In an inter-laboratory survey, the pour plate, surface spread, agar droplet and spiral plate methods were used in parallel with the surface drop method for enumeration of micro-organisms in foods. Good agreement was obtained between all surface methods of enumeration, but there was poor agreement between molten agar methods and the surface drop method. A total of 1143 samples of food that were ready for consumption at the point of retail sale were examined. Eight types of food products were chosen: meat pasties, sausage rolls, real-cream slices, synthetic-cream slices, mayonnaise-based coleslaws, faggots, patés and continental sausages. The results of this survey suggest that the upper limit for an acceptable viable count should vary according to the food product. Salmonellae were not isolated on any occasion. Potentially harmful organisms were not isolated at levels expected to constitute a public health hazard. Information concerning the nature of the product, the total viable count, the presence or absence of pathogenic, toxigenic or indicator organisms, the spectrum of the bacterial flora and the relative predominance of each organism must all be considered when assessing the microbiological acceptability of retail 'ready to eat' products.