A bicycle route and safety survey was distributed to 6,000 bicycle commuters in Ontario in 1995. The objective of the survey was to collect both collision/fall histories and detailed travel behavior information. A description is provided of the questionnaire development, which included a map for route tracing, as well as the sampling procedures that involved attaching the mail-back survey to the crossbars of parked bicycles. The resulting analyses, which address methodological issues, are presented. No differential response rates between men and women were found. No evidence was found to suggest that cyclists who had experienced accidents were more likely to respond. A slight decrease in incidents was found as one moves backward in the time, suggesting slight recall bias, but, overall, the time period over which information was collected (3 years for collisions and 1 year for falls) was deemed appropriate. The measure of travel exposure combined information from the map with estimates of commute trips per month. The aggregate overall estimate was deemed satisfactory but the ability of cyclists to recall commute trips on exact days even in the near past was inadequate. Overall, the survey was successful, and the insights should provide helpful guidance to others who seek to gather bicycle travel information.