This article was downloaded by: [UQ Library] On: 13 March 2015, At: 07:45 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Critical Reviews in Food Science and Nutrition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/bfsn20

Stable carbon isotopes and the study of prehistoric human diet a

b

Thomas W. Boutton , Mark J. Lynott & M. Pamela Bumsted

c

a

Department of Rangeland Ecology and Management, Texas Agricultural Experiment Station , Texas A&M University , College Station, Texas, 77843–2126 b

National Park Service , Midwest Archeological Center , Federal Building, Room 474, Lincoln, Nebraska, 68508–3873 c

Department of Chemistry , University of Massachusetts , Amherst, Massachusetts, 01003–0035 Published online: 29 Sep 2009.

To cite this article: Thomas W. Boutton , Mark J. Lynott & M. Pamela Bumsted (1991) Stable carbon isotopes and the study of prehistoric human diet, Critical Reviews in Food Science and Nutrition, 30:4, 373-385, DOI: 10.1080/10408399109527548 To link to this article: http://dx.doi.org/10.1080/10408399109527548

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Critical Reviews in Food Science and Nutrition, 30(3):373—385 (1991)

Stable Carbon Isotopes and the Study of Prehistoric Human Diet Thomas W. Boutton Department of Rangeland Ecology and Management, Texas Agricultural Experiment Station, Texas A&M University, College Station, Texas 77843-2126.

Mark J. Lynott

Downloaded by [UQ Library] at 07:45 13 March 2015

Midwest Archeological Center, National Park Service, Federal Building, Room 474, Lincoln, Nebraska 68508-3873.

M. Pamela Bumsted Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003-0035.

ABSTRACT: Mass spectrometric analysis of the stable carbon isotope composition (13C/12C or δ13C) of bone collagen from human remains recovered at archaeological sites provides a direct chemical method for investigating dietary patterns of prehistoric human populations. This methodology is based on the facts that (1) different food items within the human diet have distinct δ13C values, and (2) the δ13C value of human bone collagen is determined by the δ 13 C value of the diet. Studies of the development of subsistence patterns based on corn agriculture, one of the most significant developments in North American prehistory, can benefit from the use of stable carbon isotope techniques because corn has a high δ13C value relative to other components of the human diet. Measurements of δ 13 C of bone collagen from prehistoric human skeletal remains from southeastern Missouri and northeastern Arkansas indicate that intensive corn agriculture began in this region around A.D. 1000, that the incorporation of corn into the human diet was a rapid phenomenon, and that 35 to 77% of the human diet from A.D. 1000 to A.D. 1600 consisted of corn. Results from an isochronous population in southeastern South Dakota (A.D. 1400) suggest that 78 to 90% of the diet of this group consisted of corn, with no difference between males and females. Coupled with more traditional archaeological methods, stable carbon isotope analysis of bone collagen can significantly enhance reconstruction of dietary patterns of prehistoric humans.

13

12

KEY WORDS: prehistoric human diet, bone collagen, mass spectrometry, carbon isotopes, corn, C/ C ratios.

I. INTRODUCTION Human dietary information is sought frequently in archaeological studies, because it can reveal a great deal of information about other characteristics of past human populations. For example, diet may be related to population size, fertility, disease, movement and settlement pat-

terns, and economic systems. However, despite its obvious importance, dietary information has often been difficult or impossible to obtain because many food items are poorly preserved or not preserved at all. Also, food items are certainly not preserved in the same proportions they were eaten. For example, vegetable food items decompose readily and are probably underrepre-

1040-8398/91/$.50 © 1991 by CRC Press, Inc.

373

Downloaded by [UQ Library] at 07:45 13 March 2015

sented at most archaeological sites, whereas animal bones or shells are often well preserved. Thus, traditional methods of studying prehistoric diets, which rely heavily on the recovery of preserved food items and implements associated with food processing, are limited to specifying which items may have been part of the diet; they cannot establish the relative importance of these potential food items in the diet. In the past 10 years, these traditional indirect methods of studying prehistoric human diet have been supplemented by techniques based on direct chemical analysis of the elemental and isotopic composition of human bones recovered from archaeological sites. 13 These techniques capitalize on the fact that the elemental and isotopic composition of human bone is a function of the diet. Among the most successful of these direct chemical methods is the analysis of the natural stable carbon isotope ratio (13C/12C) of bone collagen by mass spectrometry. The carbon isotope composition of human bone collagen has been used to study the introduction of corn into the human diet in North and South America,4"7 and to establish human dependence on marine vs. terrestrial sources of food. 810 In this article, we will focus on the use of carbon isotopes as a tool for investigating the introduction and relative importance of corn in the diet of prehistoric humans in the eastern part of North America.

