The Dig for the Missing Caisson

There have been 2/3 substantial digs: one in Sept 1997, a series in and around the site of the former pumping engine at the Bull's Nose, and the latest in August 2003.
We'll start with the latter.

2003 Dig near Lock 5

A masive excavator was brought in to dig a hole 30 ft deep in Caisson Field. Member Richard Hignett has been largely responsible for pin-pointing another possible site nr lock 5. Some of the evidence (Richard lists 16 points in his article in the "The Weigh-House" Autumn 2003. pp 14- 19)
1. The 1804 map was traced onto a 1970 map. The caisson tunnel then appears to be near to the railway line.
2. The tunnel would pass below the forebay of lock 5.
3. There was a suspicoius spoil heap near by (see next diagram), with enough room for a second caisson lock on top of it.
4.The 1804 map showed stop gates that were missing on the 1884 map. What were the stop gates protecting if not the caisson and its tunnel(s)? 5. Assuming the caisson had a drop of 46 feet, the various levels (level of the spoil heap and upper pound water level) match near-perfectly.

so a trench was dug...
1. The tunnel is not there but could have been . A rich vein of clay may have been used to fill-in the tunnel. The bed-rock under this clay is not orientated as expected and may represent the collapsed tunnel..
2. The spoil may have been used to fill the caisson exit cutting and raise the level of the ground below lock 5 in order to facilitate the building of lock 5

There are no further plans for any more excavations.






Excavations on a site deemed to be suitable because of irregularities in the Geophysical studies alas showed little, except infilling; chippings and blue-clay mainly. The good news is that proper archaeological examination of the whole site may be possible if support is granted from English Heritage. Watch this space.

For a detailed account of the dig and all the various findings please see The Weigh House mag. No. 21 (Autumn 1997). However, Mike Chapman's article on this is reproduced below.


Following last year's geophysical survey by Geo-Quest Associates on the south side of Caisson House (in what was once part of Caisson Field),and the discovery of an early map (c.1804) showing its location in the same area, the opportunity arose to test the new findings by means of a trial excavation. This area, not far from the chestnut tree identified long ago by Dr. Hugh Torrens as being the probable site of the Caisson, had already been subjected to various unproductive 'prospective digs' over the years, none of which had been recorded or analysed. On this occasion the excavation was to be carried out by hand under controlled archaeological conditions (though on a voluntary basis) by professional staff from Bath Archaeological Trust in collaboration with the Society's own works team. This project, directed by Marek Lewcun from the Trust, was given the backing of B&NES Built Heritage Department (managed by Mary Stacey who provided for the geophysical survey) with supervision by Bob Sydes, Council Archaeologist.

The geophysical survey had detected several 'anomalies' roughly in the area of the lock site as shown on the map, one of which appeared to be a pair of parallel walls not far below the surface (the geomagnetic equipment has a limited range of about 3 feet). Being of similar scale to the known dimensions of the lock, they were interpreted (with a 70 per cent 'confidence rating') as being the walls of the Caisson chamber. Next to them was a much larger interference which extended towards the present entrance gate to the field. It was interesting to compare these results with a pre-excavation survey carried out by Adrian Tuddenham using his own ground resistance equipment. Although in this instance no 'walls' were indicated, a similar large-scale disturbance was detected towards the gate.

A trench 1 metre wide and 11 metres in length was cut across the line of these 'walls' in a roughly north-south direction, later extended a further 6 metres northward to examine the large anomaly towards the gate. The southern end of the trench terminated a few metres north of the turf line of the old boundary wall of Caisson Field (removed some time after the field was subdivided by the construction of the GWR Limpley Stoke railway in 1910) which proved a useful reference point. It was at this end of the trench that the original ground level was confirmed to be much as it is today, but that from this point northward the whole surface was found to have been lowered by stripping off between 2 to 3 feet of subsoil and underlying natural clay. This had later been filled in with a mottled clay and levelled off with subsoil. Also encountered, crossing the trench, were traces of a recent cutting; possibly the result of a previous 'Caisson-hunt.

