A modified NC membrane-based dot-ELISA was used to detect and differentiate between Trypanosoma brucei, T. congolense and T. simiae procyclics in the midguts of experimentally infected tsetse flies. The modification of the assay consisted of (a) the lysis of T. congolense or T. simiae in NC membrane applied sample dots using Triton X-114, and (b) treatment of sample applied NC membrane strips with hydrogen peroxide to remove non-specific stains. Also, T. brucei was detected in the salivary glands, and T. congolense and T. vivax were detected in the mouthparts, however, in dot-ELISA without modification. In all the assays, T. brucei and T. congolense parasites were detected directly using MoAbs specific to each of them, whereas T. simiae parasites were detected by exclusion using a T. congolense specific and Nannomonas subgenus-specific MoAbs. The sensitivity of the assay for detecting midgut infections was 90.5%, 84.6% and 94.4% in detecting T. brucei, T. congolense and T. simiae, respectively. Sample dots stored at room temperature (19-26 degrees C) under desiccated conditions did not show any loss in activity in 90 days. However, after 7 days of storage, a ring-pattern reaction appeared on some sample dots that were tested with T. brucei specific MoAb, irrespective of whether T. brucei antigens were present or not. These ring reactions, however, did not interfere with correct interpretation of the assay results. The specificity of the assay for detection of T. brucei in the salivary glands was 100% and the sensitivity was 90%. Also, T. vivax and T. congolense organisms were each detected in the mouthparts of infected tsetse flies, with 100% specificity. The sensitivity was, however, lower, 43.8% for T. vivax and 55.6% for T. congolense.