The purpose of this study was to estimate the risk of acute myocardial infarction related to smoking habits, especially the risk among smokers with a daily tobacco consumption of five grams or less. The study is based on data from the 1914 population examined by the Glostrup Population Studies at the age of 50 and 60 in 1964 and 1974. Information concerning deaths and cases of hospitalisation has been obtained from national registers up to 1985. Blood pressure, lipids, body mass index and physical activity were used as confounders. It was not possible to make a definite conclusion for the group smoking five grams or less daily as a class, since both the size of the group and the number of myocardial infarctions within it were small. When tobacco consumption was used as a quantitative variable the risk of myocardial infarction was found to increase with increasing amount but the relation was found not to be non-linear. The best description of the tobacco-related risk of myocardial infarction was a logarithmically relation to daily tobacco consumption.