It has been pointed out that much of the literature on brain mechanisms of aggression fails to distinguish between different kinds of aggressive behaviour, and that this may account for some of the apparent contradictions in different publications(1). In the experiments reported here the effects of hypothalamic lesions were measured on two different types of aggressive behaviour: defence behaviour, which consists of a defensive upright posture and boxing in the shock box; and territorial behaviour, which consists of the offensive side posture and the full attack posture with biting and kicking by a socially isolated home rat against an intruder. These postures associated with social interactions are quite stereotyped in the rat. They are also present in modified form in other rodents and have been illustrated and catalogued previously(2).

We used adult male hooded rats, raised in the laboratory. Fighting in response to foot-shock (defence behaviour) was elicited by placing a pair of rats in a 'Plexiglas' cage 20 x 15 cm with an electrified grid floor. Each trial consisted of a series of twenty 60 cycle a.c. shocks lasting 0.5 s and delivered once every second. A grid scrambling device was used to guarantee that the rats could not escape the shock by standing on certain pairs of grids. Fighting threshold was determined by the titration method; that is, shock intensity was raised in a trial after one in which the rats did not fight and was lowered after one in which they did fight. Fighting was defined as defensive upright posture and boxing by both rats in the pair. Among normal rats these were the predominant responses seen, although at high shock intensities the rats sometimes bit at each other or leapt and rolled over together biting and kicking. Behaviour similar to the offensive side posture and full attack posture was not exhibited by normal rats in the shock box.

Territorial behaviour was obtained from rats that had been socially isolated for at least 1 week in a large home cage (1.4 m2 of space on six levels with connecting ramps). Fighting, defined as offensive side posture and/or the full attack posture with biting, was exhibited against a male intruder placed in the home cage. The behaviour was quantified by determining the number of minutes in which fighting occurred during the 10 min period after the first occurrence. If no fighting occurred after 30 min, the trial was terminated.

All pairs of rats tested in the shock box showed rearing and boxing. Some rats did not fight off an intruder in the territorial experiment, however, and they were not used in the study.

Testing a rat in the shock box inhibited it from showing territorial fighting characteristics in the home cage. The rat seemed to become conditioned in the shock box to expect pain when approached by another rat. For this reason, the experimental animals were not tested pre-operatively in the shock box, but only after the initial post-operative tests had been carried out in the territorial situation. Post-operative results in the shock box were compared only with results in normal controls.

Lesions were placed in the hypothalamus by stereotaxic procedure using a radio frequency lesion maker. Surgery was performed aseptically on animals anaesthetized with 'Nembutal' (60 mg kg-1) and the rats were allowed 3 days for recovery before behavioural testing. Lesions involving the lateral hypothalamus abolished spontaneous feeding, in which case the animals were maintained by stomach tube feeding. Lesions were assessed on cresyl violet stained serial frozen sections.

Only rats with symmetrical lesions were included in the study. The experimental results (Fig. 1) depended on whether the lesion substantially destroyed the medial hypothalamus (six rats), the lateral hypothalamus (six rats) or both (six rats). The lesions were not confined to one particular structure, but lateral lesions always interrupted most of the medial forebrain bundle bilaterally. Medial lesions always destroyed the posterior hypothalamus and made substantial invasions at least into the dorsomedial hypothalamus and the posterior portion of the ventromedial nucleus.

Lesions involving both lateral and medial hypothalamus on both sides abolished territorial behaviour and simultaneously enhanced defence behaviour in the shock box, both effects being very strong. Rats which had previously attacked an intruder male within 1 min of the intrusion and persisted with biting and offensive side posture for most of the succeeding 10 min were rendered totally docile in the same situation post-operatively. The intruder might sniff at the rat, wander around him, or even lie down on him, but the rat with the lesion would ignore the intruder or sniff at him and turn away. The same animal, on the other hand, would show defence upright posture and boxing in the shock box at thresholds only one- tenth of the intensity of normal thresholds. Whereas normal rats usually fought only during presentation of the shocks, the rats with lesions usually continued to fight after the shock was discontinued and often fought spontaneously between trials.

By confining lesions to the medial and lateral hypothalamus it was possible to separate the two effects. The enhancement of defence behaviour in the shock box was caused by lesions of the medial hypothalamus; a lesion confined to the lateral hypothalamus on both sides had no effect. The abolition of territorial behaviour was caused by lesions of the lateral hypothalamus; a lesion confined to the medial hypothalamus had no effect.

This work was supported by a USPHS Biological Sciences training grant.

(1) Moyer, K., Comm. Behav. Biol., 2, 65 (1968).

(2) Grant, E., and Mackintosh, J., Behaviour, 21,246 (1963); Barnett, S., in The Rat: A Study in Behaviour, 81 (1963).