In theory, indicators represent two things: changes in habitat over time, and the status of other animal species. But in practice, the problems become apparent. How much can a butterfly, for instance, reveal about the life of a mammal? If a bird that winters in South America fails to nest in Washington state, does that say something concrete about the Northwestern landscape?

Consider the northern goshawk. The raptor is used as an indicator species for old-growth Douglas-fir in the Northern Rockies. The goshawk usually doesn't nest in dead trees, though, so it doesn't reveal much about the status of the flammulated owl or the great gray owl, which do. In some areas of the goshawk's range, it can also nest in lodgepole pines and aspen. So the absence of goshawks in a given stand of Douglas-fir doesn't say anything definite about the state of the habitat.

Or take the sage grouse, often selected as an indicator for sagebrush habitat. Idaho biologist John Connelly recommends a 15 to 25 percent canopy cover of sagebrush for sage grouse breeding habitat. But historically, the average sagebrush canopy may have been much less dense because of fire. Should managers aim to re-create the original landscape, before fire suppression became the norm? Or should they simply try for one that provides the best homes for sage grouse?

In the 1990s, scientists and forest managers began to look for management techniques based on preserving the whole ecosystem, with all of its functions, over a significant period of time, rather than tailoring the forest to meet the needs of one chosen creature.

"If we want to know what the condition of our forests is, why don't we just measure that?"  asks one longtime biologist with the Forest Service, who doesn't want his name used. "Why measure it through a surrogate when we can measure it directly?" If you want to know the condition of a stream, take the water temperature and test the water quality. If you want to know the condition of a stand of trees, look at the age and diameter and canopy closure and the number of non-native species. And manage for specific future landscape goals -- for instance, that the forest should be 20 percent old growth 50 years from now -- providing a wide variety of habitats able to withstand the inevitable disturbances caused by fire, insects and global warming.