., 1985),and disparate geochemical environments from which samples are extracted (Sillen and Kavanagh, 1982; Price et al., 1985). Vlasak (1982), using microprobe analysis of cross sections of prehistoric femora, observed homogeneous distributions of strontium throughout the bone sections and concluded that no postmortem enrichment or depletion had taken place. Calcium was more mobile, leaching out of bone. In a com3Age of menopause varies interindividually and between populations. Moreover, menopause and “age of mother at birth of last child”need not be the same; fecund women may choose not to be fertile (see Bongaarts and Potter, 19831,and infertile females can be of any age. In the absence of more precise demographic data from prehistoric populations, adult females under age 40 years are herein considered the group most likely to be fertile. STRONTIUM IN FEMALES IN PREHISTORIC AMERICA parative analysis of 10 elements in modern and prehistoric bone from the U.S. Midwest, Lambert et al. (1979) found strontium and zinc least sensitive to diagenesis; postdepositional depletion of calcium did occur. In a follow-up study, Lambert et al. (1982) assayed concentrations of strontium, calcium, and other elements in the porous, trabecular bone of ribs and the dense, cortical bone of femora from archaeological sites in Illinois. The results showed similar levels of strontium in ribs and femora, whereas calcium loss was greater in ribs than in femora (for a critical discussion, see Price et al., 1985427). Other investigators report diagenetic effects on strontium as well as on calcium. Using inductively coupled plasma atomic emission spectroscopy to assay strontium, calcium, and other elements in well preserved Sicilian skeletons, Klepinger and colleagues (1986) observed that calcium exhibited a unidirectional increase with time; the behavior of strontium was more erratic. Nelson et al. (1986) employed X-ray diffraction and infrared spectroscopy to evaluate bone mineral diagenesis in prehistoric and modern marine and terrestrial mammals and noted contamination by strontium in the prehistoric samples. Byrne and Parris (1987) measured bone strontium and calcium at the Middle Woodland Abbott Farm site in New Jersey and detected strontium enrichment by a factor of three to four and calcium depletion by a factor of 1.1-1.6. Although the data are ambiguous, the message is clear: To use strontium and calcium to infer diet or physiological processes from archaeological samples, one must control for diagenesis by using X-ray diffraction analysis, infrared spectroscopy,animal bone standards, or other appropriate means. In the present study, diagenetic differences within and between samples from the Etowah and King sites were negligible for the following reasons. 1. The geochemical environments at the locales, even during seasonal flushings, are virtually identical (S.E. Rose, personal communication); the sites are situated in the alluvial valley of the same river drainage (EtowaWCoosa),less than 50 km apart. This is consistent with the observation of Skougstadt and Horr (1966) that strontium and calcium levels in soils and water tend to be uniform in moderately sized drainage systems. 177 2. At both sites, most inhumations were primary interment^.^ 3. At both sites, all remains were above the water table. 4.At the King site, burial took place within a 50 year period. 5. At Etowah, borrow pits used to construct the mound were located in the flood plain adjacent to the village burials. 6. The pH a t both locales today ranges between 5.2 and 6.0. Price et al. (1985) point out that, as bone decays, its constituent elements gradually move into the surrounding soil matrix, and other elements permeate into the bone; the higher the pH, the less bone deterioration will occur (Gordon and Buikstra, 1981). The high pH levels at the Etowah and King sites indicate that the soil was not so acidic that it dissolved bone and thus released elements to be carried off by ground water. However, the occasional use of lime by modern farmers in the region may have slightly elevated the pH from its aboriginal level. 7. The recent use of superphosphate fertilizers, which are high in strontium (Schroeder et al., 19721, may have elevated strontium in bone recovered from cultivated fields-i.e., the Etowah village and King site-but the effect should be consistent across sexes and age groups within the samples. To reduce further the potential effects of diagenesis, the following procedures were adopted for this study. 1. All bone samples were taken from the diaphyses of well preserved femora. 