- Top of page
- Materials and methods
Plants may control their interactions with herbivores via a range of defence mechanisms (Agrawal & Karban 1999), from associational defence (e.g. Atsatt & O'Dowd 1976), which is the least reported, to chemical defence (e.g. Paul et al. 2001), the most frequently reported, and perhaps most widespread, type.
Despite their obvious protective benefits, defences may be associated with several disadvantages for their users (Agrawal & Karban 1999), including the risk of self-intoxication by defence compounds, metabolic costs of production and/or storage of chemicals, and herbivores becoming adapted to constitutive plant traits that would otherwise repel them (but see Schmitt et al. 1995; Agrawal 2000). If defence against herbivory incurs a metabolic cost, plant fitness should be enhanced by tolerating herbivory under low grazing load, but inducing defences when grazing intensity is high (Karban et al. 1999). Such an inducible response should allow deactivation or elimination of defensive traits to pre-stimulus levels after consumer activities cease (Zangerl 2003), so that adjustable, rather than permanently expressed (constitutive) defences, could reflect a selective advantage under variable grazing pressure (Agrawal 2001). Additional potential benefits of inducibility could be its lack of predictability (Hay 1996) and its suitability as a cue for predators/parasitoids of plant consumers (Zangerl 2003).
Inducible defences can have important implications in shaping ecological communities (Trussell & Smith 2000; Agrawal 2001; Van Zandt & Agrawal 2004a) and can also structure food webs, either directly, by altering the abundance and distribution of herbivores (references in Agrawal 2001), or indirectly by attracting their enemies (e.g. Turlings et al. 1995). Ultimately, even herbivore community richness may thus be controlled by inducible plant defence (Van Zandt & Agrawal 2004a). The specificity of induced defences, which is an important but so far understudied feature, can result from either the dependency of plant responses on grazer identity (Agrawal 1999) or effects being specific to those herbivores that are sensitive to particular compounds (Hay et al. 1987; Agrawal 2000; Huntzinger et al. 2004). Specificity of plant responses has been reported at the molecular (de Moraes et al. 1998; Stout et al. 1998) and metabolic pathway levels (Zangerl 2003), for both grazer-emitted (Agrawal 1999, 2000; Van Zandt & Agrawal 2004b; but see Coleman et al. 1996), and volatile plant cues (de Moraes et al. 1998; Karban et al. 2004). Temporal variability in the abundance of a herbivore, as well as in chemical plant properties (Bowers & Stamp 1993), should add an additional level of complexity to plant–animal interactions. Although the time lag of induction should match the temporal variability of consumption pressure, there is very little evidence for such patterns (Turlings et al. 1995; Doan et al. 2004).
Although inducible defences in terrestrial plants have been much more extensively studied than in aquatic plants, recent studies indicate that this phenomenon might be common, and taxonomically widespread, among seaweeds (e.g. Hemmi et al. 2004, brown algae; Ceh et al. 2005, red algae; J. Ceh, unpublished data, green algae). The very few studies testing for specificity of induced defences in macroalgae (Pavia & Toth 2000; Rohde et al. 2004), suggested that alga-herbivore combination, as well as the consumer itself, influences the ability of macroalgae to induce defences (Amsler 2001).
The present study investigated the combined effects of timing and specificity of induced defences, which were previously studied only in isolation. It was tested whether inducible defences in the habitat-structuring brown alga Ecklonia cava were (i) grazer-specific, (ii) seasonally variable, (iii) due to waterborne cues from neighbouring grazed conspecifics or non-grazing consumers, and (iv) reduced after grazing ceased.
- Top of page
- Materials and methods
This study revealed that a defence induction mechanism exists in Ecklonia, but depends on the type of mesoherbivore, and that the alga reduced this defence shortly after grazing ceased. The response varied seasonally, being present in summer and absent in autumn, and could not be induced by waterborne cues.
Previously attacked individuals were less palatable than ungrazed controls, indicating that direct Littorina grazing induced Ecklonia to express a herbivore-repulsive morphological or chemical trait. While the induced plant trait may play multiple functions (e.g. antifouling and antiherbivory Schmitt et al. 1995) the fact that it was switched on and off by grazer activity hints at its primary antigrazing role. Although a reduction of previously induced defences is suggested to be an integral part of inducible responses (Zangerl 2003), this has, to date, been reported for very few algae (Hemmi et al. 2004; Rohde et al. 2004; Weidner et al. 2004). Inducible defences under variable grazing regimes may confer a higher fitness than is possible for constitutively defended or undefended species (sensuKarban et al. 1999). Species with inducible defences are also phenotypically more variable than non-induced species and this should lower the risk of coevolution by their enemies (Agrawal & Karban 1999).
