The Cook and Medley (1954) Hostility (Ho) scale has been used in several important studies evaluating potential health consequences of hostility. A relative lack of compelling information about the construct validity of the Ho scale, however, has raised concerns about the appropriate interpretation of previous research. In this study, 60 married couples discussed a low conflict topic, a high conflict topic, and then a second low conflict topic. High Ho men responded to the high conflict topic with significant increases in self-reported anger and anxiety and overt hostile behavior, but low Ho men did not. Furthermore, compared to low Ho men, high Ho men blamed their wives more for their usual disagreements on the high conflict topic and saw their disagreement-engendering behavior as more intentional. Among women, Ho scores were weakly related only to overt hostile behavior. Finally, couples consisting of two low Ho persons displayed a uniquely agreeable interactional style.