Introduction

Populations of arthropod herbivores may show periodical outbreaks, large amplitude cycles, strongly bounded fluctuations or stable equilibria. At small spatial scales they may also show dynamics different from those at large spatial scales, and they may even go extinct. There is a need to experimentally test models predicting the dynamical consequences of the mechanisms underlying these patterns. In particular, what needs an explanation is the observation that plants retain a green appearance despite attacks by herbivores (Hairston et al., 1960; Strong et al., 1984). There are two possible explanations: (i) plants defend themselves effectively against herbivores, (ii) predators suppress herbivore populations to very low levels. Hairston et al. (1960) tacitly ignored the first and emphasized the latter in formulating their so-called ‘the world is green’ hypothesis. The two explanatory mechanisms, however, are not mutually exclusive; the plant may promote predator foraging success and the herbivore's enemies may use the facilities offered by the plant.

Price et al. (1980) stimulated a growing awareness that plants may defend themselves both directly against herbivorous arthropods and indirectly by promoting natural enemies, for example, by providing protection, food, and alarm signals to the enemies of the herbivore. Direct defenses include plant structures that hinder feeding by the herbivore, and secondary plant compounds that inhibit digestion, intoxicate the herbivore or deter feeding. Before Price et al. (1980) appeared, examples of indirect defenses were already well-known from ant–plant interactions (Janzen, 1966; Bentley, 1977; see also historical reviews by Beattie, 1985, and Jolivet, 1996).