The involvement of cellular immune reactions in the pathogenesis of many inflammatory dermatoses can be postulated from their histopathologic features, e.g., those dermatoses that show a perivascular lymphohistiocytic infiltration in the dermis in addition to their own characteristic epidermal changes. Several granulomatous dermatoses further show an additional infiltration of epithelioid cells. However, experimental production of similar changes is usually not feasible in animals. A human model is needed of delayed-hypersensitivity skin reaction induced by a definite antigen. After closely observing the histopathologic changes of the skin reactions to the purified protein derivative (PPD) of tuberculin, we noted that all patients did not show uniform histologic patterns at intradermal PPD injection sites. Therefore, we have tentatively analyzed histopathologically the pattern of tuberculin skin reactions at injection sites 48 h to 1 year after injection in eight healthy volunteers and 63 patients with various dermatoses. We classified the reactions at 48 h into three types, based on the extent of the tissue damage (i.e., their resemblance to features noted in corresponding inflammatory dermatoses): (a) the perivascular dermatitis type, (b) the basal spongiotic dermatitis type, and (c) the erythema multiforme type. There was a clear correlation between the magnitude of macroscopic skin reactions and these histopathologic patterns. Even those with clinically negative reactions showed a reaction pattern of perivascular dermatitis, whereas those who clinically developed prominent inflammatory reactions with blister formation always showed a erythema multiforme-like histologic pattern. Interestingly, in more than half of the cases, the dermal cellular infiltrate had a mixture of various numbers of neutrophils in addition to mononu-clear cells. Two-thirds of biopsy specimens taken after 14 days showed features of epithelioid cell granuloma as noted in tuberculosis, sarcoidosis, tuberculoid leprosy, and metal-hypersensitivity granulomas. In the remaining cases, however, there was only a monotonous mononu-clear cell infiltration similar to that described as typical delayed-type hypersensitivity reaction. Two biopsy specimens taken long after PPD injections, i.e., one after 4 months and another one after 1 year, respectively, showed palisading granuloma associated with collagen degeneration as noted in rheumatoid nodules. We think that the analysis of these model inflammation patterns of delayed-hypersensitivity reaction induced by intradermal PPD injection in adults provides a valuable clue for the elucidation of thus far unknown pathomechanisms of various inflammatory dermatoses that are presumed to develop based on cellular immune reactions.