The effects on the sexual behaviour of male rats of excitotoxic amino acid-induced lesions of the medial preoptic area-anterior hypothalamic area (mPOA/AHA), infusions of beta-endorphin, alpha-melanocyte stimulating hormone and naloxone into the mPOA/AHA, systemic naloxone and castration were compared using different behavioural paradigms. These included measures of unconditioned copulatory behaviour, instrumental responses for an oestrous female presented under a second-order schedule of reinforcement, conditioned place preference and partner preference. The results demonstrate that manipulations of the mPOA/AHA markedly affect consummatory aspects of sexual behaviour (mounting, intromitting and ejaculating) but tend not to affect appetitive or reward-related aspects of sexual behaviour, although intra-mPOA/AHA alpha MSH did result in a small increase in instrumental responses, while beta-endorphin infused into the mPOA/AHA also abolished preference for an oestrous over an anoestrous female. Systemic naloxone, on the other hand, reduced instrumental behaviour and a place preference conditioned by prior sexual interaction, while the same compound infused into the mPOA/AHA markedly facilitated copulatory responses but did not affect other measures of appetitive sexual responses. Castration caused an extremely rapid attenuation of conditioned place preference which was apparent before the males had experienced reductions in their copulatory performance. This treatment only slowly reduced partner preference. The results indicate that the use of several behavioural procedures can reveal discrete actions of neuroendocrine treatments on separable psychological processes which underly the integrated pattern of masculine sexual behaviour. In particular, they suggest that the mPOA/AHA is especially concerned with the copulatory responses of mounting and intromitting, but is much less important for a variety of appetitive sexual acts as well as sexual reward, as measured in the place preference procedure. The marked effects of castration on conditioned place preference taken together with the lack of effect of lesions of the mPOA/AHA on this measure indicate that testosterone affects sexual reward-related processes by an action at a site other than the mPOA/AHA. The implications of these findings are discussed.