Of 19,653 patients hospitalized in the medical divisions of two teaching hospitals, 3980 were treated with an aminopenicillin, 808 with other penicillins, 427 with a cephalosporin, 2619 with cotrimoxazole and 846 with allopurinol. The first part of the study deals only with the incidence of exanthemas definitely or probably due to a specific drug on the basis of clinical considerations. The exanthema incidence is 8.0% for aminopenicillins, 4.7% for other penicillins, 1.9% for cephalosporins and 2.8% for cotrimoxazole. The second part of the study employs a cross-tabulation to determine the incidence of exanthemas definitely and probably drug-induced, and the temporal relationship of these reactions to aminopenicillin and allopurinol exposure. The observed risks of developing an exanthema are as follows: aminopenicillin without allopurinol 10.1%, aminopenicillin combined with allopurinol 7.2%, allopurinol without aminopenicillin 3.0%, neither of the two drugs 1.5%. The increased incidence of exanthemas observed by the Boston Collaborative Drug Surveillance Program (BCDSP) in patients concomitantly treated with aminopenicillin and allopurinol was not confirmed by our results. Our hypothesis is that the time of exposure to aminopenicillins might have been shorter for patients of the BCDSP who were not treated in connection with neoplastic disease and did not receive allopurinol. The incidence of aminopenicillin induced exanthemas increases severalfold with the duration of exposure time during the first 2-3 weeks. In the CHDMB, on the other hand, exposure time does not differ between the patients treated with aminopenicillin alone or in combination with allopurinol.