Depth below Datum, Meters
Composite Section Caption Key within Soils: CF = Crushed Flints, CF+ C = Crushed Flints with Chalk, HC = Hard Chalk, P = Pottery, R = Root, S = Silty Soil. Key without Soils: SSC = Solid Soft Chalk, Arrows = finds, - - - = Trench Bottom. The LIDAR Grid line is between 8S and 20S built up of Solid Chalk with Crushed Flints and Chalk above. Remains of the Roman Road lane is a band of Hard Chalk between 2S and 6S. The Arrows above the Composite Section were the predicted positions from the LIDAR. Both are buried deeply under Crushed Flint with Roman finds right across the Composite Section, which leads to the interpretation that the Ledge on which we worked was made by the Romans. Trenches 2 & 3 were opened lower down the Ledge so the soil layers do not marry up. A full size Composite Section is available at the end of this report.
The Winchester - Chichester Roman Road was sought because if a direct route existed it is likely that the Roman settlement of Clausentum from the Antonini Itinerary VII would to be at Exton near the A32. A LIDAR image of a street grid has been found, reported in Refs 1 & 4. Excavations were conducted where the Northern Roman Road lane crossed one of the minor Grid Lines near the A32. A ledge was selected which was thought to be a river terrace, where remains would be near the surface - avoiding the flood plain which would be very deep in sediments. It turned out that the ledge is man made, burying the Roman features.
The base line was taken as the A32 boundary, and the distance from the gate into the field and the northern edge of the northern Roman Road lane was measured on the 1:5000 LIDAR, and taken as the archaeological grid origin on the ground. The Composite Section is shown above.
Trench 1, 2N to 2S was first opened. It shows soils going down to 60cm below ground surface with limited features. The section drawers show no layers, though subsoil was recorded in the excavation at 20cms below ground surface as a slightly redder colour. Subsoil in this area is often difficult to distinguish. Below 30cms below ground surface a large amount of crushed flint or flint gravel occurred. With this flint material were finds of coal, pottery and brick fragments. Two were taken as Special Finds as possible but not certainly Roman including a nail. With no features at 60cms below ground surface, a Box Section was sunk. Almost at once 20cms packed with Roman pottery came up - tile, brick and pottery fragments. No items could be definitely assigned outside the Roman period. The Box Section was continued from 80 to 1m below ground surface but showed only the same featureless soils, save for crushed flints.
This depth causes us problems with our site rules and insurance issues. The crushed flint laiden soils might be taken as the Roman Road stonework, but does not show the expected limits - it should end around 00, and also shows no bottom. 70cms of such flint soils is excessive. One meter depth of soils is also not natural - one expects around half of this on top of the geological strata - here expected to be chalk. The large catch of Roman pottery at 80cms below ground surface also implies that the soils above it have been re-deposited by man at some point after the Roman pottery date. Substantial earth movements have occurred.
Trench 4, 10S - 14S expected to be over the Grid Line was next opened. Much was obstructed by tree roots, but similar deep soils to Trench 1 was found. Subsoil with crushed flints occurred down to 60cms, with metal working and Roman building material finds. Ref 2 describes these. A Box Section was taken from 60cms, with a layer containing hard chalk and crushed flint at 65cms. At 75cms below ground surface solid soft chalk was encountered. This was trowelled into, finding brick fragment and worked flint (their position indicated by arrows on the above Composite Section) - so this layer was not natural, and was taken as the Grid Line. Later work indicates that the layer above, of hard chalk and crushed flint with subsoil, is also part of the Grid. As before, finds at this level means all the soils above are not natural but re-deposited by man.
Trench 5, 14S- 18S
At a similar time Trench 5 was opened with similar results to Trench 4. Layers were not recognised during excavation, but were by the section drawers. Crushed flint with subsoil comes up, but appears to be falling in depth as you go south (this heralds the southern edge of the Grid Line). The layer also containing chalk comes up at 70cms. Finds down to this level include coal, worked flint, brick fragments, bone, and at lower depths a lot of slag. A Box Section went through the subsoil, crushed flint and chalk layer, into the solid chalk layer which also contained some flints - which was taken as the Grid Line as in Trench 4.
Trench 6, 18S - 22S
Trench 6 was arranged to go to the depth of the natural, expected to be chalk. It was expected that the southern edge of the Grid Line would be in this Trench. The upper layers were similar to those in the other trenches. The Top Soil contained recent finds as well as worked flint and clinker. The subsoil layer coincided quite closely with the appearance of large quantities of Crushed Flint, with quite possibly finer flint gravel at its surface. From here downwards Roman pottery and building materials occurred down to the natural - not generally high density but consistent, also with worked flint. At 70cms below ground surface a hard layer came up, composed of relatively thin flat stones laid on top of each other. Tim Jeffery, a modern road engineer, said these were specially graded stones designed to produce a hard stable surface. Some of these were Roman pottery and building materials. It goes out of the Trench to the south - but does not appear at this level in Trench 5. It is however only 10cms thick.
