The indirect fluorescent antibody technique was used to stain smears of the leptomonads of 10 strains of Leishmania spp. including L. donovani, L. infantum, L. tropica, L. braziliensis braziliensis, L. braziliensis pifanoi and L. mexicana, and the Leishman-Donovan bodies of some strains. The sera used successfully and giving a strong reaction were from rabbits artificially immunized by leptomonads of all the above strains, and human sera from patients infected with Mediterranean infantile kala-azar, South American kala-azar, South and Central American muco-cutaneous and cutaneous leishmaniasis. Other sera used successfully but giving a weak reaction were from rabbits artificially immunized by Leishman-Donovan bodies of L. mexicana, human sera from patients infected with leishmaniasis tegumentaria diffusa, uta and British Honduran chiclero's ulcer, and serum from monkeys infected in the laboratory with L. mexicana and L. braziliensis. Reactions were less strong with Leishman-Donovan bodies as slide antigen than with leptomonads. No differentiation of strains was possible by serial dilution of sera used in the test, as all sera stained all antigen preparations at more or less the same intensity, indicating a group antigen-antibody reaction in reference to this method. Specific absorption of sera completely eliminated the reaction regardless of the antigen used for absorption. Sera sent from British Honduras from patients with chiclero's ulcer, non-immune subjects and others, were tested for antibodies by the fluorescent antibody technique. No correlation was obtained between disease or possible exposure and positive fluorescent antibody results. It is concluded that the fluorescent antibody test is not useful for the detection of past or present infection with L. mexicana.