This investigation is part of a program of study of the neural organization of defense behavior. Previous investigations [2, 8] have suggested that defense behavior, under the control of a defense motivational mechanism, and offense behavior, under the control of a separate motivational mechanism, are responsible for the intraspecific aggression of rats.
The primary releasing and directing stimuli  which facilitate the motor patterning mechanism of upright posture and boxing consist of tactile sensations received by the vibrissae and tactile hairs of the face. Removal or blocking of the vibrissae reduces the probability of shock-elicited boxing [4, 6] and anesthesia of the vibrissal pad is even more effective . Work in our laboratory indicates that following anesthesia of the vibrissal pad, there may be some recovery of function mediated by tactile hairs on other parts of the head, but that injections of local anesthetics which inactivate tactile sensation over the entire face and head region completely abolish the response in naive animals.
Tactile sensations from the head are not the only releasing and directing stimuli; vision may also perform these functions under certain circumstances. Pairs of animals with previous experience in the shock-induced fighting situation will continue to show the behavior despite removal of the vibrissae or anesthesia of the vibrissal pad  and, in our experience, despite inactivation of all tactile sensation from the head region. We have found that this boxing is visually-released; if the animal with the face anesthetized is then temporarily blinded by suturing the eyelids shut, the upright posture and boxing is completely abolished  .
We have investigated the neural circuitry necessary for the visually-released boxing which follows facial anesthesia. The visual pathways are cortical rather than tectal. Tectal lesions do not abolish the behavior, while lesions of the visual cortex are effective. The combination of unilateral lesions of the visual cortex and occlusion of the ipsilateral eye will abolish the response, while occlusion of the contralateral eye is ineffective and serves as a control procedure .
The present paper documents an unexpected phenomenon which we encountered in the course of tracing the neural pathways of visually-released boxing. Lesions of the ventrobasal thalamus abolished the response. This was not expected since the ventrobasal thalamus is not usually considered to be involved in visually-released behaviors, but is the main relay for transmission of tactile sensations from the head and body to the cerebral cortex . It appeared that the neural circuitry involved in the processing of visual stimuli to replace the inactivated tactile stimuli had incorporated higher levels of the normally tactile projection system into a secondary visual system releasing the behavior .