II. BASIS OF THE CARBON ISOTOPE TECHNIQUE Because the natural abundance of 13C is low and differences in the I3C/12C ratios of natural materials are small, stable carbon isotope ratios are expressed in relative terms as 513CPDB values. A 813CPDB value represents the %o (per mil, or parts per thousand) deviation of the 13C/12C ratio of the sample from that of the international PeeDee belemnite (PDB) limestone standard, the 13CPDB value of which has been arbitrarily set to 0%o.u Thus, a 8I3CPDB value of — 15%c would mean that the sample had a 13C/12C ratio that was 15 parts per thousand lower than that of the PDB standard. All 513C values in this article are relative to PDB. Small differences in 813C values exist be-

374

tween individual food items of both the past and present human diet. These small isotopic differences are due largely to carbon isotope fractionation during photosynthesis. Plants that incorporate atmospheric CO2 according to the C3 photosynthetic pathway discriminate against 13 CO2 to a greater extent than plants with the C4 pathway. As a result, C3 plants have 813C values approximately 14%0 lower than C4 plants. 813C values for C3 plants range from — 33 to — 22%c with an average of — 27%o, while C4 plants range from - 1 6 to -9%o and average -12.5%o.12-13 Most plant species possess the C3 pathway, and C3 species occur in virtually all environments. Far fewer plant species possess the C4 pathway. Due to the biochemical and physiological characteristics associated with the.C 4 pathway, C4 species are most common in warm, dry environments and are often of tropical or subtropical origin. Major food plants with C4 photosynthesis include corn, sorghum, and sugarcane. This isotopic variation in plants becomes useful in an archaeological context, because the natural stable carbon isotope composition of bone and other animal tissue is a function of the carbon isotope composition of the diet. This relationship between the 813C value of the animal and the 813C value of its diet has been demonstrated repeatedly under controlled laboratory conditions.14-15 Therefore, by analyzing the stable carbon isotope composition of human bones associated with archaeological sites, it is possible to determine whether those people subsisted on a diet of C3 plant foods, C4 plant foods, or a mixture of both. Prior to the introduction of corn into the human diet, prehistoric populations throughout much of eastern North America subsisted almost exclusively on C3 plant foods, and on animals that also fed on C3 plants.16"18 As a result, the introduction of the C4 plant, corn, into the diet of prehistoric humans in this region around A.D. 1000 is discerned readily as an increase in the 813C value of human bone. Therefore, studies of the development of subsistence patterns based on corn agriculture, one of the most significant developments in North American prehistory, can benefit from the use of stable carbon isotope techniques. We have analyzed the stable carbon isotope composition of prehistoric human skeletal

remains from eastern North America in an effort to (1) determine when corn became a significant component of the human diet in this area, (2) estimate the relative importance of this food item in the diet after it was introduced, and (3) evaluate dietary variation within an isochronous population engaged in corn agriculture. Some of the data presented in this article have been published in greater detail elsewhere.1922

Downloaded by [UQ Library] at 07:45 13 March 2015

III. DESCRIPTION OF STUDY SAMPLES To study the time of introduction and significance of corn in the human diet, bone samples from 27 individuals ranging in age from approximately 3200 B.C. to A.D. 1880 were obtained from 16 archaeological sites in southeast Missouri and northeast Arkansas.20 The location of these sites is shown in Figure 1. Samples were obtained from the Mississippi River alluvial valley and the adjacent eastern Ozark region, and were included in the study only when the date could be determined reliably from associated artifacts or by direct radiocarbon dating. These samples span three distinct cultural stages: Archaic, Woodland, and Mississippian. The Archaic stage began about 8000 B.C. and persisted for about 7000 years in this area. Human populations during the Archaic stage were mobile hunters and gatherers. However, there is evidence that squash was introduced as early as 5000 B.C., and cultivation of native plants was relatively widespread by about 2000 B.C.23 The Woodland stage began about 1000 B.C. and reflects increased sedentism. While hunting and gathering continued to be the basis for subsistence, horticulture presumably became increasingly important. The Mississippian culture was a widespread manifestation in portions of the mid western and southeastern U.S. The formative stage of the Mississippian culture is called Emergent Mississippian, and appears in various places in the midwest and midsouth between A.D. 600 and A.D. 800. The Mississippian culture first appears in fully developed form about A.D. 1000 in the Lower Illinois River drainage.24 This stage is characterized by intensive corn agriculture, large civic ceremonial centers, sedentary villages, and a complex sociopolitical organization.

To assess isotopic variation within a population practicing corn agriculture, 51 individuals (25 females, 26 males) from the Wolf Creek component of the Crow Creek site (39BF11) of the Initial Coalescent Tradition in Central South Dakota (Figure 1) were sampled.21 The Wolf Creek component was a large, fortified horticultural site occupying bottomland near the confluence of Crow Creek and the Missouri River, and was inhabited from A.D. 1300 to 1400. Around A.D. 1400 a large raiding group attacked the settlement and massacred 486 inhabitants, all of whom were buried later in a common grave. All 51 individuals were obtained from this common burial and were therefore coeval. In addition, all shared the same postmortem environment. At the time of death, the individuals sampled in this study were all adults, mostly between 25 and 35 years of age. Bone was sampled from the same skeletal position for each individual from the South Dakota site. It was not possible to sample bone from a single skeletal position for the Missouri/Arkansas samples. However, it has been shown that for animals fed the same diet, the variation in 813C between skeletal positions within a single individual is less than Ifoc.25 In addition to the human samples, faunal remains from several sites in Missouri/Arkansas and from the Crow Creek site were analyzed for 813C in an effort to understand better the isotopic composition of this portion of the human diet.

