BACKGROUND Diurnal variations in vision have been reported in radial keratotomy (RK) patients. Applicants to certain professions may meet an uncorrected vision standard at the time of testing but fail to meet that standard at another time. METHODS Ten firefighter applicants who had undergone RK and nine normal subjects attended for two morning visits, and two afternoon visits. At each visit subjective refraction, uncorrected visual acuity, best-corrected visual acuity, automated refraction, and keratometry were measured for each eye. Monocular contrast sensitivity was measured using the Pelli-Robson chart with and without a glare source (brightness acuity tester, BAT) and for both natural and dilated pupils. RESULTS The RK group showed a significant myopic shift from morning to afternoon (mean = -0.41 +/- 0.33 D; t = 3.92, p = 0.004), which was well correlated (r = -0.86) with significant steepening of the corneal curvature (mean = +0.41 +/- 0.36 D; t = -3.65, p = 0.005). The controls showed no change in either refractive error (mean change = +0.06 +/- 0.42 D) or corneal curvature (mean change = +0.05 +/- 0.08 D). No significant difference in corrected visual acuity or contrast sensitivity was found between the RK and control groups for natural pupils. For dilated pupils, the RK patients showed significantly poorer log contrast sensitivity both with and without glare (RK mean = 1.49 +/- 0.11; control mean = 1.67 +/- 0.11; t = 3.21, p = 0.005). All RK subjects met the firefighter visual acuity standard on the initial visit, three subjects did not meet the standard at an afternoon examination. CONCLUSIONS We propose that, if RK applicants are to be considered for employment, existing visual standards be amended to include visual acuity testing in both the early morning and late afternoon. Clinicians and agencies should also be aware that contrast sensitivity may be reduced at low light levels.