1. Renal haemodynamics were monitored over an average period of 19 months in 17 children being treated with cyclosporin A. Sixteen had juvenile dermatomyositis and one had chronic polyneuropathy. The dose of cyclosporin A ranged from 2.3 to 8.3 mg day −1 kg −1 (median 4.1 mg day −1 kg −1 ). 2. Glomerular filtration rate (expressed in terms of extracellular fluid volume), renal blood flow (expressed as a fraction of cardiac output) and filtration fraction were measured by using 99m Tc-labelled diethylenetriaminepenta-acetate. They were compared with the dosage and trough blood levels of cyclosporin A, and, in 15 patients receiving prednisolone in addition to cyclosporin A, with steroid dosage. 3. All 17 children had a renogram performed 6 months after starting cyclosporin A treatment. Nine of them also had a renogram before starting cyclosporin A treatment (baseline study), while 13, in addition to their renogram 6 months after starting cyclosporin A treatment, also had at least one further renogram. 4. Glomerular filtration rate/extracellular fluid volume fell slightly but significantly from 0.009 ( sd 0.0013) before starting cyclosporin A treatment to 0.0085 (0.002) min −1 ( P < 0.01) 6 months after cyclosporin A treatment in the nine children who underwent a baseline study. This was accompanied by a significant ( P < 0.001) fall in filtration fraction from 0.108 (0.015) to 0.088 (0.014). However, renal blood flow/cardiac output showed no change. 5. In the 13 children studied beyond 6 months after starting cyclosporin A treatment, there was no further significant overall change in any renal haemodynamic variable. However, throughout this period, trough blood levels of cyclosporin A, which ranged from 20 to 258 ng/ml (median 67 ng/ml), correlated inversely with glomerular filtration rate/extracellular fluid volume, but not with renal blood flow/cardiac output. Furthermore, although a weak correlation between filtration fraction and trough blood levels of cyclosporin A did not reach statistical significance, sequential changes in filtration fraction correlated significantly with corresponding sequential changes in glomerular filtration rate/extracellular fluid volume. 6. We conclude that the predominant mechanism of renal impairment at these relatively low trough blood levels of cyclosporin A is a reversible reduction in filtration fraction.