The Armaly technique, a method of screening for glaucomatous field defects was evaluated in 100 normal eyes and 38 selected central field defects using the Goldmann perimeter and the Auto-Plot tangent screen. The screening results obtained from the Goldmann perimeter and the tangent screen using the selective perimetry of Armaly were compared with the results obtained from the Friedmann central field analyser. Testing normal eyes we found the highest initial false positive rate with the Friedmann analyser and the lowest initial false positive rate with the Auto-Plot tangent screen. The specificity of the Friedmann analyser for detecting early central field changes is less than the Goldmann perimeter or the Auto-Plot tangent screen when using Armaly's technique. On the other hand, the Friedmann analyser turned out to be the easiest instrument to handle and requires least experiences for the observer. Some problems of mass screening perimetry are discussed.