Pulsed-field gel electrophoresis (PFGE) was used to analyse 147 isolates collected in two regions of Quebec province (Estrie and Montreal) between March 1998 and Feb. 1999, to determine the utility of molecular strain typing for a population-based collection of Campylobacter jejuni and to compare directly the discriminatory power of SmaI and KpnI restriction digests. With a combination of epidemiological criteria including space and time plus molecular strain typing, 49% of isolates from Estrie and 39% of isolates from Montreal were identified as belonging to a putative cluster. For 41% of the cases, sources were either missing or explicitly unknown; the remaining sources were subject to recall bias. Thus, the evaluation of sporadic cases of campylobacter enteritis by descriptive clinical investigation alone is neither sensitive nor reliable for identifying sources of infection. In the PFGE analysis, KpnI digests provided appreciably greater discriminatory power than SmaI digests. When combining the PFGE analyses with basic epidemiological criteria, 30% of the putative SmaI clusters were inconsistent with the epidemiological criteria compared with 17% of the KpnI clusters. Among the 98 isolates assigned to clusters by SmaI, only 65% gave concordant results with KpnI. In contrast, among the 81 isolates assigned to clusters by KpnI, 92% gave concordant results with SmaI. Finally, clusters that were epidemiologically related to ingestion of raw milk and specific water sources correlated better with the typing results based on KpnI than SmaI. Thus, KpnI is the enzyme of choice for molecular epidemiology studies of C. jejuni. The combination of continuous epidemiological surveillance and molecular strain typing may be useful for identifying new sources and mechanisms of transmission for community-acquired C. jejuni infection andultimately for developing new approaches to prevention.