In this study we examined Kosslyn's (1987) claim that the right hemisphere exhibits a relative superiority for processing metric spatial relations, whereas the left hemisphere exhibits a relative superiority for processing categorical spatial relations. In particular, we examined whether some failures to observe strong visual field (VF) advantages in previous studies might be due to practice effects that allowed individuals to process tasks in alternative manners (e.g., to process a metric task using a categorical strategy). We used two versions of a task previously employed by Hellige and Michimata (1989) in which individuals judge the metric (distance) or categorical (above/below) spatial relations between a bar and a dot. In one version, the position of the bar was held static. In another, the bar's position varied. This manipulation prevented participants from using the computer screen as a reference frame, forcing them to compute the spatial relationships on the basis of the relevant items only (i.e., the bar and the dot). In the latter, but not the former version of the task we obtained evidence supporting Kosslyn's hypothesis, namely, a significant right visual field (RVF) advantage for categorical spatial processing and a trend toward a left visual field (LVF) advantage for metric spatial processing. Furthermore, the pattern of results for trials on which information was presented centrally (CVF trials) was similar to that observed on RVF trials, whereas the pattern for trials in which identical information was presented in each visual field (BVF trials) was similar to that observed on LVF trials. Such a pattern is consistent with Kosslyn's suggestion that categorical processing is better suited for cells with small receptive fields and metric processing for cells with larger receptive fields.