Testing the contribution of vertebrate predators and leaf traits to mainland–island differences in insect herbivory on oaks

1 INTRODUCTION

Islands provide valuable settings to study the ecological and evolutionary drivers of biodiversity (Gillespie et al., 2008; MacArthur & Wilson, 2001; Ricklefs & Bermingham, 2007), species interactions (Spiller & Schoener, 1990; Traveset et al., 2013), and trait evolution and speciation (Barrett et al., 1997; Burns, 2019; Carvajal-Endara et al., 2020; Grant & Grant, 2007). To address insularity effects, studies have either compared islands differing in historical and physical features (e.g. island size, isolation, geological age), or insular systems with their closest mainland counterparts (Moreira & Abdala-Roberts, 2022). Within this body of research, work on plant–herbivore interactions poses that herbivory should be weaker on islands than on mainland as a result of lower herbivore abundance and diversity owing to processes such as species dispersal constraints and environmental filtering (Carlquist, 1974; Losos & Ricklefs, 2009; Ricklefs & Bermingham, 2007). This hypothesis has primarily been formulated and tested for mammalian herbivory, given the general absence of mammals in most insular systems (Burns, 2014; Cubas et al., 2019; Salladay & Ramirez, 2018; Vourc'h et al., 2001). Alternatively, insect herbivory could be potentially higher in islands than on mainland due to reduced predation pressure (Schoener et al., 2016; Terborgh, 2010), which may lead to overconsumption by phytophagous insects. In this regard, a recent meta-analysis found no overall significant differences in herbivory by invertebrates (i.e. insects and molluscs) between islands and mainland (Moreira et al., 2021). However, these results were based on a limited number of studies (only three on insects), underscoring the need for additional research on insect herbivory on islands and calling for further work to reassess predictions and test underlying mechanisms.

A critical gap in our understanding of insularity effects on insect herbivory has been the lack of experimental studies on top-down effects by natural enemies, that is predators and parasitoids (Abdala-Roberts et al., 2019; Hairston et al., 1960; Price et al., 1980). There are good examples of such tests involving inter-island comparisons (Henneman & Memmott, 2001; Holt, 2010; Kolbe et al., 2023; Spiller & Schoener, 1990), but mainland vs. island comparisons of natural enemy effects are virtually absent (Moreira & Abdala-Roberts, 2022). Analogous to studies comparing islands with different features, some authors have argued that top-down effects of predators should be weaker on islands than on mainland (Holt, 2010; Schoener & Spiller, 2010), presumably because islands have lower predator abundance and diversity or even entirely lack higher trophic levels (Holt, 2010; Terborgh, 2010). However, some vertebrate predators (e.g. birds, bats, lizards) known to exert strong top-down control of insects (Bael et al., 2008; Maas et al., 2016; Mooney et al., 2010; Van Bael et al., 2003; Whelan et al., 2008) can exhibit high population densities on islands partly due to the lack of top predators (Jones et al., 2009; Presley & Willig, 2022; Terborgh, 2023), thereby potentially intensifying the top-down regulation of insect herbivory. Yet, manipulative field studies testing for island–mainland differences in predation are needed to formally test whether top-down trophic forcing contributes to island–mainland differences in herbivory.

Herbivory patterns can also be influenced by bottom-up control through plant physical and chemical defensive traits (Agrawal, 2011; Carmona et al., 2011; Marquis, 1992; Rhoades, 1979), but studies analysing plant defences on islands are underrepresented (Moreira et al., 2021). Island–mainland variation in abiotic conditions (e.g. climate or soil) can shape differences in plant traits (e.g. effects of abiotic stress or resource availability) predictive of herbivory, including chemical defences or nutritional traits. For example, islands tend to have wetter and less seasonal climates than their mainland counterparts (Weigelt et al., 2013), that is favourable growing conditions that may result in higher growth and nutrient content at the expense of defence allocation (Coley et al., 1985). At the same time, islands may be characteristic of soil types with low productivity (e.g. ultramafic soils; Pillon et al., 2010), which could lead to predictions in the opposite direction, that is lower plant growth and higher defences leading to lower herbivory. Furthermore, the extent and nature of climate- and soil-related variability is often strongly site-specific, thus cautioning against generalizing about the direction of abiotically mediated island–mainland differences in plant traits. Although limited by a low number of studies, a recent meta-analysis was unsupportive of this abiotic forcing paradigm, as there was no overall island–mainland difference in chemical or physical plant traits putatively associated with herbivory (Moreira et al., 2021). Nonetheless, more island–mainland comparisons jointly measuring herbivory, multiple plant traits (e.g. Moreira et al., 2019) and abiotic correlates of plant–herbivore interactions are needed to robustly test these mechanisms of insularity effects on herbivory.

In this study, we explored island–mainland differences in insect leaf herbivory using 52 populations of 12 oak (Quercus) species distributed across three distant island–mainland sites, namely: The Channel Islands of California vs. mainland California, the Balearic Islands vs. mainland Spain, and the island Bornholm vs. mainland Sweden. Additionally, we tested for predator- and plant-driven island–mainland differences in herbivory. To this end, we conducted a field experiment in which we excluded vertebrate predators (birds, bats) and measured herbivory and leaf traits putatively associated with herbivory, namely: physical traits (measured as specific leaf area, a proxy of leaf thickness), secondary metabolites (phenolic compounds) and nutritional traits (nitrogen and phosphorus content). Additionally, we characterized abiotic conditions (climate and soil characteristics) for each population by using data from global datasets to explain differences in leaf traits associated with herbivory. Specifically, we asked: (i) Do levels of leaf herbivory differ between islands and mainland? (ii) Do vertebrate predators affect herbivory and does such effect contribute to explaining islands vs. mainland differences in herbivory (i.e. a top-down mechanism)? (iii) Does the expression of leaf traits differ between islands and mainland and is any such difference related to variation in abiotic conditions (i.e. climate and soil)? and (iv) do island–mainland differences in leaf traits contribute to explaining herbivory variation (i.e. a bottom-up mechanism)? We hypothesized that oak trees on islands would have lower herbivory than mainland trees. In addition, predictions on top-down and bottom-up effects can act in different ways. On the one hand, stronger herbivore predation by vertebrates on islands would be consistent with, and contribute to explaining, lower herbivory on islands. On the other hand, weaker predation on islands would favour higher herbivory and therefore counter other forces driving lower herbivory on islands, thus leading to weaker island–mainland differences, no overall difference, or even greater herbivory on islands. Lastly, a similar rationale can be applied for bottom-up effects. Less favourable climatic or soil conditions on islands would result in lower plant quality (i.e. lower nutrient content and higher defence) for herbivores and contribute to explaining lower herbivory on islands. More favourable conditions, however, would lead to higher plant quality on islands, which would promote herbivory and therefore counter other factors driving lower herbivory in islands. Overall, by jointly addressing predator effects and plant traits across different study sites, this study provides a novel test of bottom-up and top-down drivers of plant–herbivore interactions, shedding light on the mechanisms underlying differences in herbivory between islands and mainland.