bone remodeling rates and skeletal maturation in three archaeoloqical skeletal populations

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 98:161-171(1995)

Bone Remodeling Rates and Skeletal Maturation in Three Archaeological Skeletal Populations

SAM D. STOUT AND RHONDA LUECK Department of Anthropology, University of Missouri, Columbia, Missouri 65211 (S.D.S.); U.S. Army Corps of Engineers, St. Louis District, St. Louis, Missouri 63103-2833 (R.L.)

KEY WORDS Effective age of adult compacta

Cortical bone, Histomorphometry, Remodeling,

ABSTRACT Cortical bone remodeling rates for rib samples from three archaeological populations and a modern autopsy sample were determined using an algorithm developed by Frost (Frost [1987a] Calcif. Tissue Res. 3:211-237). When plotted against the relative antiquities for population samples, histomorphometric variables; i.e., activation frequency (jirc), net bone formation (,,tVf,,,,), and mean annual bone formation rate (Vf,J, exhibit a concordant trend of increased cortical bone remodeling activity levels over time. Two intensive foraging populations, Windover and Gibson, are similar for all bone remodeling parameters and have the lowest remodeling activity levels among the samples. The more recent Ledders sample, which is reported to practice agricultural subsistence, is consistently intermediate between these and a modern autopsy sample. Although there appear to be differences in bone formation rates among the populations, it is concluded that these differences cannot be attributed to differences in bone remodeling rates among the populations, but rather are reflecting different effective ages of adult compacta for their ribs. These findings suggest that the earlier populations, particularly Windsor and Gibson, appear to have reached skeletal maturity at an older age than observed for modern. o 1995 Wiley-Liss, Inc.

The skeleton, because of its mineralized structure, is the most durable organ of the body, and is usually all that remains of deceased individuals. Fortunately, rather than being an inert pillar of minerals func- tioning solely to provide structural support for the other tissues of the body, this dy- namic, metabolically active tissue responds to a variety of physiological, genetic, and en- vironmental factors. The same structure that make osseous tissue enduring also gives it the unique property of providing a durable record of past metabolic events encoded in its microstructure.

A number of methods have been developed to extract some of the biological information encoded in bone histomorphology. Most of these methods, however, have been limited to providing estimates of age at death (Ker-

ley and Ubelaker, 1978; Ahlqvist and Dams- ten, 1969; Thompson, 1979; Ericksen, 1991; Stout and Paine, 1992).

Biomedical research using in vivo labeling of bone has led to a better understanding of the metabolic basis for many of the histomor- phological structures that are observed in skeletal remains of considerable antiquity (Stout, 1978). Using the results of tetracycline labeling research, Wu et al. (1970) proposed a method by which cortical bone remodeling rates for the rib can be determined without requiring in vivo labeling. Stout and

Received November 3, 1994; accepted May 8, 1995. Address reprint requests to Dr. Sam D. Stout, Department of

Anthropology, 200 Swallow Hall, University of Missouri, Colum- bia, MO 65211.

0 1995 WILEY-LISS, INC

162 S.D. STOUT AND R. LUECK

Teitelbaum (1976) and Stout (1983) verified that the method produces reasonable estimates of bone remodeling rates when applied to archaeological skeletal remains. The method depends upon the ability to estimate accumulated osteon creations (AOC) from the total visible evidence for osteons (OPD). OPD is merely the sum of the observed intact and fragmentary osteons per unit area for a bone sample. AOC is OPD plus those osteons that are missing because subsequent modeling and remodeling have removed all evidence for them. AOC must be estimated on the basis of the relationship between observed OPD and its corresponding AOC; this has been established through tetracycline labeling studies (Frost, 1969). The existence of unrecognized adult-life modeling drifts, and the effects of cortical thickness and osteon diameter, however, can introduce error. Frost (1987a) has proposed an algorithm for estimating the missing osteons corresponding to a given OPD that accounts for the effects of cortical thickness and osteon diameter on AOC, and thus improves our ability to estimate bone remodeling rates from static histomorphometry. Stout and Paine (1994) have applied this algorithm to a modern autopsy sample and found that the resulting bone remodeling rates compared well with tetracycline-based values. This paper provides the results of research employing this algorithm to compare rib cortical bone remodeling rates among three archaeological populations differing in mode of subsistence and antiquity, and a modern autopsy sample.