IV. SAMPLE PREPARATION AND MASS SPECTROMETRIC ANALYSIS The success of the isotopic approach to prehistoric diet analysis requires that we analyze a bone fraction that is not subject to diagenetic modification of the isotopic signal derived from diet. Carbonate carbon is present in bone tissue, but is known to undergo exchange with carbonates in the environment. While some studies have shown that the carbonate associated with bone apatite may reflect diet,26-27 others have shown that carbon in this fraction undergoes diagenetic alteration of 8I3C.28"30 Additional studies are needed to clarify whether the isotopic composition of apatite carbon can be used as an indicator

375

Downloaded by [UQ Library] at 07:45 13 March 2015

1. Scatters 2. Lepold 3. Billy Moore 4.McCarty 5. Christensen Cave 6. Nevins Cairn 7. Zebree 8. Round Spring 9. Li Ibourn 10. Turner 11.Berry 12. Campbell 13. Hazel 14. Euro-American 15. Owl's Bend 16. Kersey 17. Crow Creek

FIGURE 1. Map showing locations of archaeological sites from which human remains were acquired for isotopic analyses.

of diet. Organic carbon in bone is found largely in the structural protein, collagen, which does not suffer from exchange phenomena.31 Although the amount of collagen present in prehistoric bone is typically less than that of fresh modern bone, this loss of collagen does not influence the 813C value of the remaining collagen.30 Postmortem alteration of bone collagen 813C values can occur, especially if the bones have been burned. However, it has been shown that the elemental C/N ratio of bone collagen that has undergone isotopic alteration in the postmortem environment always lies outside the 2.9 to 3.6 range typical of fresh collagen.30-32 Thus, the C/N ratio of bone col-

376

lagen can serve as an indicator of possible diagenetic alteration of 813C. We analyzed the isotopic composition of bone collagen because it is known to reflect diet, does not exchange with the environment, and is relatively resistant to postmortem isotopic modification. Depending on the condition of the bone, approximately 0.5 to 10 g is needed to provide enough collagen for isotopic analysis. Details of the collagen extraction procedure have been presented elsewhere,1921 and involve: (1) removal of soil particles by manual scrubbing and ultrasonic cleaning, (2) grinding the bone sample to a powder, (3) demineralization of the bone pow-

Downloaded by [UQ Library] at 07:45 13 March 2015

der with dilute HC1, (4) removal of soil humic substances with dilute NaOH, and (5) solubilization of the collagen in mildly acidic water. Samples from South Dakota were not treated with NaOH. Instead, soil humic substances were removed by subjecting the collagen extract to reverse-phase chromatography using XAD-2 resin.21-33 This treatment does not modify the isotopic composition of the collagen.21-33 To verify that collagen was extracted, amino acid analyses were performed on several samples from the Missouri/Arkansas sites and the South Dakota site.19-21 Total nitrogen in the collagen fraction was determined by a micro-Kjeldahl method in order to compute elemental C/N ratios. For each sample, approximately 5 mg of collagen was combusted for 2 h at 850 °C in sealed quartz tubes containing CuO.34 The combustion products were released from the quartz tubes into a vacuum system using a tube-cracking device, and the CO2 was purified and isolated by cryogenic distillation. Percent carbon was determined by manometric measurement of the volume of CO2 produced and was used to calculate elemental C/N ratios. The CO2 was transferred to a Nier-type, dual-inlet, double-collector gas isotope ratio mass spectrometer, and the 13C/12C ratio was determined on the CO2.35 Results are expressed as:

5i3cP

=r

x 103

where R is the mass 45:44 ratio. Overall precision associated with sample preparation and mass spectrometry was less than 0.2%o (1 SD).

V. AMINO ACID PROFILES OF BONE COLLAGEN The amino acid composition of the collagen recovered during the extraction procedure was representative of human bone collagen (Table 1). Both the amino acid composition and the relative abundances of the individual amino acids were characteristic of human bone collagen.36 The degree of similarity between individuals was remarkable, considering that the samples ranged in age from 1000 B.C. to A.D.1880, and that sam-

ple preparation differed slightly between the Missouri/Arkansas samples and the South Dakota samples. There was no evidence in these samples that specific amino acids were preserved or lost differentially. This stability of the amino acids is critical, because the 813C values of individual amino acids in bone collagen may differ by as much as 17%c.37

