Erythrocyte transketolase activity and the effect of adding thiamine pyrophosphate (% thiamine pyrophosphate effect) were measured in 111 subjects suspected to suffer from Leigh's disease (subacute necrotising encephalomyelopathy). From clinical evidence these subjects were divided into five groups: (1) necropsy-proved cases of subacute necrotising encephalomyelopathy, (2) cases positive for urinary thiamine pyrophosphate: adenosine triphosphate phosphotransferase inhibitor, (3) clinically likely cases of subacute necrotising encephalomyelopathy (patients still alive, or on whom no necropsy was performed), (4) cases diagnosed as diseases other than subacute necrotising encephalomyelopathy (control group), (5) cases for which no diagnosis had been made. Comparison of erythrocyte transketolase activities with and without added thiamine pyrophosphate and of the % thiamine pyrophosphate effect for each group compared with the control group showed no statistically significant differences from normal values for any of these parameters. Similarly, there were no differences between the two sexes in transketolase activity, and no correlation between transketolase activity and age. These results indicate that erythrocyte transketolase activity is not altered in subacute necrotising encephalomyelopathy and is unlikely to be of value for the diagnosis of Leigh's disease.