SKELETAL SAMPLES The skeletal samples used in this study

represent both time depth and different modes of subsistence. The oldest sample (N = 38) is from the Early Archaic Windover site from east-central Florida. The stratum in which the skeletal remains were found has been radiocarbon dated between 6,900 and 8,120 years b.p., making this site the largest skeletal sample of its antiquity in North America (Doran and Dickel, 1988). Based on archaeological evidence, the Win- dover population exploited a wide variety of resources, including deer and other upland

game, marsh vertebrates, freshwater fish, mussels, shark, manatee, and shellfish, a s well as a plentiful and diverse supply of plant resources.

Two other archaeological skeletal samples represent Woodland sites from the Lower 11- linois River Valley. The Gibson (N = 25) samples are from a population that dates to the Middle Woodland period, which in this region, has been radiocarbon dated between 50 BC and AD 400 (Buikstra, 1972). The mode of subsistence practiced by Middle Woodland populations in this region has been described as intensive harvest collecting. The re- maining samples are from the Ledders site (N = 181, which has been radiocarbon dated at AD 1000, placing it in the Late Woodland period. These sites, therefore, represent the transition from Middle to Late Woodland Pe- riods and a shift from an intensive harvest collecting economy to that of maize agriculture (Whatley and Asch, 1975). Bio-archaeological evidence suggests that significant changes occurred between Middle and Late Woodland in the Lower Illinois River Valley. Based on a study of growth, disease, and mortality seen in children of weaning age from the Ledders and Gibson sites, Cook (1984) suggests that the population increase and other changes resulting from the shift to agriculture greatly affected the health of these populations.

The modern sample of 45 sixth ribs was originally collected for an earlier study to develop an histological age estimating technique (Stout and Paine, 1992). Five of the individuals from whom the ribs were taken were black; the remainder were White. None had any evidence of metabolic bone disease.

Methods All bone samples were from the middle-

third of the rib. Although the histomorphometric method was developed for the sixth rib, the fragmentary nature of archaeological skeletal remains often made it impossible to determine exact rib number. Ribs 1 and 11-12 were excluded, however.

Undecalcified sections were prepared following routine histological procedures reported elsewhere (Stout and Teitelbaum, 1976; Stout and Paine, 1992).

Histomorphometric analysis was per-

OSTEONAL BONE REMODELING 163

formed with a standard research microscope fitted with integrated eyepiece grid (see Kimmel and Jee, 1983) to allow microscopic field delineation and perform aresl measurements. The following histomorphometric variables were determined for each rib sample. Osteons are defined as those structures that result from the tethered remodeling processes of activation, resorption, and formation within cortical bone.

Mean osteonal cross-sectional area (Ah), the average area of bone contained within the cement lines of structurally complete osteons for each rib specimen.

Mean cross-sectional diameter (Dh) of the intact osteons of a specimen, determined from mean osteonal cross-sectional area using the formula

nent x is 3.5, as suggested by Frost (1987a). The OPD asymptote for a given specimen is estimated by the formula

OPD asymptote = k((Dh)2)-',

where

(Y = OPD(0PD asymptote)-'.

It is based on the relationship between mean osteonal cross-sectional area, and the unit of measurement (1 mm2) represented by the expression (Dh)2. K is a packing factor that accounts for the fact that a unit of area of bone can actually contain more intact osteons and their fragments than predicted by a theoretical orthogonal distribution (Frost, 1987a). I t is estimated using the formula

Zntact osteon density (Pi), the number of osteons per unit area that have 90% of their Haversian canal perimeters intact, i.e., un- remodeled.

Fragmentary osteon density (PJ, the number of osteons per unit area for which 10% or more of the perimeter of their Haversian canal has been remodeled by subsequent generations of osteons.

Osteon population density (OPD), the sum of Pi and Pf.

Accumulated osteon creations (AOC =

Pi + Pf + Pmissing), the total number of osteon creations corresponding to a given OPD, is estimated. Continuous remodeling activity eventually produces an asymptote in OPD when osteonal bone occupies the entire cortex, and each new creation removes evidence for an earlier creation. The contribution of missing osteons to the AOC equation increases exponentially as this asymptote is approached. The algorithm uses a scaling operator p, which when multiplied by an observed OPD, provides an estimate of its corresponding AOC. p is defined by the equation

p = (1 - a x ) - 1 ,

where ci is an OPD normalized to it predicted asymptote as defined below, and the expo-

k = (OPD asymptote) (Dh)'.

