4. Compounds alter neuronal activity in particular regions of the CNS.

5. Hormonal changes may be correlated with alterations in localized neurotransmitters. There is ample evidence (see Brain, 1977–1979a) that hormones (especially steroids) alter local concentrations of neurotransmitters in the central nervous system and that some of these changes can induce further modifications of endocrine activity (i.e., they may, in some cases, be part of the feedback mechanisms). Hormones not only can change concentrations of neurotransmitters in adult animals but can produce biogenic amine changes in the brains of neonates and even alter neural enzyme systems in immature animals.

One should note that attempts to link hormonal actions to changes in biogenic amines and other factors are much rarer in aggression research than in studies involving sexual behavior (e.g., Meyerson, 1983). Changes in different patterns of these biogenic amines have (in some cases) been related to forms of aggressive behavior in rats and mice (reviewed in Daruna, 1978).

In line with the ethoexperimental approach, one should note that a number of motivational systems may respond to a single manipulations of an animal's endogenous hormones (Adams, 1980). Thus, patrolling-marking, male and female sexual motivational systems and aggressiveness may all be concomitantly changed by exposure to androgens and estrogens. One should also note that other "sex" hormones (e.g., hypothalamic LHRF and pituitary gonadotropins are being implicated in aggression and violence.

The many studies that have investigated the abilities of sex steroids to maintain motivation for social aggression in mice (e.g., Brain, 1979b; Brain and Bowden, 1979) are of relevance here. These hormones can be further implicated in the modulation of aggressiveness

by citing studies correlating endogenous titers with fighting and/or dominance. The motivational states underlying such responses are not, however, identical to spontaneous motivational states (e.g., those that serve locomotor activity; Schallert, 1977).

The behavioral influences of androgens in male rodents may depend on their metabolic conversion (Naftolin and Ryan, 1975). There are obvious species and strain differences, but a compelling body of evidence has been accumulated (Larsson et al., 1973; Parrott, 1975, 1976), suggesting that testicular androgens are neurally converted into estrogenic metabolites before motivating ejaculatory behavior in the rat. Brain and Bowden (1979) and Brain (1983) have provided some support for the idea that androgens are similarly neurally aromatized before having their motivational effects on fighting in "TO" strain mice. Brain et al. (1983) have reported that natural (e.g., estradiol) or (especially) synthetic estrogens (e.g., diethylstilbestrol) and aromatizable androgens (e.g., testosterone) are the most effective compounds in terms of their abilities to maintain motivation for social conflict in castrated mice. Since this study, Simon and Whalen (1986) have suggested that the subject's genotype strongly influences which sex steroids are implicated in the control of male aggression in mice. The use of intact male stimulus animals also strongly indicates that estrogen directly increases aggressiveness in mice (all rodent studies in which estrogens suppress aggression depend on treating both intact subjects with these steroids, altering their stimulus effectiveness as targets for attack). Estrogens are implicated in the control of aggressive motivation in other species or sexes of animals. Payne and Swanson (1972), for example, recorded that attacks directed by castrated male golden hamsters toward intact male opponents are restored by injection of the former animals with estradiol benzoate. It is also of interest to note that Harding (1989) has shown that androgens and estrogens interact to modulate social behaviors including aggression in songbirds such as zebra finches (Poephila guttata) and red-winged blackbirds, suggesting that the metabolic products of androgens are important in such species also.

One should reiterate that other aggression models using rodents (especially mice) give different associations between hormones and behavior. Conner et al. (1983) have shown that shock-induced fighting has hormonal correlates that are not too dissimilar from social aggression (e.g., sex differences are evident and androgens have generally stimulatory actions). Svare, in an excellent series of studies (reviewed in Svare and Mann, 1983; Svare, 1989), showed that the major impact of hormones on maternal aggression

is via their effects predisposing the animals receiving suckling stimulation. Haug and Brain (1989) have compared and contrasted the attack behavior by group-housed mice on lactating intruders with the more utilized social aggression. They find that this form of attack produces a mirror-image hormonal picture, with castration stimulating attack by males and replacement with androgens or estrogens suppressing this "female" form of attack. One can argue (with considerable justification) that we should attempt to learn as much as possible about the endocrine correlates of many forms of rodent behavior. Not only will there be a gain in theoretical knowledge, but because different kinds of aggression in many species (including our own) are motivated differently (see introductory material), we need such information.

Since this account considers agonistic behavior, it is worth mentioning that avoidance of attack (which can be regarded as the other end of the spectrum of activities that makes up agonistic behavior) seems much more influenced by ACTH and the adrenocortical hormones (see Leshner and Roche, 1977; Leshner, 1980). These studies should be extended because it is clear that "stress-related" hormones are commonly released in social encounters, and it appears that they may influence the progress and eventual outcome of interactions.

The hormonal bases of learning to be submissive or aggressive are of great relevance but far from fully evaluated. Archer (1977) suggested that the testosterone-induced increases in persistence in castrated mice may underpin the effect of this hormone on intermale fighting behavior. Brain (1979a) has reviewed some of the evidence in rodents that ß-lipotropic hormone, ACTH (more especially the 4–10 peptide sequence), melanocyte stimulating hormones, and a variety of related peptides (e.g., enkephalins and endorphins) can alter the acquisition and retention of a variety of reward- and aversion-mediated responses. There is also some evidence that adrenal glucocorticoids influence ongoing or subsequent avoidance reactions. One may consequently suggest that one of the ways in which hormones influence aggression and violence is by mediating learned responses associated with persistence, submission, and avoidance. Consequently, one should look for interactions between hormonal variations and subsequent responding. For example, markedly stressful situations (even in our own species) could predispose individuals to show avoidance, leading to social isolation and subsequent behavioral problems. Anabolic steroids (androgens) might very easily increase persistence in body builders.