2. Direct comparisons of Sr/Ca ratios and of strontium levels are made only within population samples from a single site. 3. Between-samples observations are based on relative rather than absolute element concentrations. ‘Seven individuals in the Etowah Mound C samplefrom but two burials (numbers1 and 1 5 t h a d unusually low levels of zinc. Depressed concentrationsof zinc have been associatedwith small stature (Prasad, 19661, endocrine disorders (Miller et al., 19681, lowered disease resistance (Weinberg, 19721, retarded bone development (Haumontand McLean, 19661, delayed wound healing (Pones and Strain, 1966). and elevated copper (Guggenheim and Gaster, 1973). (Because zinc and copper are antagonistic, high levels of copper ingestion inhibit the absorption of zinc.) No such patterns were noted in the skeletons from burials 1 and 15. The only discernible link, which also distinguishes them from other mound interments, is open-airexposure prior to inhumation (see Larson, 1971:65;Blakely and Beck, 1981:427). If open-air exposure does reduce zinc, the reason eludes this author. 178 R.L. BLAKELY Although the controls indicate that diagenesis was similar in the groups tested, they do not measure the extent to which diagenetic change occurred. Sillen (1988:57-58) has shown that variability in Sr/Ca ratios decreases with diagenesis. Coefficients of variability (CVs) for modern browsing and grazing animals generally range from 30%to 35%.For modern humans, CVs are reported between 32% and 40%. Archaeological human remains have yielded CVs ranging from 25-30% for a 400-year-old Dutch sample to 9.6% for a 20,000-year-old Aurignacian sample. At the King and Etowah sites, CVs of 2 6 2 6 % for Sr/Ca ratios and 23-27% for strontium levels suggest that diagenesis was not a serious problem. Atomic absorption spectroscopy With one exception, elements tested included strontium, calcium, copper, magnesium, and zinc. Calcium data are not available for Etowah. The analysis was performed by atomic absorption spectroscopy, using an Instrumentation Laboratory atomic absorption spectrophotometer (IL 250) fitted with a graphite furnace and an air-acetylene flame system. Standard wavelengths and useful ranges for determination are listed in Thompson and Reynolds (1978:33-34). For strontium, atomic absorption spectroscopy offers fewer chemical interferences than atomic emission spectroscopy, and the linear range is comparable (Lee, 1983). To reduce interferences, lanthanum chloride was added to the solution. The IL 250 atomic absorption spectrophotometer has detection limits of 10-20 parts per billion (ppb) for strontium, 3-6 ppb for calcium, 5-10 ppb for copper, 1 ppb for magnesium, and 1-3 ppb for zinc (Thompson and Reynolds, 1978). Ten grams of cortical bone were removed from the midshafts of well preserved femora using a stainless steel bone saw. Sample preparation entailed first cleaning the bone fragments with distilled water. After drying, the specimens were crushed into a fine powder with a porcelain mortar and pestle. One gram samples were ashed at 600°C for 3.5 hr. The bone ash was digested with nitric acid; the solution then was heated to dryness, resolubilized with concentrated nitric acid, and filtered to remove the silica residue. The filtered samples were diluted t o 50 ml with deionized water. This procedure, as noted by Szpunar and coworkers (1978), has the advantage of complete dissolution of bone, yielding reproducibly higher results than those obtained by digestion of bone only. Each specimen was tested twice. Average difference between replicates for strontium was 6.3%; average difference between replicates for calcium was 6.4%. Values for each individual were calculated by averaging the two readings. Following Beck (1985), Price et al. (19861, Sillen (1988), and others, a difference-ofmeans test was chosen to compare the concentrations of elements between population samples. The 0.05 level of confidence was used to reject the null hypothesis that no significant differences existed between samples. Concordant distributions of strontium between females from the Etowah village and Mound C permitted those samples to be pooled for subsequent tests. The same was true for males from the two burial grounds, and they too were combined. RESULTS AND DISCUSSION Figure 2 depicts the distributions of strontium for males and females at the Etowah site. The difference is statistically significant. When plotted against age, as shown in Figure 3, strontium among males is randomly distributed. Females, on the other hand, exhibit a clear pattern: The broken line, a best-fit curve by category means (Zeller and Carmines, 19781, shows that strontium peaks between ages 20 and 40, before declining precipitously in the postmenopausal years. Mean strontium level of reproductive-age females exceeds that of older females by almost 25%. Table 1 indicates that the difference in bone strontium between women under and over age 40 years is statistically significant beyond the 0.005 level of confidence. Strontium exhibits less variability in females under age 40 years (CV 12.8%) than in any other group, possibly because of the functional importance of the absorption of alkaline earths during pregnancy and lactation. For men, age is not a predictor of strontium content. It is not surprising that the difference between men and women under age 40 years is significant; between men and women over age 40 years, the difference is not significant. In Figure 4, Sr/Ca ratios at the King site are plotted against age. Males exhibit little patterning, except for a wide dispersion in the forties and fifties largely resulting from variation in concentrations of strontium. The coefficient of variability for strontium is STRONTIUM IN FEMALES IN PREHISTORIC AMERICA 179 40 200 300 500 400 600 91 D 80 0 700 Strontium (ppml Fig. 2. Distribution of strontium by sex among postadolescents from the Etowah site; t -1.631,49 df, P % 0.05. ETOWAH MALES M = 26) 90 . 