AIM: To see if any a-wave measure segregated normal cats from those affected by a recessively inherited Abyssinian rod/cone dystrophy more efficiently than a(max) to scotopic I(max). METHOD: A-waves to electroretinograms (ERGs) evoked by a 4 cd x s/m(2) scotopic flash were extracted from 241 ERG sessions using 108 cats. They were either wild type or from an affected Abyssinian stock. Fourty four were bred by back-crossing to have a 50% probability of being affected. Most were diagnosed by retinal appearance or by the pattern of loss in a long protocol ERG. Eight were still unclassified. The diagnostic efficiency of amplitudes at 7, 8, 9, and 10 ms and a(max), of a(max) peak time, age at testing, and the main components of principal components factor analysis were compared by scaling their ability to segregate affected and normal cats. RESULTS: Variance and overlap between the groups both decreased as time along the a-wave increased. The loading of each animal on the largest factor also gave considerable overlap. There was a small absolute separation between groups when a(max) itself was used. Age and peak time were uncorrelated with disease. The light intensity used could be calculated to be equivalent to one sufficient for about 75% of full saturation in man. CONCLUSION: A(max) is a simple measure that is already in routine clinical use. When the flash is very bright and the animal fully dark adapted, this single measure is the most efficient sign of this rod/cone degeneration and possibly of all degenerations involving rods.