The Extended Aggression Scores were developed to quantify the aggressive Rorschach imagery produced by violent Antisocial Personality Disordered (ASPD; American Psychiatric Association, 1980 ) offenders. Despite their histories of real world violence, these subjects produced few Aggressive Movement (AG; Exner, 1993 ) responses. Why didn’t violent children, adolescents, and adults produce more AG responses? Considering their expression of uncensored pleasurable affect when relating their aggressive acts during their interviews, conscious censoring ( Exner, 1993 ; Meloy, 1988 ) did not adequately explain the paucity of AG responses among sentenced adults. Why would they describe their violent acts with pride and bravado during an interview and subsequently censor AG on the Rorschach? Conscious censoring among the Conduct Disorder (CD) children and adolescents, who frequently produced sexual content, seemed an equally unlikely explanation ( Gacono, 1997 ). Earlier Rorschach research ( Holt & Havel, 1960 ; Rapaport, Gill, & Schafer, 1946 , 1968 ; Schafer, 1954 ) provided clues to understanding the discrepancies between Rorschach production and the interview/historical data. Direct or implicit aggressive content was thought to imply tensions of aggressive impulse ( Rapaport et al., 1946 \ 1968 ). Initial findings ( Gacono, 1988 , 1990 ; Heaven, 1988 ) suggested that the paucity of symbolized aggression, represented by AG movement, might be due, in part, to the ego-syntonic nature of aggression in ASPD and psychopathic subjects. The clinical logic was that AG symbolized tensions of ego-dystonic aggression when produced by violent, antisocial patients. In the absence of binding the aggressive impulse, the violent individual would, instead, act it out, thus vitiating the need to symbolize it. Existing data supported this hypothesis: Exner’ ;s (1995) character disordered sample produced lower AG frequencies than his adult nonpatients; violent children and adolescents produce lower AG frequencies than child and adolescent nonpatients; and the majority of the forensic subjects with known histories of violence produce less AG than nonpatients and the Gacono and Meloy clinical samples without histories of violence ( Gacono, 1997 ; Gacono & Meloy, 1994 ). Despite the paucity of AG responses in ASPD records, other aggression imagery was not absent. Rather, the presence of other potentially scoreable aggressive imagery (see Gacono, 1988 , 1990 , 1997 ) allowed for the development ( Gacono, 1988 ) and refinement ( Gacono & Meloy, 1994 ; Meloy & Gacono, 1992 ) of five additional scoring categories: Aggressive Content (AgC), Aggressive Past (AgPast), Aggressive Potential (AgPot), Aggressive Vulnerability (AgV), and Sado-Masochism (SM). Since their introduction ( Gacono, 1988 ), the Extended Aggression Scores have received considerable clinical interest and empirical study. As noted in the Rorschach Workshops’ Alumni Newsletter (2000) concerning the work of the Rorschach Research Council, “Another project on which there has been good progress is the special score for Aggressive Content (AgC). Council has reviewed the criteria and guidelines for its applications and has evaluated the literature concerning it .. . A more precise interpretation of AgC responses will probably hinge on findings for other variables .. .” (p. 13). Additionally, the ROR-SCAN Version 6 Rorschach Interpretive Scoring System ( Caracena, 2002 ) now includes AgC, AgPast, AgPot and SM.. What began as an attempt to expand the scoring of Rorschach aggressive imagery in CD and ASPD subjects, has evolved into a larger study of aggression on the Rorschach. In this article we present information concerning the reliability, psychometric properties, and construct validity of the scores and discuss their clinical meanings. We conclude that while AgPotential, AgVulnerability and Sado-masochism need additional research, the current research supports inclusion of the more frequently appearing AgContent and AgPast scores in the Comprehensive System.