For this study, the value for k was determined to be 1.7. Following the suggestion made by Frost (1987a), k is based on a n inde- pendent sample of rib cross sections from older individuals in which all primary lamel- lar bone had been replaced by secondary osteonal bone. The sample was obtained from dissecting room cadavers and had a mean age of 73.6 years and an age range of 60-102 years. The maximum observed OPD for these samples (36.2/mm2) served as the asymptotic value. D; is 0.047 mm and is derived from a mean osteonal cross-sectional area of 0.037 mm2 reported by Wu et al. (1970) for the sixth rib based on a large clini- cal sample. AOCs for individual rib speci- mens can then be estimated by multiplying their observed OPD by p:

AOC = p . OPD.

The net bone formation (netVf,r,t) that OC- curred over the lifetime of an individual can be estimated in terms of mm2/mm2 by multiplying AOC by the mean osteonal area for the specimen

Mean activation frequency (iirc), the mean number of osteons created annually per mmz


Fig. 1. Effects of growth (A) and cortical drift (B) and their combined effects (C) on the rib. It is because of these effects that the actual mean tissue age for the cortex of a given bone is less than its chronological age.

of bone is estimated by dividing AOC by age. Transverse cortical drifts occurring during growth (Fig. 1) remove all evidence for earlier remodeling activity. Therefore, OPD does not represent the osteon creations since birth, but rather since a later age subsequent to the period of active drift. This is referred to as the effective age of the “birth” of adult compacta, and has been estimated for modern humans to be 12.5 years of age for the middle third of the sixth rib (Wu et al., 1970). Therefore,

- prc = AOC +- (chronological age - 12.5 years).

Mean annual bone formation rate (Vf,r,l) in mm21mm21yr is estimated by multiplying the mean activation frequency by the mean osteonal cross-sectional area for a specimen, as described by the expression

Vf,r,t = p r c ’ Ah.

The statistical package SYSTAT (Wilkinson,

1989) was used for statistical analysis. Bone remodeling parameters estimated by this algorithm are based on OPD, which is strongly correlated with chronological age. Although the mean ages for the samples are not statistically different (P = 0.2461, the age distributions for the samples are not normally distrib- uted (Fig. 2). Therefore, after testing for homogeneity of slopes, an analysis of covari- ance (ANCOVA), using age as the covariate, was employed for statistical comparisons among the samples. When the overall AN- COVA was significant, the Tukey-Kramer method of post hoc pairwise comparisons for unequal sample sizes was performed to determine which of the sample means differed significantly from the others.

RESULTS Table 1 summarizes the results of the his-

tomorphometric analysis for the three archaeological samples, and a modern compar- ative sample. The means do not differ significantly between males and females; therefore, the sexes were combined for fur- ther analyses.

All population samples exhibit age-associ- ated increases in OPD, AOC, and net bone remodeling (netVLr,t), while activation rates (iirc), and bone formation rates (V,,,) decrease with age (Fig. 3A-E).

The mean values for OPD range from 18.8/ mm2 for the modern sample of 17.4/mm2 for the Windover sample (Fig. 4A). Differences are statistically significant between modern and Windover (P = 0.013), and Ledders and Windover (P = 0.046). None of the slopes for the regression of OPD against age is significantly different among the sample populations.

Corresponding AOC range from 24.71mm2 for the modern sample to 19.91mm2 for Gib- son (Fig. 4B). The difference is not significant between Ledders and Windover but is statistically significant between modern and Windover (P = 0.014).

Mean osteon cross-sectional area (Ah) ranges from 0.0401mm2 for modern to 0.0331 mm2 for the Ledders sample (Fig. 4C). They differ statistically between modern and Led- ders (P = 0.008), which have the largest and

0.4

0.3

C

m

$ 0 2

z 0 C

g 0.1 2

0.4

U U

0.3 W

z P k 0.2

g 0 C

0.1

OSTEONAL BONE REMODELING

15 0

10

5

10 40 70 1W 130

AGE (yrs.)

Glbson 1 10

1 5 2 5 3 5 4 5 5 5 6 5

AGE (yrs.1

10 25 40 55

AGE (yrs.)

0.3 C

2 C

n

P 0.2

k 2 E 0.1

15 30 45 w 75

AGE (yrs.)

