The hypothesis tested in the present experiments is that phobics show poor focused attention for phobic stimuli. A test of the hypothesis was undertaken using spider phobics' performance in a recognition memory task involving dead spiders as an index of processing. The first experiment confirmed the hypothesis of poorer processing of spider stimuli in phobics, but only for big spiders on a post hoc division of spider stimuli. These were interpreted as being the more arousing. There was also a tendency for desensitization to improve spider recognition. The second experiment investigated recognition memory under 'ordinary' and 'elaborated' processing conditions, but the hypothesis of poorer spider recognition in phobics was confirmed only in a post hoc correlational analysis for the elaborated processing condition. The hypothesis is related to findings that long-term benefit from exposure treatments is facilitated by focused attention.