An hypothesis of increased endorphinergic activity has been proposed to account for the characteristic symptoms of Rett syndrome. Cerebrospinal fluid samples from eight girls with Rett syndrome were analysed for beta-endorphin (beta-EP) immunoactivity and compared with samples from a control group of 15 children with acute leukaemia in remission. Severity of symptoms was not found to be related to beta-EP level. A group of early-treated adolescents with phenylketonuria had beta-EP levels similar to the Rett syndrome patients, but no symptoms resembling theirs. Therefore it is unlikely that increased levels of beta-EP are of primary pathogenetic significance. The conflicting findings of many earlier reports may be a result of differences between control groups.