This Article considers the lessons that may be drawn from the recent controversy created by one federal agency’s shifting approach to the use of models and surrogates.  The National Forest Management Act (NFMA) delegates to the National Forest Service (Forest Service, or Service) the responsibility to develop land and resource management plans (LRMPs) for units of the National Forest System (NFS) and to make site-specific decisions about the use of those units in a manner consistent with the plans. The NFMA charges the Forest Service with the task of issuing regulations governing the land use planning process that, among other things, “provide for diversity of plant and animal communities based on the suitability and capability of the specific land area.” The Service has encountered difficulty in carrying out its responsibility to issue and apply those regulations because of the enormous complexity of the ecosystems within the forests under its jurisdiction.  To minimize the uncertainty it faces in predicting what impact a particular proposed action, such as a timber sale, will have on the diversity of plant and animal species in the affected forest, the Forest Service has turned to the use of models and surrogates.  For years, acting under land use planning regulations issued by the agency in 1982, it identified management indicator species (MIS) that it determined were representative of the health of the ecosystem as a whole.  The MIS were supposed to act as surrogates for the impact of activities such as timber sales on plant and animal diversity.  The agency’s aim was to predict the effects of management actions on the selected MIS and to monitor the fate of the MIS after the action was taken to determine whether the action was interfering with diversity.  If it was, suitable changes in management approaches could then be made.