90 ETOWAH FEMALES ! - 30Cb 2 0 10 20 0 30 40 I 0 1 50 60 1 70 80 I: : 200 10 R( = 25) . .. * 804 , 20 . , , 30 4 0 , o: 50 = 60 ~ 70 80 Age in years Age In years Fig. 3. Strontium levels by age among postadolescent males and females from the Etowah site. Broken line is a best-fit curve. TABLE 1. Mean strontium of postadolescent females and males from the Etowah site (Statistical comparisons are between groups sharing the same superscript letter) - Females under age 40 years".b Males under age 40 yearsb.c Females over age 40 Males over age 40 yearsczd 'P < 0.005. t = 2.981, 23 df. 'P < 0.05. t = 1.942, 22 df. EP> 0.10. t = 0.038, 24 df. d P> 0.10. t = -0.304, 25 df. N X Sr (ng/mg) SD cv 15 9 10 17 685.4 569.3 550.2 566.9 87.69 155.39 127.55 139.19 0.128 0.273 0.232 0.246 180 R.L. BIAKELY . . 2.25, 2.00- 2.252.001.75. 1.759 8 $5 1.50- . - 1.25- g o 9 9 9 m 1.50- &5 ;; 1.25 vs 9 9 1.00- - 9 ,/' --=)' I x: J- 9- 9 - 1.00- 9 - 0 9 0.75-5.0 . .y - . 0.75 -45.0 10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80 Age in years Age In years Fig. 4. Sr/Ca ratios by age among postadolescent males and females from the King site. Broken line is a best-fit curve. TABLE 2. Mean strontium, calcium, and Sr/Ca ratios o f postadolescent males and females f r o m the King site X Sr X Ca X Sr/Ca N (ng/mg) SD CV (pg/mg) SD CV (X1,OOO) SD cv Males Females 24 19 413.5 464.7 116.90 70.61 0.283 0.152 326.3 318.5 19.47 18.98 0.060 0.060 1.27a 1.46a 0.334 0.166 0.263 0.242 aDifference of means: t = -2.000, 41 df, P < 0.05 over 28%; for calcium, it is 6%-within the limits of error indicated by the replicability data. Among females, Sr/Ca ratios peak in the twenties and thirties, then decline in the postmenopausal years. The difference in Sr/ Ca ratios between pre- and post-menopausal females barely failed to achieve significance because the subsample of those over age 40 years included two individuals with unusually high Sr/Ca ratios. This may be an artifact of incorrect age assessment, strontium contamination of the remains, or endogenous differences in strontiumfcalcium metabolism. Nevertheless, the mean Sr/Ca ratio offemales was over 14%greater than that of males. Table 2 shows that the difference in SrlCa ratios between the sexes at the King site is statistically significant and reflects both higher levels of strontium and lower levels of calcium among women. Pregnancy and lactation Investigators working with living populations have generally found that among adults strontium and Sr/Ca ratios either remain constant or increase modestly with age (Lengemann, 1963; Tipton et al., 1969; Schroeder et al., 1972; Tanaka et al., 1981). This finding is at odds with the data €or females from the Etowah and King sites. The discrepancy is resolved by assuming that many women at the archaeological sites were either pregnant or lactating throughout most of their reproductive years. This would result in discrimination against strontium in favor of calcium in the transport of ions to the placenta and mammary glands, decreasing the amount of calcium available to maternal bone. Normal bone demineralization during lactation would lower further the level of calcium in the maternal skeleton. Enhanced intestinal absorption of alkaline earths would increase the amount of strontium available to maternal bone, and, given the ethnographic evidence for the likelihood of taboos against eating meat a t times during the reproductive cycle (Swanton, 19221, I suspect that reproductive-age women from 181 STRONTIUM IN FEMALES IN PREHISTORIC AMERICA ETOWAH M O U N D C - -m 90 0 80- KING ETOWAH VILLAGE Females Males Females Males ( N = 9) ( N = 11) (N = 11) [ N = 12) I Males ( N = 18) 3 .0 .-C 70- L. 60- - 50- .