Early industrial complex was built for sailors, not locals
A large Roman water-mill collection in France didn’t supply its adjacent city, as assumed. Andrew Masterson reports.
A Roman water-mill located in what is likely to be one of the earliest industrial complexes in human history was not, as previously thought, critical for supplying the local population but had a more seasonal function, isotope analysis has shown.
The complex, which in its day comprised eight mill units sporting 16 wheels, is called Barbegal and located near the town of Arelate in France. It dates from the second century CE and was built under the aegis of the Roman occupation. Its foundations and other elements remain well preserved today.
Originally discovered in the 1940s, the Barbegal site is regarded as one of the earliest extant examples of how Roman industry developed in response to improvements in engineering. Writing in 1996, French archaeologist Phillippe Leveau characterised it as an example of “the role of technological change in the transition from Antiquity to the Middle Ages”.
The site was also subject to continuous improvement. A report in 2007 suggested that the range of millstones found there meant that Roman builders tested several designs to “take advantage of the increased power of hydraulically driven mills”.
The mill had an estimated production capacity of 25 tonnes of flour per day – enough to feed the Arelate population of 27,000.
Impressive though Barbegal undoubtedly was, it was not an entirely indispensable piece of Roman infrastructure, a team headed by Gül Sürmelihindi from Germany’s Johannes Gutenberg University now suggests.
In a paper published in the journal Science Advances, the researchers report the results of isotope analyses which clearly establish that the mill was only in use for part of each year.
The scientists collected 142 carbonate deposits from the complex, taking them from two distinct parts of the environment. One set was scraped from large carbonate slabs that that formed in the nine-kilometre-long channels through which water ran to turn the mill wheels. The other was collected from deposits that had formed on one of the wheels itself, left behind following decay.
Stable isotope analysis of the deposit samples showed that they were laid down cyclically rather than continuously. The mill wheels ground to a halt, Sürmelihindi and colleagues concluded, in late summer and remained stationary through autumn.
This unexpected pattern of interrupted activity was inconsistent with earlier assumptions about Barbegal’s purpose. It could not have been used to supply the needs of the people of Arelate, the researchers state, because milled grain spoils quickly and an annual shutdown would have badly impacted the locals’ food supply.
The speed at which the flour would have deteriorated also excludes the idea that it was transported for use elsewhere. There is no evidence, either, of nearby troop garrisons that might have needed feeding when not off somewhere campaigning.
The real purpose of Barbegal, the researchers surmise, was for the production of “ship’s bread”, known as Panis nauticus, a ferociously unpalatable double-baked hardtack that could be loaded onboard vessels and stored throughout long journeys.
The theory derives some support from geographical evidence: the Roman-era ports of Arles and Fossae Marianae were nearby. Other evidence arises from well attested maritime records. Roman shipping typically stopped in late autumn – meaning that an annual milling halt in late summer is exactly consistent with demand.
The distribution of watermill complexes during the Roman period is still poorly understood. Sürmelihindi and colleagues say the Barbegal findings mean that searches around ancient ports might well uncover currently unknown examples.