A trapping ELISA for the detection of circulating antigens of Trypanosoma vivax with a specific monoclonal antibody was developed by Nantulya and Lindqvist in 1989. With the African strains tested, its sensitivity proved very high. Using reagents supplied by ILRI, the sensitivity of this test was compared with that of the Woo test in two sheep experimentally infected with a Guyanese strain of T. vivax. Blind tests carried out at CIRDES and ILRI confirmed the results obtained in French Guiana. The animals were bled regularly during the first 130 and 285 days of infection. Whatever the period of infection ans whether there was patent parasitaemia or not, the mean sensitivity of the ELISA test was very low: 2.1% of positive results, far below the Woo test's 54% of positive results. Associating the two techniques did not improve the sensitivity. The mean optical densities recorded during these long periods of infection (0.010 +/- 0.004 and 0.012 +/- 0.002) are so close to the sensitivity threshold of the ELISA readers that they cannot be proposed as a cut-off threshold to improve the sensitivity of the test. A complementary study of the sensitivity of this test is in progress at ILRI. As already observed in Burkina Faso and The Gambia, this poor sensitivity may greatly affect the results of the epidemiological surveys being carried out with these reagents in Africa. New monoclonal antibodies must be developed to produce a test for T. vivax antigens with more satisfactory sensitivity.