The disposable multiple-puncture tuberculin test device, Monotest, was recently introduced in South Africa for tuberculin test screening of tuberculosis infections. Three studies were carried out to compare the intradermal Mantoux test with the Monotest. In the first study, conducted on confirmed hospitalised tuberculosis patients, 307 subjects underwent a 2 TU RT 23 (Statens Seruminstitut, Copenhagen) Mantoux test, 155 a 5 TU test of the same antigen, and 111 a Monotest. In the second study, another group of 98 confirmed tuberculosis patients was double-tested with 5 TU RT 23 by Mantoux test and Monotest. In the first study, 100% of 5 TU Mantoux tests resulted in indurations > or = 5 mm, 97% of 2 TU Mantoux indurations were > or = 10mm, and 96% of Monotests produced indurations > or = 2 mm (manufacturer's recommended cut-off point). In the second study, 99% of the 5 TU Mantoux tests measured > or = 5 mm, and 100% of Monotest indurations were > or = 2 mm. In both studies, raising the cut-off point for the Monotest to > or = 4 mm produced sensitivities of 95% and and 100% respectively. All these tests may therefore be regarded as highly sensitive. In a third study, 58 healthy schoolchildren were double tested with 2 TU RT 23 and with Monotest. Results indicated that if the positivity cut-off point of the Monotest is set a 8 mm instead of at 2 mm, sensitivity suffers slightly, decreasing from 100% to 90%, while specificity increases considerably from 8% to well over 80%. Receiver operating characteristic analysis indicated the high likelihood that the Monotest as well as the Mantoux test could discriminate between infected and uninfected subjects. For both tests the Wilcoxon statistic exceeded 0.8. The Monotest is a useful alternative to the Mantoux test. It is recommended that if tuberculosis is suspected, reaction > or = 4 mm be interpreted as positive. For screening purposes, a cut-off point of 8 mm seems to distinguish best between the infected and the uninfected individual.