VI. ISOTOPIC EVIDENCE FOR MAIZE AGRICULTURE IN MISSOURI/ ARKANSAS 813C values and C/N ratios of bone collagen extracted from the Missouri/Arkansas samples are given in Table 2. All samples had C/N ratios between 2.9 and 3.6, suggesting that there was no diagenetic modification of the 813C values.32 Samples 1 through 11 represent the time period from 3200 B.C. to A.D. 1000 and have 813C values ranging from —21.7 to — 19.1%o with an average of — 20.7%o. Because human bone collagen is approximately 5.1%o more enriched in 13 C than the diet,38 and because the NaOH treatment to remove soil humic substances increases the 813C value of the collagen by l%o,19 a bone collagen value of — 20.7%e translates into a value of - 26.8%o for the diet. An estimated 813C value of — 26.8%o for the diet suggests strongly that, prior to A.D. 1000, people in this region consumed only C3 plant foods and animals that also fed exclusively on C3 plants. There is no detectable C4 component to the diet prior to A.D. 1000. Although corn has been documented at archaeological sites in this region as early as A.D. 500,17-23 the isotopic evidence indicates that it could not have been a significant component of the human diet prior to A.D. 1000. Around A.D. 1000, 8:3C values of collagen became significantly larger, as exemplified by samples 12 through 27 (Table 2). After A.D. 1000, 813C values of bone collagen range from —15.8 to — 9.8%o with an average value of - 1 2 . 6 %o. This shift to higher 813C values in collagen around A.D. 1000 indicates that the diet changed to include a food or foods with a higher 13 12 C/ C ratio. Since all but a few known wild and domestic food plants in the eastern half of the U.S. prior to the introduction of corn have the

377

TABLE 1 Amino Acid Composition of Bone Collagen Extracted from Human Remains from Sites in Missouri and Arkansas, and from the Crow Creek Site in South Dakota

Downloaded by [UQ Library] at 07:45 13 March 2015

Missouri/Arkansas Amino Acid

23BU26 1000 B.C.

3PO467 A.D. 200

Glycine Alanine Proline Hydroxyproline Aspartic acid Glutamic acid Arginine Serine Lysine Leucine Valine Three-nine Phenylalanine Isoleucine Histidine Hydroxylysine Methionine Tyrosine

398.3a 116.0 116.0 NMb 71.5 73.0 41.6 31.9 31.9 24.6 44.0 16.4 16.4 9.7 4.3 NM 2.9 1.4

381.0 115.7 129.8 NM 86.3 81.1 44.5 30.3 30.8 26.1 27.0 18.9 12.1 10.5 3.6 NM 2.2 0.0

South Dakota

23PM5 Christensen Euro-Am. 58-14 A A.D. 1600 A.D. 1600 A.D. 1880 A.D. 1400 330.8 254.5 103.9 NM 67.3 63.8 38.5 30.4 25.5 22.3 18.2 15.0 10.1 8.6 4.2 NM 4.7 2.1

398.6 95.2 132.2 NM 85.7 81.2 47.1 34.4 31.5 27.5 19.6 18.5 8.2 10.8 5.3 NM 4.2 0.0

361.7 139.5 125.9 NM 79.9 78.4 47.2 35.3 30.1 27.6 19.6 18.2 11.3 10.1 6.2 NM 5.8 3.2

345.8 129.5 111.9 95.9 57.9 75.2 47.7 31.6 16.2 21.4 24.9 12.7 9.7 8.7 3.0 2.6 4.8 0.7

65-8 A.D. 1400

141-1A A.D. 1400

331.3 117.0 112.0 86.3 84.8 77.6 50.2 31.7 18.2 23.9 21.3 12.2 9.8 9.4 6.6 3.0 0.0 2.0

337.6 117.4 95.7 82.5 112.3 76.3 48.0 32.5 20.1 17.9 19.4 14.1 9.3 6.9 5.8 3.2 0.0 1.3

Results given as residues per 1000. NM = not measured.

C3 pathway of photosynthesis, the dramatic shift in collagen 813C values seen in these prehistoric bone samples can best be interpreted as the result of significant corn consumption. Intensive cultivation of corn as a component of the Mississippian culture is well established in the archaeological literature, so this direct chemical evidence for corn consumption is consistent with previous archaeological evidence. It is interesting to note that while samples from the Kersey site date to A.D. 1000 and had high 813C values, indicative of corn consumption, sample 11 from the Round Spring also dates to A.D. 1000 but had a low 813C value, suggesting complete reliance on C3 plant foods. This difference between sites suggests that contemporary human populations in close proximity may have had significantly different food habits. In this case, it is clear that humans at the Kersey site were heavily dependent on corn, while a contemporary population at Round Spring only 160 km away was not. Allowing for the + \%o fractionation from

378

NaOH treatment and the +5.1%o fractionation between diet and collagen, the average 813C value of the human diet after A.D. 1000 was approximately — 18.7%o. When a diet consists of two isotopically distinct food categories for which 813C values are known, it becomes possible to estimate the relative proportions of each by applying an isotopic mass balance equation: D = C3(P) + C4(l - P)

(1)

where D is the 8I3C value of the diet, C3 is the 813C value of C3 food items in the diet, C4 is the 813C value of C4 food items in the diet, P is the proportion of C 3 food in the diet, and 1 — P is the proportion of C4 food in the diet. Assuming mean 813C values for C3 and C4 plants of - 27.0%o and — 12.5%o, respectively, a human diet with a 813C value of — 18.7%o consists of approximately 43% C3 food and 57% C4 food (i.e., corn). Estimates for individuals after A.D. 1000 in Table 2 range from approximately 35 to 77% corn in the diet. Thus, after A.D. 1000, more than half

TABLE 2 Descriptions of Human Bone Samples and Results of Carbon Isotope Analyses Sample

no. 1 2 3

Downloaded by [UQ Library] at 07:45 13 March 2015

4

Site Scatters (3RA19) Lepold (23RI59) Billy Moore (23BU26) McCarty (3PO467)

