A method is described which in principle is a quantitative spot analysis; 0.5-5 .mu.l of a protein solution in the concentration range between 0.01-10 mg/ml is used for 1 determination. The sample is taken up by capillary attraction in a 0.5-, 1-, 2- or 5-.mu.l capillary and transferred to a cellulose acetate strip. The protein is fixed and stained simultaneously by dipping the cellulose acetate strip into a solution of Amido Black or a benzoxanthene derivative (Hoechst 2495) dissolved in methanol/acetic acid. After elution of the excess of dye (3 .times. 5 min), the quantitative evaluation can be performed in different ways. The sample is fixed and made transparent by incubation in dioxane/1-butanol and evaluated densitometrically (Amido Black 10B) or the evaluation is performed in situ by spot fluorometry (Hoechst 2495). The sample can be dissolved together with the acetate layer completely in dioxane, dimethylsulfoxide or N,N-dimethylformamide and evaluated photometrically or fluorometrically. Highest sensitivity is reached if the fluorochrome (Hoechst 2495) bound to the protein is eluted with 15% NH4OH and measured fluorometrically. There is a linear correlation with a correlation coefficient of 0.999 between the fluorescence and a protein amount of 10 ng to 20 .mu.g. In addition to its simplicity, the method has the advantage of being independent of or well-defined against other external influences, e.g., sodium dodecyl sulfate, mercaptoethanol, Triton X-100. The stainability of a protein with Amido Black is influenced stoichiometrically by sodium dodecyl sulfate (not by mercaptoethanol) while the staining with Hoechst 2495 was not at all affected. As there is a linear correlation between the area of a spot on an acetate layer and the volume applied in the range between 0.5 and 5 .mu.l (only influenced stoichiometrically by the protein concentration in that volume, which is measured by staining with Amido Black), then with a simple iterative calculation on the basis of suitable calibration curves, it is easily possible to determine a protein concentration in mg/ml even in an unknown volume between 0.5 and 5 .mu.l.