Roman water infrastructure in post-Roman Britain. Part 1: Roman infrastructure

Reviewing Pompeii and the astonishing engineering in the Roman water supply system serving the Bay of Naples prompted me to ferret out my notes from when I was researching Roman water supply in Britain. This was turning into a long post even by my standards, so I've split it into two. Conveniently, it more or less divides along the lines of 'evidence' (part 1) and 'interpretation' (part 2).

Roman Britain never had the magnificent aqueducts found in Roman Gaul, such as the Pont du Gard serving Nemausus (modern Nimes) (the top row of arches carry the aqueduct channel; scroll down the page for close-up photos) or the aqueducts serving Lugdunum (modern Lyon) with their sophisticated closed siphon systems for crossing deep valleys (I remember being fascinated by the models of these aqueducts when we visited the Musee Gallo-Romaine in Lyon on the way back from the Alps, and wishing I had enough French to buy the French-language survey publication. This description of the Gier aqueduct will give you an idea; the idiosyncratic English needs a little concentration but it's well worth it), or the aqueducts serving Rome itself.

This probably reflects at least three major factors: climate, terrain and degree of urbanisation. In most districts of Britain, water can be relied on to fall out of the sky on a regular basis and the water table is within reach of wells, so there would be less need to bring water long distances from mountain sources to arid plains. On the whole, Britain doesn't have the deep gorges and valleys of the Massif Central in France, so aqueducts can be routed unobtrusively round contours and don't need spectacular arcades. Roman Britain was never as urbanised as Roman Gaul or Italy, and even the largest British cities such as Londinium (modern London) and Verulamium (modern St Albans) never reached the size of major Gallic cities such as Lugdunum (modern Lyons) - which was a lot bigger than Paris at the time, as the museum in Lyon is keen to point out. Thus, there would have been less need for massive infrastructure to bring large quantities of water to major population centres.

Nevertheless, there is evidence of sophisticated water engineering from several towns in Roman Britain. Perhaps the most spectacular is the water-lifting apparatus excavated in London in 2001 (4th paragraph of the ‘City’ section in the link). Two large timber-lined wells were discovered, each containing a chain of wooden buckets that lifted water continuously from the well to a nearby cistern, from which it could have been used to supply the nearby baths and/or distributed across the city. I saw a Time Team TV programme on this discovery soon after the excavation and it was suggested that the bucket chain could have been powered by a treadmill, either human- or animal-powered, and could have supplied enough water for several thousand people. Tree-ring dating dated the first well to AD 63 and showed it had partly collapsed in AD 71. The second well was dated to the early second century (possibly a replacement for the first?) and appeared to have been damaged or destroyed by fire in the late second century.

Archaeological excavation in Colchester has shown that the Roman town had a pressurised water supply from its earliest days. Wooden water mains consisting of straight sections of bored-out log connected by iron collars have been found. A ‘water-works’ excavated in 1928 had a room (or tank?) that would have filled with water and some indication of a lifting system, perhaps a water wheel (or a bucket chain like those in London?) to raise water to a header tank for distribution. Surplus water flowed out of the tank down a large overflow drain, about 2 feet high judging from the sketch, and into the town ditch. The water source is thought to have been a timber aqueduct that came in over the city walls, but this is not known for certain (Crummy 1997).

In York, environmental analysis of the silt in the Roman sewer system excavated under modern Church Street found large amounts of tree pollen from mixed deciduous woodland, including species that generally favour limestone soils, and pollen from moorland plants such as heather and pine. This pollen is unlikely to have come from the city itself, and is consistent with a water supply brought in by aqueduct from upland areas some distance from the city (Buckland 1976). The report suggests one aqueduct coming from a well-wooded area with moorland on higher ground nearby; I wonder if there might have been two aqueducts from different locations, one from moorland and one from limestone country? No traces of an aqueduct have (yet) been found at York, but as noted above, British aqueducts don't need dramatic arches for crossing dramatic Mediterranean gorges. Where identified, Roman aqueducts in Britain were typically large ditches containing a ceramic or concrete water pipe, laid along natural contours so that the water flowed by gravity. Such a structure wouldn't be obvious above ground and is only likely to be discovered if someone happens to dig in just the right place. Lead water pipes in stone-lined trenches, and a substantial stone street fountain, show that at least some of the Roman city had a piped water supply. Timber-lined Roman wells have also been excavated in York (Ottaway 2004).

How much of this infrastructure survived into the post-Roman period and for how long? See part 2.....

References:
Buckland PC, The environmental evidence from the Church Street Roman sewer system. York Archaeological Trust, 1976.
Crummy, P. City of Victory, Colchester Archaeological Trust, Colchester, 1997.
Ottaway P. Roman York. Tempus Publishing, 2004.