Since 1985 Dr. C. R. Salisbury has, in conjunction with the Leicestershire Archaeological Unit (LAU), been undertaking a survey of the floodplain archaeology of the River Trent around Castle Donington. The survey has included the detailed recording of ancient river channels and associated structures as they have become visible in the exposed faces of the Ennemix gravel quarry at Hemington near Castle Donington (grid reference SK 459302).
The quarry is located near the centre of the floodplain formed by the confluence of the rivers Trent, Soar and Derwent and is sited to exploit the thick deposits of sand and gravel which overlie deposits of Mercia Mudstone. The lower sands and gravels were probably deposited after the maximum glaciation of the last Ice Age (Devensian) between about 15,000 and 10,000 BC. The upper sands and gravels have accumulated since the start of the Flandrian period (around 10,000 BC) during which time the Trent shrank and settled into its present meander belt. During this time the lateral migration of the river has deposited between 2m and 5m of sand, gravel and alluvial silt and clay. The dynamism of the river's migration has led to the excellent survival of riverine structures in the old channels and the anaerobic conditions have ensured the survival of organic remains. The thickness of the overburden deposited by the river makes conventional archaeological prospection difficult and constant monitoring of quarrying is necessary to ensure the recording of both archaeological sites and other, natural, features.
The granting of planning permission for gravel extraction at Hemington Fields in 1984 was followed by the arrangement of a watching brief which was carried out by the Burleigh Archaeological Fieldwork group on behalf of Leicestershire Museums. In 1985 they reported the discovery of a large timber structure revealed by the quarrying. Subsequent salvage excavation by LAU revealed the site to be a Norman mill dam dated to c 1120 with evidence for an associated mill. At this point Dr. Salisbury started his weekly monitoring of the quarry working. Since 1985 the watching brief and resulting salvage excavations yielded a considerable amount of information about the use of the river and its margins. There have been further finds relating to the milling industry (timber wheel breasting, axle cover and wheel paddles) and the fishing industry (43 post alignments, mostly of Saxon and medieval date, thought to represent fishweirs) and 128 grooved 'anchor stones', probably used to weigh down nets or baskets. A series of palaeochannels have been recorded across the quarry mostly dating between the 8th-9th century to the 13th-14th century, but with some evidence for the fragmentary survival of Neolithic channels in the relatively stable meander course of the river. The palaeochannels have produced additional structural evidence in the form of displaced timbers (bridge base plates, wheel paddles), worked stone (e.g. Saxon cross fragments and Roman architectural pieces) and environmental evidence from the organic deposits (pollen, beetles, plant remains). Recent work to the north-east of the quarry has produced evidence for earlier Saxon channels.
The monitoring and recording work carried out by Dr. Salisbury culminated in August 1993 with the discovery, during overburden stripping, of the timber and masonry foundations for two bridge piers. Following a site inspection by members of the LAU, a programme of excavation and recording was planned and this was carried out in September 1993.
During the final week of fieldwork a second bridge (Site 2) was discovered when a dragline struck two massive oak beams 2m down in the working face, some 50m to the north-east. Preliminary cleaning showed that the timbers extended beneath the section and were jointed to others. Negotiations with the owners of the site and with English Heritage secured a six week extension to allow the excavation and recording of the timbers which proved to be the remains of an earlier bridge.
A third site was discovered during the stripping of the overburden between sites 1 and 2. Excavation was necessarily limited as the area was the only section of the quarry available for working during the excavation of site 2. The locations of the posts were recorded using an Electronic Distance Measurer (EDM) and detailed recording of the timber carried out after their removal by dragline.
The quality of preservation and the significance of the bridges was such that funding was made available to the LAU from English Heritage and Leicestershire County Council to allow the excavation and recording of all three sites.
The research project which focused on the bridges was drawn up with reference to the research priorities outlined in the English Heritage publication Exploring our Past. These were;
Methodology and objectives
Following permission from the owner of the site to allow access for the purposes of excavation, work began on September 6th 1993. The intentions were:
Results
The surviving structure of the latest bridge comprised four pier bases constructed using two different methods. The northernmost pair were of stone construction and lay adjacent to the bank or in the shallows of the river. Pier base 1, which survived virtually intact, consisted of a hexagonal sandstone plinth measuring 9.6m by 4m with finely chamfered blocks facing a rubble core. The blocks were fastened across their tops with iron straps sealed with lead. The plinth was set upon a larger platform of sandstone blocks up to two courses high.
