One hundred and twenty-two women who had not conceived after more than one year of unprotected sexual relations underwent selective salpingography. All had been shown to have proximal obstruction of the fallopian tube by standard hysterosalpingography. The women underwent the procedure as outpatients during the follicular period. They were given prophylactic antibiotic treatment and salpingography was performed by a radiologist in an X-ray room. We used the commercial sets of Cook and Zorn. The mean age of patients was 34 years (range 28 to 46) and the mean duration of infertility was 7.7 years (range 2.2 to 14 years). Two hundred and forty tubes were examined, of which 213 were cleared of obstruction (88.7%), with both tubes affected in 66% of cases and only one tube affected in 33% of cases. All patients had at least one patent tube after the procedure. The patent tube was normal in 73% of cases (177 tubes) and abnormal in 15% (36 tubes). Catheterization failed in 11.3% of cases (27 tubes). Forty-nine women conceived spontaneously during the 2- to 12-month follow-up period, and 39 of these women delivered healthy babies. No extrauterine pregnancies or serious complications were reported. The most frequent side effects were slight pain, bleeding and nausea. One perforated tube was reported and none of the patients died. The greatest difficulties encountered were a lack of cooperation in patients with previous painful experiences of hysterogram and catheterization problems in women with a strongly flexed or distorted uterus. Complications were minimal and the one case of tubal perforation at the beginning of the series had no serious consequences. Thus, selective salpingography should be used more widely because it is simple and more cost-effective than the surgical management of tubal obstruction and artificial insemination. Our results suggest that this procedure should be the first-line treatment of infertility caused by proximally obstructed fallopian tubes.