T HE DISSIMILARITY between persons who attempt suicide and those who compl6te the act has been demonstrated with respect to many important personal and social characteristics (1,3). For example, men, older persons, and whites are overrepresented in completed suicide, while women, younger persons, and nonwhites are overrepresented in attempted suicide. Nevertheless, the most reasonable assumption about the two groups is that the poputlation of completed suicides is drawn from that of attempted suicides but, as suggested by Crocetti (4), not on a random basis. Two complementary hypotheses are involved: (a) persons who have previously attempted suicide are more likely to commit suicide, that is, are at greater risk, than persons who have not, and (b) among attempted suicides, the more closely individuals approximate those who commit suicide in personal and social charaleteristics, the greater the likelihood of their succeed'ing in killing themselves, that is, the greater their risk. The first hypothesis receives some suipport from data on histories of suicidal behavior (attempts, threats, or communication of intent) in studies of completed suicide (5-7); the second hypothesis cannot be tested by these studies because the pertinent data are not provided, but the studies do provide clues regarding the charalcteristics of individuals within a group of attempted suicides most likely to kill themselves in a suibsequent attempt. The studies have important limitations as far as their use in testing the two hypotheses is concerned: (a) it is unclear whether persons not classified as having made previous attempts in fact did not do so or merely represented those for Whom no inlformation was available, and (b) it is not always possilble to determine what percentage of the sample made previous attempts since attempts and threats are sometimes combined into one category. T'he present study avoids these limitations by following a sample of attempted suicides with respect to the number of deaths from suicide within a 1-year period. This makes it possible to compare the suicide risk of attempted suicides and of su!bgroups within the sample with analogous groups in the general population. For the purposes of this study, the general population has been considered a group that has not attempted suicide. Several studies of attempted suicide have presented data on subsequent mortali'ty from suijcides (8-10), but none used general population data as a baseline. Moreover, these studies overlook the concept of risk and tend to minimize the problem by stating that the number of suicide deaths represents a very small proportion of those making an attempt. Such interpretations confuse the issue of risk by equating a small percentage of deaths with little risk or significance.