OBJECTIVE To compare the threshold variability of blue-on-yellow (B/Y) perimetry with white-white(W/W) perimetry, and evaluate the reproducibility of B/Y perimetry. METHODS The B/Y perimetry and W/W perimetry in the Octopus 101 perimetry were used to examine the visual fields of 12 normal subjects (24 eyes), 16 cases (32 eyes) of primary open angle glaucoma (POAG), and 7 cases (14 eyes) of suspected POAG respectively. The B/Y perimetry and W/W perimetry were repeated to examine within 2 weeks. The point by point threshold variability of the two perimetries were compared and analysed. RESULTS The total mean threshold variability for B/Y perimetry (2.61 +/- 0.94) dB was greater than that for W/W perimetry (2.11 +/- 0.90) dB in all subjects, but it had no significance (P = 0.6244). The total mean threshold variability for B/Y perimetry (2.07 +/- 0.54) dB was significantly greater than that for W/W perimetry (1.50 +/- 0.34) dB in normal subjects (P = 0.0006), while no significance was found in suspected POAG and POAG groups between the two perimetries (P = 0.0523 and 0.9371). The threshold variability in the areas of some eccentricities for B/Y perimetry were significantly greater than that for W/W perimetry in normal subjects and suspected POAG group, but no significance was found in POAG group for all eccentricities between the two perimetries. CONCLUSIONS The threshold variability for B/Y perimetry was greater than that for W/W perimetry in normal subjects, but no significance was found in suspected POAG and POAG between the two perimetries.