Roadside survey is the most reliable method to study not only the profile of drunken drivers but also the overall involvement of alcohol in traffic. This presumes that the method used is adequate to which much attention was paid in the present study. The same team consisting of two police investigators and medical doctors (the authors) with the aid of 6-12 policemen and interviewers conducted all road blocks. All drivers were breathalysed with an alcolmeter-r (pst-m1). In suspected cases a blood sample was taken and the stage of drunkenness was examined by the physician on the road. Relevant data of the traffic were recorded and some of the drivers were interviewed as recommended by the OECD working group. More than 40000 drivers were investigated in six different areas of the Province of Uusimaa in 1979. The frequency of the drivers having ingested alcohol was only a few per cent. However, only in a minority of these cases the BAC was over 0.50% per thousand (lower statutory limit in Finland). Significant differences were found in the rate of drunken driving, e.g. The highest rates were found in the mornings. In many aspects the profile of drunken driving based on roadside surveys differed clearly from the profile of arrested drivers described in the first part of this study. (TRRL)