Publication SAA 2001

Kathryn Puseman, Paleo Research Labs, Golden, Colorado
and
Ralph E. Klinger, Bureau of Reclamation, Denver, Colorado

ABSTRACT

Bulk soil samples often are submitted for radiocarbon dating; however, bulk soil has the potential for containing large amounts of modern carbon. Various studies have demonstrated the effect of modern carbon contamination on measured radiocarbon ages. Identification of charcoal or other archaeological carbon prior to radiocarbon dating provides an opportunity to date specific materials, resulting in more accurate dates, while concomitantly providing paleoenvironmental data. This paper will assess the results of studies that have identified materials prior to dating and detail the modern carbon commonly identified in bulk soil samples.

Bulk soil samples are commonly used for radiocarbon analysis from archaeological sites, as well as in other areas of research including paleoflood studies, paleoseismology, and paleoclimatology. There are several reasons why bulk samples are used. Often, no apparent charcoal or other charred organic material is observed. A bulk soil sample charged at a conventional radiocarbon analysis rate is less expensive than a date obtained using AMS (accelerator mass spectrometry) radiocarbon analysis on a small amount of charred material. However, several problems exist in using bulk soil for radiocarbon analysis (Matthews 1980). These include 1) uncertainties surrounding the time between the formation of the material being analyzed and the point at which it was deposited, 2) determining the exact relationship between the datable material and the stratigraphy from which it was recovered, and 3) post-depositional contamination. It is better to submit a specific type of material for radiocarbon analysis (i.e. bone, charcoal, other charred organic material, shell, etc.) rather than a bulk soil sample.

Not only is it important to recover a specific type of material for dating, it is important to identify the material being dated. The separation and identification process must be performed under strict conditions of cleanliness to prevent contamination. Identification of charcoal and other charred plant material prior to radiocarbon analysis provides the opportunity to choose the material that would yield the best date possible. For example, a mixed charcoal sample might not yield as good a date as a single identified species. Identification of material is a recommended pretreatment strategy. Paleoenvironmental data and information concerning plant resources available to and utilized by the site occupants also can be obtained by identifying charcoal and other charred plant material prior to radiocarbon dating.

DISCUSSION

Although bulk soil samples commonly are used for conventional radiocarbon analysis by archaeologists and other researchers, they are very low in the "recommended sample material for radiocarbon dating" order. Trumbore (2000:43) notes that "¹⁴C dating is applicable to organic matter formed from photosynthetically fixed carbon within the last 50,000 to 60,000 years." When determining a potential usage for radiocarbon dating, both the cultural/contextual and the biophysical/biochemical characteristics must be considered for a particular sample type. Charcoal and charred organic material (including macrofossils and bone) are considered the most reliable type of sample material for radiocarbon analysis. When an insufficient amount of charcoal or charred organic material is available for dating using conventional decay-counting methods, an attempt is made to obtain a date on the bulk organic matter found in a soil. "Radiocarbon activity of soil organic fractions is extremely variable and the usefulness of using

such values to infer age in archaeological applications is generally quite limited except under special conditions" (Taylor 1987:62). One of these conditions includes using bulk samples collected from buried soils that are beyond the range of bioturbation. This would limit the input of organic material and restrict the potential for contamination.

Bulk soil samples are not recommended for radiocarbon analysis because a soil sample can incorporate either old or modern carbon depending on environmental conditions, the type of material, and the degree to which the sample is closed to contamination. Older material can be eroded and reworked or incorporated into younger deposits. Soils also are noted to receive continual input of new carbon (Birkeland 1999; Hsieh 1992, 1993). Younger material is commonly introduced through bioturbation such as insect, earthworm, or burrowing mammal activity. Seeds, leaves, and grasses often are carried into the subsurface as foodstuffs and bedding, and these animals introduce fecal material into the soil. In addition, some seeds have special features that allow them to work their way deep into the ground. Erodium cicutarium (filaree) has a corkscrew-shaped awn that drives the seed into the ground. This species was introduced into central California in the late eighteenth century by Spanish missionaries. Erodium seeds have been found more than a meter below the ground surface in central California archaeological sites known to be several thousand years old (G. J. West, personal communication, 1997).

