I. Chromatographic and spectrophotometric studiés have been carried out on thionine, toluidine blue and azure A. Thionine contains blue and red components, in the ratio of approximately two parts of blue to one part of red. The blue fraction has strong metachromatic properties; the red component stains all tissues, particularly those which are basophilic, in a bright red colour. Toluidine blue and azure A both contain metachromatic and non-metachromatic fractions but there is no question of the metachromatic colours which they produce being due to differential staining as, unlike thionine, they contain no red fractions. II. The dye concentrations and staining times necessary for optimum results have been studied and discussed. III. The treatment of sections after staining, particularly the vexed question of alcoholic dehydration as opposed to the examination of wet sections, has been the subject of particular attention. It follows from our findings that when the metachromatic staining reaction is used for histochemical purposes it is essential that steps be taken to differentiate histochemically significant irreversible metachromasia from the non-specific reversible metachromasia which can be exhibited under certain conditions by an extremely wide range of tissue elements. This can be done either by mounting the sections after staining in Apáthy's medium or syrup of laevulose, as recommended by Lison (1935, 1936), but far more effectively by alcoholic dehydration. IV. The findings of test tube experiments are not directly applicable to histological material in that although the metachromasia obtained on mixing solutions of metachromatic dyes and chromatrope substances is readily reversed by alcohol, the metachromasia seen histologically in sections of cartilage, mast cells etc., is never changed back to the orthochromatic colour by alcohol or drying. V. In the light of the experimental findings a technique for carrying out the metachromatic staining reaction for histochemical purposes is suggested.