The makers of the Humphrey perimeter have developed a rapid thresholding program, FASTPAC, to address the problems of fatigue and poor test performance associated with prolonged perimetry. The performance of FASTPAC was compared with the standard threshold program of the Humphrey visual field analyser (program 24-2) in 44 glaucoma patients. FASTPAC reduced the mean test time by 35% (from 12.6 to 8.2 minutes), owing to a reduction in the number of stimulus presentations. FASTPAC underestimated the mean deviation (MD) (Wilcoxon, p = 0.007) and corrected pattern standard deviation (CPSD) (Wilcoxon, p = 0.005). The sum of the differences between FASTPAC and the standard program was -1.19 (SD 2.37) dB for MD and 0.97 (2.14) dB for CPSD. The measured difference between the two methods was independent of the value of either MD or CPSD, indicating that the measurement error was just as likely to occur with either early or advanced visual field loss. No difference was noted in short term fluctuation or in the reliability indices of test performance. These results indicate that FASTPAC, while considerably faster, is not as accurate as the standard threshold program at measuring retinal sensitivity in glaucoma. This inaccuracy may be offset in practical terms by greater patient acceptability and by increasing the number of patients capable of performing reliable threshold perimetry because of the reduction in duration of the FASTPAC program.