In a population based series of 4070 children with acute lymphoblastic leukaemia treated in Britain during 1971-82, the five year survival rate improved from 37% for those treated in 1971-3 to 66% in 1980-2. During the same period there was an increase in the proportion of children included in the Medical Research Council trials and an increase in the proportion who were treated at centres with an average of at least six new patients per year. Survival rates were significantly higher for children who were included in the trials than for those who were not. Among children not in the trials, the survival rate was highest at centres treating at least six children per year and lowest at those with fewer than one per year. The increase in the numbers of children treated according to standardised protocols, particularly within controlled clinical trials, has had a major impact on the mortality from childhood acute lymphoblastic leukaemia.