In areas where there is a low prevalence of schistosomiasis mansoni, faecal examination is a relatively insensitive method of detection and infected people may also be missed because most show only mild morbidity. In such settings, serology may be a more useful diagnostic tool than microscopy. In the present study, the clinical and biochemical characteristics of individuals who were stool-positive for Schistosoma mansoni eggs were compared with those of individuals, from the same low-prevalence area of Brazil, who were stool-negative but seropositive for the parasite. Overall, 269 subjects were checked both for schistosome eggs in their faeces (using Kato-Katz smears and Lutz sedimentation) and for anti-S. mansoni IgG in their sera (using an ELISA). Although 128 (48%) of these subjects were found seropositive, only 26 (10%) were found to be egg excretors and two of the egg excretors were seronegative. Compared with the seropositive egg-negatives, the egg excretors had significantly higher frequencies of fatigue, melaena, jaundice and swelling of the abdomen. The egg excretors also had higher frequencies of hepatomegaly (20% v. 16%) and splenomegaly (4% v. 1%). In both groups of subjects, mean concentrations of serum proteins and haemoglobin and mean leucocyte counts were in the normal range whereas most blood concentrations of alanine aminotransferase and many of those of aspartate aminotransferase were slightly elevated. Although the egg excretors tended to have low-intensity infections, it seems possible that the seropositive nonexcretors had even milder infections that could not be detected by faecal examination. The high frequency of cure observed when the egg excretors were given praziquantel at 40 mg/kg (94%) is probably another indication that most had light infections when they were treated.