Male CBA/J mice, sensitized by cutaneous inoculation with viable Candida albicans blastospores, were used to study in vivo and in vitro cellular immune responses. Three antigens of C. albicans, viz., a cell wall glycoprotein (GP), a membrane extract (ppt-HEX), and soluble cytoplasmic substances (SCS), were used in vitro in a lymphocyte stimulation assay, whereas the GP and ppt-HEX were used in vivo to detect delayed hypersensitivity by the footpad assay. Delayed hypersensitivity to GP and ppt-HEX was transferred from sensitized donors to naive recipients with peritoneal exudate cells and not with serum. Moreover, the transfer of the reactivity to ppt-HEX was abrogated by the prior treatment of the transfer suspension with anti-theta 1.2 serum and complement. The in vitro lymphocyte response to GP and ppt-HEX correlated qualitatively with the in vivo responses. SCS, a preparation shown to be ineffective in vivo previously, did stimulate lymphocytes from sensitized animals in vitro. The in vitro response to Candida antigens, as well as phytohemagglutinin, was abolished by treatment of the lymphocyte suspension with anti-thymocyte 1.2 serum before assay, whereas anti-immunoglobulin serum had less effect on these responses. The in vivo and in vitro reactivity to the Candida antigens, therefore, was dependent upon viable T-lymphocytes. Preliminary specificity studies were carried out in the lymphocyte stimulation assay, using lymphocytes from mice infected with C. albicans tested against ppt-HEX preparations extracted from two other species of Candida, C. tropicalis and C. guillermondii, and from two other pathogenic yeast forms, Histoplasma capsulatum and Blastomyces dermatitidis. Significant cross-reactivity was observed with C. tropicalis only, a species which is known to be serologically related to C. albicans.