-0’ 40- 5 0 v) I m 0 30- I t 20. a n 1 0- t ” f Fig. 5. Percent carious teeth per individual between the sexes at Etowah Mound C, Etowah village, and King site. Range equals 1SD on either side of the mean. the Georgia sites periodically supplemented their diet by consuming large quantities of food rich in alkaline earths. Diet, diagenesis, and differences in fertility schedules, rather than differences in reproductive physiology or age alone, probably account for most of the variation in strontium observed in reproductive-age females from archaeological sites.5 Diet Among food resources said to be highest in strontium are fish (Bisel, 1980); freshwater molluscs (Schoeninger and Peebles, 1981); nuts; certain medicinally important herbs; leguminous plants, such as beans; and cereals, including corn (Alexander et al., 1956; Comar and Wasserman, 1964; Gilbert, 1977, 1985). Fish probably can be ruled out as a source of dietary strontium, since the edible 5Lambert and coworkers (1982) found bones with thinner cortices more subject to diagenesis than bones with thicker cortices. Since females generally have thinner cortices than males, diagenetic enrichment by strontium and leaching of calcium could partially account for the elevated Sr/Ca ratios of females, but it does not explain why premenopausal females have higher Sr/Ca ratios than postmenopausal females. Hormonal changes with age, however, cannot be ruled out entirely. At menopause, estrogen levels decline, leading to more frequent and more severe cases of osteoporosisamong other things (Perzigian, 1973). At the King site, cortical thickness and other measures of bone rohusticity do decrease with age but not significantly so. Although far from conclusive, these indices of bone maintenance do not suggest that estrogen-based changes contributed appreciably to age-related differences in Sr/Ca ratios. portions store little strontium (Boroughs et al., 1956). Molluscs also were an unlikely source because the EtowaWCoosa drainage was ecologicallyunsuited for shellfish (Wauchope, 1966). Beans and medicinal plants were consumed but probably not in quantities sufficient to explain elevated strontium in reproductive-age women a t Etowah and King (Swanton, 1922). Corn, as in so many communities throughout the Southeast in the Mississippian and Historic periods, formed a dietary staple at both sites-more so at the Etowah site than at the King site (Kestle, 1988). Corn adversely affects oral health, because the sucrose in corn, like sugar in candy today, leads to decay (Mormann and Muhlemann, 1981), and the “sticky” consistency of porridges made from corn thwarts attempts to clean the teeth (Cook, 1984; Powell, 1985). Since the incidence of caries covaries with the amount of corn consumed, caries can be used to estimate the relative proportion of corn in diets (Rose et al., 198416 ‘Many studies have asserted a positive correlation between maize and dental caries (see, for example, Cassidy, 1972; Cohen and Armelagos, 1984; Rose et al.,19841, but corn was not solely responsible for the prevalence of caries in aboriginal farming groups. Other cariogenic agents include genetic susceptibility; a wide variety of foods and techniques of food preparation; and levels of such elements as calcium, strontium, fluorine, selenium, and molybdenum in the soil and water (Schroeder et al., 1972). 182 R.L. BLAKELY As a crude measure of the percent carious consider reproductive processes in paleoditeeth per individual in the Georgia samples, etary research employing strontium. The present study suggests that pregcarious teeth plus teeth lost antemortem (most of which were surely carious) were nancy and lactation as well as diet contribdivided by the number of teeth present plus ute substantially to SrfCa ratios and stronthose lost antemortem. The results are pre- tium levels in prehistoric samples. It may sented in Figure 5. The high incidences of eventually prove possible to use Sr/Ca ratios caries among females from the King site and to estimate rates of fertility and age of weanEtowah villageboth statistically different ing, as well as dietary preferences attendant from their male counterparts-suggest that to pregnancy and lactation, in ancient popucorn, although sometimes mineral-poor lations (Sillen and Kavanagh, 1982; Sillen (Buikstra et al., 19811,may have contributed and Smith, 1984; Price et al., 1986). Another caveat deserves mention: Stronto bone strontium in those groups. Corn did not significantly elevate strontium levels in tium and calcium metabolically interact Mound C females a t Etowah, and it may not with other elements used in paleodietary have been the principal source of strontium research. For example, linked to diet as well as to strontium and calcium are barium (Linin any of the populations. I believe that nuts provided much of the ieki, 19711, cadmium (Jones and Fowler, dietary strontium. Ethnobotanical and eth- 19801, copper (Guggenheim and Gaster, nohistoric evidence indicates that nuts made 19731, magnesium (Solomon and Styner, up a substantial part of the diet throughout 1969), and zinc (Guggenheim and Gaster, the region (Hudson, 1976; Hatch and Geidel, 1973). To use these elements to infer paleo1983; Beck, 1985). Nuts are high in magne- diets, their relationship to strontium and sium, which among other things aids in the calcium must be firmly established and fitabsorption of other alkaline earth metals ted to information on Sr/Ca ratios in