THE important contributions to the theory of renal secretion made by students of dilution diuresis began with Starling's [1899] consideration of the influence of the colloid osmotic pressure of plasma proteins on the rate of glomerular filtration. Despite Magnus's [1900] impression, based on extensive observations, that the diuresis was too great to be attributed solely to a change in glomerular activity, the view that tubular activity was not directly affected gained ground and was supported by Barcroft & Straub's [1910] finding that dilution diuresis was unaccompanied by increase in metabolism of the kidney and was therefore a "mechanical" diuresis. This view was reinforced by Knowlton's [1911] comparison ofthe effects of saline with gum-saline and gelatine-saline solutions, which identified the diuretic stimulus as a reduction in colloid osmotic pressure of the plasma. The purely "mechanical" view of dilution diuresis was strongly advocated by Cushny [1917] and has since been generally accepted as evidence in favour of the filtration theory of glomerular function in the mammal.Cushny [1917] concedes that "the change in the amount of urine in Magnus's experiments is out of all proportion to the change in the concentration of the colloids", but he felt able to explain away the discrepancy on the grounds that the proportion of filtrate reabsorbed in the tubules would be lower during the diuresis, and that this would magnify the change in glomerular filtration rate which he apparently took to be proportional to the reduction in concentration of plasma colloids.More recently, observations on perfused kidneys have directed attention to the disproportionately great diuresis produced by a small reduction in colloid osmotic pressure of the plasma in comparison with that 1 Assisted by a grant from the Medical Research Council.