Discoloration around restorations may lead practitioners to suspect recurrent caries. Dentists, using conventional optical and tactile criteria, often fail to render the enamel-dentine junction completely caries-free during cavity preparation. The aim of the present laboratory study was to test the hypothesis that such residual caries could take up stain from tea and hence affect clinical judgement on the presence of recurrent disease. Cavities were prepared in freshly extracted carious teeth and restored with amalgam. Specimens were then temperature cycled in tea and chlorhexidine or in tea alone to encourage leakage. Subsequent removal of restorations showed staining of parts of the enamel-dentine junctions and histological examination showed these stained areas to correspond to areas of demineralization. Similarly prepared caries-free teeth showed no such staining. Thus residual caries in teeth with leaking restorations can, in the laboratory, take up stain. If this were to happen in vivo, stained residual caries could subsequently be misdiagnosed as recurrent caries.