The nature of object-centred (allocentric) neglect and the possibility of dissociating it from egocentric (subject-centred) forms of neglect are controversial. Originally, allocentric neglect was described by and in patients who reproduced all the elements of a multi-object scene, but left unfinished the left side of one or more of them. More recently, however, Karnath, Mandler, and Clavagnier (2011) have claimed that the severity of allocentric neglect worsens when a complex "object" shifts from an ipsilesional to a contralesional egocentric position. On the basis of these and of other clinical data, showing that allocentric and egocentric neglect are strongly associated, they have questioned the possibility of dissociating these two forms of neglect, suggesting that egocentric and allocentric neglect constitute different manifestations of the same disturbed system. Since these statements were inconsistent with the clinical findings which had prompted the construct of object-centred neglect, we checked in a group of right brain-damaged patients, who had copied the original multi-object scene, if the degree of neglect for the left side of figures varied as a function of their position on the horizontal axis. Furthermore, we reviewed all papers where copies of other multi-object scenes had been reported. Results of both studies failed to confirm the assumption of a relationship between spatial location of the stimulus and severity of object-centred neglect. This discrepancy between our data and those obtained by Karnath et al. (2011) could be due to the characteristics of stimuli and of procedures used to evaluate "object-centred" neglect. If the stimulus is complex and the task requires its thorough exploration, the spatial location of the stimulus will influence the severity of "object-centred neglect". If, on the contrary, the stimulus is simple and can be identified with few eye fixations, the spatial location of the stimulus should not influence the severity of "object-centred neglect". In any case, our data confirm the possibility of dissociating allocentric from egocentric neglect. (Contains 2 tables.)