The effectiveness of ethamsylate in the prevention of periventricular haemorrhage (PVH) in very low birthweight infants was evaluated by means of a multicentre, placebo-controlled, double-blind trial. In 330 infants without evidence of PVH on initial cranial ultrasound examination there was little difference between ethamsylate and placebo groups with respect to subependymal haemorrhage, but intraventricular and parenchymal haemorrhages developed in 30/162 infants (18.5%) in the treated group, compared with 50/168 (29.8%) in the control group (p less than 0.02). The incidence of intraventricular and parenchymal haemorrhage in survivors was 20/137 (14.6%) in the ethamsylate group and 37/146 (25.3%) in the controls (p less than 0.05). In 30 infants with evidence of PVH on the initial scan, ethamsylate treatment seemed to limit parenchymal extension. Analysis of the total cohort of 360 infants showed that the proportion of infants in whom an increase of two or more grades of severity of PVH was recorded during the trial was lower in the treated than in the placebo group (p less than 0.01). No adverse effects were attributed to ethamsylate therapy. The reported incidence of patent ductus arterious was lower in the treated than in the placebo group (p less than 0.02). Mortality was similar in the two groups.