Source of Roman stone for Aquae Sulis (Bath, England): field evidence, facies, pXRF chem-data and a cautionary tale of contamination

The Roman town of Bath (Aquae Sulis), renowned for its Temple to Minerva and thermal baths complex, is estimated here to have required around 500,000 m3 of stone for its construction. This huge amount of stone was likely to have been supplied from quarries within 5 km of the town, located towards the tops of the hills around Bath. Observations at the many old quarries show few features indicating Roman exploitation except for one Lewis bolt-hole and reports of chisel marks. The features of the majority of the stone in the Roman Baths-Temple Complex all suggest that the stone was sourced largely from the Combe Down Oolite Member (CDO; Jurassic, Bathonian stage), rather than the Bath Oolite Member (BO), of the Great Oolite Group. A portable X-Ray Fluorescence (pXRF) instrument, used to determine the trace element geochemistry of Bath Stone for comparison with the Roman stone, shows that the CDO and BO are very similar, except for different contents of Si, Al, Fe and Mn. These likely reflect variations in clay and organic matter content. However, with regard to the Roman stone and sculptures in the Baths-Temple Complex, all analyses of surfaces show enrichment in virtually all elements, but especially in P, Si, K, Al, S, Cl, Fe, Pb, Zn, Nb, and As. This contamination is largely attributed to the buried nature of the site (5�8 metres) from the 5th century AD until the end of the 19th century, during which time the stone would have been affected by groundwater, mostly derived from the hot-springs, with its high content of many elements. Analyses of cores cut into blocks of Roman stone show that the contamination is absent after 1�2 cm. This study demonstrates that care must be exercised in using geochemical analyses of ancient building materials for provenance studies, and that fresh surfaces of the material may well be required.