5. Finds Presentation

5.1 Introduction

The finds recovered comprised almost entirely pottery and tile, although there were a few lithic objects and a number of pieces of metal slag (View clickable map of all fields walked). In 1995, a preliminary study of the ceramics was made by Catriona Gibson, a PhD student from Reading University. On the basis of her provisional analysis, it was decided that most information would be gained by undertaking a detailed, fabric-based assessment and a comparison with excavated assemblages from the region. This study, undertaken by Dr Steven Willis, is presented in section 5.4. The ceramic evidence was also used in order to identify possible settlement sites, and this analysis is presented first.

5.2 Pottery and tile analysis

In total, 9871 sherds of pottery were collected (46.023kg), together with 3672 fragments of tile (191.380kg). The distribution of this material was not even across the fieldwalked transects. Indeed, marked variations in finds densities were noted during fieldwalking. Factors which affected this include not only the derivation of material from settlement sites, but also its redeposition as a result of episodic deep ploughing, terrace construction and ploughsoil erosion resulting from rainfall and flooding.

Analysis was undertaken systematically, with the principal aim of mapping concentrations of pottery and tile across each of the fieldwalked transects, based on their densities per hectare. Data analysis was undertaken using the following procedure (cf. Carreté et al. 1995):

Select the 'more detail' link to examine further these stages in the analysis.

  1. The pottery and tile were examined, classified and quantified. Where possible they were also identified and dated (cf. 5.4). [More detail]
  2. The quantities of each individual pottery and tile fabric were converted to values per ha. [More detail]
  3. The dated pottery was grouped by fabric into broad chronological phases for analysis. Fabrics of uncertain chronology were assigned to probable phases wherever possible (and thence analysed separately). [More detail]
  4. The median and percentile values were established for each chronological group, and also for selected individual fabrics. [More detail]
  5. Maps showing the density of the various pottery and tile distributions were created using the percentile values as scales. [More detail]
  6. High density concentrations of pottery and tile relative to the background distributions were identified on these maps, allowing the definition of probable sites. [More detail]
  7. The distributions of fabrics of uncertain phase were compared with those of the dated material in order to establish whether there was any correlation between them. [More detail]
  8. The finds assemblages from the probable sites identified on the maps were examined in order to characterise them on the basis of diagnostic tile and pottery.

5.3 Methodological limitations

Although the methodology described above provides a clear and systematic method of quantifying and identifying significant concentrations of pottery, a number of factors affect or limit the degree to which results of the analysis can be relied upon without a more intuitive appraisal of their significance.

First, the results are based on aggregate densities of ceramics collected by individual fieldwalkers in a variety of climatic and agricultural conditions. It is apparent that the impact of potential biases on the rate of material collection will vary depending on weather conditions, surface visibility, lighting and the types of crop being grown (see fieldwalking record). These factors have to be considered when individual concentrations are evaluated (cf. Haselgrove 1985). Further problems exist in basing the density scales on percentile values, especially where the quantities of material in particular fabrics or from individual phases is small. Although such a method of scaling compensates for variations in field size, the problems of small sample size made it necessary to review the significance of ceramic densities in each individual case. This work resulted in the decision to sub-divide the range of densities contained in the top percentile in order to help us assess the significance of potential site locations. Finally, fabrics of known date) comprised only a very small percentage of the overall total of material collected (Table 5.8). As it was necessary to base the ceramic density scales purely on the quantities of the dated pottery and tile, we must remain aware of this limitation.

Iron Age fabrics Roman fabrics Medieval fabrics Total ceramics per hectare
Quantity by sherd number per ha 693.881 2524.168 307.369 28046.465
Percentage of overall total by sherd number 2.474% 9.000% 1.09% 100%
Quantity by weight (g) per ha 2381.220 113995.237 1291.744 245222.091
Percent of overall total by weight 0.971% 46.487% 0.527% 100%
Table 5.8 Quantities of pottery and tile collected by period

Further analysis of the dating of certain fabrics was also restricted by the small size of the fields in the survey area, which limited the total number of sherds of pottery available to be collected. Thus, whilst some 'Possibly Roman' fabrics could be associated with dated Roman material, the general distribution of pottery and the limited size of the samples across the study area restricted the potential to associate 'Possibly prehistoric' and 'Possibly medieval' fabrics with material definitely of those phases. This problem was increased by our desire to apply stringent criteria in dating the pottery since many individual fabrics had a range of possible dates or were simply of uncertain age. These problems, together our wish to construct phase by phase density maps based on the most reliable data, limited the number of fabrics that could be utilised for the distribution analysis.

Success in the identification of sites through clusters of high densities of dated fabrics relies first on a systematic analysis of the quantities of material and their distribution, and second on intuitive judgements about the significance of the concentrations observed. Whilst the methods discussed do not present a perfect solution to the problem of identifying concentrations of artefact densities by cultural phase, it is hoped that they provide a reliable method for identifying potential sites which are then subject to further archaeological consideration. The storage of the fieldwalking data in a digital form also allows reinterpretation of the material by others using different criteria.


© Internet Archaeology URL:
Last updated: Sat Dec 30 2000