Between the stone pier bases was a timber trestle structure comprising a transverse base-plate some 5.28m long with two upright posts some 3.80m apart. The eastern upright was mortised into position and had an external lateral brace to the base plate. The western upright was half jointed to the side of the base plate and penetrated the river bed. The trestle appears to have served to support the bridge decking between the main piers. It resembled a large trestle which formed part of the earlier Bridge Site 2 and which employed a similar combination of earthfast and timber frame techniques. A structure with a possibly similar function but a rather different structure was recorded between pier bases 2 and 3. This was a rectangular arrangement of five timber piles (with a possible lost sixth) measuring 4.40m by 3.20m.
Pier bases 3 and 4 extended into the mid-stream of the contemporary river and employed a somewhat different foundation method to the landward piers, reflecting the greater difficulties involved in constructing and maintaining a pier in the deeper section of the river. The better preserved pier base 4 provided details of this construction, a hexagonal circuit of oak piles driven up to 3m deep into the bed of the river to form an enclosure measuring 10m by 4 m. The piles were joined by a ring beam consisting of overlapping oak timbers secured to the tops of some piles by mortise and tenon joints and resting on others. The enclosure was filled with roughly hewn sandstone blocks forming a platform. Surrounding the pier bases was a layer of sandstone blocks dipping into the scour channels between the pier bases and the downstream scour pool. It is uncertain whether this was a deliberately constructed platform designed to protect the piers bases from erosion or whether it was the remains of the collapsed superstructure of the bridge. Definite evidence for a collapsed stone superstructure was present in the form of ashlar blocks, several of which retained their lead sealed iron straps. A single chamfered ashlar block suggested that the mid stream piers probably had plinths similar to those found on the masonry pier base 1.
Immediately upstream of pier base 4 was a wooden structure which apparently served to protect the cutwater of the pier from damage either by erosion, floating debris or both. The position of the river bed beneath the structure suggests that it was added sometime after the construction of the bridge at a time when erosion of the bed around the bridge had already begun.
It seems likely that the bridge continued to the south, beyond the present limits of the quarry. Towards the end of quarrying operations in this area the limit of excavation lay some 5m to 6m south of pier base 4. A single driven post was visible in the section surrounded by a spread of sandstone rubble and ashlar blocks. These were interpreted as the remains of a further inter-pier structure and of pier base 5.
The dendrochronological analysis of the timbers from the bridge has not yet been completed, but early results reveal that the piers were constructed from timber felled in the late 1230s, possibly early 1240s.
It was clear from the initial examination of the exposed timbers that the wooden structure revealed by the dragline was substantially different to Bridge 1. The objectives of the excavation were thus similar to those defined for Bridge 1 but with variations consequent upon the differences between the two structures. The intentions were:
Initial machine clearance and hand excavation revealed a large box-like structure, a trestle frame and pile posts. A second box structure was found beneath the trestle and extending beneath the quarry face. A provisional date within the 11th century was assigned on the basis of evidence from the pile posts which had first been observed during a watching brief in 1991. This early date was corroborated by Dr. G. Milne following an examination of the carpentry. A more detailed dating of the structural timbers is currently in progress and methods are being developed for non-destructive sampling.
It was soon realised that these remains represented a second bridge. The two box-like structures were the remains of pier bases, the collapsed trestle was a mid stream support to the decking and the posts formed a double row (extending for some 27m across the river channel) which would also have supported the decking, perhaps representing a second phase of bridge building. The north-western end of the bridge had been destroyed by quarrying during gravel extraction in 1991 although four upright posts recorded at the time undoubtedly represented the continuation of the double line of posts located during the excavation. The southern extent of the bridge was destroyed as the river meandered southwards.
The two bridge piers consisted of two similarly framed wooden bases upon which the bridge superstructure rested (see sketch I). Each pier was constructed on a bed of four timber base plates, diagonally lapped to form a lozenge shaped foundation. Each frame was then braced across its inside internal corners. Four vertical corner posts were then fitted into large square mortise holes cut through the intersecting base plates. Each post was then secured externally with diagonal bracing which ran down to the ends of the base plates, and internally with a more random set of braces. Planks were then edge lapped round the inside of the posts forming the walls to a large box. Assembled, the internal dimensions of each box measured 2.4 by 4.5m by at least 1.2m deep. Once assembled in the river each box was then filled with approximately five tonnes of sandstone rubble which would have acted both as ballast to secure the structure to the river bed and as a platform to support the superstructure. Dendrochronological dating suggested that the timbers from pier base one had a final felling date of 1097 A.D. Similarities in carpentry details, and more particularly, identical tool signatures found on timbers from both pier bases, suggest that the two structures were actually built at the same time.
