The present paper describes how the formation of artefactual nitrosodimethylamine (NDMA) may jeopardize attempts at determining volatile nitrosamines in cheese by the mineral oil technique (MOT) or the glycerol isolation technique (GIT). When such isolation techniques were performed at an initial pH of about 11 and a final temperature of 100 degrees C (compared to 70 degrees C), up to about 10 micrograms NDMA per kg was formed. In a meat product, which was examined for comparison, even more extensive artefact formation (up to 70 micrograms NDMA per kg) was found. The addition of 10 mg/kg morpholine, as a readily nitrosatable substrate, to this meat product resulted (under the most unfavourable conditions) in formation of 250 micrograms/kg artefactual nitrosomorpholine. Alkalization of samples before nitrosamine analysis is a widely practised method to prevent artefact formation and has been emphatically recommended by an international working group, convened by the International Agency for Research on Cancer (Lyon). The mechanism of this peculiar artefact formation, which is strongly temperature- and pH-dependent, is not as yet known. It may however involve transnitrosation by C-, S- or O-nitroso compounds, e.g. nitrosophenols, nitrosothiols, alkyl nitrites, sugar nitrite esters, or lipid-nitrite reaction products. Several parameters of initial pH, final temperature and use of inhibitors under which the phenomenon will or will not occur are discussed. The artefact formation described here occurs mostly at pH values which are common to laboratory situations, but which are rare as for as the preparation of food in the household is concerned. Therefore, the effect does not appear to imply a risk to the consumer's health.