Windover

Fig. 2. Age distributions for the four samples used in this study. Samples include a modern autopsy sample (A) and three archaeological skeletal populations: the Late Woodland Ledders (B), the Middle Woodland Gibson (C) sites from the Lower Illinois River Valley, and the Early Archaic Windover (D) site from Florida.

TABLE 1. Descriutive statistics for samales'

165

3

>

$ 8 3 5 2

1

15

10 : 5

5

Variable Modern Ledders Gibson Windover

Age (yr) 34.9 i 3.09 32.5 f 3.067 39.0 ? 2.76 40.4 f 2.33 OPD (#/mm2) 18.8 i 1.07 18.6 f 1.21 17.8 2 1.22 17.4 i 0.89

(16.2 i 0.71) Ah (mm2) 0.040 5 0,001 0.033 2 0.001 0.036 i- 0.001 0.035 i 0.001

Dh (mm) 0.223 2 0.004 0.205 f 0.003 0.210 2 0.004 0.210 -t 0.003 AOC (#/mm2) 24.7 i 2.30 21.3 i 1.96 19.9 ? 1.50 20.2 I 1.44

(25.3 i 1.47) (22.7 2 2.38) (19.0 i 2.03) (18.6 2 1.63) 2.4 2 0.48 2.0 i 0.57 1.0 ? 0.144 1.1 f 0.13 pvc (#/mm2/yr)

(2.3 ? 0.29) (1.5 i 0.47) (1.1 ? 0.40) (1.2 i 0.32) netVr,,,, (mm2/mm2) 0.937 i 0.091 0.702 i 0.072 0.676 ? 0.046 0.713 f 0.053

(0.950 2 0.0633) (0.753 t 0.1019) (0.653 ? 0.0872) (0.677 f 0.0700) V,,,, (mm2/mm2/yr) 0.102 i 0.023 0.065 f 0.019 0.038 i 0.006 0.039 2 0.005

(0.096 2 0.0133) (0.050 f 0.0214) (0.041 i 0.0183) (0.044 i 0.0146)

(19.2 i 0.64) (19.5 f 1.04) (17.1 ? 0.89)

(0.039 i 0.0010) (0.033 f 0.0016) (0.035 i 0.0013) (0.036 f 0.0010)

-

' Parenthetical numbers are the least-squares age-adjusted means and their standard errors.

166

' A I I I ' 0 I I I

0 .7

ij. E

E E

g 0 . 6

- p 0 . 5

J W n 0

a 3 0 . 4

I- W z

0.3

0 . 2

1 2 3 4 5

AGE (yrs)

Gibson t I 2 3 1 5

AGE (yrs)

0.1298

Ledders

0.0392

0.0090 10 20 30 40 50

AGE(yrs)

2.5

1.0

- r - 2 1..

i; 1.1

E z -

0.7

0.2

1

2 3 4 5

AGE (yrs)

, I I 1 2 3 4 5

AGE (yrs)

Fig. 3. Graphic representation of the association of the histomorphometric variables with chronological age for the four skeletal samples. A Osteon population density (OPD). B: Accumulated osteon creations (AOC). C: Net remodeling (,,,V,,J. D: mean annual activation frequency (L,). E: Mean annual bone formation rate (VzJ The variables have not been age-adjusted for these illus- trations.


2 0 -

19

18

17

16

167

-

-

-

-

T

om5

0.040

f - I

0.035

0.030

15 I I I I , I

0 1 2 3 4 5

POPULATION

I I I I

0 1 2 3 4 5

POPULATION

1

25

g 5 3 15

05 I I I

0 1 2 3 4 5

POPULATION

I

Fig. 4. Graphic representation of the mean least squares age-adjusted means for each sample are plotted against their relative antiquity: 1, modern autopsy sample, 2, Late Woodland Ledders sample (ca. AD 1000); 3, Middle Woodland Gibson sample (55 B C - ~ 400); 4, Early Archaic Windover sample (circa 6000 BC). A-F

3

I 1 I I

0 1 2 3 4 5

POPULATION

I

1.0

0.9

g 0.8 E E E 2 0.7

5

a - - c

0.6

0.5

T

0 1 2 3 4 5

POPULATION

0.110

0.087

0.064

0.041

F

0018 I I I 1 I

0 1 2 3 4 5

POPULATION

illustrate osteon population density (OPD), accumulated osteon creations (AOC), mean osteon cross-sectional area (Ah), net remodeling (.,,V,,,,), mean annual activation frequency (L,). and mean annual bone formation rate (V,,,), respectively.


smallest mean osteon cross-sectional areas, respectively. Since osteon diameters were derived from cross-sectional areas, similar results obtained for mean osteonal diameter (Dh).

