Analysis of Cryptosporidium occurrence in six watersheds by method 1623 and the integrated cell culture-PCR (CC-PCR) technique provided an opportunity to evaluate these two methods. The average recovery efficiencies were 58.5% for the CC-PCR technique and 72% for method 1623, but the values were not significantly different (P = 0.06). Cryptosporidium oocysts were detected in 60 of 593 samples (10.1%) by method 1623. Infectious oocysts were detected in 22 of 560 samples (3.9%) by the CC-PCR technique. There was 87% agreement between the total numbers of samples positive as determined by method 1623 and CC-PCR for four of the sites. The other two sites had 16.3 and 24% correspondence between the methods. Infectious oocysts were detected in all of the watersheds. Overall, approximately 37% of the Cryptosporidium oocysts detected by the immunofluorescence method were viable and infectious. DNA sequence analysis of the Cryptosporidium parvum isolates detected by CC-PCR showed the presence of both the bovine and human genotypes. More than 90% of the C. parvum isolates were identified as having the bovine or bovine-like genotype. The estimates of the concentrations of infectious Cryptosporidium and the resulting daily and annual risks of infection compared well for the two methods. The results suggest that most surface water systems would require, on average, a 3-log reduction in source water Cryptosporidium levels to meet potable water goals.