Internet Archaeol. 21. Pitts. Data description and quality

Two phases of excavation carried out in 1993 and 1994-5 at a known late Iron Age and Roman settlement at Heybridge (Atkinson and Preston 1998; forthcoming) yielded a large number of stratified pottery groups from the period c. 50 BC-AD 200. The morphology of the site shifted from being largely dispersed in the 1st century BC to a nucleated settlement at the start of the 1st century AD. By this time three functional zones had developed (agricultural, domestic and religious), corresponding to the northern, southern and central areas of the settlement respectively by the time of the conquest (Atkinson forthcoming; see Atkinson and Preston 1998, 93, fig. 4; and Figure 1 here). This layout was largely maintained into the Roman period.

Phase	Date-range	No. of assemblages	Quantity of pottery (EVE)
1	50 - 15 BC	18	27.85
2	15 BC - AD 20	20	62.26
3	AD 20 - 55	27	70.01
4	AD 55 - 80	27	131.51
5	AD 80 - 125	9	37.19
6	AD 125 - 170	14	62.4
7	AD 170 - 210	14	38.63

Table 1: Details of ceramic data obtained from Elms Farm by analytical phase

Over 100 stratified pottery groups were selected for analysis, all dating to the period c. 50 BC to AD 200 (Table 1). These had all been recorded digitally by the Camulodunum (Hawkes and Hull 1947, updated in Hull 1958; Bidwell and Croom 1999) and Chelmsford (Going 1987)) pottery type-series. Moreover, the specific form types had been quantified by EVEs (estimated vessel equivalents), which facilitated the creation of a high-quality dataset for the period c. 50 BC-AD 200. EVEs are the least biased means of pottery quantification (Orton et al. 1993), being unaffected by factors such as fragmentation, relative weight and subjectivity in counting incomplete vessels (as opposed to other measures of quantification, such as sherd count, sherd weight and minimum number of vessels respectively). EVEs are calculated by finding part of a pot that can be measured as a fraction of some whole, usually by measuring rim sherds. These percentages are totalled to give the number of estimated vessels in an assemblage. However, some pottery forms, especially those with small rims relative to the rest of the vessel (e.g. amphorae) are less suitable for quantification by EVEs. The amphorae at Elms Farm had only been quantified by weight and sherd count (Sealey forthcoming), although it was possible to account for this discrepancy in the ensuing analysis conducted here. Finally, the pottery assemblages had been helpfully coded by excavated area, corresponding to functional zones described in the discussion of site morphology, which is an added bonus for intra-site analysis.

2.1. Data description and quality