Between 1990 and 1996 Oxford Archaeology examined sites and landscape features dating from the 5th century AD to the post-medieval period in the ARC (now Hanson Aggregates) gravel extraction pit between Yarnton and Cassington, Oxfordshire. Early Saxon settlement was first identified on the edge of an Iron Age and Roman occupation site at Yarnton. Subsequent work revealed the presence of middle Saxon settlement immediately to the east, prompting the first excavation of a rural site of this period in Oxfordshire. Investigations further afield uncovered evidence for early and middle Saxon settlement adjacent to the neighbouring modern hamlet of Worton, and middle Saxon buildings were uncovered among Iron Age pits and postholes at Cresswell Field. These discoveries have had important implications for understanding settlement patterns in the 7th and 8th centuries AD, but have also highlighted the difficulties of locating middle Saxon settlements comprising posthole buildings and few durable objects.

The wider landscape context of these settlements was investigated by fieldwalking, geophysical survey and trenched evaluation, revealing manuring scatters, field boundaries, ploughsoils and trackways. Former river channels on the floodplain provided valuable sources of information for reconstructing the landscape and changes in agricultural practices over time, particularly the presence of pollen and waterlogged macrobotanical and invertebrate remains.

Romano-British settlements are present at both Yarnton and Worton. Within this context it has been possible to study the period of transition from the late Roman to the early Saxon periods, particularly on the more intensively excavated Yarnton site. Early Saxon buildings appear to have respected late Roman features, although no obvious traces of sub-Roman activity were recovered and the character of settlement is strikingly different to that of the Roman period. A similar pattern of contiguous but discontinuous occupation emerges from a survey of the sites in the immediate area. Despite the paucity of evidence for habitation through the 5th century the environmental indicators suggest that the land continued to be farmed, though by less intensive methods than those employed in the Roman period. One explanation for permanence of settlement location may be the retention of farming units and land boundaries by Saxon settlers.

In the middle Saxon period the character of settlement became less dispersed, with more organised and intensive use of space on habitation sites and a greater diversity of building forms, including timber halls. At the same time, there is evidence of agricultural intensification in the form of increased quantities of charred cereals, a change in crops grown, including new varieties to Yarnton such as rye and legumes, the presence of horticulture and the importance of flax production. Initial depletion in soil fertility seems to have been countered by more effective weeding techniques and, probably, manuring. Hay meadow was created on the Thames floodplain, probably within substantial double-ditch-and-hedge boundaries.

It is suggested that the middle Saxon evidence represents indications of settlement nucleation and embryonic village development on the one hand, and the origins of the open-field farming system in this area on the other. The dating evidence, although not very precise, suggests that these changes began in the 8th century but gathered pace in the 9th.

As a story of changing settlement and landscape, the record for Yarnton is truly remarkable, for it can be seen within a trajectory of over five millennia of human activity in the area. From the earliest evidence of settlement on the floodplain in the early part of the 4th millennium BC, it is possible to trace increasing permanence of occupation and clearance of the landscape. Yarnton can be seen to have shifted gradually eastwards from its Iron Age location to that of the middle Saxon period, and then north-eastwards towards the medieval village of Yarnton which appears to have been clustered around the Norman church.

A Hiberno-Viking stud from Yarnton

Objects from the Viking world have been found as far inland as central England. A Hiberno-Viking enamelled stud of 9th- or possibly 10th-century date, and probably part of a horse bridle, was found by metal detectorist, Mike Shott, in ploughsoil just north-east of the Yarnton middle Saxon site. It was identified by Susan Youngs of the British Museum as a finely-crafted copper-alloy object based on a central equal-armed cross; the background would probably have been filled with red enamel. It probably came from Ireland, or possibly Iona, and similar objects have been found in the Balladoole ship burial on the Isle Man and at Kolset in Norway.