Burial number 1

Location Randolph Co., AR

—

Ripley Co., MO

1

Butler Co., MO

—

Poinsett Co., AR

5

Christensen Cave

—

Pulaski Co., MO

6

Nevins Cairn (23PU200) Zebree (3MS20)

2

Pulaski Co., MO

3F

Mississippi Co., AR

8

Zebree (3MS20)

6

Mississippi Co., AR

9

Zebree (3MS20)

3G

Mississippi Co., AR

Owl's Bend (23SH10) Round Spring (23SH19) Kersey (23PM42) Kersey (23PM42) Kersey (23PM42) Kersey (23PM42) Kersey (23PM42) Kersey (23PM42) Round Spring (23SH19) Lilbourn (23NM38) Zebree (3MS20)

1

Shannon Co., MO

1

Shannon Co., MO

7

Pemiscot Co., MO

9

Pemiscot Co., MO

10

Pemiscot Co., MO

17

Pemiscot Co., MO

22

Pemiscot Co., MO

29

Pemiscot Co., MO

2

Shannon Co., MO

16D

New Madrid Co.,

7

10 11

12 13 14 15 16 17 18

19 20

21

Turner (23BU22A)

MO 4

21A

Mississippi Co. AR

Butler Co., MO

8"C

C/N ratio

Late Archaic 3200 B.C. Late Archaic 1980 B.C. Late Archaic 1000 B.C. Early Woodland 300 B.C. (Tchula Occupation) Woodland A.D. 500 (Meramec Springs complex) Woodland A.D. 500 (Meramec Springs complex) Mississippian A.D. 900 (Big Lake phase) Mississippian A.D. 900 (Big Lake phase) Mississippian A.D. 900 (Big Lake phase) Mississippian A.D. 900 Woodland A.D. 1000 (Meramec Springs complex) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1000 (Hayti phase) Mississippian A.D. 1200

-21.1

3.39

-21.7

3.30

-20.5

3.17

-21.7

3.27

-19.9

2.91

-20.1

3.16

-21.2

3.34

-20.5

. 3.21

-21.2

3.15

-19.1

3.30

-20.7

3.36

-10.6

2.98

-11.4

3.16

-10.9

3.12

-9.8

3.03

-11.5

3.06

-9.9

2.93

-15.6

3.28

Mississippian A.D. 1200 Mississippian A.D. 1200 (Lawhorn phase) Mississippian A.D. 1300 (Powers phase)

-14.9

3.22

-13.0

3.11

-15.8

3.18

Temporal placement

379

TABLE 2 (continued) Descriptions of Human Bone Samples and Results of Carbon Isotope Analyses Sample

Location

Site

22

Turner (23BU22A)

28

Butler Co., MO

23

Turner (23BU22A)

36B

Butler Co., MO

24

Berry (23PM59) Campbell (23PM5)

22

Pemiscot Co., MO

52

Pemiscot Co. MO

26

Hazel (3P06)

56

Poinsett Co., AR

27

—

—

Butler Co., MO

25

Downloaded by [UQ Library] at 07:45 13 March 2015

Burial number

no.

of the human diet may have consisted of corn in this geographic area. These results are in excellent agreement with other isotopic reconstructions of prehistoric human diet in eastern North America.4-6-38-39 Based on plant remains and vegetation reconstructions at our study sites, we are reasonably confident that corn is the only C4 plant food that could have produced the observed increase in human bone collagen 813C values after A.D. 1000. The animals that these people used for food should reflect isotopically the C3 nature of the plant communities in which they occurred. To be more certain of this, we have begun to assess the isotopic composition of animal remains associated with our study sites. Deer typically represent the most common vertebrate remains at our sites. Nine deer samples from three sites had 813C values that reflect a diet based entirely on C3 plants (Table 3). Even after corn is a wellestablished component of the human diet, deer do not seem to have been feeding to a significant extent in the corn fields (Table 3) and could not have been a factor in the increased carbon isotope ratios observed in the humans. Fish remains are also commonly associated with these human sites due to close proximity to several major river systems. 513C values of muscle tissue from two com-

380

Temporal placement Mississippian A.D. 1300 (Powers phase) Mississippian A.D. 1300 (Powers phase) Late Mississippian A.D. 1600 Late Mississippian A.D. 1600 (Armorel phase) Late Mississippian A.D. 1600 (Parkin phase) Nineteenth century Euro-American ca. 1880

813C

C/N ratio

-13.2

3.06

-14.1

3.32

-13.5

3.41

-10.4

3.46

-12.9

3.01

-13.3

3.20

mon fish species caught recently were also similar to the C3 plant resource base (Table 3), in agreement with other studies.40 Thus, it is unlikely that these major terrestrial and aquatic resources were responsible for the observed increase in 8I3C values of human bone collagen after A.D. 1000. Additional isotopic measurements of potential dietary items are needed to provide further confirmation.

