Processing of Milk Products in Pottery Vessels through British Prehistory

Article excerpt

Detecting dairy fats in pottery vessels Early farming communities exploited domesticated animals for a wide range of products, including their meat, fats, milk, wool, skins, etc. The exact timing of the utilisation of domesticated ruminant animals (cattle, sheep and goat) for their 'secondary products' (including non-carcass products such as milk) is unknown, but is believed to have occurred a substantial period of time after the initial domestication (Sherratt 1981, 1983, 1997). Until recently, the evidence for dairying in prehistory was principally restricted to perforated pottery vessels (Bogucki 1984) which are presumed to be 'cheese-strainers', or through the analysis of faunal remains enabling herd structures to be reconstructed (Payne 1973; Legge 1981) and classified as being dairy, meat or traction orientated (or indeed a combination of any of these economic strategies). Both these lines of evidence essentially provide indirect support for dairying in prehistory. A more direct, chemical method has been developed to detect dairy fats preserved in walls of pottery vessels used in the processing of dairy products (Dudd & Evershed 1998).

Chemical criteria for the detection of degraded dairy fats include the identification of triacylglycerols (TAGs) displaying characteristically wide distributions (with the number of carbon atoms varying from 40 to 52). However, this method can be problematic, since the TAG distributions from highly degraded dairy fats may transform to resemble those derived from degraded ruminant carcass fats (Dudd & Evershed 1998). An alternative method using stable isotopes has proven to be extremely robust. Differences in the [[delta].sup.13]C values of the principal fatty acids present in lipid residues, namely, palmitic acid ([C.sub.16:0]) and stearic acid ([C.sub.18:0]), reflect their origins in meat or milk ([[delta].sup.13][C.sub.18:0] values from ruminant carcass fats are c. 2.3[per thousand] higher than in the dairy fats: Dudd & Evershed 1998; Copley et al. 2003). Thus, by comparing the [[delta].sup.13]C values of fatty acids found in pottery to those of reference fats from modern animals raised on strict [C.sub.3] diets (isotopically comparable with the diet available to prehistoric domesticates; Copley et al. 2003), it is possible to classify the lipid residues to specific commodity group(s).

This method is enabling us to investigate the significance of dairying to pottery-using peoples in prehistory. We present here analyses of the lipids from potsherds from prehistoric Britain, and compare them with economic reconstructions based on the faunal evidence at each of the sites concerned.

Sites, pottery and methodologies

A total of 437 sherds were analysed from six Neolithic sites (Copley et al. 2005a; 59 sherds from Abingdon Causewayed Enclosure, 88 sherds from Eton Rowing Lake, 72 sherds from Hambledon Hill, 57 sherds from Runnymede Bridge, 70 sherds from Windmill Hill and 91 sherds from Yarnton Floodplain). A further 256 sherds were selected from four Bronze Age sites (Copley et al. 2005b; 59 sherds from Black Patch, 59 sherds from Brean Down, 69 sherds from Potterne and 69 sherds from Trethellan Farm). Lastly, a total of 237 sherds were sampled from four Iron Age sites (Copley et al. 2005c; 69 sherds from Danebury Hillfort, 54 sherds from Maiden Castle, 65 sherds from Stanwick and a further 49 sherds from Yarnton Cresswell Field). Thus, over 900 sherds from 14 sites were analysed from all three prehistoric periods in Britain (Figure 1). Inland sites were selected to avoid complications from isotopic variation due to marine effects.


The analytical protocol followed has been described in detail in earlier papers (e.g. Evershed et al. 1990; Charters et al. 1993; Dudd & Evershed 1998; Copley et al. 2003). Lipid concentrations are highest in rim or upper-body sherds, so these were selected wherever possible from vessel types likely to have been used for processing food. …