Homicide-suicides (HS) are relatively infrequent events. Yet, they are of great concern because they often result in the death of family members, young children, and cause additional morbidity, family disruption and childhood psychological trauma. The aims of our study were (a) to examine the sociodemographic, clinical, and autopsy characteristics of HS in Paris and its suburbs from 1991 through 1996, and (b) to analyze the psychodynamic determinants leading up to the onset of HS. Our findings are compared with those obtained in other international studies. For the purpose of the present study. HS was defined as a violent event in which an individual committed homicide and subsequently committed suicide within a few hours. The main results are as follows. During the six-year study period, 56 HS involving 133 victims were investigated at the Institute of Forensic Medicine of Paris. Seventeen events occurred in Paris and 39 in its suburbs. Paris has a population of approximately 2,200,000 while its suburbs are home to another 8,500,000. Of the 56 offenders, 48 (85%) were males. The mean age of offenders was 51 years in males (range, 24 to 83) and 40.5 years in females (range, 33 to 56). In 45 events (80%), the offenders used a gun for both the homicide and suicide. A knife was used in only 4 murders, strangulation in 4 other cases, with poisoning, arson, or beating occurring in one case each. In 9 cases, the offender used a different weapon for homicide and suicide. Among firearms, handguns were more likely to be used (26 cases) than shotguns (6 cases) or rifles (13 cases). In 40 cases the offender killed one victim, in 11 cases 2 victims, and in 5 cases 3 victims. The homicide victims consisted of 34 children (21 boys), mean age 8 years (range, 1 to 16), 29 spouses (26 females), 2 girlfriends, 10 strangers, and 2 relatives. Five pets were killed. HS were most likely to be committed in the home. A suicide note was found near the victims in 29 cases (50% of events). In 42 cases the offender was found to be severely depressed. Familial HS were the most frequent events followed by suicidal pacts. When male sexual proprietariness and amorous jealousy were involved, the HS perpetrators acted often impulsively in carrying out the HS. There had been a chronically chaotic relationship and frequent physical violence and verbal threats. The triggering event was often the female withdrawal or estrangement. Only in a few cases, other motivations such as occupational or money problems were involved. Chronic alcoholism was found in 16 cases. In conclusion, the offender was more likely to be a male, severely depressed, violent and jealous who killed his spouse, and often his children, with impulsiveness, but after numerous threats. Our findings suggest that HS differ from both homicide and suicide and thus occupy a distinct epidemiological domain requiring specific prevention programs.