Results of a test of the suitability of the conventional methods utilized in travel-cost demand estimation for estimating the demand for wilderness recreation are reported. The Desolation Wilderness Area in California was selected as a case study and three models were estimated using the 1972 visitation experience data. Results of the test indicated that neither model provided a reasonable representation of the behavior patterns described by the data. This finding is surprising, given the apparent goodness of fit of each of the models. It would tend to indicate that the fairly loose specification of the travel-cost model is not theoretically discriminating enough for the modeling of wilderness recreational behavior patterns due to differences in level of use. Hence, these visits cannot be simply aggregated as if they were homogeneous. It may also indicate that origin zone socioeconomic characteristics do not reflect the characteristics of wilderness users.