In early 1992 we identified an outbreak of cryptosporidiosis in Oregon and sought to identify and control its source. We used a series of studies to identify risk factors for illness : (i) a case-control study among employees of a long-term-care facility (LTCF); (ii) a matched case-control study of the general community; (iii) a cohort study of wedding attendees; and (iv) a cross-sectional survey of the general community. Drinking Talent water was associated with illness in the LTCF (OR = 22·7, 95% CI = 2·7–1009·0), and in the community (matched OR = 9·5, 95% CI 2·3–84·1). Drinking Talent water was associated with illness only among non-Talent residents who attended the wedding (P < 0·001) and in the community (RR = 6·5, 95% CI 3·3–12·9). The outbreak was caused by contaminated municipal water from Talent in the absence of a discernible outbreak among Talent residents, suggesting persons exposed to contaminated water may develop immunity to cryptosporidiosis.