The purpose of this study was to establish the extent of variability in the visual fields of patients with glaucoma using a suprathreshold perimetric test strategy. The visual fields of 104 patients (126 eyes) were measured twice, on separate days, with a Henson CFA3000 central visual field analyser. All eyes had a diagnosis of glaucoma and had a minimum of three stimuli missed at 12 dB above their threshold estimate derived at the beginning of the first suprathreshold examination. The perimeter tested 132 locations within the central 25 degrees with a multiple stimulus suprathreshold strategy. Data were analysed on the basis of the difference in the number of recorded missed stimuli between the two visits. Large amounts of variability were found in the visual field results. Variability was largest when all levels of defect were included (> = 5 dB) and was found to be independent of the overall extent of loss. Results from the superior hemifield were found to be more variable than those from the inferior hemifield. The results highlight the problem of differentiating real change in visual field status from the enhanced variability seen in patients with glaucomatous visual field loss. The results from this study can be used to establish return-referral criteria for shared care schemes. The advantages of duplicate measures of the visual field are highlighted.