In this investigation of the copper loss in utero from the Copper-T, the release rate was estimated by iodometric titration and the corrosion was microscopically examined in ground plastic casts. The release rate was determined in 28 preweighed Copper-Ts with a surface area of 100–400 mm2 during the first month, and in 60 Copper-T-200s (TCu-200) which had been used for 5.3 to 40.3 months. The depth of corrosion was measured in 68 TCu-200s used for 3.2 to 52.0 months.During the first month the copper release from preweighed Copper-Ts with a surface area of 100,200,300 or 400 mm2 was 26.3, 44.4, 75.2 and 73.6 μg/day, respectively. The release rate increased with increasing surface area up to 300 mm2, but not thereafter. The average release rate of the TCu-200s was 43.8 μg/day from the 21st up to the 41st months of use. However, the release rate was significantly higher in devices removed because of bleeding. In these, it was twice or even four times as high as in the devices removed from symptom-free patients. Metallurgic micrographs revealed both inter- and intraindividual variations in corrosion of the wire. Wire breakage occurred as early as in the 12th month of use, and the risk of breakage was about 30 per cent during the fourth year.Owing to the corrosion, a replacement of TCu-200 seems indicated in symptom-free patients every three years, in patients with abnormal bleeding even more often. The life time of copper IUDs could be prolonged by increasing the thickness of the wire and by supplementing it with a non-corrodable (e.g. silver) core, which would eliminate the fragmentation due to corrosion. Moreover, an increase of the wire surface to 300 mm2 would probably reduce the pregnancy rate.