The combination of clinical data plus one positive immunopathologic test provide the best combination of sensitivity and specificity (98%), and seem to be most appropriate in defining patient populations for study purposes. The relationship between the classes of immunoreactants should be better evaluated with reference to the site of skin biopsy. It may be suggested, however, that the likelihood of BP existence is very low when in vivo C3 is absent or of lower intensity of fluorescence than the concomitant Ig(s).