In April 1997, a large city in the northeastern United States changed their drinking water treatment practices. The city, which previously provided only chlorination for their surface water sources added filtration in addition to chlorination. To assess whether Cryptosporidium infections rates declined following filtration, we tested serological responses to 15/17-kDa and 27-kDa Cryptosporidium antigens among 107 community college students 1 month before and 225 students 5 months after filtration. Results suggest that levels of Cryptosporidium infections did not decline following water filtration. However, seasonal changes in other exposures may have confounded the findings. Swimming in a lake, stream or public pool and drinking untreated water from a lake or stream predicted a more intense response to one or both markers. Residence in the city, not drinking city tap water or drinking bottled water, gender, travel or exposure to pets, young pets, diapers or a household child in day care were not found to be predictive of more or less intense serological responses for either the 15/17-kDa and 27-kDa antigen.