Three different models of intrauterine pessaries are compared with each other in respect of tolerance and safety. They were worn for different lengths of time, partly for up to ten years. The comparison shows that the Pearl index in its conventional form is not a suitable indicator. It does provide a pointer if the individual collectives are reduced to the same wearing period. The Life-Table method yields better results. It relates the events for a total of 100 women in each case to the same interval periods. This shows that the most favourable rate of complications and quota of failures is obtained with Cu-T 200. During the first three years this quota decreases steadily. Hence it is recommended to effect intrauterine contraception according to this model wherever suitable and to allow Cu-T to be retained in position for at least three years if it is well tolerated.