JOurn.31 ofr iLn.err.3dua.3- Re--search
..Oluin.e 5, issue -6 - M.1arch 2i::1114
PIXE and Paleodiet: Reconstructing Subsistence of Florida's Middle
Archaic Using a New Method of Trace Element Analysis
Trace elements are naturally occurring elements found in organic and inorganic materials. Archaeologists
are interested in trace elements because it has been demonstrated that elemental concentrations within bone
may reflect elemental concentrations of diet. This research studies trace element concentrations of human and
non-human bone collected from Nona's Site, a Middle Archaic locality (7,000-5,000 years BP) in Sarasota
County, Florida. Sample analysis focuses on the use of Particle-Induced X-ray Emission (PIXE), a recently
introduced ion beam analytical method. Comparison of the data obtained from PIXE with those derived from ICP-
MS, a more traditional method of trace element analysis, indicates that PIXE is a valid method to determine
trace element concentrations in archaeological bone remains. However, the data generated suggest a
diagenetic signal rather than a dietary one, based on the elevated concentrations of strontium and calcium in
the Nona's Site bone samples. This is likely the result of inundation in an aqueous environment.
The Trace element analysis of bone has demonstrated that the elemental composition of skeletal remains
correlates with subsistence patterns of the sampled population (Radosevich 1993). This research examines
the elemental composition of prehistoric bone material utilizing Particle-Induced X-ray Emission (PIXE), a
relatively new spectroscopic technique, to determine if PIXE is an effective tool for paleodiet reconstruction.
Sample analysis is restricted to elements demonstrated to correlate with diet: strontium (Sr), calcium (Ca),
barium (Ba), magnesium (Mg), manganese (Mn), iron (Fe), zinc (Zn), and copper (Cu). To date, PIXE has
been minimally used to infer paleodiet in prehistoric remains. Examination of remains from sites dating to the
Florida Middle Archaic (i.e. Nona's Site) provides benchmark data which can be used to redefine the ecology
and settlement patterns of this time.
Particle-induced X-ray Emission (PIXE) is an ion beam analytical method introduced in the 1970s (Jankuhn
1997; Sandford 1993). Bombardment at high energy levels changes the sample's atomic structure by
displacing electrons from their current energy level, prompting other electrons to move in to replace them.
This electron shift from one energy level to another releases a photon in the X-ray region of the
electromagnetic spectrum. The emitted X-rays are collected by an Si(Li) detector, converted into electrical
impulses, sent to a multi-channel analyzer, and displayed in a graph as the number of X-rays per energy. The
area under each X-ray peak correlates to the element's concentration within the sample. Quantitative analysis
was conducted using RobWin analytical software and are reported in parts per million, ppm (Figure 1). The 1.7
MV tandem accelerator used for this PIXE investigation is housed in the University of Florida's Department of Physics.
I I I I I I I I
,o,.. II AI A
Figure 1. Spectrum created from PIXE analysis. Peaks on graph represent X-rays per energy level.
Area under each peak is then interpreted as an elemental concentration.
Samples for this study are from Nona's Site in Sarasota County, Florida and date to the Middle Archaic period,
ca. 7000-5000 years BP (Figures 2a, 2b). These remains, discovered in the 1980s, were found on the
southeastern edge of a year-round pond (Luer 2002). The site is important as it is situated in close proximity to
both coastal habitats of the Gulf and to Florida's upland interior. This diversity of exploitable resources creates a
wide-range of potential subsistence patterns available to the prehistoric inhabitants (DeLeon 1998).
N.. A 1
~'-r .~ I
Figure 2a. Nona's Site (from Luer 2002). Map of Sarasota County, Florida, location of Nona's
Site. Demonstrates the close proximity to the coast and Florida's upland interior.
Figure 2b. Nona's Site (from Luer 2002). Plan view of excavations at Nona's Site. Illustrates layout
of topography and various points of excavation.
PIXE analyses of ground bone include a total of ten samples, as listed in Table 1. Human samples were chosen
based on bones with good cortex, mainly long bone fragments. Non-human samples were analyzed as
an independent assessment of diagenesis, or alteration of bone chemistry. Sample preparation procedures for
PIXE analysis included those used in previous PIXE experiments involving ground material and required
minimal modifications (Kravchenko et al. 2003).
