This paper discusses the practical difficulties experienced in registering and coding coronary events. The populations being monitored for fatal and non-fatal coronary events in the World Health Organization MONICA study are too large for surveillance of individuals. Routine medical and medico-legal sources have to be used to indentify potential events, which are then coded and categorized according to standard criteria. Methods are dependent on, and have to be adapted to the local system of medical care. Non-fatal cases in hospital are identified and registered either through their admission, "hot pursuit", or through their discharge, "cold pursuit". Each method has its own advantages and disadvantages. Local legal and ethical constraints are also responsible for differences between MONICA Collaborating Centres. Adequacy of investigation of events, and the availability and completeness of medical records are major determinants of the ease and quality of registration. Changes in medical care could cause spurious changes in event rates, and so potential biases need to be monitored and allowed for.