A serological study of candidiasis was conducted with use of (1) sera sent to the laboratory by clinicians who suspected infections due to various organisms including Candida species, and (2) sera available in our serum bank from patients with candidiasis proven at autopsy and from those with documented candidemia. With this population of patients, we were able to evaluate both potential false-negative and false-positive results. Microimmunodiffusion and slide agglutination tests were used. In many cases, serial specimens were available for measurement of rises or falls in titers of agglutinating antibody. Sera from less than one-half of the patients with disseminated or invasive gastrointestinal candidiasis exhibited positive immunodiffusion reactions, titers of agglutinating antibody of greater than 1:16, or fourfold rises in titer of agglutinins. Sera from several patients with pharyngitis due to Candida and from several who were only colonized with Candida or for whom clinical or cultural evidence of candidiasis was lacking showed positive immunodiffusion reactions or agglutinin titers of greater than 1:16, and some showed fourfold rises in titer. Conversions from negative to positive immunodiffusion reactions were not consistently correlated with invasive candida infection. Using the methods described, we have not found immunodiffusion tests and titers of agglutinating antibody to be reliable indicators of invasive candida infection, since false-positive as well as false-negative reactions occur. Greater specificity for invasiveness as well as greater sensitivity in immunosuppressed patients are necessary before these tests can become important adjuncts to the evaluation of patients with suspected invasive candidiasis.