Preparation Procedures for PIXE Analysis of Ground Bone Samples
Description Dry Weight DIUF Added Yttrium pL Yppm
ï¿½i ' 'It
I ~. ï¿½/* 1' -. 'l .
* .. ' *
j. t ,'
*n" -^ 11
FS9- snake 0.3461 150
PNS0203 FS10 - alligator 0.2891 200 300 1038
PNS0310 FS10 - turtle 0.3073 150 300 975
PNS0452 FS19 - turtle ss 0.2012 150 200 994
PNS0541 FS18 - human 0.536 300 500 932
PNS0654 FS19-human 0.2185 150 200 915
PNS0757 FS31 - human 0.3238 250 350 1018
PNS0862 FS47 - human 0.1363 150 100 733
PNS0933 FS14 - deer 0.2777 150 350 900
PNS1023 FS10 - human 0.3784 200 350 924
Bone ash analysis: Inductively Coupled Plasma - Mass Spectroscopy (ICP-MS) and PIXE
Six samples were also examined using a more standard method of trace element analysis known as ICP-MS
to provide a comparative baseline with which to interpret the PIXE results (Sandford 1993). ICP-MS protocol
requires ashing the bone samples prior to analysis. This process removes organic material and decreases
the probability of sample contamination. Therefore, additional preparation included baking ground bone samples in
a muffle furnace at 720f C for 8 hours.
The six samples, including 5 human and 1 deer, were ashed and sent to The University of Wisconsin's Laboratory
for Archaeological Chemistry. Samples were diluted in nitric acid, and an aqueous solution of known
concentration was analyzed. The six bone ash samples were also subjected to PIXE analysis. Results of ground
bone PIXE analysis and bone ash PIXE analysis could then be compared to other published studies as well as
data obtained from ICP-MS analysis.
The PIXE analysis of ground bone yielded highly variable results. Table 2 outlines Pate's (1995) range of
elemental concentrations in relation to diet. In summary, diets relying on marine resources tend to have high Sr
and Mg, but low Ba concentrations. Terrestrial diets generally result in low concentrations of Sr and Mg, but
high concentrations of Ba. Herbivorous diets are high in Sr, Ba, and Mg, but have low concentrations of
Zn. Carnivorous diets reflect low levels of Sr, Ba, and Mg, but have high concentrations of Zn.
Accordingly, omnivorous diets have intermediate concentrations for Sr, Ba, Mg, and Zn.
(From Pate, 1995) Chart of Bone Elemental Concentrations in Relation to Diet
Diet Sr Ba Mg Zn
Herbivorous High High High Low
Omnivorous Intermediate Intermediate Intermediate Intermediate
Carnivorous Low Low Low High
Marine High Low High N/A
Terrestrial Low High Low N/A
Elemental concentrations for the 5 human samples are shown in Figure 3 . Previous research has concluded
that PIXE analysis is highly effective in determining concentrations of heavier elements, including Ca and Fe,
but poorly reflects concentrations of lighter elements, such as Mg (Kravchenko et al. 2003).
I' : i
Ba Ca Cu F# Mg
Figure 3. Elemental concentration variability of PIXE
*.: r L
Mn Sr Zn
ground human bone alanlysis for each of the
The PIXE data was highly variable when compared to other published ranges utilizing trace element analysis.
PIXE data interpretation relies on incorporation of quantitative analysis of raw data combined with appearance
of each element on the spectrum graph. Ba and Mg quantitative levels matched those seen in other studies,
however, since those levels are not apparent on the graph, data for these elements cannot be used for
subsistence reconstruction. When compared to similar studies involving trace element analysis of bone, Zn and
Mn concentrations seem comparable while highly elevated levels of Ca and Sr are apparent (Figure 4a). Sr and
Ca are especially useful in paleodiet studies; however, the extremely high levels observed in Nona's Site samples
do not correspond with other published ranges and are off by over lOOx. This strongly suggests diagenesis
rather than a reflection of dietary influence in the samples analyzed.
