Homicide followed by suicide has been a neglected area in criminological theory and research. Also, available work is marked by a series of methodological limitations, such as small samples and lack of systematic multivariate analysis. This study analyzes data on 16,245 homicides in Chicago. A multivariate logistic regression analysis largely confirms the thesis that the closer the ties between offender and victim in a homicide, the greater the ensuing guilt and the greater the likelihood of suicide. Controlling for sociodemographic variables, killing an ex-spouse/lover increased the risk of suicide the most, 12.68 times. The odds of suicide are increased 10.28 times for killing a child, 8.00 times for killing a spouse, 6.11 times for killing a girlfriend or boyfriend, and 1.88 times for killing a friend. The act of homicide-suicide is conceptualized as stemming from a frustrated intimate relationship and a blend of self- and other-blame in one's attribution style. Notes, tables, appendix, references