repro(Solomons and Styner, 1969). This may ex- duction. Because one seldom knows beforeplain why females at both archaeological hand what diets may be encountered in sites exhibit elevated levels of m a g n e ~ i u m , ~paleoanthropological research, this argues but nuts were probably a preferred food for the use of an array of elements, including among pregnant and lactating women be- but not limited to the alkaline earth metals. Paleodietary reconstruction based solely cause they are so rich in proteins and carboon trace elements-without macroscopic or hydrates. microscopic corroboration from bone, teeth, CONCLUSIONS hair, or mummified tissue-remains a risky The results of this study suggest that an- business, and intersite comparisons are thropologists have often overlooked a com- fraught with difficulties. Only after we conmon cause of variation in Sr/Ca ratios. Even trol for diagenesis, age changes, element in the absence of a concomitant dietary interactions, and gender-related physiologichange, pregnancy and lactation elevate ma- cal differences can we confidently associate ternal bone strontium and depress maternal trace elements with diet.8 The hypothesis tested h e r e t h a t pregbone calcium. In an experiment using rats fed a low-strontium diet, Price and col- nancy and lactation elevate SrfCa ratios in leagues (1986) found that pregnancy and archaeologically recovered human bone lactation account for roughly a 6%increase -should be tested at other sites. The results in strontium, a 2%decrease in calcium, and a of this study should be replicable, since we 9%increase in the Sr/Ca ratio. These figures are not applicable to humans because of differences in the rates of bone mineral turn&rhe same problem confronts those using trace elements to over between rats and humans, nor do the diagnose disease (Brown et al., 1979; Price et al., 1985). It is axiomatic that, if we are to recognize the "abnormal,"we must authors address the effects of repeated preg- first be able to recognize the normal. Strontium and calciumcan nancy and lactation events on Sr/Ca ratios. be used as diagnostic or corroborative tools in paleopathological Although aboriginal populations seldom lived under However, their work provides the clearest research. chronic conditionsof strontium or calcium deprivation,depressed experimental evidence to date of the need to levels have been observed (Schroederet al., 1972). Hypocalcemia 'Mean concentrations of magnesium at the Etowah site are 1,019.6 ppm for females and 937.6 ppm for males. At the King site, mean concentrations of magnesium are 1,064.2 ppm for females and 1,036.0ppm for males. and hypercalcemiacan be indicative of osteoporosisand tuberculosis, respectively (Shils and Young, 1988). Strontium and calcium are dependent on vitamin D for proper absorption through the intestinal wall, implicating both elements in osteomalacia and rickets (Wasserman, 1963). Strontium toxicity is also known but appears to have been uncommon in preindustrialpopulations (Underwood,1977). STRONTIUM IN FEMALES IN PREHISTORIC AMERICA are dealing with reproductive processes common to all human groups. If such studies can delimit more precisely the contribution of pregnancy and lactation to bone Sr/Ca ratios, then paleodietary research will be advanced. ACKNOWLEDGMENTS A preliminary version of this paper was presented at the fifteenth annual meeting of the Paleopathology Association in conjunction with the fifty-seventh annual meeting of the American Association of Physical Anthropologists, March 23, 1988, in Kansas City, Missouri. I am grateful to Lane Beck and Antoinette Brown, who collaborated with me on earIier trace element anaIyses of the Etowah and King samples, respectively. H. James Bronaugh, Department of Chemistry, Georgia State University, kindly carried out the atomic absorption spectroscopy. I am indebted to David Hally for loan of the King site skeletons and to Lewis Larson for loan of the Etowah skeletons. The thoughtful comments of Lane Beck, Ray Crook, Bettina Detweiler-Blakely, and anonymous reviewers for the Journal materially improved this work. This research was supported in part by National Science Foundation grants BNS8217377 and SOC-7503833 and Georgia State University grant 8504702. LITERATURE CITED Alexander GV, Nusbaum RE, and MacDonald NS (1956) The relative retention of strontium and calcium in bone tissue. J . Biol. Chem. 218:911-919. Atkinson PJ, and West RR (1970) Loss of skeletal calcium in lactating women. J . Obstet. Gynaecol. Br. Comnwlth. 