This fabric varies in colour but is generally oxidised in ranges of orange (2.5YR 6/6–6/8, 5YR 6/4), orange-brown (5YR 6/6) or red-brown (2.5YR 5/6, 5YR 5/4) with similar surfaces (5YR 7/6–6/6, 2.5YR 5/6), sometimes more light brown (7.5YR 6/4) or yellow (10YR 8/4) and mottled. Reduced examples tend to be grey (5/0–4/0) with brown-grey (10YR 6/1–6/2, 5YR 6/1–5/1) or pale brown (10YR 7/3) surfaces. Oxidised sherds occasionally have reduced cores and vice versa. The fabric is hard, normally with a hackly fracture and harsh or occasionally soapy surfaces. Brown (1994, 56) notes that the normal surface is soapy, with harsh ones resulting from overfiring.
The vessels are likely to be handmade, although petrographic studies provided no positive support for this (Woods 1994, 102–4). ‘Belgic’ style jars, large storage jars, reed-rim bowls, ledge-rim and hook-rim jars, and flanged bowls with rilling are typical. Of decorative motifs, fine scored lines are typical during the early period, thumbing from the second half of the 2nd century onwards.
Although a variety of fabrics were identified from the site (Woods 1994, 99–103), all the samples represented here belong to one general category, characterised by ill-sorted shell and limestone, ranging in size between 0.2–0.7mm. Abundant fossil shell dominates, with most fragments measuring 0.3–2.5mm; common limestone (occasionally fossiliferous) is frequently 0.3–1.0mm and serves to distinguish the fabric from other shelly wares housed in the collection. Other inclusions rarely exceed 0.5mm and are sparse, but include matrix-coloured clay pellets and multicoloured iron-rich fragments.
Mr John Cooper writes that the shells are crushed to fine, medium and coarse states, either deliberately or by sieving/screening naturally occurring shell fragments. One sample contained an oyster shell identified as Praeexogyra hebridica, by Mr Philip Palmer, who also suggested that the clay could derive from the Blisworth Clay (a Middle Jurassic, Upper Bathonian clay) found about three miles south of Northampton and twelve miles west of Harrold. It is likely that all the oyster fragments in the samples are of the genus Liostrea and also of the Blisworth Clay, or a similar shelly clay bed.
The sample is characterised by abundant ill-sorted fossil shell and limestone, frequently fossiliferous, set in a clean clay matrix. Two main size ranges can be distinguished, both common to abundant: 0.1–0.3mm and c 0.5–2.0mm. Sparse opaques (<0.2mm) are also present.
Nine kilns are known at the production site two kilometres from the village of that name (Brown 1994). As Cooper’s analysis suggests, the shells may have originated west of Harrold and thus the clay, or possibly the shells, may have travelled some distance.
Bedford Museum; Verulamium Museum, St Albans (likely from published descriptions of pottery)
Brown, A E, 1994 A Romano-British shell-gritted pottery and tile manufacturing site at Harrold, Beds, Bedfordshire Archaeol 21, 19–107
Marney, P T, 1989 Roman & Belgic pottery from excavations in Milton Keynes 1972–82, Buckinghamshire Archaeol Soc Monogr Ser 2
Wilson, M G, 1984 The other pottery, in Verulamium excavations 3 (S S Frere), Oxford Univ Comm Archaeol Monogr 1, 200–66
Woods, A, 1994 Petrological report on ceramics from Harrold, in A Romano-British shell-gritted pottery and tile manufacturing site at Harrold, Beds (A E Brown), Bedfordshire Archaeol 21, 99–104