Several Scandinavian clinics are now practising intracytoplasmic sperm injection (ICSI), and our data largely resemble that published from the Brussels group (Van Steirteghem et ai., 1993). In recent meetings, Scandinavian workers have come to some mutual decisions: (i) not more than two embryos will be replaced; (ii) recombinant follicle stimulating hormone (FSH) has been introduced into ovarian stimulation protocols and, unsurprisingly, it works effectively in ICSI treatment cycles (Table I). We have also paid especial attention to the rate of malformations in the children conceived by ICSI. In our first 110 deliveries after ICSI, we have had no triplets, and only 20 twin deliveries. We have identified one major malforma­ tion, an aortic valve stenosis, with a reduction of blood flow of 19%. That child is doing very well and an operation is not planned. It should perhaps be classified as a minor malformation, although our paediatricians have classified it as a major malformation. It is important to keep in mind that ICSI is merely a part of the technique of achieving fertilization in vitro. Data from Jacques Cohen's group (Cohen, 1991), indicating an increased risk for malformations, were based on a combination of ICSI and the utilisation of toxic agents, such as pentoxifylline, 2-deoxyadenosine, or acid Tyrode's solution. Incidentally, the one major malformation we have recorded above arose in fact through a combination of ICSI and assisted hatching using acid Tyrode's. One case cannot tell us much, but a combination of ICSI with toxic agents may involve an increased risk. It may be possible to lower malformation rates, if chemical additives are avoided. Pharmaceutical grade media are of importance. Finally, I wish to comment on injecting spermatozoa into oocytes of different maturity (Figure 1 and Table II). We have deliberately injected some oocytes with a germinal vesicle (GV), others in metaphase I and most in metaphase II. In all programmes, certain oocytes are difficult to classify with certainty. Data in Table II show that fertilization and cleavage can sometimes be achieved when oocytes with a germinal vesicle are injected. When metaphase I oocytes are injected immediately instead of waiting for maturation to be completed, a fertilization rate of 31 % is obtained with metaphase I oocytes, which is one-half of that achieved with metaphase II oocytes. If fertilization does occur, cleavage