Sketch 2: Reconstruction of trestle (unshaded timbers are conjectural)
While the structure itself was very solid, it eventually collapsed into a scour pool which had formed as a result of turbulence caused as the flow of water met the solid mass of the pier bases. Directly over the collapsed remains of pier base II the broken, but still jointed remains of a massive timber trestle was found (see sketch 2). The trestle had undoubtedly provided a support to the bridge decking between the two pier bases, although its precise original position is not known. The trestle survived as four of a probable six timbered structure. The huge base plate (over 10m long) would have rested on the river bed parallel with the flow. Two main posts were mortised through the base plate, with the tips of each post projecting below the base plate by 1.2m. Each post was then stabilised by bracing which ran down to the ends of the base plate. The articulation between the main posts and the brace comprised of a combination of angled mortise and tenons, hitherto unattested in Saxo-Norman carpentry.
The presence of an edge peg hole in the side of one of the main posts but unrelated to the brace jointing mechanism suggests the positioning of a collar for the trestle. The collar would undoubtedly have supported the decking of the bridge. It can therefore be assumed that the length of the collar (distance between the two through mortises on the base plate) also represents the width of the decking (2.7m) and the height of the decking above the river bed (4.5m).
(Fig.3 Plan of the late 11th century trestle)
Dendrochronological dating of the trestle gave a date range spanning the supposed final felling date of the pier base timbers (1097 AD). Similarities in woodworking techniques and a high t-value match with one of the pier base timbers suggests that the structures were contemporary.
As well as the pier base and trestle structure the bridge decking also appeared to have been supported by a double row of posts. Fifteen upright pile driven posts were recorded from the Site 2 bridge, which collectively formed two parallel rows of posts spanning the river channel. The two rows were 2.9m apart and were observed for 27m, although the northern extent of the bridge had been lost during an earlier phase of gravel extraction. Seven pairs of posts were observed, spaced irregularly at intervals of between 4.4 and 6.4m. Majority of the posts appeared to have been braced down to the river bed but, as no articulation survived, it is also possible that the bracing related directly to the bridge decking.
Dendrochronological dating of these posts proved inconclusive with a possible maximum date range of almost a century (1051 - 1148 AD). However, two posts provided absolute dates of 1099 and 1103 AD suggesting that the posts succeeded the pier base phase, either as reinforcements to the pier base phase or as a replacement bridge structure.
The recording of Bridge 3 was undertaken on a rescue basis. Timbers from the bridge were encountered piecemeal from the time of the initial stripping of the overburden. The first timbers to be discovered were recorded as individual finds as their function was uncertain. Further timbers were recovered during the excavation of Bridge 2 and it soon became apparent that they formed part of a relatively simple bridge, essentially a double row of earthfast posts. Parts of the structure had been recorded and sampled during quarrying in 1990 and it was on the basis of the analysis of these samples that a date within the early part of the 13th century was assigned.
The bridge consisted of a double row of earthfast posts extending for some 56m across the river channel, on a north-west to south-east alignment. The southern end of the bridge appeared to continue beyond the limits of the quarry. The posts were not evenly spaced (varying form between four and six metres apart), and the posts often appeared as clusters of two to four posts. These may represent repairs to the bridge or an attempt to increase the stability of the structure. All the posts were pile driven and the remains of three iron pile shoes were found still attached to the tips of posts.
A gap in the centre of the double row corresponded with the deepest part of the channel and may have been the position of a navigable arch. It is expected that when the dendrochronological analysis is completed it will be possible to resolve some of the questions concerning the maintenance and repair of the bridge.
The analysis of the timbers recovered from the Hemington Gravel quarry is far from complete. In addition to the further study of the dendrochronological data, research is continuing into forestry management, carpentry techniques, notably analysing methods of jointing and the types of tools used, and the engineering aspects of bridge construction. In addition there are plans for the conservation and eventual display of the timbers, together with interpretative material, at the Snibson Discovery Park near Leicester.
