This study tests two hypotheses regarding factors affecting arrest of the perpetrator in domestic violence incidents. Black's relational-distance thesis is that the probability of arrest increases with increasing relational distance between perpetrator and victim. Klinger's leniency principle suggests that the probability of arrest is lower for male perpetrators assaulting female intimate partners, compared with other scenarios. The authors employed marginal logistic regression models using incident-based data from the National Survey of Violence and Threats of Violence Against Women and Men in the United States, 1994-1996, to test both effects. They found support for Black's thesis: The likelihood of arrest was lower when the perpetrator was an acquaintance, a relative, or a romantic partner of the victim, versus a stranger. However, the authors' results failed to support Klinger's hypothesis. They found that men were more likely to be arrested when assaulting a female--regardless of relationship status--compared with assaulting another male. (Contains 2 tables.)