Visual field examinations were performed serially on 20 patients with pseudotumor cerebri using a modified Armaly-Drance visual field strategy with a Goldmann perimeter as well as an automated perimeter (Octopus). Visual loss was found in 75% of eyes using the manual strategy and in 77.5% of eyes with automated threshold perimetry. This incidence of visual loss is 50% greater than any previously reported series. All major defects detected were present with both types of perimetry. Both strategies were more sensitive for documenting visual loss than previously described strategies. Since therapy for pseudotumor cerebri is determined by the degree and progression of visual loss, a specific sensitive strategy, rather than routine screening perimetry, should be used for determination of visual loss.