A method for quantitation of intact proteoglycans as GAGs in biological fluids (blood plasma, synovial fluid) or 4 M guanidine extracts of tissues has been published previously (S. Björnsson,Anal. Biochem.210, 282–291, 1993). The method is based on the specific interaction between sulfated polymers and the tetravalent cationic dye Alcian blue at pH 1.5 in 0.4 M guanidine-HCl and in the presence of 0.25% Triton. The absorbance assay has a measuring range of 1–20 μg of glycosaminoglycan (GAG) which is not sensitive enough to measure the low contents of proteoglycans in blood plasma, urine, or wound fluid. A dot blot assay is now described in which the Alcian blue–GAG complexes are collected on a polyvinylidene fluoride membrane, by filtration in a dot blot apparatus, and the stain is quantitated as reflectance by scanning and densitometry. The assay requires 10 μl of sample and has a measuring range of 10–800 ng of GAG, corresponding to a concentration of 1–80 mg/liter, suitable for proteoglycans in biological fluids. The procedures for chemistry, scanning, densitometry, and curve fitting were each evaluated separately. The error contributed by chemistry accounted for a minor portion of the imprecision. The imprecision contributed by scanning was the most important source of within-run and between-run imprecision, and was caused by inequalities of the charge-coupled device along the scanning arm. Unexpectedly, curve fitting was also a major source of total imprecision in dot blot quantitation and differed with the type of equation used. The between-run imprecision calculated as CV (SD/mean · 100) was 13.0% at 8 mg/liter. The response of the assay was identical for six different commercial preparations of GAGs (chondroitin-4-sulfate, chondroitin-6-sulfate, dermatan sulfate, keratan sulfate, heparan sulfate, and heparin) despite differences in degree of sulfation known to exist. There was no positive or negative interference by blood plasma, apart from a slight negative interference on the quantitation of heparan sulfate. Analysis of 319 paired blood plasma and urine specimens from hospitalized patients showed a variation of plasma GAGs of 0.1–17.6 and urine–GAGs of 0.0–45.6 mg/liter. There was no correlation between plasma and urine GAG concentrations.