PIXE Ground Bone Results (ppm)
Ranes ELEMENT PIXE RESULTS & RANGES
2s - Ba 4 19-117
3271-341 Ca T 35,354 - 524,492
31j.- 8O Cu 3. - 10
3 73-4239 Fe t 76-10,460
S713- 53 Mg t 3129 - 10,400
3 -81 Mn = 90-126
3201-622 Sr 10,519 12,067
21 -23S7 Zn - 70-105
Source I Katt~e rg (19M) Radoovwichi (10031 3WIImnme (123
Figure 4a. PIXE and ICP - MS results. Elemental concentrations resulting from analysis of ground
bone using PIXE analysis in comparison to other published studies using trace element analysis.
PIXE analysis of bone ash shows similar results to the PIXE results of ground bone (Figure 4b). ICP-MS analysis
of bone ash, when compared to published ranges (Figure 4c) show similar results for three of the four
elements compared, including Ba, Mg, and Mn. Comparison of bone ash results from both PIXE and ICP-MS
analysis (Figure 4d) demonstrates a wider range of results when using PIXE analysis. ICP-MS ranges for Ba, Mg,
and Sr fall within the PIXE ranges, while Mn concentrations are somewhat high.
PIXE Bone Ash Results (ppm)
Ranges ELEMENT PIXE RESULTS & RANGES
65- 275 Ba + 34-76
3271 - 341 Ca 1ff 281,652 - 43,057
31.8-6.0 Cu 9-34
1280 - 1986
73- 4230 Fe It 399- 901
S 0 Mg t e938-9862
2 797-4443 Mg
a 68-811 Mn 68-113
3201 5l2 Sr tt+ 7743-14,706
1 52-271 Zn 62-116
1acM ' KtAnnberg (19aM) * Ra.daovcti (1983) 'WMijam (1993)
Figure 4b. Elemental concentrations resulting from analysis of bone ash using PIXE analysis
in comparison to other published studies using trace element analysis.
ICP-MS Bone Ash Results (ppm)
Rawnae ELEMENT ICP-MS RESULTS & RANGES
2as-275 Ba = 63-115
3271 -41 Ca NA
1, - s0 Cu NA
11280 - 1ie6
S73-423 Fe NA
3 7. 4443 Mg = 17Ts -E3
1248 - 542 .
i G-11 MnR 131-163
201-S92 Sr 9420 - 12.749
215w 207 Zrn NA
Source: I Kietnbetg (1M9) 2 Riadwich (1 3) 3Wil iam 1g93n)
Figure 4c. Elemental concentrations resulting from analysis of bone ash using ICP-MS analysis
in comparison to other published studies using trace element analysis.
PIXE vs. ICP-MS Bone Ash Results
RESULTS & RANGES (ppm)
ELEMENT PIXE ICP-MS
Ba 34-76 63-115
Ca NA NA
Cu NA NA
Fe NA NA
Mg 938-9862 1750 - 3400
Mn 68 - 113 131 -163
Sr 7743 - 14,706 9420 - 12,749
Zn NA NA
Figure 4d. Comparison of elemental concentrations from resulting analysis of bone ash using PIXE
versus analysis of bone ash using ICP-MS.
DISCUSSION AND CONCLUSION
The variable results obtained from PIXE analysis, specifically the extremely high concentrations of Sr and Ca,
suggest fluctuations due to factors other than diet. Sample selection and preparation as well as diagenetic
variables have possibly affected the elemental concentrations. Samples were selected from various locations
within Nona's Site under the pretense that mean elemental concentrations would be more accurate when
considering fluctuations due to diagenic influences. In trying to obtain an overall balance of diagenic variation,
sample selection procedures limited the likelihood that bone specimens were from the same individual,
thus increasing the variability within the study.
