One of the main strengths of the stream analogy is its ability to deal with overall patterns of lethal violence and account for regularities that cannot be addressed by traditional studies of suicide or homicide in isolation from one another; however, there are certain homicidal and suicidal acts to which the model may not be applicable at the microlevel. So-called felony homicide, in which the victim "gets in the way" during the commission of some other crime (such as robbery) belong in this category. It could be argued, on the other hand, that the victim in such cases becomes a new source of frustration by offering active resistance or non-cooperation, but the imputation of blame in such cases is momentary and situated rather than a product of culturally based patterns of attribution. Further research pertinent to the stream analogy should focus on the development of valid, general-use, macrolevel measures of frustration and the culturally based attribution of blame to self or others. Another fruitful avenue for research would be the inclusion of a third stream of lethality that pertains to collective violence. Political assassinations and deaths as a consequence of riots, strikes, insurgencies, and wars would be included in this third channel of lethal violence. For now, it can be concluded that the stream analogy has remained remarkably robust in contributing to the understanding of the use of violence against self and others.