Functional importance vs keystoneness: Reformulating some questions in theoretical biocenology

Two concepts relating to the influence of individual species on the biocenoses in which they occur are reviewed. The first, the general functional importance of a species, is defined as the sum, over all species, of the changes (sign ignored) in productivity which would occur on removal of the particular species from the biocenosis. General functional importance is calculated as: [GRAPHICS] where P-j is the productivity of the jth species before (t = 0) and after (t = 1) removal of the particular (ith) species being evaluated. Though I-i values cannot be determined empirically, this concept raises provocative questions for theoretical biocenology. The second concept reviewed is that of the keystone species. Never having been precisely or operationally defined, 'keystone' has come to mean little more than 'important for something.' Moreover, there is no empirical or theoretical foundation for the idea that there exists in any biocenosis a natural dichotomy corresponding to the verbal one of keystone and non-keystone species. Some investigators have implied that such a dichotomy is suggested by the frequency distributions of experimentally determined values of interaction strength. The patterns they refer to are, however, artifacts resulting from small sample sizes and the plotting of frequency distributions on arithmetic rather than logarithmic scales. As a casual metaphor 'keystone species' was appealing and harmless; but the pretence that it is a well-defined concept or phenomenon has had a stultifying effect on ecological thought and argument.