Sera of individuals from Burundi excreting eggs of Schistosoma mansoni (prevalence 35%; 178 subjects) and of similar individuals from Maniema, Zaire (prevalence 95%; 99 subjects), and of 159 Dutch and 81 Zairean non-infected controls, were screened by enzyme-linked immunosorbent assay for the presence of schistosome circulating anodic antigen (CAA). No false positive results were obtained. The sensitivity of the test was 75% in Burundi and 93% in Zaire, a significant difference (P less than 0.05). However, in matched egg output classes the test results did not differ significantly; 60% and 67%, respectively, of those excreting 1-100 eggs per gram of faeces (epg), 86% and 100% of those excreting 101-400 epg, and 100% of those excreting over 400 epg were detected. The efficiency of the assay was 91% in Burundi and 93% in Zaire. The Spearman rank coefficient of correlation between antigen titre and egg output (determined by 3 consecutive Kato egg counts) was 0.61 in Burundi and 0.82 in Zaire. The sensitivity of the test compared well with a single egg count. In addition, preliminary data showed that occasionally CAA was detectable in serum of individuals not excreting schistosome eggs. As CAA is found only in the presence of living worms, such cases reflect active infections.