VII. CARBON ISOTOPE VARIATION IN AN ISOCHRONOUS PREHISTORIC HUMAN POPULATION In most isotopic studies of prehistoric human diet, the number of samples that can be obtained at a given site is often limited. For example, only a single individual could be sampled from most of the sites discussed above. How well does that single individual represent the isotopic composition of the larger population from which he or she was sampled? To answer this, 25 females and 26 males from an isochronous population at the Crow Creek site (39BF11) in central South Dakota were sampled. This population dates to A.D. 1400 and was clearly engaged in agriculture with corn and other plant foods.

TABLE 3 813C Values for Bone Collagen from Deer and Flesh from Fish from Southeastern Missouri: Values for C/N Ratios and 813C Values are Means Standard Deviations

Site Owl's Bend (23SH10)

Gooseneck (23CT54)

Downloaded by [UQ Library] at 07:45 13 March 2015

Turner (23BU21a)

Glaze Creek (Ripley Co., MO)

Glaze Creek (Ripley Co., MO)

Species Odocoileus virginianus (white-tailed deer) Odocoileus virginianus (white-tailed deer) Odocoileus virginianus (white-tailed deer) Micropterus salmoides (largemouth bass) Lepomus macrochirus (bluegill)

Temporal placement

Sample size

C/N ratio of collagen

800 AD

3.11 ± 0.2

-20.2 ± 0.4

800 AD

3.13 ± 0.0

-20.1 ± 0.5

3.19 ± 0.0

-20.1 ± 0.2

1350 AD

2

1986 AD

-23.2

1986 AD

-23.1

A frequency distribution of 813C values for this population is presented in Figure 2. The mean and standard deviation for the entire population is - 11.3 ± 0.8%o. Females had slightly higher 813C values and less variability than males, although neither difference was statistically significant. 813C values for the entire population ranged from -13.5%o to -9.6%o. The small standard deviation (0.8%o) associated with the entire sample and the relatively small range (3.9%o) of values suggests that results from a single individual should be reasonably representative of the larger population from which he or she was sampled. The high mean 813C value of - 11.3%o for this population indicates that a substantial proportion of the diet was derived ultimately from C4 plants. Allowing for the + 5. l%o fractionation between diet and collagen, the average 813C value for the diet of the Crow Creek people was approximately -16.4%o. Although this site was located in a riparian habitat where all plant species were C3, this riparian corridor was embedded in a matrix of mixed prairie containing a variety

of C4 grass species. While all plant foods recovered from this site were C3 (except corn), faunal remains show conclusively that plains animals (e.g., bison and antelope) were hunted and consumed by these people.21 If these plains animals fed on the C 4 grass species present in the mixed grassland, then corn would not have been the sole source of food with a high 813C value. To investigate the isotopic contribution from the meat portion of the diet, 8I3C values of bone collagen from bison, deer/antelope, and canid/ wolf were determined. A single bison had a 813C value of - 15.7%c (Table 4), indicating that meat from this animal has a higher 813C value than the C3 plant portion of the diet. Similarly, two canid/ wolf samples had high 813C values (Table 4). These may have been domestic dogs, which scavenged leftovers and waste from the human diet and were themselves often part of the human diet. High 813C values for dogs associated with prehistoric human populations have been reported by other investigators.40-41 The deer/antelope samples had 813C values reflecting the largely C3 plant diet that they are known to consume (Table

381

-11.2 ±0.6 (N=25) -11.4 ±0.9 (N=26) ALL-.-11.3 + 0.8 (N=5I)

-12

-II

-10

SI3C %o vs. PDB

Downloaded by [UQ Library] at 07:45 13 March 2015

FIGURE 2. Frequency distribution of 813C values of bone collagen from individuals recovered from the Crow Creek site in South Dakota.

TABLE 4 813C-Values of Some Potential Dietary Components at the Crow Creek Site in South Dakota Sample size Bison Canid/wolf Deer/antelope

1 2 3

813C ± SD

-15.7 -12.0 ± 1.9 -18.8 ± 0.8

4). Thus, deer and antelope probably did not contribute to the high 813C values obtained from this human population. Given that bison had relatively high 813C values due to grazing on C4 grasses, meat from this animal would have contributed to the high 813C values of bone collagen from the Crow Creek people. Clearly, determination of the proportion of corn in the Crow Creek diet is not possible using the simple two-component mass balance model applied to the Missouri/Arkansas data. Instead of two isotopically distinct food classes, we now have three: C3 vegetables, C4 vegetables (i.e., corn), and meat. Thus, it is no longer possible to derive a single estimate of the relative proportions of each food category in the diet. However, it is possible to use a three-component mass balance approach to derive ranges of values

382

for the relative proportions of each food category in the diet using the expression: D = C3(P3) + C4(P4) + M ( P J

(2)

where P 3 , P 4 , and Pm are the relative proportions of C3 foods, C4 foods, and meat, respectively, in the diet; and C3, C4, M, and D are 813C values of C3 foods, C4 foods, meat, and the average human diet, respectively. Using this model, the composition of the Crow Creek diet has been estimated as 0 to 10% C3 vegetables, 0 to 22% meat, and 78 to 90% corn.21-22

VIII. PROBLEMS AND LIMITATIONS Several aspects of the isotopic technique for prehistoric diet reconstruction require further attention. One of the major problems is that we are not certain what portion of the human life span is represented by the 813C value of the bone collagen. In other words, does the bone collagen 813C value integrate the average isotopic composition of the diet over the entire life span of an individual, or does it reflect a more restricted portion of that person's life? This question cannot be answered precisely until we know the turnover rate of bone collagen in humans or a suitable

