reduced by such exposure. Dixson (1980) has also reviewed the available data, suggesting that in some infrahuman primates the stress of defeat produces a marked reduction in circulating androgen levels. Archer (1990) points out that stressful experiences (including aggression) can lead to reduced testosterone levels in humans. It is not at all unlikely that such stressors influence as wide a range of hormonal factors as those evident in rats and mice.

There is clearly a complex impact of alcohol on levels of violence in U.S. and U.K. populations—the associations are generated by alcohol influencing a variety of processes and perceptions (Brain, 1986). Acute or chronic alcohol ingestion has major impacts on neurophysiology and endocrinology (Berry and Brain, 1986). With particular relevance to the present study, ethanol changes the levels of endogenous opioids and other peptides, lowers blood calcium level, alters carbohydrates metabolism, increases pituitary adrenocortical function, and profoundly reduces the secretion of LHRF, LH, and testosterone. The last finding suggests that alcohol does not increase aggression by augmenting androgen release, but combinations of alcohol and androgens may be especially lethal. It is likely that certain alcohol-related increases in aggression and violence are more concerned with inappropriate processing of information or signaling rather than disinhibiting aggression. There is evidence that some endocrine disorders predispose individuals to ingest alcohol.

These are exciting times for studies attempting to relate hormones to aggressive and hostile behaviors. The approaches are moving away from the highly simplistic view that particular hormones switch aggressiveness on and off, to acceptance that these chemical factors play complex roles at a variety of stages of development in particular species and in particular contexts by altering the predisposition to produce activities that are likely to receive the labels aggressive or hostile. Although such associations are much more plastic than the old truisms, they are no less important.

Genuine progress in this area (which has lagged in sophistication behind developments in the study of sexual behavior) is likely

if the quality of the behavioral analysis (which should be as direct and as detailed as possible) is matched to the sophistication of the endocrine manipulations and measures. The techniques for hormone assays are well developed for most current purposes, but it is obvious that we should look at a much wider range of hormones. Although still a long way from realization, nuclear magnetic resonance and magnetometry techniques might eventually hold the key to noninvasively estimating what is going on in vivo in clinical studies. In terms of research directions in the area of hormones and violence, it is suggested that studies on nonprimates, infrahuman primates, and humans all have utility to our understanding of the phenomena encountered in this research area. In nonprimates, it is suggested that one can examine the range of behaviors receiving the label aggression and can use molecular biologic and neurophysiologic techniques (rarely possible in other animals) to consider chemical transformations and hormone-brain interactions in greater detail. Such animals also provide valuable pilots for assessing the impact of currently little-investigated hormonal systems on aggression, as well as looking at associations among genes/hormones/behavior, drugs/hormones/behavior, and environment/hormones/behavior. Given current technologies, it should be possible to noninvasively examine heart rate as an indicator of arousal in many of these models. It should not be necessary to remind the panel that such studies should not be interpreted as providing support for a mechanism of violence. Studies with infrahuman primates are important not only because such animals have an evolutionary affinity to our own species, but because they enable us to examine the involvement of hormones over the life span in organisms that operate in complex social organizations. It is strongly advocated that early influences of hormones on subsequent behavioral potential be reexamined by using such species in well-controlled investigations that employ reasonable numbers of animals and look at alternative explanations (e.g., do primate mothers treat their androgenized female offspring differently?). The impact of hormonal changes around puberty should also be reinvestigated along with a more systematic evaluation of the impact of a range of hormonal and antihormonal treatments. So far as humans are concerned, there is an urgent need to perform multiple hormonal measurements (preferably using saliva samples to reduce stress effects) in well-evaluated, homogeneous groups of individuals. Multiple behavioral measurements concerning ongoing behavior are distinctly advantageous. Measuring hormones over the life span might be a valid aim, but

it is difficult to see how this could be achieved without great personal interference and ethical problems (do you jail a person for excess hormone production?).

Finally, it should not be assumed that one can consider the impact of hormones on such behavior without being cognizant of the contributions of environmental factors, social experience and other biological factors (e.g., genes, neural circuits, and drugs) to the generation of those activities that (rightly) cause such current concern.

The major public policy implication of the review is that there is not a simple relationship between any hormone and behaviors that will receive the labels aggression or violence. ''Raging hormones" do not cause violence. This having been said, there are diverse and subtle influences of hormones on the developing and developed individual, which can alter the predisposition for showing particular responses; authorities should be prepared to consider the involvement of these powerful messenger molecules in particular human processes (e.g., development of gender differences in behavior and onset of puberty) and recognize the possibility of mitigation in individuals with profound pathologies of the endocrine system. We should, however, move away from the highly simplistic view that hormones simply switch aggression on and off.

Another policy implication (because of the powerful and lasting effects of some hormones on morphology, physiology, and mood) is that we should be concerned about the impact of some clinical treatments (e.g., therapies of pregnant women) and borderline "misuse" (e.g., employment of anabolic steroids and human growth hormone by sports people and others). In the latter case, we should be especially concerned about hormone and alcohol combinations.

Because of their chronic effects, hormones and hormone derivatives offer the possibility of the development of therapies to assist individuals in coping with overwhelming behavioral problems. There is an urgent need to establish (a) which therapies are useful and (b) which subgroups of conditions are appropriate for treatment (we have already noted that rape, homicide, and violence are not homogeneous categories). Hormone treatments should never be considered as alternatives to other therapies but as an option. It is felt that appropriate hormone-based treatments should be combined with psychotherapy, counseling, and empathy training to achieve lasting results. The simple view of the relationship between hormones and violence leads to such therapies being regarded as a form of punishment or as a means of curing a