Abstract
1. In unrestrained locusts, in which responsiveness of the descending contralateral movement detector (DCMD) unit fluctuates after the initial habituation, spontaneous or induced antennal cleaning has a very high probability of coinciding with low DCMD responsiveness. 2. First visual stimuli which coincide with cleaning still evoke an undiminished DCMD response, which in turn does not interrupt cleaning; the two are not mutually inhibitory. Their inverse probability of occurrence must derive from a common factor(s) with reciprocal effects. 3. Lesions to the CNS which decrease sensory input increase the probability of grooming reflexes, similar to antennal cleaning, and decrease the probability of high DCMD responsiveness. 4. The main factor affecting DCMD responsiveness is known to be neural, driven (probably directly) by a variety of sensory inputs. The responsible system thus appears to affect also the otherwise unrelated activity of antennal cleaning. 5. These observations are consistent with the hypothesis of a general arousal system in the insect CNS, deriving from sensory input and influencing several aspects of nervous function and behaviour. 6. Passive movement of the head brings about recovery of DCMD response, due the recruitment of a new peripheral visual pathway. Movement of the head during antennal cleaning does not do this. Proprioceptive reafference therefore plays no role. 7. As the DCMD is not inhibited during cleaning (2), the CNS therefore seems to compensate for the shift on the retina of a familiar stimulus; as first visual responses are high during cleaning, this mechanism does not operate for unfamiliar stimuli. The implications of this are discussed, and it is suggested that the site-specific habituation properties of the afferent pathways to the DCMD could provide an appropriate ‘memory’ for this integration.