Mean net bone remodeling (,,,V,,,) ranges from 0.937 mm2/mm2 for the modern sample to 0.676 mm2/mm2 for Gibson (Fig. 4D). The means are statistically different when modern is compared to Windover (P = 0.023), or Gibson (P = 0.034).

Mean activation frequency (p,) ranges from 2.4/mm2/yr for modern to l.0/mm2/yr for Gibson (Fig. 4E). The difference is significant between modern and Windover (P = 0.043).

Mean annual bone formation rate (V,,,) ranges from 0.102 mm2/mm2/yr for the modern sample to 0.038 mm2/mm2/yr for Gibson (Fig. 4F). Although the overall ANCOVA is significant (P = 0.031), none of the pairwise comparisons was significant.

DISCUSSION The results of this study indicate that the

algorithm developed by Frost (1987a) to estimate missing osteons can be used to provide reasonable estimates of cortical bone remodeling rates for skeletal samples of varying antiquity. When plotted against the relative antiquities for the population samples, the variables show a concordant trend of increased cortical bone remodeling activity levels over time (Figs. 4A-F). Windover and Gibson are similar for all bone remodeling parameters and have the lowest remodeling activity levels among the samples. The Led- ders sample is consistently intermediate between these and the modern autopsy sample, which exhibits the highest levels of cortical bone remodeling activity.

The Windover and Gibson populations, which exhibit very similar patterns of cortical bone remodeling activity, are also the most similar in subsistence base; archaeological data indicate that they were both intensive foragers. Cook (1984) compared the incidence of stress indicators, such as growth retardation, growth disruption (Harris lines), bone density changes (Nordin’s In- dex), and periostitis, for subadults from the Ledders and Gibson populations with extant

populations. Her results tentatively identify the effects of nutritional stress as the main causative factor for the growth retardation in the Gibson population, while episodes of malnutrition and treponemal infection are indicated for the Ledders population. It is possible that the differences in cortical bone remodeling between Ledders, as compared to Gibson and Windover, also reflect the dif- ferential effects of these kinds of stress factors.

Cortical bone remodeling rates are determined by several factors. The rate at which new foci of remodeling are created (activation frequency) can increase or decrease, or the amount of bone remodeled per remodeling unit (osteon size) can vary.

Mean activation frequencies (FrC) differ only between the modern and Windover samples. Since none of the slopes for the regression of OPD against age are different, the rate of creation for cortical BMUs must not have differed among the sample populations. I t is not likely, therefore, that differences in prc can explain the bone formation rate differences observed among them. Os- teon size differs between the modern and Ledders samples, which exhibit the largest and smallest osteon cross-sectional areas (Ah), respectively, but these samples do not differ significantly in their estimated bone formation rates. I t must be concluded that the differences in bone formation rates among these populations are due to their AOCs, regardless of F,, or Ah.

Errors in the age-at-death estimates for the archaeological skeletons may account for these results. This would not explain the difference between Ledders and the other archaeological samples, since any bias in aging archaeological samples should be consistent.

A better explanation for the findings is the existence of differences in the effective ages of the birth of adult compacta among the populations; i.e., osteon creations have accumulated over longer or shorter periods of time. This could occur if the duration and rate of skeletal growth and cortical drifts differed among the populations. The lower OPDs and AOCs for the archaeological populations compared to modern may indicate that skeletal maturity occurred a t a later age resulting in an older effective age of


0.w

0.08 - 0.07

E E 2 0.06 - -. f 0.05

4.0

3 5

3.0

2 5

F 1 2.0

1 1 5

1 a

0 5

i -

-

- -

0'04 t I

[ f

I e

169

I I I I I I I 1 I I 0 1 2 3 4 5 0 1 2 3 4 5

0.03 I

POPULATION POPULATION

Fig. 5. Graphic illustration of the relative least-squares age-adjusted means for (A) mean activation frequency (F,), and (B) mean annual bone formation rate (V,,,), when 16 years is used as the effective age of the birth of adult compacta for the archaeological samples, rather than 12.5 years determined for the modern sixth rib.

the birth of their adult compacta, i.e. >12.5 years. In this situation, the apparent mean annual activation rate (Fr,) would be de- creased because the denominator used in the formula to calculate it would be inflated (see equation for E, above), and the resulting mean annual bone formation rates (V,,,) would be reduced proportionally.

