network does not allow ‘resource sharing’ among linked plants. It is probably irrelevant to the botanical Various claims have been made about the ecological components of a community, but it may be fundasignificance of plant-to-plant carbon movement mental for fungal members. The ‘mycocentric’ view is through common mycorrhizal networks (CMNs). Most that fungal structures within roots are parts of suggest that resource competition among intercon- extended mycelia through which fungi move carbon nected plants should be less important than previously according to their own carbon demands, not those of thought. If true, that would profoundly alter our per- their autotrophic hosts. ception of how plants interact among themselves and with their environment. However, there are difficulties Key words: Arbuscular mycorrhiza, carbon, common in quantifying the amounts of resource transferred via mycorrhizal network, ectomycorrhiza. CMNs, ensuring that transfer is genuinely through hyphae, not soil, and understanding its control. Carbon movement has not been quantified in many of the Introduction published studies. Where it has, its likely functional role has not been clarified. Some recent, well- The 7 August 1997 issue of Nature was headlined ‘The publicized research suggests that carbon transferred wood-wide web’, announcing the publication of the latest to trees via an ectomycorrhizal (EcM) network may be report (Simard et al., 1997a) of plant-to-plant carbon (C ) physiologically and ecologically important. Our view, transfer via a common mycorrhizal network (CMN ). The however, is that the evidence for this remains equi- paper itself, highlighted by Read’s (1997) commentary, vocal. Appropriate controls for the possibility of carbon left Nature readers in no doubt that this intriguing transfer via soil were not used under field conditions. phenomenon was important. Plants connected by a CMN In laboratory experiments, controls failed to clarify the could, it was claimed, share C. role of EcM links in carbon transfer. To resolve some This claim was not new. For example, Grime et al. areas of uncertainty, abundances of 13C have been (1987), suggested that grassland herbs connected by a measured to estimate carbon transfers via an arbuscu- CMN formed by arbuscular mycorrhizal (AM ) fungi lar mycorrhizal (AM) network connecting grasses and allowed C to be transferred from ‘donor’ to ‘receiver’ forbs of the same or different species. Permeable bar- species in laboratory microcosms. But Simard et al. riers to roots and hyphae allowed any direct carbon (1997a) were the first to suggest that the same was true transfer via soil to be detected. Large amounts of for ectomycorrhizal ( EcM ) networks connecting diVerent carbon (typically 10% of that in roots) were transferred tree species in the field, and that the transfer of C was between linked plants via the CMN. Transferred carbon bidirectional. This, clearly, is big news. If plants in a was never transported into shoots of ‘receiver’ plants. community can, via a CMN, really share C, this would It remained in roots, probably inside fungal structures ‘short-circuit’ one of the main constraints to the acquisiand, therefore, unavailable to the plants into which it tion of C by neighbouring plants, namely, competition. Interactions between neighbours would then be less of a was apparently transferred. Carbon transfer via an AM