Transcranial magnetic stimulation (TMS) of the posterior parietal cortex increases the latency of memory-guided saccades (Müri et al., 1996), as well as of visually-guided saccades (Müri et al., personal communication). Animal studies (Gnadt and Mays, 1995) show that the parietal area is also involved in vergence and combined saccade and vergence movements. This study examines in normal humans by means of the TMS the role of parietal cortex for saccades, vergence and combined movements. Single-pulse magnet stimulation (MegaStim 200) was applied on the right posterior parietal cortex of nine subjects; the stimulator output was set at 80% of the motor threshold. The targets of visually-guided eye movements were 6 LED's, one lit at a time. Three LED's were placed one at the centre and two at ±20° at a distance of 20 cm from the subject; the other three LED's were placed at a distance of 150 cm (at the centre and ±20°). The target LED was switched on 200 ms after extinction of the central LED (gap interval, aiming to produce short latencies). The TMS was applied 80 ms after the target LED was switched on. Each type of eye movement was repeated 10 times. Blocks with no TMS were also performed. Eye movements were recorded binocularly with the IRIS device. For pure saccades, ipsilateral or contralateral to the stimulated side, the latency was significantly (p < 0.05) longer with TMS than without TMS for individual subjects and for the group (the group mean difference was 35 ms). For pure vergence, the latency increased significantly with TMS (the mean difference was 25 ms for convergence and 45 ms for divergence). For saccades combined with convergence, the group mean latency was longer with TMS (25 ms) showing a tendency for significance. In contrast, for saccades combined with divergence, there was no difference for the group between the TMS and no TMS conditions. The results on pure saccades confirm prior studies. The new finding is that TMS also increases the latency of vergence movements and of saccades combined with convergence. This indicates the involvement of parietal cortex in all these types of eye movements. The finding of no group effect of the TMS for saccades combined with divergence, could be due to the longer latency of these movements. Perhaps, TMS is less effective when occurring outside of a certain time window during the process of computation of target location. Alternatively, the parietal cortex could differently be involved for different types of gaze shifts in space. E. Isotalo was supported by the Fondation pour la Récherche Médicale and Foundation of Maud Kuistila.