Shellfish obtain carbon for shell building from two sources. Organic or metabolic carbon comes from marine or terrestrial plant material, including diatoms and protozoa, kelp and phytoplankton, or humus and peat added to estuaries and rivers (Keith, Anderson and Eichler, 1964; Tanaka, Monaghan and Rye, 1986; Little, 1993). Inorganic carbon comes from ocean water bicarbonate (which has a d13C value of close to -2 per mille wrt PDB), atmospheric carbon dioxide (-7 per mille) and fresh water bicarbonate (usually less than -8 per mille) (Keith et al., 1964). Dissolved, metabolised or oxidised CO2 from terrestrial plant and humus reservoirs affect the organic sources for molluscs in food webs and the bicarbonate pool (Keith et al.., 1964). CO2 from the respiration of land plants and the oxidation of humus in soil reaches the groundwater and is fed into lakes and rivers. Its effect is to lower the average d13C content of the bicarbonate that exists in continental waters (this is known as the ground water effect). The oxidation of humus within river and lake sediments also releases CO2 for uptake by mollusc populations. The sources of carbon from terrestrial and marine systems are of variable age and d13C, so an analysis of the extent of fractionation within them is vital for radiocarbon dating shell from these environments.
Keith et al.. (1964) measured typical d13C values in shell from marine, lacustrine and riverine environments. They showed that the value of d13C is influenced primarily by the isotopic composition of the shells' immediate environment (Keith et al., 1964; Head, 1991). Marine shells possess a delta13C between +4.2 and -1.7 per mille wrt PDB, river shells between -8.3 and -15.2 per mille and shells from large clear lakes between -2.4 and 6.0 per mille (Keith et al., 1964).
One of the principal advantages of dating shell is that the 'lag' or dilution effect of the oceanic reservoir upon the Suess 'wiggles' means that the calibration curve is smoother than for terrestrial samples. It is therefore possible to obtain more precise calendar ages.
There are a number of uncertainties for dating shell. First, there has been uncertainty over exact reservoir corrections. Second, there are local errors of varying magnitude introduced by dissolved bicarbonate from calcareous rock formations. Third, there is the problem of upwelling.
During the early years of the radiocarbon method, discrepancies were noted between samples of charcoal and shell which were excavated from the same stratigraphic layers of the same site and were therefore assumed to be the same age. Shell dates were older than the terrestrial samples from between 300 and 500+ years. This is due to the reservoir effect and is caused both by the delay in exchange rates between atmospheric CO2 and ocean bicarbonate, and the dilution effect caused by the mixing of surface waters with deep ocean waters (Mangerud 1972). Deep ocean water is depleted in radiocarbon so it reduces the activity of surface waters. This is termed the upwelling effect. Radiocarbon spends a comparatively short time in the atmosphere. According to Taylor (1987) the residence time of C14 in the atmosphere ranges between 6 and 10 years. On the other hand, residence times of C14 in the ocean may be up to 1000's of years. The effects of the upwelling zone upon shellfish 14C ages is particularly well documented along the coasts of California and Mexico (Berger, Taylor and Libby, 1966; Druffel and Williams, 1991). The hard water effect refers also to the dilution of radioactivity in marine reservoirs but this time the diluting factor is dissolved bicarbonate originating from geological deposits of infinite age. Usually, the rocks are carbonaceous sedimentary rocks such as limestone, sandstone or mudstone.
Shell material may undergo a post-depositional recrystallisation which may affect its true radiocarbon age. This is described in the pretreatment section of these pages.