During the last 25 years, a series of epidemiological studies in North-Western Europe and U.S.A. have demonstrated a negative association between social class and the incidence of cardiovascular disease (CVD), that is, an increasing incidence the lower the social class. In studies where possible explanations of this negative gradient have been analyzed, it was concluded that the traditional individual risk factors, such as elevated blood pressure, high serum cholesterol, and smoking, could explain about one half of the differences demonstrated. In a prospective study of a cohort of 504 men from the County of Copenhagen, the participants were examined when 40 and 51 years old. At both examinations the social class of the participants was recorded in addition to a number of cardiovascular risk factors. The latter included both the traditional risk factors and some not previously analyzed in relation to social class. At the 51-year examination we found statistically significant negative associations between social class and the following risk factors: plasma fibrinogen (p less than 0.001), short height (p less than 0.001), smoking (p less than 0.05), physical inactivity in leisure time (p less than 0.01), shift work (p less than 0.05), job strain (p less than 0.05), living alone (p less than 0.01), and having a poor social network (p less than 0.05). Two factors showed a significant opposite association with social class: Type A behaviour (p less than 0.001) and physical inactivity at work (p less than 0.001). In the last 10-15 years, a tendency has been demonstrated in many countries towards a strengthened association between social class and cardiovascular risk factors. This tendency was not found in our cohort. It has been discussed whether some of the social inequalities observed could be due to selection, so that people with a favourable cardiovascular risk profile socially were upward mobile. We found no support for such a selection hypothesis in our study.