Working northward to the point where the 'walls' were expected, the lowered ground surface was found to be covered with a thick layer of dense blue clay (presumably canal puddling clay) containing large quantities of nails and other canal ironware. No walls were found, and it was immediately apparent that it was the large iron content of this clay which had given such a strong reading, suggestive of walls, during the magnetometer survey. Also found among the small finds in this layer was a George III half-penny clearly dated 1807. Half-pennies of this kind, with milled edge, were only minted in the two year's of 1806 and1807 and although somewhat corroded, this example showed little wear, indicating that it was lost soon after the latter date.

Some 12 feet further on, approaching the northern end of the trench, the blue clay increased in thickness to just below turf level. Immediately below the clay at this point, showing through the earlier ground surface, a band of packed chippings and mortar about 2 feet in width and vertically edged was found crossing the trench on an east-west axis. This band was not detected until a late stage owing to a number of large unworked stones and boulders which lay on top of it, and it was therefore not possible to determine its depth or course downward in the time available. Although a quantity of the blue clay mixed with lumps of mortar occurred for a further 4 feet, the band of chippings evidently marked the southern edge of a deep excavation containing a massive infill of rubble and stoney clay. The depth of this infill could not be ascertained, although its edge was followed downward a further 2 feet below the earlier ground surface to a maximum depth of 5 feet below ground level.

It was at this stage that it was decided to extend the trench northward to examine the stoney rubble and clay infill of the deeper excavation, a rough calculation of the width of the Caisson being made to ensure that the trench was of sufficient length to detect the opposite edge of the chamber (assuming that its southern edge existed in the vicinity of the band of chippings). It was found that the surface of rubble infill, which included a thin layer of yellow clay levelled off with overspill from the blue puddling clay quickly rose to turf level before falling off towards the gate. Although the trench was taken down to 4 feet below the turf, the depth of the infill could not be ascertained, nor was there any evidence of the northern edge of the chamber. At the end of the trench, some 18 feet from the gate, the surface of the infill had fallen to over 2 feet below ground level but evidently extended some way further northward, perhaps beyond the field gate. It was presumably this rubble which produced the large anomaly on the two geophysical surveys, a feature which was initially ignored owing to the general observation that gateways tend to accumulate such disturbed material over time. It was not then realised that the gate and hedge had only been installed relatively recently, after the old boundary wall of Caisson Field had been removed as mentioned above.

Although such a large (and apparently deep) excavation was to be expected on the site of the Caisson, as indicated exactly on the canal map, the only structural element found was the band of chippings at its southern edge. This however, is a feature that might be expected in association with dismantled masonry - supporting the opinion that, following the removal of the superstructure (but before backfilling), much of the stone lining of the upper part of the lock chamber was removed for use elsewhere. Certainly the amount of blue puddling clay lying beyond it (about 12 feet in extent) would have been sufficient to provide the basis of the waterproof embankment (or 'angle of response') needed to support the superstructure of the lock.

The presence of so much ironwork in the puddling clay still requires explanation, but the finding of the half-penny in this material suggests that the site was backfilled in 1807 or soon after in agreement with the mapping evidence. The Caisson was apparently still open when the lock flight and pump were installed early in 1805, as shown on the early c.1804 map, but had disappeared by the time of the next datable map of the lock flight, c.1809. The rubble infill was barren of human artifacts and was presumably construction material excavated from elsewhere along the canal.

Following the inspection of the site by the Council Archaeologist it was decided that, with these findings, a more exact exploration of the site is now possible. Archaeological examination should also be made at sites nearby which related to the Caisson, an overall project which might well be given support by English Heritage. Such sites could include the Inclined Plane, the Engine House, the Caisson Exit Tunnel etc. and in particular the Terminal Basin and its adjoining subterranean chamber. Although a few preliminary levels were taken on this occasion for the trench, it would also be necessary to establish the level of the canal bed in the basin to obtain a clearer picture of its relation to the Caisson. It was also concluded that, if permission could be obtained, considerable time would be saved in future excavations if machinery was employed for surface stripping.

This article is re-printed from "Weigh-House" - Autumn 1997. (The newsletter of the Somersetshire Coal Canal Society)

Mike Chapman

Mike has since published drawings of the "suggested sequence of demolition and infill". Please see the "Weighouse" no.24 (Summer 1998).

The writer is also a keen member of 2 Trusts engaged in keeping the memory of the Somerset & Dorset Railway open.

The writer's e mail is rtjstevens@btopenworld.com