More information : Roman Aqueduct at Lanchester.
The first reference to the aqueduct serving Lanchester Roman fort (NZ 14 NE 4) is by Hodgson in 1822 (a), who enclosed a survey of its course made `several years ago' by White and Fenwick. This plan shows a north and south aqueduct, the former extending eastwards from the reservoir (Low Dam - NZ 14 NW 27) to skirt the northern slopes of Humber Hill, and the latter from a point (at about NZ 1175 4652) just north of a cottage now called Forest Lodge; these watercourses were shown to converge at NZ 1509 4645, and continue as one to a reservoir immediately south-west of the fort (NZ 14 NE 3). Hodgson stated that in many places, especially in the woods on the Woodlands estate, both lines are `as visible as the day they were made'.
The greater part of the feature was no longer visible in 1857 when the Ordnance Survey surveyed and published the 1st edition 6" map. Only the sections within the plantations, specifically 1.8kms of the north channel and 930m of the south channel within Sawmill Wood.
In 1937, Steer(b) sectioned the upper part of the north aqueduct, the Low Dam and one point on the south aqueduct. A series of levels taken at the dam and along the north channel indicated that the course for about 1km was uphill; Steer postulated a syphon system here, the water being conveyed by pipes, but no trace of these were found.
Valuable fieldwork was undertaken in the early 1970s by Reed and Austin (c) who suggested an alternative course for the aqueduct. They were able to trace the north watercourse from Low Dam eastwards as far as Long Edge Lane (NZ 1274 4643) as `twin sand tracks' in the plough, from which point it was destroyed by opencast operations until it reappeared alongside the road at Colepike Farm (NZ 1471 4598). The authors considered that the north branch which looped around the north side of Humber Hill was not fed by Low Dam as indicated on plan in Hodgson, but was a feeder in its own right collecting water from a number of springs. About 1970 they excavated two trenches along the line of the north watercourse (Section B see NZ 14 NW 35) and also a section on the south aqueduct where a minor V-shaped ditch was revealed, which Austin considered to be less convincing as a Roman work.
Since the work of Steer in 1937, the majority of the eastern part of the system, including the loop north of Humber Hill has been obliterated by former opencast mining, and in Woodlands Park the trees have been felled and the land turned over to agriculture.
The revised line described by Reed and Austin, which is a more direct route utilising a broad saddle in the vicinity of Woodlands Park, is convincing, but the `sand tracks' are no longer visible making verification impossible. Mr Austin (d) pointed out on the ground the position of his excavations (see above) and also the location of surviving sections of the system.
The remit of the RCHME Newcastle as part of the Durham SAMs Project was to examine and survey the portions of the aqueducts which were extant or scheduled, and also to check the anomaly of the uphill course east of Low Dam noted by Steer (see below). The north channel was located in four separate areas (Section A - NZ 14 NW 34, Section B - NZ 14 NW 35, Section C - NZ 14 NW 36, Section D - NZ 14 NW 33) and the south channel (Section E - NZ 14 NW 37)
The OS 6" 1st edition map of 1857 shows and annotates in two places (NZ 1404 4670 and NZ 1451 4672) the remains of a `North Roman Aquaduct'. There is no evidence of the latter, but vague traces of a ditch can still be seen by the road side at the former location; these points are at least 250m south of Hodgson's putative line, and if they do represent the true course of the North aqueduct, then Hodgson's survey must be suspect in this area at least. The remains of the ditch by the road are about 242m OD, some 15m higher than the upper channel as it leaves Low Dam, so a connection between the two makes no sense.
The problem of the uphill course of the aqueduct from Low Dam eastwards for 1km, highlighted by Steer in 1937, has not been resolved satisfactorily. West of section A (NZ 14 NW 34), levels
were taken using the Wild total stations theodolite and EDM at those points where the course of the channel, as depicted on OS 25" 1975 as a pecked line, crosses modern field boundaries. At NZ 1115 4715 the channels are at 227.86m OD and 224.34m OD; thus the upper leat continues to rise (1.3m), whereas the lower, more or less maintains the same level. The channels are shown to converge at NZ 114 470, close to a dried-up stream gully. The next level at NZ 1154 4785, just east of the convergence, is 227.35m OD, and at the excavation by Reed and Austin by the roadside at NZ 1185 4709, was 223.69m OD. Close to the former position there is a low ploughed down stream gully running north-south; the stream is now culverted but is shown on the OS 6" 1st edition map. It is possible that the previously mentioned upper channel was in fact a feeder from this stream to the dam.
The lower channel drops from 224.54m OD at the dam to 224.34m OD at the wall crossing (NZ 1115 4715). However, at the wall crossing at NZ 1154 4703 it is 3m higher. All these heights are based on the assumption that the Ordnance Survey course is accurate, for nothing is visible on the ground. From this last point towards the fort, various levels were taken where the aqueduct was visible on the ground or where its position had been pointed out by local informants (a). The fort itself lies at approximately 182m OD and so the total fall from the dam is approximately 45m.
If the proposal by Steer of a syphon is accepted, then Low Dam would have to have been at least 4m higher than its present highest point. The ground configuration is such that this is feasible, but there is no trace of wings to the dam. A level was also taken to the headwaters of the south aqueduct at NZ 1214 4642. At this point it is 204.68 OD, some 23.5m above the highest point of the fort. (1)
NZ 108 474 - NZ 149 462. Remains of Roman aqueduct. Scheduled. (2)
Documented. (3) |