Iron Age origins of European Alpine dairy farming

Chemical analysis of organic residues from pottery reveals lipids

Dairy produce from the high Alps is today of enormous economic and cultural importance to the region. The recent history of dairy farming is well-documented, but there is very little in the way of archaeological evidence to attest to its origins. Transhumance (seasonal migration of livestock between pastures) leaves few traces, and the problem is compounded by acidic soils that lead to the deterioration of faunal remains.  

Milk production in lowland settings in Europe is documented from the Early Neolithic, and ceramic sieves for separating curds and whey are evidence for the production of cheese and fermented milk products by populations that were still predominantly lactose intolerant. In the Alpine lowlands, there is evidence for dairy farming in the form of lipid residues on pottery vessels from around 6,000 BC. However, it was at this stage part of a generalised mixed farming economy that also included meat production.  

From around 3000 BC, it has been speculated that dairy farming intensified as there was a greater reliance on ‘secondary’ animal products such as wool and milk, and greater utilisation of poorer and less accessible land. As noted above, the limited supporting archaeological evidence is not unexpected. Seasonal occupation of high-altitudes intensified from around 2500 BC to 1000 BC and large dry stone enclosures were constructed during this period. They are thought to be livestock pens, but the scarcity of artefacts or faunal remains means that there is no definite clues as to their function. The only ceramics to have been recovered from these sites are small, highly fragmented potsherds.

To investigate further, researchers obtained 30 securely-dated potsherds from six highland archaeological sites of the Engadin region of southern Switzerland. The sites dated from 5000 to 1000 BC. The region is typical of the alpine environment, with valley bottoms above 1,000 m above sea level and high seasonal pastures ranging from around 2,000 m to 2,800 m above sea level. Five of the sites were more than 2,000 m above sea level. They include early Neolithic and Bronze Age rock-shelters and a later Iron Age stone enclosure and hut.

Lipids were successfully extracted in varying amounts from the potsherds, with all the Iron Age pots yielding much higher quantities. Analysis of the lipids using gas chromatography mass spectrometry revealed profiles typical of degraded animal fats. Several Bronze Age and Iron Age potsherds contained ketones with chain lengths consistent with heating of saturated fatty acids, suggesting that the pots were used for heating animal products.

Carbon stable isotope analysis was then applied to fatty acids obtained from 28 of the potsherds. The milk of ruminants and to a lesser extent carcass fat of ruminants, is depleted in 13C relative to other fatty acids. Values obtained were then compared with those obtained from dairy, ruminant and non-ruminant sources. It was found that values for lipids obtained from the Iron Age potsherds were consistent with dairy products. By contrast, the earlier Neolithic and Bronze Age potsherds yielded values consistent with ruminant and non-ruminant animal fats. Dairy farming was identified at all of the Iron Age sites included in the study.

Thus it appears that while low-altitude dairy farming was present from the Neolithic onwards, specialist Alpine dairy farming was a later development. Iron Age alpine pastoralists would have had to face adverse and unpredictable weather, and a significant reduction in the yield and quality of milk. It is likely that pressure on lowland pastures and an increased demand for alpine cheese were motivating factors. This in turn was probably triggered by social and economic changes, deterioration of the climate, and demographic growth during the Late Bronze Age and the Early Iron Age.

Reference:x
Carrer, F. et al., Chemical Analysis of Pottery Demonstrates Prehistoric Origin for High-Altitude Alpine Dairying. PLoS One 11 (4), DOI:10.1371/journal.pone.0151442 (2016).

x