The aim of the study was to compare the sensitivity of short-wavelength and conventional automated static threshold perimetry for the psychophysical detection of abnormality in patients with clinically significant diabetic macular oedema. The sample comprised 24 patients with clinically significant diabetic macular oedema (mean age 59.75 years, range 45-75 years). One eye of each patient was selected. Exclusion criteria included the presence of lenticular opacity. The sensitivity of the macular visual field of each patient was determined with programme 10-2 of the Humphrey Field Analyser on two occasions, using both short-wavelength and conventional stimulus parameters; the results of the second session were analysed to minimise learning effects. A pointwise horizontal hemifield asymmetry analysis was derived for short-wavelength perimetry (thereby negating the influence of pre-receptoral absorption); the pointwise pattern deviation probability plot was analysed for conventional perimetry. Abnormality was defined as 3 or more contiguous stimulus locations with negative asymmetries (short-wavelength) or reduced sensitivity values (conventional) that resulted in a statistical probability level of p less than 0.05. The fields of 8 patients were abnormal as assessed by conventional perimetry while all were classified as abnormal using short-wavelength perimetry. In the 8 patients who exhibited both abnormal conventional and abnormal short-wavelength perimetry results, the extent of field loss was generally greater using short-wavelength perimetry. The position of the localised field loss (i.e. as distinct from field loss that was generalised across the visual field) assessed by short-wavelength perimetry corresponded with the clinical mapping of the area of diabetic macular oedema but the extent of this loss was generally greater than that suggested by clinical assessment. Short-wavelength automated perimetry offers improved sensitivity for the psychophysical detection of clinically significant diabetic macular oedema.