Ponto-geniculo-occipital waves and rapid eye movements (eye saccades) are two prominent phasic events of paradoxical sleep which occur in conjunction. Although they have been studied intensively, the neuronal link between these two events is still poorly understood. On the basis of our previous results, combining brainstem transections and carbachol microinjections, we postulated that the oculomotor and ponto-geniculo-occipital systems do not work in series, but in parallel, and that the caudoventral pontine tegmentum might represent a structure controlling and/or co-ordinating the simultaneous production of the two phenomena. This hypothesis was further supported by the demonstration that, during paradoxical sleep, the instantaneous velocity of eye saccades in bursts is higher than that of isolated ones which, in turn, are more rapid than waking saccades. This acceleration of eye saccades in bursts also seems to be under the cholinergic control of the caudoventral pontine tegmentum. In order to test the hypothesis that this area may be a prime mover leading to the simultaneous appearance of these two phasic events as a whole, we investigated, in the present study, the effects of pharmacological stimulation (with carbachol) and inhibition (with atropine) of the caudoventral pontine tegmentum on the production and the characteristics of eye saccades and ponto-geniculo-occipital waves. Cats' eye movements were recorded using the technique of the scleral search coil in a magnetic field, together with sleep-waking parameters. We found that: (i) unilateral microinjections of carbachol (0.4 microg) induced, during waking, a majority of long bursts of ponto-geniculo-occipital waves (i.e. bursts containing at least five waves) which had intra-burst intervals similar to natural ones (48-259 ms) and decreased the frequency of isolated ponto-geniculo-occipital waves; (ii) unilateral microinjections of atropine (2.4 microg) strongly decreased, during paradoxical sleep, the frequency (number/min) of eye saccades in bursts directed contralaterally to the side of the injection (by 48-54%) and reduced the velocity of these saccades to that of isolated eye saccades. Atropine also significantly reduced the frequency (by 60%) of all types of bursts of ponto-geniculo-occipital waves, with a maximal effect (80% reduction) on long bursts of ponto-geniculo-occipital waves, while it increased the frequency of isolated ponto-geniculo-occipital waves. However, atropine did not change the value of intra-burst intervals. These findings support the hypothesis that eye saccades in bursts and associated ponto-geniculo-occipital waves are generated as a whole by a common structure and that this structure is at least partly defined by the caudoventral pontine tegmentum.