Programmes of suicide prevention require for their planning accurate epidemiological surveys. Doubt has been cast on the accuracy of many existing surveys because of the realization that suicide is under-reported and because of the lack of consistency in the procedure for suicide ascertainment. Two studies are described in this paper which attempt to examine the problem. The first is part of an international study supported by the World Health Organization. In it, Denmark and England are compared, and it is shown that there are striking differences in suicide ascertainment procedure between the two countries. Next, on a blind basis, coroners and their opposite numbers in Denmark examine a sample of each other's case records. It is found that the Danes consistently report more suicides than do the English coroners on the same case material; thus considerable doubt is cast on the supposed difference in suicide rates between the two countries. In the second study, deaths by poisoning are examined for certain coroners' districts in England and Wales. It is shown that there is considerable variation from one district to another in the relative proportions of these deaths which achieve an accident, an open or a suicide verdict. This suggests that in England and Wales coroners may not be consistent in their suicide ascertainment criteria. Hypotheses attempting to account for differences in suicide rate based on such epidemiological surveys should be viewed with great caution.