Of 110 consecutive open tibial fractures 90 were reviewed and analysed retrospectively with particular reference to wound closure, method of stabilisation, infection rate and the incidence of non-union. There were 41% Gustilo type I, 39% type II and 20% type III injuries. The incidence of deep infection was 20% after primary wound closure compared with 3% after delayed closure, and eight of the nine non-unions followed primary closure. We conclude that primary wound closure should be avoided in the treatment of open tibial fractures.