A hypothesis that statistically infrequent role sets foster relatively high suicide rates is tested. An analysis of gender differences among American labourers confirms this perspective for females who had a suicide rate of 38.6/100,000 or 7.6 times that of females in general. Overrepresentation of males in this occupation did not decrease their suicide risk. An economic theory is consistent with the high rates observed. No support was provided for a Durkheimian theory of social class & suicide.