Two thousand paint samples taken from buildings in the South Wales, Mid-Wales and Monmouthshire areas were examined to discover if any from different sources were indistinguishable. Over 98% were differentiated by microscopy and solvent tests; the remainder were further examined by spectrography and pyrolysis on gas chromatography. An analysis of the results showed it was possible to calculate the chance against two random samples agreeing, both when only microscopy and solvent tests had been carried out, and when more sophisticated methods were used.