Little information is available either for the clinical value of many ophthalmic tests performed preoperatively in the evaluation of patients for cataract surgery or for variation in ophthalmologists' use of such tests. To assess variation in ophthalmologists' use of ophthalmic tests, we conducted a national survey of American Academy of Ophthalmology members. Thirty-three percent, 17%, 37%, and 19% of the respondents reported that they "frequently" or "always" perform glare testing, contrast sensitivity testing, potential acuity measurement, and specular microscopy, respectively, in patients being considered for cataract surgery who have no history of eye disease other than cataract. In contrast, 27%, 54%, 24%, and 48% of respondents reported that they never perform each of these four tests in such patients. Two ophthalmologist characteristics--a surgical volume of greater than 200 cataract extractions per year and performance of surgery in an ambulatory surgical center or private office (as opposed to a hospital)--were independently associated with an increased probability of performing each of these four tests frequently or always. Ten percent or less of the respondents reported that they frequently or always perform electroretinography, visual evoked response testing, photography of fundus or anterior segment, B-scan ultrasonography, formal color vision testing, and formal visual field testing in such patients. Thus, there is considerable variation in ophthalmologists' use of glare testing, contrast sensitivity testing, potential acuity measurement, and specular microscopy. A small percentage of ophthalmologists may be overusing several other tests in the evaluation of patients being considered for cataract surgery.