A plethora of new concepts for managing production forests so as to preserve biodiversity have found their way into management procedures without much testing to make them most effective. The general framework for a new approach has, in most regions, been ecosystem management, although small reserves are still used, as in the Swedish key habitat sites. Ecosystem management usually entails a focus on processes rather than on species directly, and the key processes in boreal forests are almost always seen as disturbances. Further, although such disturbances as insect and pathogen outbreaks, windstorms, and snow avalanches can be crucial, the concern has been almost wholly with fire. The idea that production forestry should mimic natural disturbance (fire) regimes is a step in the right direction but too simplistic. What needs to be determined is exactly which aspects of natural disturbance regimes are important to which species. Because different types of disturbance interact with one another, understanding the impacts on biodiversity of complex disturbance regimes will require experiments on a scale relevant to regional conservation, as well as intensive monitoring. So far, the effort on experimentation and monitoring is not nearly sufficient to generate detailed management prescriptions, although many management procedures already in place (e.g., retention of green trees and coarse woody debris) are probably useful. © 2001 Taylor & Francis.