Roentgenography and ultrasonography have been recommended for location of an intrauterine contraceptive device (IUCD) when the strings are not visible at the cervical os. Real-time ultrasound may have limited application compared to gray-scale B-scanning in certain clinical applications. Because of previous difficulty in locating an IUCD in 2 gravid and 2 nongravid patients, a blind study of 68 women was conducted. Seven patients were eliminated from the study because of inadequately filled bladders and failure to return for follow-up visits. There were 37 patients with IUCDs and 24 patients in the control group who used other forms of contraception. An incorrect diagnosis was made in 12 of 61 patients. Ten of 37 patients with an IUCD were identified as having no IUCD and 2 of 24 patients without an IUCD were identified as having an IUCD present. The accuracy of correctly identifying the presence or absence of an IUCD appears to be related to the type of device used. All of the 20 patients with a Copper-7 were correctly identified as having an IUCD, although the type was in error in 11 of the 20. Five of 7 patients with a Lippes Loop and none of the patients with a Saf-T-Coil were correctly identified as having an IUCD. Prior literature is reviewed, and possible sources of error are discussed.