Reconstructing a Medieval Welsh Landscape: Exploring the Evidence of Nineteenth Century Field-Names and Land-Use Data in Landscape Archaeology

In 2010 a computer-based 'predictive model' of early medieval settlement location was developed as a way of addressing the difficulties of identifying early medieval secular settlements in Wales (Seaman 2010). The model, which focused on a 100km2 area in the eastern Vale of Glamorgan, aimed to identify 'core settlement zones' where prospection techniques, such as geophysics, could be most profitably targeted. Initial analysis suggested that the model could also have applications for exploring patterns of historic land-use. The aims of this project were to make the results of the model freely available, and to use field-names and patterns of agricultural land-use recorded in nineteenth century Tithe surveys as proxy-data with which to explore its applicability. Field-names have long been acknowledged as vital sources by landscape historians, but archaeology has been slow to draw upon their evidence, particularly in Wales where, given the lack of alternative sources, their potential as indicators of archaeological features is perhaps greatest. Mapping of nineteenth century agricultural land-use information also provides a valuable data-set through which the organisation of agrarian resources within a landscape can be explored. The project was funded by the British Academy/Leverhulme Small Research Grant scheme. This collection includes the results of the predictive model and a fully georeferenced database of field-names and land-use information derived from twelve Tithe surveys. The data pertains to a 90km2 area in the eastern Vale of Glamorgan. The results of the predictive model consist of a network of 'settlement zones' labelled 1 to 4 in descending order of their likelihood of containing evidence for early medieval domestic settlement (1 most likely, 4 least likely; see Seaman 2010 for detailed discussion of the methodology of the model). The database of field-names and land-use information is derived from information held within Tithe survey apportionments (dating to between 1838 and 1841) for twelve parishes within the study area. Two further parishes, Porthkerry and Barry, were excluded because their Tithe apportionments included neither field-names nor land-use information, and Highlight (a part of Wenvoe parish) was not included as its field boundaries were not recorded on the Tithe survey plans. A total of 4609 entries listing the name and/or state of cultivation of fields were transcribed from the apportionments and geocoded the British National Grid via a network of field-boundaries based upon the First Edition Ordnance Survey 6 inch 1 to a mile maps. Changes to a small number of field boundaries between the dates of the Tithe surveys and the Ordnance Survey maps meant that some fields could only be reconstructed to their approximate original proportions, but individual errors are unlikely to be greater than 10 meters. There were also inconsistencies in the amount information provided in the Tithe apportionments, and whilst both field-names and state of cultivation information was provided for 1212 fields, 930 fields had only field-names, and 2467 fields had only entries for the state of cultivation. The names of settlements, houses, farms and associated structures have not been included in the database. Where possible the state of cultivation information and field-names has been used to assign individual fields to one of six agricultural land-use categories: arable, pasture, meadow, mud flat, wood or orchard. A system of coding is used to indicate where field-names provide evidence for features of potential archaeological significance within or near to a particular field, or where contradictions between the state of cultivation and field-name indicate a change in land-use.