Downloaded by [UQ Library] at 07:45 13 March 2015

animal model. If human bone collagen has a very slow turnover rate, then 813C values of bone collagen may largely reflect diet during the skeletal growth period. If collagen has a rapid turnover rate, then 813C values may best reflect diet over the past few years of life prior to death. Evidence based on uptake and decay of radiocarbon derived from nuclear weapons testing suggests that human bone collagen turns over very slowly.42-43 However, the long half-life of 14 C does not lend itself to detection of rapid turnover that may occur over relatively short time intervals. Recent evidence suggests that collagen in skin, muscle, organs, and connective tissue has a very rapid turnover rate,44 and that bone collagen may be more metabolically active than these other types of collagen.45 Furthermore, significant seasonal changes in human bone resorption and formation46 should result in some bone collagen turnover. Thus, the widespread belief that bone collagen is metabolically inert and does not have an appreciable turnover rate deserves to be scrutinized closely. Another potential problem is lack of good information regarding the actual isotopic composition of food consumed by prehistoric people. In most mass balance calculations (Equation 1) aimed at determining the proportion of corn in the human diet, mean 813C values, such as - \2.5%o for C 4 plants and - 27%o for C 3 plants, are usually used. However, are these mean values appropriate? As an example, we have measured five corn cobs from the Crow Creek site and obtained a mean and standard deviation of —10.6 ± 0.5%o. Similarly, nine 813C measurements of corn grain and cobs from several archaeological sites in the southwestern U.S. yielded a mean and standard deviation of - 10.7 ± 0.6%c.47 Use of these higher 813C values for the C4 component of the diet would clearly produce estimates of the relative proportion of corn in the diet different from those calculated above. Similar discrepancies may exist for the C3 plant portion of the diet. There is also limited information regarding the isotopic composition of the meat component of the prehistoric diet. As demonstrated for the Crow Creek site above, some human populations may have hunted and eaten animals such as bison, which derived a significant proportion of

their diet from C4 plants. This would be an especially important consideration at archaeological sites located within the Great Plains or the southwestern desert regions, where C4 plants comprise a significant proportion of the plant species and biomass.48 In addition, although it is known that marine and terrestrial foods differ in 8I3C value, little is known about 813C values of freshwater plants and animals. To further define the carbon isotopic composition of components of the human diet, we recommend that isotopic measurements be made not only on human remains, but also on the faunal remains frequently associated with archaeological sites. Despite some uncertainty surrounding the actual carbon isotopic composition of components of the prehistoric human diet, which may limit quantitative estimates of the proportion of corn in the diet, 813C values of bone collagen remain as extremely useful indicators of past diet.

IX. FUTURE PROSPECTS FOR STABLE ISOTOPES IN ARCHAEOLOGY It is likely that the application of stable isotope techniques to the study of prehistoric human diet will continue to be refined and expanded. As described in this article, 813C values of the bone collagen can be used to study the introduction of corn into human diets based largely or exclusively on C3 plants. Other archaeological studies with stable carbon isotopes have demonstrated that it is possible to resolve relative inputs of food from marine vs. terrestrial souces at coastal sites. 810 Elements other than carbon have also demonstrated isotopic distributions that can provide useful dietary information. For example, 8I5N values of human bone collagen can be used to resolve marine from terrestrial food inputs,49 legume from nonlegume plant foods,3 and pastoralists from agriculturalists.50 It is becoming apparent that 13C and 15N measurements on the same bone sample can often provide considerably more dietary information than might otherwise be accessible through the use of either element alone.50 More accurate and powerful reconstructions may also result from linking stable isotope analyses with trace element analyses. I>21'39 Because of difficulties inherent in re-

383

constructing prehistoric human diet from food remains and tools for their preparation, it seems likely that isotopic techniques and other direct chemicals methods of bone analysis will continue to be developed and applied in a variety of archaeological contexts.

