The direct fluorescent antibody test and two culture methods were compared for accurate diagnosis of chlamydial infections. Using the same samples, 109 were found to be positive in the microtitre method with the direct confirmation test without subpassage, whereas 66 were positive in the vial method with Giemsa staining and subpassage. The direct test was evaluated for accuracy using cervical and male urethral specimens. Specimens for culture were obtained prior to sampling for the direct test. For cervical samples the sensitivity of the direct test, with the vial method taken as reference, appeared to be 72.2% with a specificity of 93.5%. With the microtitre method as standard, these values were 55.9% and 91.3%, respectively for females, and for male patients 49% and 95.6%, respectively. For cervical samples, in which sampling for the direct test was carried out prior to sampling for culture, the values were 46.3% and 93.2% respectively. Both culture method and study population influenced the sensitivity of the direct test. According to our findings, the direct test cannot replace the culture method for diagnosis of chlamydial infections.