The College of American Pathologists and the American Society of Human Genetics offer a proficiency testing program in clinical cytogenetics. Two hundred twenty-five laboratories now provide data for this survey, which was begun in 1986. Challenges have consisted of photographed metaphases, fixed lymphoblastoid cell suspensions, fresh peripheral blood, and disarranged karyotypes. The "correct" response was based on 80% or greater consensus among either the referees or the participants. Referee laboratories performed better than participants. More laboratories were able to report accurate recognition of abnormalities by using a coded list than could write the interpretation in standardized nomenclature. Deletions, unbalanced translocations, and inversions were more difficult challenges than balanced translocations or trisomies. Prenatal and lymphocyte challenges were more likely to result in consensus than were bone marrow challenges. Participants performed best on whole-blood challenges. Fixed cell suspensions were less satisfactory. Excellent quality case material is essential for a successful challenge. A grading system has been devised to separate artifacts of the survey process from proficiency variables.