Downloaded by [UQ Library] at 07:45 13 March 2015

REFERENCES 1. Lambert, J., Simpson, S., Szpunar, C , and Buikstra, J., Acc. Chem. Res., 17, 298, 1984. 2. Price, T., Schoeninger, M., and Armelagos, G. J., Human Evol., 14, 419, 1985. 3. De Niro, M., Amer. Sci. 75, 182, 1987. 4. Vogel, J. and van der Merwe, N., Am. Antiq., 42, 238, 1977. 5. DeNiro, M. and Epstein, S., Geochim. Cosmochim. Acta, 45, 341, 1981. 6. Bender, M., Baerreis, D. and Steventon, R., Am. Antiq., 46, 346, 1981. 7. van der Merwe, N., Roosevelt, A., and Vogel, J., Nature, 292, 536, 1981. 8. Tauber, H., Nature, 292, 332, 1981. 9. Chisholm, B., Nelson, D., and Schwarcz, H., Science, 216, 1131, 1982. 10. Walker, P. and DeNiro, M., Am. J. Phys. Anthropol., 71, 51, 1986. 11. Hayes, J . , Spectra, 8, 3, 1982. 12. Smith, B. and Epstein, S., Plant Physiol., 47, 380, 1971. 13. O'Leary, M., BioScience, 38, 328, 1988. 14. DeNiro, M. and Epstein, S., Geochim. Cosmochim. Acta, 42, 495, 1978. 15. Tieszen, L. and Boutton, T., Stable Isotopes in Ecological Research, Rundel, P. W. et al., Eds. Springer-Verlag, New York, 1989, 167. 16. Lewis, R., Mississippian Exploitive Strategies: A Southeast Missouri Example. Missouri Archaeological Society Research Series No. 11, Columbia, MO, 1974. 17. Asch, D. and Asch, N., Prehistoric Food Production in North America, Ford, R. I., Ed., Museum of Anthropology, University of Michigan, Anthropological Paper no. 75, 1985, 149. 18. Parmalee, P., Paloumpis, A., and Wilson, N., Animals Utilized by Woodland Peoples Occupying the Apple Creek Site, Illinois. Illinois State Museum, Reports of Investigations no. 23, 1972. 19. Boutton, T., Lynott, M., Klein, P., Price, J., and Tieszen, L., Stable Isotopes in Nutrition, Turnlund, J. and Johnson, P., Eds., ACS Symposium Series, no. 258, American Chemical Society, Washington, D.C., 1984, 191.

384

20. Lynott, M., Boutton, T., Price, J., and Nelson, D., Am. Antiq., 51, 51, 1986. 13 21. Bumsted, M. P., Human Variation: δ C in Adult Bone Collagen and the Relation to Diet in an Isochronous C 4 (Maize) Archaeological Population, Ph.D. thesis, University of Massachusetts, Amherst, MA, 1984. 22. Bumsted, M. P., J. Human Evol., 14, 539, 1985. 23. Conard, N., Asch, D., Asch, N., Elmore, D., Gove, H., Rubin, M., Brown, J., Wiant, M., Farnsworth, K., and Cook, T., Nature, 308, 443, 1984. 24. Bareis, C. and Porter, J., Ed., American Bottom Archaeology, University of Illinois Press, Urbana, IL, 1984. 25. DeNiro, M. and Schoeninger, M., J. Archaeol. Sci., 10, 199, 1983. 26. Sullivan, C. and Krueger, H., Nature, 292, 333, 1981. 27. Thorp, J. and van der Merwe, N., S. Afr. J. Sci., 83, 712, 1987. 28. Schoeninger, M. and DeNiro, M., Nature, 297, 577, 1982. 29. Schoeninger, M. and DeNiro, M., Nature, 301, 177, 1983. 30. Nelson, B., DeNiro, M., Schoeninger, M., De Paolo, D., and Hare, P., Geochim. Cosmochim. Acta, 50, 1941, 1986. 31. Wyckoff, R., The Biochemistry of Animal Fossils, Scientechnica, Bristol, England, 1972. 32. DeNiro, M., Nature, 317, 806, 1985. 33. Stafford, T., Brendel, K., and Duhamel, R., Geochim. Cosmochim. Acta, 52, 2257, 1988. 34. Boutton, T., Wong, W., Hachey, D., Lee, L., Cabrera, M., and Klein, P., Anal. Chem., 55, 1832, 1983. 35. Craig, H., Geochim. Cosmochim. Acta, 12, 133, 1957. 36. Estoe, J., Biochem. J., 61, 589, 1955. 37. Tuross, N., Fogel, M., and Hare, P, Geochim. Cosmochim. Acta, 52, 929, 1988. 38. van der Merwe, N., Am. Sci., 70, 596, 1982. 39. Ambrose, S., Emergent Horticultural Economics of the Eastern Woodlands, Keegan, W., Ed., Center for Archaeological Investigations, Occasional Paper no. 7, Southern Illinois University, 1987, 87. 40. Katzenberg, M. A., J. Archaeol. Sci., 16, 319, 1989. 41. Burleigh, R. and Brothwell, D., J. Archeol. Sci., 5, 355, 1978. 42. Libby, W., Berger, R., Mead, J., Alexander, G., and Ross, J., Science, 146, 1170, 1964. 43. Stenhouse, M. J. and Baxter, M. S., Radiocarbon Dating, Berger, R. and Suess, H., Eds., University of California Press, Berkeley, 1979, 324. 44. Laurent, G., Am. J. Physiol. (Cell Physiol.), 252, C1, 1987.

48. Boutton, T., Harrison, A., and Smith, B., Oecologia, 45, 287, 1980. 49. Schoeninger, M., DeNiro, M., and Tauber, H., Science, 220, 1381, 1983. 50. Ambrose, S., J. Human Evol., 15, 707, 1986.

Downloaded by [UQ Library] at 07:45 13 March 2015

45. Cundy, T., Bartlett, M., Bishop, M., Earnshaw, M., Smith, R., and Kanis, J., Metab. Bone Dis. Rel. Res., 4, 297, 1983. 46. Hyldstrup, L., McNair, P., Jensen, G., and Transbol. I., Bone, 7, 167, 1986. 47. Creel, D. and Long, A., Am. Antiq., 51, 826, 1986.

385