Different strategies for studying ecological aspects of systemic acquired resistance (SAR)

I am pleased that my original paper (Heil 1999) has prompted discussion on how the ecological aspects of responses of plants to pathogens (particularly systemic acquired resistance, SAR) should be investigated (Hatcher & Paul 2000; H & P hereafter). Our points of view di€er in several aspects and further debate is to be encouraged. H & P stress the potential interactions between SAR and other types of induced plant responses ± for instance, against herbivores ± and a large number of studies (including many conducted by Hatcher and co-workers, see H & P and Hatcher 1995) have indeed demonstrated important interactions between plants, herbivores and fungi. The two most important signalling pathways in induced plant defence are those mediated by salicylic acid (SA) and by jasmonic acid (JA). While some papers have suggested that there may be a positive interaction, most studies have revealed inverse relationships (Felton et al. 1999; Thaler et al. 1999). These `tradeo€s' (Thaler et al. 1999) may be the result of `signal con ̄icts' (Bostock 1999), or even of direct inhibition. As far as I am aware, experimental evidence for a direct inhibition of SA-induced gene products by the JA signalling cascade or vice-versa is still lacking. Even if a direct inhibition could be demonstrated, the question of why this inhibition may have evolved remains unanswered. This question becomes even more pertinent given the numerous occasions on which plants growing in their natural habitat face multiple attacks by di€erent herbivores and pathogens (Hatcher 1995). I very much doubt that strong `changes in conceptual frameworks' (H & P) will be needed in order to incorporate most aspects of SAR into current cost± bene®t models of herbivore resistance (as outlined in Heil 1999). At least in annual species, ®tness costs can be determined from reductions in seed set, whereas allocation costs (which themselves often result in ®tness costs) can be measured by quantifying how limited resources are allocated. Nor do I think that an integrated approach considering all types of induced defence is required from the earliest stages of an investigation into SAR. For example, plants growing under limiting conditions require an e€ective defence, although they are constrained in what they are able to invest. According to the `resource availability' hypothesis (Coley et al. 1985) and the `growth-di€erentiation balance' hypothesis (Herms & Mattson 1992), plant species that are adapted to such conditions should invest relatively more in mobile or inducible, as compared to constitutive, forms of defence. This could be investigated by comparing the potential defence mechanisms of plant species that have di€erent life forms and natural habitats. On the other hand, the constraints imposed when plants are actually exposed to limiting conditions will a€ect investment in both defence and growth, and might be the cause of the reported con ̄icts between di€erent defensive pathways. Recent studies of both molecular (Longemann et al. 1995; Felton et al. 1999) and ecological (Heil et al. 2000) aspects, using the induction of defence in plants grown with only limited resources, have begun to provide evidence for such constraints and for an inhibition of primary metabolism. Results from studies using the cost±bene®t approach are likely to suggest ecological and even evolutionary explanations for many aspects of SAR, which at the moment can only be described without any causal understanding. Consideration of these questions becomes dicult if all the various forms of induced defence are to be included and the increasing complexity of the system may make comparative studies on di€erent species impossible. Hatcher and co-workers choose the ®rst of two possible approaches, whereby they concentrate on one system and try to explore all the interactions that occur. Although the entirety of the selected system can be investigated, such studies are extremely labour-intensive. Comparing di€erent plant species is nearly impossible and it is dicult to draw conclusions concerning general mechanisms. My own approach is therefore to concentrate on more general aspects of a restricted number of *Present address and correspondence: Ce ntre d'E cologie Fonctionelle et Evolutive (CEÁ FE, CNRS), Route de Mende, 34293 Montpellier, Ce dex 5, France (e-mail Martin.Heil@lycosmail.com). Journal of Ecology 2000, 88, 707±708