77.555-560. Beck LA (1985) Trace elements and dietary variation during the Mississippian period in North Georgia. Paper presented at the Forty-SecondAnnual Meeting, Southeastern Archaeological Conference, Birmingham, Alabama, November &-9,1985. Bisel SLC (1980) A pilot study in aspects of human nutrition in the ancient eastern Mediterranean, with particular attention to trace minerals in several populations from different time periods. PhD dissertation, University of Minnesota. Blakely RL, and Beck LA (1981) Trace elements, nutritional status, and social stratification at Etowah, Georgia. In A-M Cantwell, JB Griffin, and NA Rothschild (eds):The Research Potential of Anthropological Museum Collections. Ann. N.Y. Acad. Sci. 376: 417431. Blanusa M, Harmut M, and Kostial K (1970) Comparative strontium and calcium metabolism in lactation. Arh. Hig. Rada Toksikol.21t125-127. Bongaarts J, and Potter RG (1983) Fertility, Biology, and Behavior. New York: Academic Press. Boroughs H, Townsley SJ, and Hiatt RW (1956) The metabolism of radionuclides in marine organisms. 111. 183 The uptake of calcium45in solution by marine fish. Limnol. Ocean. 1:28-32. Bronner F, and Lemaire R (1969) Comparisonof calcium kinetics in man and the rat. Calcif. Tissue Res. 3:23a248. Brown Al3 (1973) Bone strontium content as a dietary indicator in human skeletal populations. PhD dissertation, University of Michigan. Brown AB, and Blakely RL (1985) Biocultural adaptation as reflected in trace element distribution. J. Hum. Evol. 14:461468. Brown SS, Mitchell FL, and Young DS (eds) (1979) Chemical Diagnosis of Disease. New York: Elsevier/ North Holland. Buikstra JE, Lambert JB, and van der Menve NJ (1981) Elemental variability in paleopopulationstudy. Paper presented at the Forty-Sixth Annual Meeting, Society for American Archaeology, San Diego, April, 1981. Byme KB, and Parris DC (1987) Reconstruction of the diet ofthe Middle Woodland Amerindian population a t Abbott Farm by bone trace-element analysis. Am. J. Phys. Anthropol. 74:373-384. Cassidy CM (1972) A comparison of nutrition and health in pre-agricultural and agricultural Amerindian skeletal populations. PhD dissertation, University of Wisconsin. Cohen MN, and Armelagos GJ (eds) (1984) Paleopathology at the Origins ofAgriculture. New York: Academic Press. Comar CL (1963) Some over-all aspects of strontiumcalcium discrimination. In RH Wasserman (ed): The Transfer of Calcium and Strontium Across Biological Membranes. New York: AcademicPress, pp. 405-418. Comar CL, and Wasserman RH (1964) Strontium. In CL Comar and F Bronner (eds): Mineral Metabolism, Volume 2. New York: Academic Press, pp. 523-572. Cook DC (1984) Subsistence and health in the Lower Illinois Valley: osteological evidence. In MN Cohen and GJ Armelagos (eds): Paleopathology at the Origins of Agriculture. New York: Academic Press, pp. 235-269. Elias RW, Hirao Y, and Patterson CC (1982) The circumvention of the natural biopurification of calcium along nutrient pathways by atmospheric inputs of industrial lead. Geochim. Cosmochim. Acta 46:2561-2580. Fournier P, Dupuis Y, and Fournier A (1971) Effect of lactose on the absorption of alkaline earth metals and intestinal lactose activity. Israel J. Med. Sci. 7:389-391. Gilbert RI (1975) Trace element analysis of three skeletal Amerindian populations a t Dickson Mounds. PhD dissertation, University of Massachusetts. Gilbert RI (1977) Applications of trace element research to problems in archeology.In RL Blakely (ed): Biocultural Adaptation in Prehistoric America. Southern Anthropological Society Proceedings, No. 11. Athens: University of Georgia Press, pp. 85-100. Gilbert RI (1985) Stress, paleonutrition, and trace elements. In RI Gilbert and JH Mielke (eds):The Analysis of Prehistoric Diets. New York: Academic Press, pp. 339-358. Gordon CC, and Buikstra JE (1981) Soil pH, bone preservation, and sampling bias at mortuary sites. Am. Antiquity 46:566-571. Guggenheim K, and Gaster D (1973) The role of manganese, copper and zinc in the physiology of bones and teeth. In I Zipkin (ed):Biological Mineralization. New York: John Wiley and Sons, pp. 443-462. Hally DJ (1988) Archaeology and settlement plan of the 184 R.L. BLAKELY King site. In RL Blakely (ed): The King Site: Continuity and Contact in Sixteenth-Century Georgia. Athens: University of Georgia Press, pp. 3-16. Hartsook EW, and Hershberger TV (1973) Strontiumcalcium discrimination during placental transfer and fetal uptake in rats: Effects ofgestation duration. Proc. SOC.Exp. Biol. Med. 143:343-349. Hatch JW, and Geidel RA (1983)Tracing status and diet in prehistoric Tennessee. Archaeology 36.56-59. Haumont Z, and McLean FC (1966) Zinc and the physiology of bone. In AS Prasad (ed): Zinc Metabolism. Springfield, IL: Charles C. Thomas, pp. 169-186. Hudson C (1976) The Southeastern Indians. Knoxville: University of Tennessee Press. Hutchin ME, and Vaughan BE (1968) Relation between simultaneous Ca and Sr transport rates in isolated segments of vetch, barley and pine roots. Plant Physiol. 