Over an eight-month period, microhysteroscopy was performed on 180 women, candidates for in-vitro fertilization (IVF). One hundred and two of them were suspected of having uterocervical pathology (group I), and 78 were not (group II). Group I had microhysteroscopic abnormalities in 60.8% and group II in 21.8%. Microhysteroscopy showed 30.6% of 36 positive hysterographies to be false positive and 37.5% of 144 negative hysterographies to be false negative, producing a confirmation rate of 63.9%. Admission to the IVF programme depended on microhysteroscopic normality. One hundred and one women were accepted directly, and 20 refused ('primary decision'). Fifty-nine were deferred with treatable hysteroscopic abnormalities, pending medical (22) and surgical (37) treatment, 14 via the microhysteroscope. Of the 59 women deferred, 51 were finally admitted and 8 rejected after a microhysteroscopic check ('secondary decision'). Finally, 152 of the 180 candidates (84.4%) were admitted. Microhysteroscopy helped to determine the optimal route for embryo transfer in 23 women with cervical pathology. Term pregnancy rates per embryo transfer were similar with hysteroscopic findings primarily normal, or normal after successful treatment (11.2 and 10.8% respectively). Microhysteroscopy seems useful for selection of cases for IVF, accurately determining utero-cervical pathology or normality, aiding in surgical correction of abnormalities and facilitating the process of embryo transfer. We suggest that microhysteroscopy should be performed routinely on all IVF candidates.