1.2 Stable isotope signals of aquatic foods

Microscopic and biomolecular analyses provide qualitative evidence of the function of pottery; isotopic analyses produce quantitative data, which can provide information about the exploitation of different food resources to the extent that there is variation in isotope values between foods. In palaeodietary studies, the stable isotope ratios δ13C and δ15N in human bone collagen are often used to estimate the contribution of aquatic food sources within the diet, as both δ13C and δ15N are generally higher in marine fish than in terrestrial foods; δ34S is also increasingly used for this purpose. Freshwater fish also tends to be enriched in δ15N, while δ13C values are often lower than in terrestrial foods, so in principle stable isotopes can be used to estimate the dietary contribution of freshwater resources. In animals, isotopic fractionation (differential retention of different isotopes) leads to higher δ15N values at each level in the food chain (which is why the longer aquatic food chains can be distinguished isotopically), but δ15N is relatively consistent in different tissues. Values of δ13C vary significantly between bone collagen and other tissues, but when collagen δ13C values alone are compared, only minor isotopic enrichment from dietary species to consumers is evident.

Many cooking pots retain amorphous black surface deposits that are usually described as burnt food crusts. Very rarely do such deposits contain identifiable remains (e.g. fish scales), but starch grains and phytoliths show that terrestrial plants were sometimes among the ingredients (Saul et al. 2012). Lipids extracted from food crusts can also be attributed to foods cooked or stored in the same pots, and may contain biomarkers of aquatic or terrestrial animal products (Craig et al. 2011). Isotopic techniques could be used to analyse food crusts, provided that fractionation during cooking, charring and burial is insignificant. In addition, the concentrations of C and N, which are often measured during stable isotope analysis, can be used to interpret food crusts to the extent that they vary between ingredients and are not altered by cooking or diagenesis.

Experiments with modern ingredients show that measurements of δ13C and δ15N and C and N concentrations in burnt food crusts are reasonably consistent with values expected from those in the original uncooked ingredients, indicating that charring does not destroy the isotopic and elemental abundance signals of different food groups (Philippsen 2013a; Philippsen et al. 2012; Philippsen in press). Moreover, the pattern of stable isotope data seen in a large number of archaeological food crusts from the south-west Baltic region suggests that isotopic signals of food ingredients are not significantly altered during burial (Figure 1): food crusts from coastal sites often give 'marine' isotope values (enriched δ13C and δ15N), while those from inland sites often give 'freshwater' values (depleted δ13C, enriched δ15N). The clear division between coastal and inland sites is also evidence that fish were cooked regularly, since terrestrial food resources would not be isotopically distinct.

Figure 1
Figure 1: δ13C and δ15N measurements on bulk food-crust samples from late Mesolithic and early Neolithic pottery, sites in eastern Denmark and Schleswig-Holstein, Germany (based on Craig et al. 2011, fig. 2)

Biomolecular analysis confirms that freshwater resources were prepared in pottery. At the inland site of Ringkloster in Jutland, for example, 14% of the vessels had aquatic (freshwater) lipid biomarkers (Craig et al. 2011). Pottery from the Åmose contained aquatic biomarkers (Craig et al. 2007), and the isotopic values of the fatty acids suggest that the residues were derived from freshwater fish (Heron et al. 2007). Biomolecular analysis underestimates the proportion of pots used to cook fish, however, as distinctive biomarkers are not always preserved, and biomarker presence/absence data provide little indication of how much fish was cooked relative to terrestrial food resources. Furthermore, pottery from sites in the Trave and Alster valleys has not been analysed by these methods.

Human remains are scarce in this region during the late Mesolithic, but finds of freshwater mollusc shells as well as very negative δ13C values in human and dog bones suggest an especially high dependence on freshwater food during the Mesolithic-Neolithic transition (Fischer et al. 2007).