Treatment of time in travel cost models has been a source of contention among economists. The debate persists because welfare estimates, which are the principal objectives of these studies, are highly sensitive to the treatment of time. The present study examines the dual role of on-site time using evidence from two wilderness areas in Alabama. The empirical results comply with the theoretical expectation that on-site time is both a source of utility and cost. The exclusion of on-site time from demand functions results in biased parameter estimates. In particular, it yields smaller own-price coefficients and higher welfare estimates.