Abruptly presented items capture attention automatically so they constitute the first items to be examined [Yantis and Jonides, Journal of Experimental Psychology: Human Perception and Performance, 1984;10:601; Jonids and Yantis, Perception and Psychophysics, 1988;43:346; Theeuwes, Perception and Psychophysics, 1992;51:599; Theeuwes, Journal of Experimental Psychology: Human Perception and Performance, 1994;20:799]. This attentional priority can be controlled in a top-down manner by directing attention towards the locus of interest [Yantis and Johnson, Journal of Experimental Psychology: Human Perception and Performance, 1990;16:812; Theeuwes. Perception and Psychophysics, 1991;49:83; Miller, Perception and Psychophysics, 1989;45:567; Folk et al., Journal of Experimental Psychology: Human Perception and Performance, 1992; 18:1030]. The premotor theory of attention [Rizzolatti et al., Neuropsychologia 1987;25:31; Rizzolatti et al., Attention and Performance XV, 1994, p. 231] assumes that the mechanism responsible for the attentional shifts is strictly linked to that responsible for eye movements, and several studies [Corbetta et al., Society of Neuroscience Abstracts 1997;23:122.12; Nobre et al., Brain 1997;120:515; Theeuwes et al., Journal of Experimental Psychology: Human Perception and Performance, 1999;25:1595] suggested that the premotor cortex plays a role in the control of attention. However, the nature of this involvement is still unclear. We have been asking a patient (RJ) with a damage of the right premotor cortex to decide whether a target had a discontinuity on its right or left side. The absolute location of the target was pre-cued. In Section 2, an interference was observed when a sudden onset occurred in the visual space, suggesting that RJ was not able to control attentional capture. The possibility to attribute this interference to an insufficient focalization of attention or a grouping effect were discarded by Sections 3 and 4, respectively. Section 5 revealed that this interference followed exclusively the onset occurring in the hemifield opposite the one containing the target (meridian effect [Rizzolatti et al., Neuropsychologia 1987;25:31]). The results suggest that the control of attentional capture may be achieved by keeping constant the parameters of the appropriate oculomotor program.