43:1913-1918. Ingersoll RJ, and Wasserman RH (1971) Vitamin D,induced calcium-binding protein: Protein characteristics, confirmational effects, and other properties. J . Biol. Chem. 246:280%2814. Jones HS, and Fowler BA (1980) Biological interactions of cadmium with calcium. In OA Levander and L Chene (eds): Micronutrient Interactions: Vitamins. Minerals, and Hazardous Elements. Ann. N.Y. Acad: Sci. 355:309-317. Katzenberg MA (1984) Chemical Analysis of Prehistoric Human Bone From Five Temporally Distinct Populations in Southern Ontario. National Museum of Man, Mercury Series. Archaeological Survey of Canada, Vol. 29. Kelly AR, and Larson LH (1957) Explorations at Etowah, Georgia, 1954-1956. Archaeology 10:3%48. Kestle S (1988) Subsistence and sex roles. In RL Blakely (ed): The King Site: Continuity and Contact in Sixteenth-Century Georgia. Athens: University of Georgia Press, pp. 63-72. Klepinger LL, Kuhn JK, and Williams WS (1986) An elemental analysis of archaeological bone from Sicily as a test of predictability of diagenetic change. Am. J. Phys. Anthropol. 70:325-331. Kostial K, Gruden N, and Durakovic A (1969) Intestinal absorption of calcium-47 and strontium-85 in lactating rats. Calcif. Tissue Res. 4:13-19. Lambert JB, Simpson SV, Szpunar CB, and Buikstra J E (1984) Copper and barium as dietary discriminants: The effects of diagenesis. Archaeometry26:131-138. Lambert JB, Simpson SV, Szpunar CB, and Buikstra J E (1985) Bone diagenesis and dietary analysis. J . Hum. Evol. 14:477-482. Lambert JB, Szpunar CB, and Buikstra J E (1979) Chemical analysis of excavated human bone from Middle and Late Woodland sites. Archaeometry 21:115-129. Lambert JB, Vlasak SM, Thometz AC, and Buikstra J E (1982)A comparative study of the chemical analysis of ribs and femurs in Woodland populations. Am. J . Phys. Anthropol. 59:28%294. Larson LH (1971) Archaeological implications of social stratification at the Etowah site, Georgia. In JABrown (ed):Approaches to the Social Dimensions of Mortuary Practices. Memoirs of the Society for American Archaeology No. 25. Washington, DC: Society for American Archaeology, pp. 5 H 7 . Lee J (1983) Calcium matrix effects in multi-element analysis of animal bone by ICP emission spectrometry. Anal. Chim. Acta 152141-147. Lengemann F W (1963) Over-all aspects of calcium and strontium absorption. In RH Wasserman (ed): The Transfer of Calcium and Strontium Across Biological Membranes. New York: Academic Press, pp. 85-96. Lengemann FW, Wasserman RH, and Comar CL (1959) Studies on the enhancement of radiocalcium and radiostrontium absorption by lactose in the rat. J. Nutr. 68:443-456. Lenihan JMA,Loutit JF, and Martin J H (eds) (1967) Strontium Metabolism. New York Academic Press. Linieki J (1971) Kinetics of calcium, strontium, barium and radium in rabbits. Health Phys. 21:367-376. Lough SA, Rivera J, and Comar CL (1963) Retention of strontium, calcium and phosphorus in human infants. Proc. SOC.Exp. Biol. Med. 1123331-636. Marcus CS, and Lengemann FW (1962) Absorption of Ca45 and SrS5from solid and liquid food at various levels of the alimentary tract of the rat. J. Nutr. 66: 155-160. Menczel J , and Mor E (1972) The effect of thyrocalcitonin on calcium binding protein and on strontium binding protein. In Proceedings of the 2nd International Symposium on Strontium Metabolism, Glasgow, Scotland. Miller ER, Luecke RW, Ullrey DE, Baltzer BV, Bradley BL, and Hoefer JA (1968) Biochemical skeletal and allometric changes due to zinc deficiency in the baby pig. J. Nutr. 95:278-286. Mormann JE, and Muhlemann HR (1981) Oral starch degradation and its influence on acid production in human dental plaque. Caries Res. 15:16&175. Nelson BK, DeNiro MJ, Schoeninger MJ, De Paolo DJ, and Hare PE (1986)Effects of diagenesis on strontium, carbon, nitrogen and oxygen concentration and isotopic composition of bone. Geochim. Cosmochim. Acta 50:1941-1949. Pate D, and Brown KA (1985) The stability of bone strontium in the geochemical environment. J. Hum. Evol. 14:483-491. Perzigian AJ (1973) Osteoporotic bone loss in two prehistoric Indian populations. Am. J . Phys. Anthropol. 39:87-96. Pories WJ, and Strain WH (1966) Zinc and wound healing. In AS Prasad (ed): Zinc Metabolism. Springfield, I L Charles C. Thomas, pp. 37g394. Powell ML (1985)The analysis of dental wear and caries for dietary reconstruction. In RI Gilbert and J H Mielke (eds): The Analysis of Prehistoric Diets. New York: Academic Press, pp. 307-338. Prasad AS (ed) (1966) Zinc Metabolism. Springfield, IL: Charles C. Thomas. Price TD (1988) Changes in bone strontium: Diet or diagenesis? In TD Price (ed): Chemistry of Prehistoric Human Bone. Cambridge: Cambridge University Press. Price TD, and Kavanagh M (1982) Bone composition and the reconstruction of diet: Examples from the midwestern United States. Midcontinental J. Archaeol. 7:61-79. Price TD, Schoeninger MJ, and Armelagos GJ (1985) Bone chemistry and past behavior: An overview. J. Hum. Evol. 14:419-447. Price TD, Swick RW, and Chase E P (1986) Bone chemistry and prehistoric diet: Strontium studies of laboratory rats. Am. J. Phys. Anthropol. 