We found a human reovirus-like agent in the stools of 42 per cent of 143 infants and young children hospitalized with acute gastroenteritis between January, 1974, and June, 1975. Half the patients studied by electron microscopy and serologic technics had evidence of infection with the agent. The infection had a seasonal pattern: 59 per cent of those admitted during the cooler months (November to April) shed the agent, with a peak of 78 per cent in December, 1974, and January, 1975, combined. None of the patients admitted during the warmer months (May to October) shed the agent. None of 275 Escherichia coli isolates from 32 patients with diarrhea produced heat-labile enterotoxin, whereas 17 of the 32 had evidence of infection with the reovirus-like agent. In addition, 14 of 40 parents of 37 patients with diarrhea associated with the reovirus-like agent were also infected, but most infectious were inapparent. This agent appears to be the major cause of diarrheal illness in the young during the cooler months.

Reoviruslike particles were visualized by electron microscopy in stool filtrates prepared from stools of infants and young children with severe acute gastroenteritis. Patients who had such particles in their stools and whose paired acute and convalescent serums were tested developed an antibody response to the reoviruslike agent, which was measured by immune electron microscopy and by complement fixation. The reoviruslike agent was antigenically related to the epizootic diarrhea of infant mice virus and the Nebraska calf diarrhea virus.

An immune adherence hemagglutination assay (IAHA) and a modified enzyme-linked immunosorbent assay for antigenic characterization of human rotaviruses were developed. The designations of type 1 and type 2 were identical to those established previously by specific complement fixation, enzyme-linked immunosorbent assay, and immune electron microscopy. By IAHA (and modified enzyme-linked immunosorbent assay) certain animal rotaviruses were found to be closely related to human rotavirus type 1. The pattern of IAHA reactivity and the cell culture neutralization serotype were found to be distinct properties. The separation of neutralization and IAHA reactivity was apparent when animal rotaviruses which were distinguishable from each other by neutralization assays were found to share IAHA specificity. Further evidence for the dissociation of the neutralization and IAHA specificities was found in studies of human and bovine rotaviruses which underwent genetic reassortment during coinfection. Thus, it appeared that the IAHA and neutralization antigens were coded for by different genes. In view of these findings, we suggest that the term serotype be reversed to identify the antigen that reacts with neutralizing antibodies as is customary for other viruses and that the term subgroup (instead of serotype) be used for the specificity detected by specific complement fixation, enzyme-linked immunosorbent assay, and now IAHA.