The qualitative microhemagglutination assay for detection of antibodies to Treponema pallidum and the automated fluorescent treponemal antibody-absorption tests were directly compared as confirmatory tests for syphilis. They were compared on 2,790 specimens from patients of New York City Social Hygiene Clinics. Overall agreement between the two assay methods was 89.6%. Only 1.5% of the total specimens tested were reactive by the automated fluorescent treponemal antibody-absorption test, and nonreactive by the microhemagglutination assay. The majority of specimens which disagreed were reactive only by the microhemagglutination assay. These comprised 8.9% of the total specimens tested. When such specimens were further tested by the manual fluorescent treponemal antibody-absorption test, 63% were also reactive in that test, and 16.7% gave borderline reactions. Fifty-four per cent of those reactive specimens were from patients suspected of being in a latent stage of syphilis, 10% from patients suspected of being in other stages, and 36% from patients not suspected of having syphilis. Because of the increased sensitivity of the microhemagglutination assay, its use is recommended instead of the automated fluorescent test. However, the manual fluorescent test should be performed when the result of the microhemagglutination assay is completely inconsistent with the clinician's impression.