Hysteroscopic evaluation of 187 patients with abnormal uterine bleeding is reviewed. One hundred fifty-one patients had recent tissue sampling by a blind method. The hysteroscope with selected endometrial sampling was more accurate in 9.1% and less accurate in only 0.5% of patients. The specificity of both techniques with selected biopsies was 100%, but the sensitivity of hysteroscopy was greater (98%) than that of D&C (65%). Endometritis was the only condition missed by hysteroscopic viewing and selected biopsy. Those lesions missed by blind tissue sampling were uterine fibroids and endometrial polyps. Among 91 patients whose hysteroscopic view was negative, an abnormality was identified by tissue sampling in only one patient with endometritis.