Epidemiological studies of shigellosis in Bangladesh have demonstrated that surface-water sources can act as foci of infection. Studies of laboratory microcosms have shown that shigellae become nonculturable but remain viable when exposed to environmental samples of water. The present study was carried out to detect viable but nonculturable Shigella dysenteriae 1 from laboratory microcosms by the polymerase chain reaction and the fluorescent-antibody techniques. S. dysenteriae 1 was inoculated into laboratory microcosms consisting of water samples collected from ponds, lakes, rivers, and drains in Bangladesh. The survival of S. dysenteriae in microcosms was assessed by viable counting on MacConkey agar. After 2 to 3 weeks, S. dysenteriae 1 became nonculturable but remained viable. After 6 weeks, this nonculturable but viable S. dysenteriae 1 was detected by both the polymerase chain reaction and the fluorescent-antibody methods. The viable but nonculturable state of S. dysenteriae 1 demonstrated in this study may be important for understanding the epidemiology of shigellosis.