Evaluation of sequential paired staining with the unlabeled antibody peroxidase-antiperoxidase (PAP) method showed that the brown reaction product of diaminobenzidine (DAB) concealed both enzyme and antigen-antibody sites in the reagent sequence. The blue reaction product of the alternative substrate, 4-chloro-1-naphthol (CN), exerted no such blocking effect. Hence, to avoid interactions between the two PAP sequences, DAB had to be used for the first and CN for the second antigen. Complete blocking required that the DAB color reaction be of sufficient strength. When two antigens were present in the same cell, the DAB deposits inhibited staining of the second antigen unless the brown color was decreased by progressive dilution of the initial primary antibody. A mixture of brown and blue could thus reflect either concomitant staining of the two antigens or unwanted interactions between the two PAP sequences. Double staining of individual cells was, therefore, equivocal and conclusions had to be based on comparative single staining results in adjacent tissue sections. Tests carried out in several model systems showed that paired direct immunofluorescence with fluorochrome conjugates of contrasting colors (green and red) was much less time-consuming, more reliable, and of higher detection sensitivity for analyses of unbalanced mixtures of two antigens in the same cell.