In a double-blind study 236 nulligravidae were randomized to insertion of standard models of two commonly used intrauterine devices, Multiload and Gravigard, or shorter versions of these devices. Before insertion the total axial uterine length and the length of the endometrial cavity were determined. After 12 months the total continuation rate was 57.2%. The clinical performance of Multiload models was not statistically different from that of the Gravigards, and the short IUD-types did not do better than the longer ones. Among the longer models a significant higher continuation rate was observed with the Multiload. The frequency of side effects was not increased when the total uterine cavity length was less than 7 cm. As the incidence of side effects was higher than normally reported after insertion of IUD's in parous women it is concluded that insertion of IUD's in nulligravidae should be performed only when other methods of contraception are unacceptable. The ratio between the lengyh of the uterine cavity and the length of the IUD seems to be of no significance for the clinical events following insertion of IUD's in nulligravidae.