Recent publications revealed a high percentage of false positives when the direct fluorescent antibody test (DFA) was used for the diagnosis of pertussis. The aim of this work was to analyse the results obtained in our laboratory when the DFA was used for pertussis diagnosis, in two child-populations. The first group consisted of 267 children, from 15 days to 7 years old, having clinical diagnosis of whooping cough, etiology to be determined. A nasopharyngeal swab was obtained frorn each child of this group. The control group consisted of 50 normal children between 3 and 28 months old, from each of whom a nasopharyngeal swab was obtained. Sample processing included culturing in Regan- Lowe agar and DFA using Bordetellaperiussispolyclonal antibodies and a control strain. Control samples were also cultured in sheep blood agar plates and chocolate agar. B. periussis was isolated in culture in 36 of the 267 aspirates studied and the DFA was reactive in 46 of them. In 30 cases both tests were positive. Specificity for DFA was 93%. In the control group, B. pertussis was not encountered in culture, but DFA was reactive (though with low reactivity) in 8 cases (16%) (specificity 84%). False positives were confirmed by Gram stain as Bacillus sp. in 4 cases and by culture as Haemophilus influenzae and Streptococcus pneumon~aein 2 cases each. Our data confirmed that DFA could be used in the pertussis diagnosis if the interpretation of the results are based on microscopic morphology of the microorganisms as well as in the intensity of the fluorescence. The use of a trained technician and control strains are thus highly recommended.