From October 1981 to September 1984, the authors conducted a three-year longitudinal study of diarrhea among infants and toddlers attending day care centers in Maricopa County, Arizona. In the third year of study, they evaluated the effects on diarrhea rates of staff training without external monitoring and of active surveillance conducted throughout the study. From 21 study day care centers, they randomly selected 10 ("intervention day care centers") to receive staff training in procedures to reduce transmission of infectious diarrhea. Continuing active surveillance in the 10 intervention and 11 control day care centers found no difference between diarrhea rates in intervention day care centers in the pre- and posttraining years and no difference between diarrhea rates in the two groups of centers either before or after the training intervention. Biweekly family-based surveys during the two months after training also demonstrated no difference between infant-toddler diarrhea rates in intervention and control day care centers. These surveys found the 21 study day care centers to have significantly higher diarrhea rates than did day care homes or households not using day care, but significantly lower rates than day care centers not included in the active surveillance. Continuous surveillance without training was associated with a significant decrease in diarrheal illness during the course of longitudinal study. One-time staff training without subsequent monitoring did not result in additional decreases and did not lower day care center diarrhea rates to the levels observed in day care homes and households not using day care.