To illustrate the fact that a bulk soil sample often will consist of a variety of material, one bulk sample collected during a paleoflood study from a stream terrace along Lost Creek in northeast Utah was submitted to a "bucket float" process used to examine archaeological macrofloral samples (Puseman 1997). The floated sample was examined to determine the material present after the sediment smaller than 0.25 mm in size was removed. From the original 2.3 liters of sediment present, a light fraction weighing 24.52 g was recovered. Of this amount, less than gram of charcoal and other charred plant material was recovered (Table 1). The sample contained charred and uncharred seeds, numerous uncharred rootlets from modern plants, four identified charcoal types, a piece of animal tooth, a few uncharred bone fragments, insect chitin fragments, mollusk and snail shell, and sclerotia. Sclerotia are the resting structures of mycorrhizae fungi, such as Cenococcum graniforme, that have a mutualistic relationship with tree roots. They are found with a variety of coniferous and deciduous trees including Abies (fir), Juniperus communis (common juniper), Larix (larch), Picea (spruce), Pinus (pine), Pseudotsuga (Douglas-fir), Acer pseudoplatanus (sycamore maple), Alnus (alder), Betula (birch), Carpinus caroliniana (American hornbeam), Carya (hickory), Castanea dentata (American chestnut), Corylus (hazelnut), Crataegus monogyna (hawthorn), Fagus (beech), Populus (poplar, cottonwood, aspen), Quercus (oak), Rhamnus fragula (alder bush), Salix (willow), and Tilia (linden) (McWeeney 1989:229-130; Trappe 1962).

Because the organic matter in soils is a mixture of materials of different ages and because the proportions of old and modern carbon incorporated into subsurface deposits are unknown, radiocarbon dates obtained from bulk soil samples represent composite ages (e.g. an average age for all of the organics in the sample). Depending on the number of factors that control the accumulation and decay of organic matter in a given deposit, the proportions of young to old carbon can be highly variable and large uncertainties in the measured ages become inherent. Because of these large uncertainties, bulk ages are questionable at best, and measured ages might not accurately represent the true age of a deposit. Contamination of a bulk sample with younger carbon has a greater effect on the resulting age than does contamination with older carbon (Polach et al. 1981, Rosholt et al. 1991). Studies by Andrews and Miller (1980) demonstrate that addition of only 5 percent modern carbon into a sample can give a true age of 20,000 years an apparent age of 16,500 years, and give a true age of 5,000 years an apparent age of 4,650 years. When 20 percent modern carbon is introduced, a true age of 10,000 years gives an apparent age of about 7,000 years (Figure 1).

The identification of specific material to be dated is particularly advantageous and allows the researcher or archaeologist to know precisely what material is submitted for radiocarbon analysis. When dating organic material, charcoal and other charred plant remains that have been specifically identified can help resolve issues concerning stratigraphic relationships between the sample and the stratum from which it was collected. For example, in fluvial deposits the identification of local riparian flora versus distant or exotic species can be particularly helpful in interpreting the deposition context. More accurate dates also can be obtained by submitting only specific types of charcoal or other charred plant material for dating. It would be preferential to date a local species rather than a foreign one, to date a single species rather than a mixture of several types, and to date the plant type with the shortest life span, such as dating a charred corn kernel rather than wood charcoal, or dating charcoal from a shorter-lived shrub rather than a longer-lived tree. Identification of the material also is a recommended pretreatment strategy prior to radiocarbon analysis. Taylor (1987:41) notes that "whenever possible, the proper scientific nomenclature for species of plant and animal sample material should be obtained even if the fragmentary nature of the sample permits only genus or even family level designations."

\When the amount of identified material for radiocarbon analysis is small (less than one or two grams for charcoal), the accelerator mass spectrometry (AMS) method is used for obtaining a date. Conventional radiocarbon analysis is based on the production, distribution, and decay of ¹⁴C, the radioactive isotope of carbon, and involves measuring the residual content of ¹⁴C present in a sample. A very small fraction of the ¹⁴C atoms present in a sample actually are measured. Alternatively, AMS radiocarbon analysis involves acceleration of ¹⁴C atoms in the form of ions to higher energies in particle accelerators, separation of ¹⁴C ions from other isotopes and molecules, then counting the individual ¹⁴C ions present. AMS radiocarbon analysis can be done using a much smaller amount of material than conventional radiocarbon analysis--as little as 5 milligrams of charcoal. The AMS methodology produces a more precise date than conventional radiocarbon analysis due to smaller analytic laboratory errors. For this reason, some researchers will choose AMS radiocarbon analysis even when enough material is present for a conventional date. Similar precision can be achieved with extended counting (measurement) time during conventional radiocarbon analysis, although it still requires a greater amount of the material being dated than the AMS method.

