The immunochemical quantitation of dextran by single radial immunodiffusion (SRI) is described. Antidextran antisera were raised in rabbits and sheep immunized with dextran-protein conjugates. To enhance precipitation and minimize antibody consumption, 1 % polyethylene glycol 6,000 was incorporated into the agarose-antidextran gel. Measures shown to increase sensitivity include staining of faint precipitates and repeated application of dextran solution to antigen wells. Reproducibility studies showed that the error of the method is small, the standard deviation of a single determination corresponding to 2.3 % of the mean. Presence of serum does not influence dextran quantitation; presence of urine caused a slight (1–5%) reduction of precipitate size. At high sensitivity levels, the lowest concentration of dextran Mw 71,000 measurable was 2.5 μg/ml or 6 ng, with measuring range up to 40 μg/ml in the direct method (SRI). At lower sensitivity levels, the corresponding figures are 5–10 μg/ml or 12–25 ng, the measuring range extending to 10,000 μg/ml. For correct results, the molecular size distribution of the dextran sample to be quantitated should correspond to that of the reference dextran used as standard. Otherwise, the molecular size distribution of the sample has to be determined by a different method. Isomaltodecaose was the smallest molecular size species of dextran capable of forming precipitates in SRI, requiring large amounts of antidextran in the agarose gel. Studies on the correlations between molecular size of dextran fractions and precipitate area in SRI revealed a linear relationship between the logarithm of ‘relative precipitate forming capacity’ and the logarithm of number average molecular size (Mn).