This study examines the utility of objective histopathological studies in the evaluation of adult patients with erythroderma. A series of 56 skin biopsies, from 40 erythrodermic patients, was reviewed sequentially by 4 Canadian dermatopathologists who were unaware of clinical details of the cases. The final diagnosis (gold standard), in each instance, had already been determined by others, based on clinicopathologic data and response to therapy. Direct comparison revealed that the mean accuracy of the histopathological diagnoses was 53% (range: 48-66%), a favorable result in view of the difficulty of the task at hand. Additional points of information which evolved from the study are as follows: (i) identification, by microscopy alone, of spongiotic dermatitis, cutaneous T-cell lymphoma and psoriasis, as underlying causes of erythroderma was more successful than that of drug eruptions and pityriasis rubra pilaris; (ii) the epidermotropism which characterizes cutaneous T-cell lymphoma may be mistaken for inflammatory interface changes seen in drug eruptions and vice versa, thus constituting a pitfall in diagnosis; (iii) finally, it appears that submission of multiple simultaneous biopsies, rather than a single specimen, from patients with erythroderma would be likely to enhance the accuracy of histopathological diagnosis.