License plates were used as markers for capture-recapture estimates of vehicle numbers on a 12,547-ha study area. License plate numbers were recorded during random surveys. An attempt to estimate vehicle numbers, using open-population model JOLLY, failed due to problems with equal catchability caused by users leaving the area, then returning. Chapman's Lincoln-Peterson index provided acceptable estimates of vehicle numbers. Numbers of vehicles were calculated for each sample day, then extrapolated to the entire study period. Average number of people per vehicle, area of use, and use activity were calculated using information obtained from mail and personal interview questionnaires.