Diagnosis of Gambian sleeping sickness is problematic because of the very low levels of parasitaemia encountered in the field. A PCR method developed for the sensitive detection of Trypanosoma brucei was used to diagnose parasitologically negative suspects in a recent survey in Cameroon. Individuals were screened in two foci (Mbam and Fontem), firstly with the card agglutination test for trypanosomiasis (CATT) as a primary serological test, together with palpation and puncture of enlarged cervical lymph glands. Any suspects found positive by CATT (CATT+) and any clinical suspects were then subjected to several parasitological tests (examination of thick blood films and use of hematocrit centrifugation, mini-anion-exchange chromatography and a commercial kit for in-vitro isolation). Overall, 43 of the 1703 subjects screened in the Mbam focus were CATT+ and three (two of whom were CATT+) had enlarged glands. In Fontem, 56 of the 1210 subjects screened were CATT+, 78 (24 of whom were CATT+) had enlarged glands and two (both CATT+) had trypanosomes in their gland juice. However, all the suspected cases of sleeping sickness, including the two gland-positives, gave negative results in the secondary, parasitological tests. Blood samples from 28 suspects from Mbam and 30 from Fontem were selected for PCR analysis on the basis of high CATT response or clinical grounds. For each suspect, DNA was prepared from 0.5 ml blood by phenol extraction or differential lysis and then amplified by PCR using specific primers for T. brucei ssp. Four samples from Mbam and nine from Fontem, including the two gland-positives, were found positive by PCR. Compared with the other parasitological techniques, therefore, PCR was the most sensitive diagnostic method in this study, with an estimated sensitivity of 25 trypanosomes/ml blood. Although PCR analysis is too expensive for routine diagnosis, it could be very useful in determining which sleeping-suspects should be closely followed up.