old. Old carbon may contaminate a sample in two ways. First, detrital carbon, such as that from coal of pre-Quaternary age which contains no carbon-14 or that from humic soil material which may have an age of several thousand years, may be incorporated in a sample. Second, plants and animals living underwater can incorporate lower activity carbon from CO2 in the water either because the water is old or because it has bicarbonate from old rocks. For marine mollusks from the North Atlantic, the age of CO2 in ocean water increases ages by 400 to 750 yr (Mangerud and Gulleksen, 1975).
Contamination with Young Carbon Contamination with recent carbon can alter ages greatly. Samples whose ages are beyond the range of carbon dating (>75 ka) that are contaminated with only half a percent of recent carbon will yield an age of about 40 ka (Figure 13.5, lower right half). Even the most exacting of analyses can be in error: a carbon-14 enrichment age of 71 ka (representing only 0.014 percent of the original carbon-14 activity) was thought to date the Salmon Springs glaciation of northwest Washington State (Stuiver et al., 1978) until associated ash deposits were fission-track dated at 700 and 800 ka (Easterbrook et al., 1981).
Contamination with younger carbon may be responsible for many of the finite carbon-14 ages in the 20- to 70-ka range. Contamination of carbon samples with recent carbon produces effects that are not generally appreciated (Figure 13.5). Examination of this nonlinear effect also should provide a caution about assuming that “consistency” in age results necessarily is an argument for the validity of carbon-14 ages. For example, the consistency of many dates falling in the 25- to 40-ka range may only reflect contamination of samples older than about 50 ka with the equivalent of 0.5–2 percent of recent carbon (Figure 13.5). Several hundred carbon dates in the 25- to 40-ka range have been obtained from coastal deposits in the eastern United States. Many researchers (referenced in Bloom, 1983) have concluded that these samples date a mid-Wisconsin high stand of sea level, in part based on the apparent “consistency” of a large number of dates in this age range. After thorough analysis of this problem, Bloom (1983, pp. 215–218) concluded that these ages are invalid. Amino-acid racemization studies on mollusks asssociated with samples yielding caron-14 dates in the 25- to 40-ka range also indicate that the carbon-14 dates are erroneously young (Belknap, 1984).
Samples may be contaminated by in situ additions of younger carbon. Soil carbonate, mollusks, or corals are particularly susceptible to addition of young carbon, especially when subject to repeated wetting or drying. Contamination by microorganisms incorporating young carbon may occur either before or after a sample has been collected. Marine cores stored for 5 yr were found to be contaminated by enough terrestrial bacteria to account for 5–10 percent contamination with modern carbon (Geyh et al., 1974). In-place contamination of samples is not well studied, but it may be possible if microbial activity consumes CO2 from air or from water with a higher carbon-14 activity than the age of the sample. For example, methanogens participate in terminal stages of the degradation of organic matter, living on and presumably incorporating carbon dioxide and hydrogen produced by anaerobic bacteria into their tissue (see Maugh, 1977).
Extra care should be taken in order to minimize contamination with recent organic material during sampling and sample preparation. In addition, samples should be examined for visible contamination, particularly by roots. Removing modern roots will not, of course, remove contamination caused by older, largely decayed roots. Common sample pretreatment before carbon-14 dating removes base-soluble fulvic acids and acid-soluble humic acids, leaving a residue called