If 16 years, rather than 12.5 years, is used as the effective age for the birth of adult compacta for the archaeological samples, the age adjusted means Fre for the three samples become 2.7 mm2/yr, 1.7 mm2/yr, and 2.4 mm2/ yr for Ledders, Gibson, and Windover, respectively (Fig. 5A). Their resulting mean annual bone formation rates (V,,,) then become 0.092 ? .0319 mm2/mm2/yr, 0.063 ? .0275 mm2/mm2/yr, and 0.095 f .0218 mm2/ mmZ/yr (Fig. 5B). The results of an ANCOVA test are now nonsignificant for either Ere ( P = 0.43) or V, , , (P = 0.741) among the samples.

Although a number of studies have suggested that dietary and biomechanical factors may produce population differences in bone remodeling activity (e.g., Ericksen, 1973; Richman et al., 1979; Stout, 1983; Burr e t al., 1990), factors that would affect the age of the birth of adult compacta remain obscure. There is evidence to suggest that, in the growing skeleton, mechanical usage affects bone growth and modeling (Frost,

1988) and that the types and levels of mechanical loads on bones are reflected in their shapes and sizes (Ruff, 1992). Moreover, shape and size differences reflecting different activity levels relating to modes of subsistence have been reported for archaeological skeletal samples (Bridges, 1989; Burr et al., 1990; Ruff et al., 1984). The pattern of skeletal growth and development is also affected by health-related factors, such as nutrition (Simmons, 1990), and genetic differences in the skeleton's response to mechanical usage during growth may cause population differences in bone mass (Frost, 198713).

It is not clear how the above factors would influence effective age of adult compacta. Al- though they can affect the shape and size of bones, the effective age of adult compacta is determined by the age at which skeletal maturity is attained and when growth and cortical drifting have essentially ceased.

Since adult bones with different effective ages of adult compacta will exhibit different OPD regardless of bone remodeling rates, in order t o compare bone remodeling rates determined by this algorithm, or any static histomorphometric method based on OPD, differences in effective ages for the birth of adult compacta must be identified. The algorithm should therefore be applied to rib samples from additional populations represent-


ing a greater variety of subsistence strategies and time periods in order to cor- roborate these findings and should begin to identify factors that can produce differences in the effective birth of adult compacta among human populations.

CONCLUSIONS Cortical bone remodeling rates for rib

samples from three archaeological populations and a modern autopsy sample were determined using an algorithm developed by Frost (1987a). When plotted against the relative antiquities for population samples, histomorphometric variables, i.e., activation frequency (Q, net bone formation (netVf,r,t), and mean annual bone formation rate (V,,,), exhibit a concordant trend of increased cortical bone remodeling activity levels over time. The two intensive foraging populations, Windover and Gibson, are similar for all bone remodeling parameters and have the lowest remodeling activity levels among the samples, while the more recent Ledders sample, which is reported to have practiced agriculture, is consistently intermediate between these and a modern autopsy sample. It is concluded that these differences cannot be attributed to differences in bone remodeling rates among the populations, but rather reflect different effective ages of adult com- pacts for their ribs. These findings suggest that the earlier populations, particularly Windover and Gibson, appear to have reached skeletal maturity a t an older age than observed for modern.

ACKNOWLEDGMENTS Rib samples from the Ledders and Gibson

sites were provided by Dr. Jane E. Buikstra, Department of Anthropology, University of Chicago, those from Windover by Dr. Glen Doran, Department of Anthropology, Florida State University, and the modern autopsy samples by Dr. Jay Dix, Department of Pa- thology, University of Missouri School of Medicine.

REFERENCES Ahlqvist J , and Damsten 0 (1969) Modification of Ker-

ley’s method for the microscopic determination of age in human bone. J. Forensic Sci. 14~205-212.

Asch NB (1975) Woodland subsistance: Implications for

demographic and nutritional studies. Am. J. Phys. Anthropol. 42:288 (abstract).