Postmortem changes can alter the elemental composition of bone, creating difficulty when interpreting
subsistence patterns. Soft tissue decomposition, such as autolysis and putrefaction, combined with
groundwater interaction alter chemical dietary signatures of human bone by forming acid by-products
(Sandford 1993; Pate 1995). Further, elements are leached away in soils, thus altering the chemical signature
that reflects diet (Pate 1995). The advantage of ashing bone is that organic material is removed, decreasing
the possibility of sample contamination.
The results in this study demonstrate that PIXE data is comparable to ICP-MS analysis. Thus, PIXE, a fast
and accurate method to analyze multiple elements simultaneously, can be included in the list of more
traditional techniques, such as ICP-MS and AAS, which are currently used for trace element analysis.
Sample preparation is relatively quick and straightforward. However, studies of human bone using PIXE
require analysis of 3-5 targets to provide enough data to ensure accurate statistical range, therefore increasing
the time and energy spent obtaining results. Further, PIXE analysis is considerably more specialized and
expensive than other methods listed above.
Diagenesis greatly impacted the data obtained from the human samples, especially evident in the elevated levels
of strontium and calcium. In addition, inconclusive Ba and Mg data create difficulties for subsistence
reconstruction. However, it is important to note that the levels of Sr, Mn, and Zn all reflect a reliance on
marine habitats. PIXE analysis of human bone ash, when compared to data obtained from ICP-MS analysis of
human bone ash, show a wider range of data that is less precise in determining mean elemental
concentrations. However, since nearly all of the ICP-MS data are within the range of the PIXE data, it is safe
to assume that PIXE is providing accurate results. If, in future studies, diagenetic effects can be accounted for,
the use of PIXE as an analytical method of trace element analysis will provide another helpful tool to contribute to
the reconstruction of prehistoric subsistence.
1. DeLeon, V. B. (1998). Stable Isotope Analysis and Paleodiet at the Bay West Site, Collier County,
Florida. Unpublished MA Thesis, Anthropology Department, University of Florida.
2. Edward, J. B. and Robert A. Benfer (1993). The effects of diagenesis on the Paloma skeletal material. In (M.
K. Sandford, ed.) Investigations of Ancient Human Tissue: Chemical Analyses in Anthropology. New York: Gordon
and Breach, pp. 183-268.
3. Jankuhn, St., T. Butz, R. H. Flagmeyer, T. Reinert, J. Vogt (1997). Ion beam analysis of ancient human bone. In (J.
L. Duggan and I.L. Morgan, eds.) Application of Accelerators in Research and Industry. New York: AIP Press, pp.
4. Kravchenko, I.I., R.F. Kelly, F.E. Dunnam, H.A. Van Rinsvelt (2003). PIXE Study of Lacustrine Sediments Including
a Sediment Core From Lake Maicuru, Para, Brazil. In (J.L. Duggan, I.L. Morgan, eds.) Application of Accelerators
in Research and Industry: 17th Int'l Conference. American Institute of Physics.
5. Luer, George M. (2002). Three Middle Archaic Sites in North Port. In (George M. Luer, ed.) Archaeology of
Upper Charlotte Harbor, Florida. Tallahassee: Florida Anthropological Society, Publication No. 15, pp. 3-33.
6. Pate, F. D. (1995). Bone chemistry and paleodiet. Journal of Archaeological Method and Theory. 1:161-209.
7. Radosevich, S. C. (1993). The six deadly sins of trace element analysis: A case of wishful thinking in science. In
(M. K. Sandford, ed.) Investigations of Ancient Human Tissue: Chemical Analyses in Anthropology. New York:
Gordon and Breach, pp. 269-332.
8. Sandford, M. K. (1993). Understanding the biogenic-diagenetic continuum: Interpreting elemental concentrations
of archaeological bone. In (M. K. Sandford, ed.) Investigations of Ancient Human Tissue: Chemical Analyses
in Anthropology. New York: Gordon and Breach, pp. 3-57.
9. Scharf, W. H. (1988). Particle Accelerators-Applications in Technology and Research. New York: John Wiley
Back to the Journal of Undergraduate Research
College of Liberal Arts and Sciences I University Scholars Program I University of Florida I| l UNIVERSITY of
ï¿½ University of Florida, Gainesville, FL 32611; (352) 846-2032.