The ability to provenance crop remains from archaeological sites remains an outstanding research question in archaeology. Archaeobotanists have previously identified the movement of cereals on the basis of regional variations in the presence of cereal grain, chaff and weed seeds (the consumer-producer debate), and the use of weed seeds indicative of certain geological areas, principally at Danebury hillfort. Whilst the former approach has been heavily criticised over the last decade, the qualitative methods of the latter have not been evaluated. The first interregional trade in cereals in Britain is currently dated to the Iron Age hillfort societies of the mid first millennium BC. Several centuries later, the development of urban settlements in the Late Iron Age and Roman period contained residents reliant on food produced elsewhere. Using the case study of central-southern Britain, centred on the oppidum and civitas capital of Silchester, this paper presents the first regional quantitative analysis of arable weed seeds in order to identify cereal origin. Analysis of the weed seeds present alongside the fine-sieving by-products of the glume wheat Triticum spelta (spelt wheat) shows that the weed floras of samples from diverse geological areas can be separated on the basis of the geological preferences of individual taxa. A preliminary finding is that rather than being supplied with cereals from the wider landscape of the Hampshire Downs, crops were produced within close proximity of Late Iron Age Silchester. The method presented here requires further high quality samples to evaluate this, and other instances of cereal movement in the past.