The investigation revealed late Iron Age and early Roman occupation of the 1st century across the whole of the evaluation area. This took the form of pits and ditches (with one post hole), but no stratigraphy was found to survive between these features having been entirely ploughed away. This modern plough soil is consistently about 0.3m thick. Finds from the features include a range of pottery; some samian, mortaria, Gallo-Belgic type wares and amphorae. One or two fragments from querns and triangular loom-weights were also found, and several brooches. Only a small amount of bone survived, mostly in the deeper pits, but this was in poor condition. More common were finds of animal teeth, but even these were poorly preserved. A strap-end, horse-harness fitting and a number of hob-nails, found on the south-west of the site are probably military items. A very few sherds from the upper levels of one or two of these features attest some limited activity into the 2nd century (or later), but it is clear that by this time the nature of the area had radically changed, and only two features can be ascribed to this period. A water-main, previously seen on a site just to the north, continues across the eastern side of the evaluation area. This was constructed by jointing lengths of timber pipe with iron collars, which were hammered into the ends of each length. A short section of the main was excavated in one of the trenches, where two of the collars were exposed, upright in their original positions, 1.66m apart. The line of this water-main has now been traced for about 300m, and it appears to be heading toward a large L-shaped cropmark south-east of the evaluation area, near the Roman temple complex. A test pit, excavated into this cropmark feature, showed it to be about 1m deep, and though it had been dug into here in the 19th-20th century (possibly a previous unrecorded excavation) it appears to be of Roman date. A reasonable quantity of Roman tile was recovered, including two combed flue-tile fragments, together with a small quantity of building materials. If supplied with water from the main, then this feature could represent a lowered floor area in a building to accommodate a hypocaust, suggesting a Roman bath-house. A waterworks is another possibility. Though only a small quantity of probably prehistoric material was found in residual contexts, in the from of worked and burnt flint, a large pit of late Bronze Age date was located towards the centre of the evaluation area. Recent excavations have shown hat prehistoric features are area to the north of the park, and only one was found during that work: a probable late Neolithic pit close to Oliver's Lane. The late Bronze Age pit here contained a small assemblage of pottery and worked flints, with pieces of burnt flint including one large nodule. The fill of this feature was a pale brown sandy silt similar to the subsoil, and contrasting with the darker fills of the late Iron Age and Roman features. This probably reflects the soils of this period in the Gosbecks area. The implication seems to be that, as the general soil structure of the pit had been little altered from that of the present subsoil, the soils in the areas at the time had been subject to little or no cultivation in relation to those of the late Iron Age and Roman periods. No other features of archaeological significance were found during the evaluation. Although the stratigraphy across the site has been entirely ploughed away, it was demonstrated that there is some archaeological potential in the topsoil, as finds have not been significantly moved from the position of their original context. Therefore any finds material which is not subject to complete destruction when exposed to agricultural processes and weathering could indicate areas of activity and features which have otherwise been obliterated. A problem was identified in excavating within the subsoil as the features are mostly of very poor visibility in this material, and often only clearly reveal themselves where they are cut into the natural sands and gravels below. This subsoil deposit varies in thickness, but is generally up to 0.3m over most of this area, though it can be as much as 0.5m. Excavation below the surviving field boundary bank on the west of the area showed that the subsoil outside of this protected strip has been reduced by up to 0.2m. Much of the surviving archaeology of the site is preserved in this deposit (some features were found to be entirely contained within it), and despite the difficulties it should not removed other than by archaeological excavation and recording. In order to understand the subsoil deposit better, and to investigate the potential of scientific aids in locating and defining features, it has been subjected to a short programme of scientific evaluation, both on site and with the collection of samples (from the subsoil deposit and features) for phosphate and pollen analysis. The results of some of this work are still awaited. Many of the linear features present on the site can be successfully defined using a magnetometer (gradiometer). However, several ditches found during the evaluation are not visible on the magnetometer survey. The reason for this is not clear, but they appear to be features which are mostly contained within the subsoil, with little penetration into the natural sands and gravels below. The certain correlation of pits founds during the evaluation, and small anomalies on the magnetometer plot which may represent them, is more difficult, but the large Bronze Age pit (see above) appears to be clearly visible on the magnetometer survey. <1>