Ecklonia individuals responded differently to the two herbivore species used, reflecting a threat-dependent response by the alga. Higher in situ densities of Littorina than Haliotis (Yang 2002) suggest that natural grazing pressure of the former is more important, although identical consumption rates in most feeding assays demonstrate that both grazers could exert similar feeding pressure on Ecklonia. While rigorous experimental evidence of grazer-specific induction of defences in macroalgae is scarce (Pavia & Toth 2000; Rohde et al. 2004), more support for specificity of inductive responses and effects comes from studies on leaf-chewing insects (e.g. Stout et al. 1998; Agrawal 1999, 2000). Here, specific induction is triggered by salivary constitutes (Mattiacci et al. 1995; Alborn et al. 1997) or the type of feeding damage (Stout et al. 1998). Similarly, Pavia & Toth (2000) hypothesized that differences between the snail L. obtusata and the isopod Idotea granulosa in their ability to induce herbivore defences in the brown alga Ascophyllum nodosum might stem from the relatively more concentrated damage caused by snail attacks leading to a higher risk of losing complete thallus parts. However, Rohde et al. (2004) demonstrated an equal ability by L. littorina and I. baltica to induce defence in Fucus vesiculosus, suggesting that the specificity of induction may differ between algal species. The present study adds a new level of specificity, because Ecklonia induction differed between grazer species with the same grazing mode.
A specific response to waterborne grazer cues might represent a mechanism by which plants could anticipate impending grazing and therefore induce defences before incurring damage. Incubation with non-grazing snails and isopods failed to induce defences in F. vesiculosus, while waterborne cues from grazing isopods, but not snails, were effective (Rohde et al. 2004). In the present study, palatability of control pieces was not significantly different to Ecklonia individuals that had been exposed to waterborne cues. However, this interpretation might be confounded by seasonal effects, as responses of waterborne cues were only tested in autumn, i.e. the season when direct exposure to Littorina no longer induced defence.
Littorina induced defences in the summer but not in the autumn experiment, while Haliotis did not induce defences in either season. Such a temporal variability in the ability to induce defences might reflect a seasonal pattern, although this conclusion has to be viewed cautiously as only one experiment could be conducted at each season. Seasonal patterns in the ability of macroalgae to induce defences have not been previously studied, although examples are known from terrestrial (Turlings et al. 1995; Doan et al. 2004; Van Zandt & Agrawal 2004a), freshwater (Stibor & Lampert 2000) and marine habitats (Raimondi et al. 2000). On a smaller temporal scale, diurnal induction patterns controlled by herbivore activities have been reported from insect-damaged plants (Turlings et al. 1995). Several seasonally variable factors may explain the observed temporal variability of inducible defences in Ecklonia. Littorina density is fourfold lower in autumn than in summer, while that of Haliotis is very low year round (Yang 2002). Coupled with lower water temperatures, grazing intensity of Littorina in the field is strongly, and that of Haliotis moderately, reduced in autumn relative to summer months. Our set-up exposed Ecklonia to identical grazer densities at both seasons and grazing intensity was presumably similar, as Littorina consumption rates did not differ significantly between summer and autumn feeding assays. The general lack of defence induction by Haliotis might reflect that response to the notoriously low grazing pressure by this species as compared with other herbivorous molluscs (e.g. Scheibling 1994) has minor selective value. The same grazing intensity of Littorina induced defences in summer but not in autumn, apparently contradicting models that suggest that a fitness-related threshold of grazing impact is required to induce defences (Karban et al. 1999) and suggesting that factors other than feeding pressure must be considered, at least in the Littorina–Ecklonia interaction. Water temperature at the study site decreases from 24.9 °C in August to 16.4 °C in November (unpublished data, J.H. Kim), very likely affecting metabolic rates in algae and grazers, as reported for Patella species (Jenkins et al. 2001). Such changes may affect defence response patterns (Cronin 2001). However, identical water temperatures were maintained during the summer and autumn experiment, eliminating this as a potential driver of season-related responses. Growth rates of Ecklonia pieces were higher in the autumn compared with the summer experiment, but this could be the result of a trade-off against lower investment in defence when herbivore pressure is reduced.
The influence of seasonal patterns of grazing pressure on grazer-specific induction of antiherbivore defence, together with rapid reduction of defences after grazing ceases, strongly suggests that Ecklonia might use inducible defence responses as a cost-saving mechanism. The occurrence of inducible defences in Ecklonia might reflect a finely adjusted algal response, adding new levels of complexity to plant–animal interactions, presumably with strong ecological implications for food web structure and community composition of herbivores.