The control of leishmaniasis depends on a knowledge of the magnitude of the disease and of exposure to it. Delayed-type hypersensitivity testing can detect prior exposure to the parasite, but there is little agreement regarding the choice of an antigen for such testing. New and Old World leishmanins were tested in a study of patients with confirmed prior cutaneous leishmaniasis (CL), patients with confirmed prior American visceral leishmaniasis (AVL), and controls from areas in Espírito Santo, Brazil, where leishmaniasis is not endemic. Biobrás antigen (a suspended mixture of Leishmania braziliensis guyanensis, Leishmania mexicana amazonensis, and Leishmania mexicana promastigotes) detected 100% of prior CL infections and thus was superior to the Old World antigen, Leishmania major, which detected only 19% of these infections (P < .00001). Soluble New World antigens of Leishmania chagasi evoked a response in 96% of cases of prior AVL, whereas the Old World counterpart, Leishmania infantum, evoked a response in 71% of cases (P < .042). Testing of family members of patients with prior AVL showed greater utility of the New World leishmanins and suggested subclinical exposure of a large number of healthy people in the hyperendemic region. New World skin-test antigens should be used in future epidemiological studies in the Americas to more accurately determine the extent of exposure.