Below this graded layer is earth relatively free of stones, but with a small amount of Roman pottery and worked flint. At the south of the Trench over 30cms rests on natural chalk, which was trowelled into but produced no finds. At the time it was thought this surface was rammed by the Romans, as there was no intermediate layer generally found on chalk surfaces. Since then we have found in other excavations in this area that the chalk natural comes up suddenly often without much intermediate layer. In the north of the Trench there is only a few cms of this earth layer before coming onto the Chalk and Crushed Flint layer with some subsoil. As can be seen in the Composite Section drawing above this Chalk and Flint layer is rising up to meet the same layer found in the Box Section of Trench 5. As stated in Trench 4 this layer is found to be the upper part of the Grid. In Trench 6 it contained some hard pottery shards which were probably Roman.
It thus appears that the region 17S - 19S represents the southern edge of the Grid, where its components rise up some 50cms from the natural chalk. It was expected that the edge would be a couple of meters further south - but it is found that the A32 boundary has a slight curve in it which explains the difference. We have found that the Roman practice was to remove the soils, and build their Roads on the Natural. We can see that this is what they did here - the Grid is built on the chalk natural. Then at some point they added a deep layer of Crushed Flint above it - but this continues south beyond the limits of the Grid - and looks like the construction of the ledge on which we have been working - through or over which the Grid Line passes.
We can see that the graded stone layer at 70cms below ground surface ended at the Grid, but continued south out of the Trench. It was built on the deep bed of re-deposited soil without much stone, and formed the base of the deep layer of Crushed Flint. It's not possible to say what was the purpose of this arrangement - it may have been an early surface of the ledge, which became redundant when the deep layer of crushed flint was added.
Trench 3, 5S - 10S
We then moved to Trench 3, which was expected to contain the northern edge of this Grid Line. But in order to reduce the depth of the excavation, and to learn more about this ledge, the Trench was opened further down the slope of the ledge at 4mW of the base line (those described above were 1W). We now do not expect the layers to marry up with their counter parts in neighbouring trenches. The upper layers were similar to the other trenches described. The Top Soil contained a lot of coal and clinker down to 15cms, modern pottery and glass and some square nails. The subsoil contained a great deal of Crushed Flint clearly visible in Section - and surprisingly this layer was thicker than seen in some of the previous trenches. Roman pottery occurred.
Then we came down to the layer containing chalk as well as Crushed Flint with some subsoil. The chalk came up suddenly with the shape in the above Composite Section, sloping down to the north. The chalk was generally soft. Below this was a hard packed layer, generally of chalk flint, sometimes small lumps stuck together with a lot of Roman pottery. The spoil heap contained a lot of Roman cement - and its likely Roman mortar was used to bind this layer together. The two layers come to a point at 8S, below them is solid soft chalk which slopes down to similar natural chalk at the north of the trench, which is 1m below ground surface. Here there is a 20cms band of darker soil without Crushed Flint between the Chalk natural and the Crushed Flint laden subsoil. which buries the Grid Line. A similar arrangement was seen in Trench 6 at the southern edge of the Grid.
Thus we have seen both edges of the Grid Line, sloping down to the Chalk Natural, building up to c 50cms above the Natural. The width is 8 -10m on the Composite Section - but this is not going across the Grid at right angles. The 2 dark lines of the LIDAR are estimated to be 8m apart - which is reasonable agreement. It is stressed in Ref 3 that LIDAR being a new tool we do not have experience of what it is picking up.
Trench 2, 2S - 6S
We finally returned to the Roman Road Lane which we had failed to find in Trench 1. As in Trench 3 we moved further down the sloping ledge for Trench 2, to 5W - 6W. The Top Soil and subsoil were similar to the other trenches. The subsoil surface seems also the boundary of a large amount of Crushed Flint, with a few Roman finds. After 40cms of this subsoil, white speckles as the tell tail signs of entering chalk came up. The Trench was cleared to this level, which is shown as the Trench bottom for much of the Trench. Then a Box Section was sunk to the chalk natural.
There was 20cms of hard chalk lumps mixed with soil. As in the Shavards Farm machine trench, these chalk lumps are not the local, natural of soft, cream coloured material - but must have been brought in from elsewhere. There were 4 Roman pottery fragments, and a large number of small fragments, with a square nail, and possibly iron working residue. Below this was near 30cms of sandy coloured material, probably silt, with substantial Roman material. This included a Roman rim, 3 vessel fragments, 4 brick fragments, 2 large iron clinkers, and 5 Roman cement fragments. Then the soft creamy chalk natural occurred at 1m below ground surface.
