The selectivity of proteinuria was assessed in 25 patients by a gel filtration technique utilising a G 200 Sephadex column. The results were compared with those obtained using gel diffusion and immunoprecipitation to assess selectivity. In five patients, where the protein excretion was less than 1.0 g. per day, the results obtained by gel filtration through Sephadex were found to be suspect. In the remaining 20 patients, where the protein excretion was over 1.0 g. per day, the relationship between the gel filtration selectivity and that determined by immunoprecipitation could be expressed as a straight line passing through the origin. The correlation coefficient for this line (0.84) was highly significant statistically. The selectivities obtained on gel filtration were assessed in the light of the findings on renal biopsy. Two of the twenty patients with proteinuria of over 1.0 g. per day had minimal lesion glomerulonephritis. These patients also had the two highest selectivities in the series. Three of the patients studied had chronic renal failure. These patients had the lowest, the second lowest, and the fifth lowest selectivities of the series. Good renal function and minimal glomerular damage appearJto be associated with selective proteinuria, while poor renal function and severe glomerular damage appeared to be associated with unselective proteinuria. Five patients with selective proteinuria were treated with steroids. In two cases the proteinuria disappeared. In a third it fell to trace amounts. Ten patients with intermediate or unselective proteinuria were treated with steroids. In only one case was there a significant reduction of proteinuria.