Tine and Mantoux tuberculin tests were compared in 193 hospitalised White patients with a variety of illnesses. With the Mantoux test 52 (27%) were tuberculin-positive and 19 (9.8%) were positive with the tine test. Had the tine technique alone been used, 33 of 52 tuberculin reactors would have been missed. Additional findings were (i) a 73% positivity for Monilia antigen; (ii) the optimal time for reading the Mantoux tests was on the third, rather than the second day; (iii) tuberculin reactivity was not suppressed by either pyrexia or elevation of the sedimentation rate; and (iv) the use of avian PPD demonstrated very few reactors in this population. It is recommended that the tine test be replaced by the intradermal Mantoux test with stabilised freeze-dried PPD tuberculin supplied in test kits with diluent. This matter should be given high priorty in South African general hospital practice and population surveys.