The total protein concentrations of six selected candidates for an international standard (IS) of birch pollen were previously estimated (Ann Allergy 1987;58:71-77). The results of analyses showed large variations depending both on the method and the reference protein used. More than 2-fold variations in the protein content were obtained using Bradford's protein binding method. In the present communication the protein contents of these birch IS-candidates were reestimated by Lowry, a modified Lowry technique, Bradford's protein-dye binding, bicinchoninic acid reagent and amino acid compositions. Great variations were obtained for the protein contents, depending on the technique employed, emphasizing the difficulties of choosing a proper method of protein analyses of pollen allergens. The Lowry and the bicinchoninic acid reagent methods gave the highest obtained protein estimates, while the modified Lowry and Bradford's protein-dye binding methods showed the lowest values. The amino acid analysis gave concentrations similar to those obtained by Lowry and bicinchoninic acid reagent methods. The Lowry method seemed to be adequate for the analysis of plant proteins. The amino acid analysis is currently the most precise method for estimation of the protein concentration irrespective of additional prosthetic groups. A serious shortcoming of the method is that it requires an amino acid analyser or other liquid chromatrographic systems, which are not commonly in laboratory use.