One example illustrates why taxonomic identification of macroflora is important before radiocarbon dates are determined. The case involves nine charcoal radiocarbon samples from features in a unit pueblo in northern New Mexico (Roomblock S at site LA 20266). Sediment from these same features was analyzed for macrofloral remains; however, the charcoal submitted for radiocarbon analysis was recovered prior to submission of the macrofloral samples for flotation. The radiocarbon dates obtained from these charcoal samples ranged in age from the middle of the Pueblo I period to after abandonment of the area at the end of the Pueblo III period and were not considered useful in determining the time of occupation. These charcoal samples were not identified prior to submission for conventional radiocarbon analysis. The macrofloral record from the same features that were dated contained a variety of remains (Table 2, Table 3). Many of the samples contained charred corn cupule and/or kernel that could have been submitted for AMS radiocarbon analysis. The charcoal record often consisted of a variety of charcoal types, including longer-lived pine and juniper, as well as shorter-lived shrubs such as sagebrush.

Saltbush (Atriplex) charcoal also was present in these samples. Most woody plants have a C3 photosynthetic pathway, except for saltbush, which has a C4 photosynthetic pathway. The tissues of C4 plants have slightly more ¹³C in their tissues than C3 plants have. The average d¹³C value for C3 plants is , -24� PDB, whereas C4 plants have an average d¹³C value of -14.5� PDB. The presence of saltbush charcoal in a mixed charcoal assemblage will change the d¹³C value by several per mil depending upon the percent of saltbush present. Consequently, to correctly determine the carbon isotope composition of the C3/C4 plant mixture, the d¹³C values must be measured directly on the gas resulting from the combustion of the mixed species sample. The amount of correction will depend on how much saltbush charcoal is present. When obtaining conventional radiocarbon dates, it is even more crucial to obtain this value because the error produced is doubled. When dating a mixture of charcoal, especially if saltbush is known to be present or expected to be present and if conventional radiocarbon analysis is used, the d¹³C value should always be determined at the time of the radiocarbon analysis to correct the radiocarbon age reported for sample (Darden Hood at Beta Analytic, personal communication, September 14, 2000; Tom Stafford at Stafford Research Laboratories, personal communication, April 2001). The d¹³C value is automatically determined when samples are submitted for AMS radiocarbon analysis, but not when samples are submitted for conventional radiocarbon analysis--it must be requested. The radiocarbon dates obtained from Roomblock S at site LA 20266 might have been more useful if the charcoal had been identified prior to submission for radiocarbon analysis to determine the best charcoal type to date, if the d¹³C value had been determined, and/or if the charred corn had been submitted for AMS radiocarbon analysis rather than charcoal for conventional radiocarbon analysis.

Identification of charcoal and other charred organic material prior to submission for radiocarbon analysis also can provide paleoenvironmental data and/or information concerning use of individual plants. Assuming that subsurface disturbance is not too great, charred organic material from non-cultural deposits most likely represents plants growing in the area that were burned in a past fire. Charred plant material from cultural contexts most likely represents resources utilized by the site occupants. Identification of this material provides information concerning plants available to the occupants at the time of occupation, as well as specific plant resources utilized.

SUMMARY

In archaeological applications and other areas of research, it is best to submit identified material, especially charcoal and other charred organic material, for radiocarbon analysis rather than bulk soil samples. Bulk soil samples can contain reworked older material and/or introduced younger material; therefore, bulk soil samples might not accurately date the deposit. Wood charcoal and charred organic material, including bone, are believed to be the most reliable types of samples for radiocarbon analysis. Determining the d¹³C value, using the AMS methodology, and/or using extended counting produces a more precise date. The material to be dated should be identified prior to submission for radiocarbon analysis. This can aid in determining the best material for dating. It also can provide information concerning plants present in the past environment and resources available to and utilized by the site occupants.