The testing of material in order to determine its conformity to a microbiological standard involves two main processes; selection of the samples of material to be tested, and the testing of this material. For the procedure to be reliable, the sampled material must be representative of the whole and the test must give a reliable result. The errors arising from the sampling stage are indicated in this paper, and sampling schemes which have been recommended for this purpose are discussed. The chief test methods (colony and total counts, dilution methods, presence/absence tests, and chemical methods) are then discussed, both from the point of view of their underlying statistical reliability and the assessment of operator errors. Finally it is concluded that the initial sampling stage is the most critical part of any test procedure and where representative samples are difficult to obtain, the enforcement of a code of practice in manufacture is preferable to attempts to control quality by subsequent testing.