The interval distribution for exclusive, mixed-use bicycle paths that include pedestrians, skaters, and joggers is examined. Interval distributions are used to generate arrivals, an essential input to a traffic simulation model. Different time headway distributions have been fitted for vehicular traffic flow based on the type of facility, traffic composition, and flow rate. However, only one study reports on the arrival patterns of bicycles approaching an intersection, and no studies examine the interval distribution for bicycles on exclusive bicycle paths. Bicycle traffic data collected for two bicycle paths in Denver, Colorado, were analyzed. Two populations were considered to describe arrival patterns on bicycle paths. The first consisted of the combined population of bicyclists, pedestrians, skaters, and joggers. The second population consisted of only bicycles, which are considered separately. With bicycles classified as unconstrained and constrained in either a combined population of bicyclists, pedestrians, skaters, and joggers or a separate population of bicyclists only, a Schuhl's composite distribution of negative exponential distribution was found to model interval distributions. The composite distribution's fit was significantly better than a negative exponential distribution, as demonstrated by Kolmogorov-Smirnov tests. The distributions may be used to generate not only the interval between consecutive arrivals but also the type of user.