High altitude spruce fir forests are typical around the world and are often subjected to multiple forms of recreational use. In this paper, we use household and recreation group data for a spruce fir forest high in the Appalachian Mountains of the U. S. to evaluate the benefits from forest protection (i. e., from improving the forest condition). Our benefits estimation procedures use the referendum-type, contingent valuation (CV) approach of Cameron (1988). We modify the usual practice of obtaining a single willingness-to-pay (WTP) value by using alternative questionnaire scenarios and conducting tests to examine i) household and recreation group value sensitivity to forest condition, and ii) recreation group differences in WTP for forest protection. A first sample of southeastern U. S. households was asked to value a forest protection program for a spruce-fir forest showing no impact from insect disturbance or atmospheric deposition. The second sample was asked to value a protection program for a forest already experiencing impact from insect infestation and air pollution. Logit analysis of the two samples revealed no statistically significant difference in household WTP between the two forest protection programs. Further analysis indicated that consumptive forest users (i. e., hunters and anglers) held forest protection values that were sensitive to a change forest condition, while nonconsumptive forest users (i. e., campers and hikers) held values that were insensitive to the same condition change. Recreation group comparisons revealed that consumptive forest users also held lower values for forest protection than nonconsumptive recreationists. These results demonstrate the importance of estimating public values for forest protection in terms of heterogeneous groups rather than as a homogeneous whole.