Forest management must change radically to maintain biodiversity. 'Biodiversity' has many components, but only one has been measured unambiguously - species richness - although there is recently much emphasis on structural, process, and functional diversity. So we must determine exactly what aspects of biodiversity to seek, and why. A battery of suggestions about how to achieve this re-focus on biodiversity (e.g., ecosystem management and the 'new forestry') are catchwords rather than guides on how to manage forests on the ground. These suggestions stem from an ill-defined concept of 'forest health,' which can be seen variously depending on the desired role and state of a forest. Ecosystem management for some versions of forest health may even decrease some forms of biodiversity. A decline in species richness need not lead to a decline in the process diversity or rates. Evidence that species richness contributes to ecosystem maintenance and function is scant. Thus, effective management for biodiversity (generally species richness) entails a frank commitment to maintain biodiversity as an end, not as a means. Some suggestions to maintain forest biodiversity while still allowing timber production, such as uneven-aged stand management and various burning regimes, are focused squarely on species richness per se, but they are hypotheses, not scientifically validated procedures. Existing empirical measurements on such techniques are usually on the amount and sustainability of timber harvest, not on how well they maintain species richness. A wealth of scientific research is needed, involving landscape-level field manipulations and careful natural historical observations on the effects on various species. The idea that forests can always serve multiple uses, including wood production and maintenance of all species, is an untested hypothesis. It may be incorrect; maintaining some species may require extensive pristine tracts. The major requirement for almost all research needed to manage forests for biodiversity is extensive and intensive monitoring. The concepts of umbrella and indicator species as management shortcuts are barely tested. Their utility can be validated only by intensive field study. Valuable umbrellas and/or indicators may exist for some forest systems. However, management procedures should not evolve towards management of indicator species, as the indicator might cease to indicate the status of other species. By contrast, managing an umbrella species is not an inherent contradiction in terms, but different umbrella species may shelter different sets of species, so management for one might be inimical to the other. The concept of keystone species may be useful in forest management. If the fates of particular species determine those of many others, managing for such keystones may effectively maintain species richness. But recognition of a keystone species requires well-designed experiments. (C) 1999 Elsevier Science B.V. All rights reserved.