It might be argued that this 20cms band soil and foreign chalk might not prove the existence of a Roman Road. However this Northern Lane has been proved in the trial machine trench at Shavard Farm in March 2019. In the upper layers this foreign chalk was mixed with soil, but the chalk got more compacted with depth, and was built on the chalk natural. Clearly the depth of soils in Trench 2 is not natural, and with the Roman materails found at great depth again indicates the whole has been re-deposited. The two trenches are on the LIDAR line attribituted to the Northern Roman Lane, and the evidence suggests this is what remains of the Roman Lane after the substantial earth movements which have buried the more complete remains of the Roman Grid Line. It is not understand why a band of soil, without much stone, separates the chalk layer from the chalk natural.
We have concluded for each Trench that the deep soils are not natural, and from the subsoil downwards containing Roman material must have been re-deposited. Ref 1 shows a wide scatter of Roman surface pottery - so soils from just the other side of the A32 could have been the source. It is concluded in Trench 6 that the Crushed Flint subsequently buries the completed Grid Line - and as the Crushed Flint appears to extend beyond the scope of this Grid Line it is likely that this Ledge is constructed of this Crushed Flint. It is also apparent that this Ledge is constructed on the Natural Chalk - which means that the original soils have been taken up - this must be so as the soils laid on this Natural Chalk in Trenches 6, 3 (and 2) have Roman finds in them. This may be the explanation for not finding the Roman Lane in Trenches 1 & 2: the Roman Lanes may have fallen out of use when the Grid was constructed (they are not in use today), and so were taken up when the Ledge was constructed (the band of soil on the Natural Chalk in Trench 2 has Roman finds in it). Possibly some of the original hard chalk of this Roman Lane got put back again and was seen in Trench 2. The report in Trench 3 that the Grid was probably cemented together could be the reason it was not taken up, but buried.
Further relevant evidence
What was this Ledge constructed for?
Reports on the finding of the LIDAR Grid and its ground evidence, Refs 1 and 4, describe the Principal E - W Grid Line as running on a Causeway across the river flood plain. This Causeway is still visible to the west of A32, 2m high above the flood plain until truncated by the old course of the river Meon - a distance of 25m. The width is huge - 85m - and 2 minor Grid Line also run on it. The argument that the Causeway continued at a similar height across the flood plain is that starting out for this distance it is likely to continue, the LIDAR image continues for the Principal and one of the minor Grid Lines, and an optical swelling has been detected on part of the flood plain. A Causeway of this height can be expected to control an irrigation system - and it is known another was constructed in the 1700s when the course of the river was moved to the west side of the flood plain - the irrigation ditches are still visible on the LIDAR and when the grass is short. Lack of maintenance of the drainage at the end of the Roman period would cause a lake behind the Causeway to overflow - washing the Causeway away - a general fate of Roman causeways which are thought to be quite common on Roads crossing valleys.
It appears that the Ledge runs onto the Causeway. It likely therefore to have been constructed as a Quay or edge to the water feature created by the Causeway, the height of the Ledge designed to match the height of the Causeway - and the maximum height of the water feature which could be created. Crushed Flint may have been selected to construct the Quay for several reasons. Crushed flint would be more stable than ordinary earth to erosion by the water - it is noted that the Romans removed all the soils down to the Natural Chalk. The Crushed Flint would provide a hard standing on the edge of this water feature - and on burying the Grid Lines (and Roman Lanes) the sloping edge would provide a stone paved route to continue along these features. Likewise there would be a stone paved route along the Quay, and at any point down onto the flood plan. The Crushed Flint now appears as mixed with soils - but these may have formed since the Ledge was constructed: The water feature would introduce sediment into the Crushed Flints, in post Roman times vegetation would introduce roots which will later turn to soil, while worms pull vegetable matter down. There may be a puzzle why at each point we have reached the Natural Chalk there is quite a thick layer of earth without stones on top of it. At these points the water of the water feature will be some way away - we expect the Crushed Flint layer to extend to reach the Natural Chalk for all points near the water. The presence of Roman finds probably results from using Roman rubbish and any kiln waster dumps to bulk out the Crushed Flint.
There is an issue of how far north does this water feature extend. For our data on the Roman Grid comes from LIDAR covering one OS Grid Square - the ledge staring near its northern boundary. One purpose of producing a dam is to store water up-stream. Other uses for such water could include a mill. An enquiry into these matters was mounted.
1. LIDAR Map of Exton & the Roman Grid NEHHAS FAB e - News 19 Winter 2019 www.nehhas.org.uk/clamap.htm,
2. J M Daguebert, Excavations beside river Meon, NEHHAS e News 17 Summer 2017, www.nehhas.org.uk/rrch10b.htm.
3. Richard Whaley, What is LIDAR picking up? NEHHAS FAB e - News 19 Winter 2019 http://www.nehhas.org.uk/lidar0.htm
4. Exton towards Old Winchester Hill, NEHHAS Journal 8 No. 10. 2017 www.nehhas.org.uk/rrch10.htm,