Bridges PS (1989) Changes in activities with the shift to agriculture in the southeastern United States. Curr. Anthropol. 30t385-394.

Buikstra J E (1972) Hopwell in the Lower Illinois River Valley: A regional approach to the study of biological variability and mortuary activity. Unpublished Ph.D. dissertation. University of Chicago.

Burr DB, Ruff CB, and Thompson DD (1990) Patterns of skeletal histologic change through time: Compari- son of an archaic Native American population with modern populations. Anat. Rec. 226:307-313.

Cook DC (1984) Subsistence and health in the Lower Illinois Valley: Osteological evidence. In MN Cohen and GJ Armelagos (eds.): Paleopathology at the Ori- gins of Agriculture. San Diego: Academic Press, pp. 235-269.

Doran GH, and Dickel DN (1988) Padiometric chronol- ogy of the archaic Windover archaeological site (8BR246). Florida Anthropol. 41:365-374.

Ericksen MF (1973) Age-related bone remodeling in three aboriginal American populations. Unpublished Ph.D. dissertation. Washington, DC: George Washing- ton University.

Ericksen MF (1991) Histological estimation of age at death using the anterior cortex of the femur. Am. J . Phys. Anthropol. 84~171-179.

Frost HM (1969) Tetracycline-based histological analysis of bone remodeling. Calcif. Tissue Res. 3r211-237.

Frost HM (1987a) Secondary osteon populations: an algorithm for determining mean bone tissue age. Yrbk. Phys. Anthropol. 30~221-238.

Frost HM (1987b) Bone “mass” and the “mechanostat”: A proposal. Anat. Rec. 219:l-9.

Frost HM (1988) Structural adaptations to mechanical usage. A proposed “three-way rule” for bone modeling. Comp. Vet. Orthop. Trauma 2:80-85.

Kerley ER, and Ubelaker DH (1978) Revisions in the microscopic method of estimating age a t death in human cortical bone. Am. J . Phys. Anthropol. 49; 545-546.

Kimmel DB, and Jee SS (1983) Measurements of area, perimeter, and distance: Details of data collection in bone histomorphometry. In RR Recker (ed.): Bone Histomorphometry: Techniques and Interpretations. Boca Raton, FL: CRC Press, pp. 89-108.

Richman EA, Ortner DJ, and Schulter-Ellis FP (1979) Differences in intracortical bone remodeling in three aboriginal American populations: Possible dietary factors. Calcif. Tissue Int. 28:209-214.

Ruff C (1992) Biomechanical analyses of archaeological human skeletal samples. In SR Saunders and AK Kat- zenberg (eds.): Skeletal Biology of Past Peoples: Re- search Methods. New York: Wiley-Liss, pp. 37-58.

Ruff C, Larsen CS, and Hayes WC (1984) Structural changes in the femur with the transition to agriculture on the Georgia coast. Am. J. Phys. Anthropol. 64: 125-136.

Simmons DJ (ed.) (1990) Nutrition and Bone Develop- ment. New York: Oxford University Press.

Stout SD (1978) Histological structure and its preserva- tion in ancient bone. Curr. Anthropol. 19~601-603.

Stout SD (1983) The application of histomorphometric analysis to ancient skeletal remains. Anthropos (Greece) 10~60-71.

OSTEONAL BONE REMODELING 171

Stout SD, and Paine RR (1992) Histological age estimation using rib and clavicle. Am. J. Phys. Anthropol. 87:lll-115.

Stout SD, and Paine RR (1994) Bone remodeling rates: A test of an algorithm for estimating missing osteons. Am. J. Phys. Anthropol. 93:123-129.

Stout SD, and Teitelbaum SL (1976) Histomorphometric determination of formation rates of archaeological bone. Calcif Tissue Res. 21:163-169.

Thompson DD (1979) The core technique in the determination of age a t death in skeletons. J. Forensic Sci. 24:902-9 15.

Wilkinson L (1989) SYSTAT The System for Statistics, SYSTAT, Inc., Evanston, IL.

Wu K, Schubeck KE, Frost HM, and Villaneuva A (1970) Haversian bone formation rates determined by a new method in a mastodon, and in human diabetes melli- tus and osteoporosis. Calcif. Tissue Res. 6:204-219.

bone remodeling rates and skeletal maturation in three archaeoloqical skeletal populations

Documents