70:365-375. Rose JC, Burnett BA, Nassaney MS, Blaeuer M Y (1984) Paleopathology at the origins of maize agriculture in the Lower Mississippi Valley and Caddoan culture areas. In MN Cohen and GJ Armelagos (eds): Paleopathology at the Origins of Agriculture. New York Academic Press, pp. 393424. Rosenthal HL (1981) Content of stable strontium in man STRONTIUM IN FEMALES IN PREHISTORIC AMERICA and animal biota. In C Skoryna (ed): Handbook of Common Strontium. New York: Plenum, pp, 503-514. Sanzharova NI (1978) Change in the ratio of strontium90 to calcium during uptake from soil by herbaceous plants. Mosc. Univ. Soil Sci. Bull. 33:42-43 (in English). Schoeninger MJ (1979) Diet and status at Chalcatzingo: Some empirical and technical aspects of strontium analysis. Am. J . Phys. Anthropol. 51~295-310. Schoeninger MJ (1982) Diet and the evolution of modern human form in the Middle East. Am. J . Phys. Anthropol. 58r37-52. Schoeninger MJ, and Peebles CS (1981) Effect of mollusc eating on human bone strontium levels. J . Archaeol. Sci. 8:391-397. Schroeder HA, Tipton IH, and Nason AF' (1972) Trace metals in man: strontium and barium. J . Chronic Dis. 19.545-571. Shils ME, and Young VR (1988) Modern Nutrition in Health and Disease, 7th ed. Philadelphia: Lea and Febiger. Sillen A (1981) Strontium and diet at Hayonim Cave. Am. J. Phys. Anthropol. 56~131-137. Sillen A (1988) Elemental and isotopic analyses of mammalian fauna from southern Africa and their implications for paleodietary research. Am. J . Phys. Anthropol. 76r49-60. Sillen A, and Kavanagh M (1982) Strontium and paleodietary research: A review. Yrbk. Phys. Anthropol. 25:67-90. Sillen A, and Smith P (1984) Weaning patterns are reflected in strontiudcalcium ratios of juvenile skeletons. J. Archaeol. Sci. 11~237-245. Skougstadt MW, and Horr CA (1966) Occurrence and Distribution of Strontium in Natural Water. United States Geological Survey, Circular 420. Washington, DC: U.S. Government Printing Office. Smith M (1987) Archaeology of Aboriginal Culture Change: Depopulation During the Early Historic Period. Ripley P. Bullen Monographs in Anthropology and History No. 6. Gainesville: University of Florida Press. Solomons CC, and Styner J (1969) Osteogenesis imperfecta: Effect of magnesium administration on pyrophosphate metabolism. Calcif. Tissue Res. 31~31%-326. Spencer H, Warren JM, Kramer L, and Samachson J (1973) Passage of calcium and strontium across the intestine in man. Clin. Orthop. Rel. Res. 91 :225-234. Swanton JR (1922) Early History of the Creek Indians and Their Neighbors. Bureau of American Ethnology, Bul. 73. Washington, DC: U.S. Government Printing Office. Szpunar CB, Lambert JB, and Buikstra J E (1978)Anal- 185 ysis of excavated bone by atomic absorption. Am. J . Phys. Anthropol. 48:199-202. Tanaka G-I, Kawamura H, and Nomura E (1981) Reference Japanese man-11: Distribution of strontium in the skeleton and in the mass of mineralized bone. Health Phys. 40:601-614. Thompson KC, and Reynolds RJ (1978) Atomic Absorption, Fluorescence and Flame Emission Spectroscopy, 2nd ed. New York: John Wiley and Sons. Thompson RC (1963) Factors and conditions modifying the absorption and retention of chronically ingested radiostrontium. In RH Wasserman (ed): The Transfer of Calcium and Strontium Across Biological Membranes. New York: Academic Press, pp. 393-404. Tipton IH, Stewart PL, and Dickson J (1969) Patterns of elemental excretion in long-term balance studies. Health Phys. 16r455-462. Toots H, and Voorhies MR (1965) Strontium in fossil bones and the reconstruction of food chains. Science 149:854-855. Turekian KK, and Kulp J L (1956) Strontium content of human bones. Science 124r405-407. Twardock AR (1963)Studies on the movement ofcalcium and strontium across the bovine mammary gland. In RH Wasserman (ed): The Transfer of Calcium and Strontium Across Biological Membranes. New York: Academic Press, pp. 327-340. Underwood E J (1977) Trace Elements in Human and Animal Nutrition, 4th ed. New York: Academic Press. Vlasak SM (1982) Elemental analysis of excavated human bone: A study of post-mortem deterioration. PhD dissertation, Northwestern University. Waring AJ, andHolder P (1945)Aprehistoric ceremonial complex in the southeastern United States. Am. Anthropol. 47:l-34. Wasserman RH (1963) Vitamin D and the absorption of calcium and strontium in vivo. In RH Wasserman (ed): The Transfer of Calcium and Strontium Across Biological Membranes. New York: Academic Press, pp. 211-228. Wasserman RH, Lengemann FW, and Comar CL (1958) Comparative metabolism of calcium and strontium in lactation. J . Dairy Sci. 41.312-821. Wauchope R (1966) Archaeological Survey of North Georgia. Memoirs of the Society for American Archaeology No. 21. Washington, DC: Society for American Archaeology. Weinberg ED (1972) Infectious diseases influenced by trace element environment. In HC Hopps and HL Cannon (eds): Geochemical Environments in Relation to Health and Disease. Ann. N.Y. Acad. Sci. 199:274-284. Zeller RA, and Carmines EG (1978) Statistical Analysis of Social Data. Chicago: Rand McNally.