Normally specific treatment of chronic Chagas' disease begins only after a positive parasitological diagnosis has been established. Xenodiagnosis, hemoculture and complement mediated lysis were associated, and repeated, as an attempt to increase the number of selected candidates for specific treatment. Thirty six chronic chagasic patients were submitted to two series of the above tests, with a minimal interval of 60 days. In the first series of tests sensitivity of xenodiagnosis and hemoculture were 30.5% and 8.3% respectively. Processing of a second sample increased sensitivity to 36.1% (xenodiagnosis) and 19.4% (hemoculture); 47.2% were shown to be positive by at least one of these tests. From the positive cases, 29.4% were consistently negative in the complement mediated lysis test, and 17.6% exhibited discordant results, positive on one occasion and negative on the other. Among patients with negative parasitológica! tests, 47.4% had negative complement mediated lysis tests, 31.5% exhibited discordant results and 15.8% were positive. We conclude that complement mediated lysis test is not a method of choice in the selection of candidates for specific treatment of Chagas' disease in view of the observed variability of results. At this moment, parasitological tests, in spite of a low sensitivity, are a safer tool for the clinician.