TREE vol. 13, no. 4 April 1998 133 known about the pollinators, seed disper- sal, seed banks or gene flow for the plants. However, she reported that most plans in- clude the gathering of more information about pollinators, breeding system and seed dispersal. The number of plans listing knowledge of pollinators as a high-priority task has increased significantly. While this is encouraging, no plan included manage- ment of pollinators in the recovery of the endangered species. All participants agreed that plant–pol- linator interactions will often be important in the restoration or re-establishment of damaged ecosystems. Direct effects of de- graded pollination systems are often easily recognized and are receiving more atten- tion from restoration and conservation bi- ologists. It is the less obvious effects of de- graded pollination systems on population and genetic structure that are in need of more attention. Perhaps most importantly, we need to shift the emphasis in resto- ration biology from individual species to community interactions and ecosystem function. Paul R. Neal Dept of Biology, Cox Science Building, University of Miami, Coral Gables, FL 33124, USA (pneal@fig.cox.miami.edu) NEWS & COMMENT Copyright © 1998, Elsevier Science Ltd. All rights reserved. 0169-5347/98/$19.00 PII: S0169-5347(97)01293-7 O ver the past several years, ecologists have begun to recognize that the role played by positive plant–plant interac- tions in communities is underappreciated. In a recent paper in Oikos 1 , Brooker and Callaghan offer an interesting perspective and new insights to this ongoing cathar- sis. They argue that plant interactions are best characterized as a balance between positive and negative components and that this balance is tipped in favor of net posi- tive or negative effect by environmental factors. This argument has also recently been made by Callaway and Walker 2 and Holmgren, Scheffer and Huston 3 . Clearly, this is an idea whose time has arrived. Brooker and Callaghan’s perspective on this balance, however, is unique in that they see the issue through the eyes of arc- tic plant ecologists, and envision physical disturbance rather than environmental stress as the trigger that shifts the balance between positive and negative effects. Brooker and Callaghan begin by noting the lack of attention paid to positive plant– plant interactions by ecologists over the past three decades, and argue that a num- ber of recent studies suggest that positive interactions among plant neighbors are commonly important. They then observe that recent examples of positive interac- tions playing an important role in commu- nity structure have two things in common. First, most examples are from relatively harsh physical environments. Second, the environmental factor constraining plant growth or survival in these examples can be alleviated by the physical presence of other plants. They then argue that posi- tive components in plant neighbor inter- actions are probably common in all habi- tats, and that, in general, plant interactions are best characterized as a balance be- tween positive and negative components. Fine so far. The paper then takes a bit of a twist. They define environmental severity as a combination of both physical stress and disturbance (following Grime 4 ) and con- cede that since the relationship between the roles played by competition and physi- cal stress in structuring communities is contentious, they will focus instead on the role played by disturbance in tipping the balance between positive and negative plant neighbor effects. This is perhaps surprising, since they had just made such a good case for a strong relationship be- tween positive interactions and physical stress. Indeed, few of their examples were unambiguously associated with disturb- ance or independent of physical stress. Brooker and Callaghan suggest that while it is widely accepted that increased disturbance reduces the intensity and im- portance of competition in plant commu- nities, as a by-product of reducing the level of competition, disturbances must increase the importance and intensity of positive neighbor interactions. They argue that since disturbance decreases the intensity of competition and that plant–plant inter- actions are a direct balance between nega- tive and positive effects, disturbance must increase the relative importance and in- tensity of positive interactions. (But surely any balance between negative, competitive and positive, ameliorating components of interactions must have a mechanistic ba- sis? Disturbances remove competitors, but only remove cooperators if there is some- thing to cooperate about.) They then ar- gue that recent examples of positive inter- actions in plant communities are driven by environmental severity which includes disturbance. In their discussion of these examples, however, physical stress al- leviation implicitly remains prominent. They conclude that ‘community manipu- lation experiments are more likely to reveal positive plant–plant interactions in highly disturbed communities … and as disturb- ance is reduced either through space or time, competition becomes more impor- tant and in parallel, positive interactions become less important, so that competition becomes the dominant selective force in plant–plant interactions’. The paper closes strongly with an in- teresting and thoughtful discussion of the selective impact of positive plant interac- tions drawn primarily from their work on arctic plant communities. Decreased seed dispersal distances, compact plant mor- phologies, limited responses to improved conditions, increased frequency of mycor- rhizal associations and increased domi- nance of clonal plants are all suggested to be traits selected for in extreme environ- ments where plants can ameliorate the harsh conditions. This is an interesting, thought- provoking, if sometimes perplexing paper. Brooker and Callaghan challenge ecolo- gists to get a better grasp on the role played by positive interactions in commu- nities, and correctly, I think, characterize plant–plant interactions as a balance be- tween negative and positive components. Their examples, and discussion, however, are more consistent with physical stresses that are ameliorated by plant cover trig- gering the dominance of positive interac- tions in communities than disturbances triggering a shift to net positive effects 5 . I suspect that their views are strongly in- fluenced by the arctic plant communities they study where plants can protect one another from disturbance, but that disturb- ances also typically create stressful physi- cal conditions that plant cover can amelio- rate. In physically benign habitats, where disturbances lead to safe sites and intense competition among recruits, their disturb- ance-based model should break down. In focusing on disturbance and simplicity, their discussion also entirely ignores con- sumers and the prominent positive roles played by plant neighbors and associ- ational defenses 6,7 . Any realistic model of the role played by direct positive interac- tions in the structure of plant communities Searching for the role of positive interactions in plant communities