In this study we examine the variation in Schistosoma japonicum egg counts caused by differences in worm loads between individuals and the variability of egg counts within individuals with a given worm load. Six villages were selected from an area moderately endemic for Asian schistosomiasis in the Dongting Lake region (Hunan province), China. From a total population of 3451 individuals (53.8% male; x=31 years), 163 subjects were identified as consistently stool egg-positive based on three successive positive Kato-Katz (KK) smears (41.67 mg/smear) obtained from one stool specimen. The distribution of eggs among individuals was found to be strongly aggregated (k=0.27), but the distribution within the three smears was found to be only slightly aggregated (k=2.59), indicating only minor clustering of eggs in stools. The relatively slight clustering of eggs within stool specimens suggests that a single KK smear may be quite adequate for detecting individuals moderately to heavily infected (>100 eggs/g stool (epg)), as needed for a strategy of morbidity control. However, for estimating the true prevalence of infection in a community, or for obtaining an accurate estimate of egg excretion for research studies, multiple KK smears are warranted. Copyright (C) 1998 Elsevier Science B.V.