A variety of principles have been proposed to describe the organisation of visual representations in the human brain. One simple proposal is that representations are divided by category, with focal cortical areas selectively processing stimuli from a given class. Currently there is fMRI evidence for three strongly category-selective visual areas: the fusiform face area, the parahippocampal place area, and the extrastriate body area. Are these areas exceptional, or are similar focal, selective activations seen for other categories? In order to provide a strong test of selectivity, each category tested must be compared against a wide range of stimuli from other classes. To this end, 12 subjects were scanned in a blocked-design fMRI study (whole-brain, 1.5T) while passively viewing 40 colour photos of stimuli from each of 20 categories: birds, bodies, cars, chairs, clothes, crystals, faces, fish, flowers, fruits & vegetables, insects, musical instruments, mammals, microbes, prepared foods, reptiles, scenes, spiders, tools, and weapons. We performed 3 kinds of contrast in SPM: 1) each category vs all others; 2) category pairs, e.g. mammals vs. tools; 3) sets of categories, e.g. living vs. non-living things. (We will also discuss anatomically-defined regions of interest, and correlation analyses of distributed patterns of category-related activity). Many contrasts revealed activated clusters, but closer examination of the timecourses typically showed that these areas were not strongly selective. For example, in many cases a category not involved in the contrast activated the cluster more strongly than those used to define it. Other clusters responded more strongly to one subset of categories than to the others, but with no apparent organising principle. In general, while there is great variability across the visual system in the response to these stimuli, the majority of categories do not elicit focal, selective activations: strong category-selectivity is rare.