Although the effects of plant diversity on herbivores are contingent upon herbivore traits and the source of plant diversity (e.g. intra‐ and interspecific), most studies have analyzed these effects separately. We compared the effects of genotypic diversity of big‐leaf mahogany Swietenia macrophylla with that of tree species diversity on two specialist caterpillars (Hypsipyla grandella stem borers and Phyllocnistis meliacella leaf miners) and three generalist leafhoppers (Cicadellidae) feeding on mahogany in a large‐scale (7.2 ha) forest diversity experiment in southern Mexico. The experiment consisted of fifty‐nine 21 × 21‐m plots, with 64 tree saplings each (3‐m spacing between plants). Plots were either mahogany monocultures or species polycultures of four species (including mahogany) and – within each of these two plot types – mahogany was represented by either one or four genotypes. Throughout a five‐month period, beginning six months after planting, we measured mahogany growth and monitored herbivore and predator (spider) abundance. We found no effect of mahogany genotypic diversity on either specialist caterpillars or generalist leafhoppers, and this result was consistent across levels of tree species diversity. In contrast, species diversity had significant effects on both specialists but neither of the generalist herbivores. Specifically, species diversity lowered H. grandella attack at the middle of the sampling season, but increased attack at the end of the season, whereas P. meliacella abundance was consistently reduced. Such effects were not mediated by effects of species diversity on plant growth (of which there were none), but rather through resource heterogeneity. Diversity did not influence spider abundance. This study is one of few to directly compare sources of plant diversity, and uniquely compares such effects among herbivores with contrasting life histories (e.g. diet breadths). Overall, we demonstrate that plant species diversity effects outweigh those of genotypes, and our results suggest that such effects are stronger on specialist than generalist herbivores.

Identifying the mechanisms of compensation to insect herbivory remains a major challenge in plant biology and evolutionary ecology. Most previous studies have addressed plant compensatory responses to one or two levels of insect herbivory, and the underlying traits mediating such responses remain elusive in many cases.• We evaluated responses associated with compensation to multiple intensities of leaf damage (0% control, 10%, 25%, 50%, 75% of leaf area removed) by means of mechanical removal of foliar tissue and application of a caterpillar (Spodoptera exigua) oral secretions in 3-month-old wild cotton plants (Gossypium hirsutum). Four weeks posttreatment, we measured plant growth and multiple traits associated with compensation, namely: changes in above-and belowground, biomass and the concentration of nutrients (nitrogen and phosphorus) and non-structural carbon reserves (starch and soluble sugars) in roots, stems and leaves.• We found that wild cotton fully compensated in terms of growth and biomass allocation when leaf damage was low (10%), whereas moderate (25%) to high leaf damage in some cases led to under-compensation. Nonetheless, high levels of leaf removal (50% and 75%) in most cases did not cause further reductions in height and allocation to leaf and stem biomass relative to low and moderate damage. There were significant positive effects of leaf damage on P concentration in leaves and stems, but not roots, as well as a negative effect on soluble sugars in roots.• These results indicate that wild cotton fully compensated for a low level of leaf damage but under-compensated under moderate to high leaf damage, but can nonetheless sustain growth despite increasing losses to herbivory. Such responses were possibly mediated by a re-allocation of carbohydrate reserves from roots to shoots.Plant Biology 21 (2019) 805-812

The induction of defences in response to herbivory is a key mechanism of plant resistance. While a number of studies have investigated the time course and magnitude of plant induction in response to a single event of herbivory, few have looked at the effects of recurrent herbivory. Furthermore, studies measuring the effects of the total amount and recurrence of herbivory on both direct and indirect plant defences are lacking. To address this gap, here we asked whether insect leaf herbivory induced changes in the amount and concentration of extrafloral nectar (an indirect defence) and concentration of leaf phenolic compounds (a direct defence) in wild cotton (Gossypium hirsutum).• We conducted a greenhouse experiment where we tested single event or recurrent herbivory effects on defence induction by applying mechanical leaf damage and caterpillar (Spodoptera frugiperda) regurgitant.• Single events of 25% and 50% leaf damage did not significantly influence extrafloral nectar production or concentration. Extrafloral nectar traits did, however, increase significantly relative to controls when plants were exposed to recurrent herbivory (two episodes of 25% damage). In contrast, phenolic compounds increased significantly in response to single events of leaf damage but not to recurrent damage. In addition, we found. that local induction of extrafloral nectar production was stronger than systemic induction, whereas the reverse pattern was observed for phenolics.• Together, these results reveal seemingly inverse patterns of induction of direct and indirect defences in response to herbivory in wild cotton.

No abstract

Previous studies have frequently reported effects of abiotic factors on herbivore‐induced plant defences based on effects on single plant traits. However, plants commonly express multiple defences simultaneously and these traits are often correlated. Thus, a fuller understanding of abiotic‐context dependency in plant defence requires measuring multiple traits and addressing their patterns of correlated expression. We evaluated the effects of soil salinity on the expression of direct (phenolic compounds, gossypol gland density) and indirect (volatile organic compounds, extrafloral nectar) defensive traits in wild cotton Gossypium hirsutum. Specifically, we asked whether soil salinity affects the induction of these traits, and whether it shapes trait correlations potentially underlying altered patterns of trait induction. We conducted a factorial experiment with 16 cotton genotypes where we manipulated soil salinity and defence induction by applying artificial leaf damage (25% mechanical damage and caterpillar oral secretions) and measured defence levels at different time points post damage. Leaf damage induced most traits except gossypol gland density, whereas salinity did not have a mean effect (across constitutive and induced levels) on any of the measured traits. Nonetheless, salinity prevented the induction of phenolic compounds (condensed and hydrolysable tannins), and also affected trait correlations. Specifically, phenolic compounds were negatively associated with nectar production only under salinized conditions, an apparent trade‐off that could affect the induction of phenolic compounds. In addition, positive correlations between phenolic compounds and gland density and root biomass observed under control conditions were lost under salinized conditions. Synthesis. By investigating the effects of soil salinity on the expression of multiple direct and indirect defensive traits and their underlying correlations, these findings build towards a better understanding of how abiotic context dependency shapes plant allocation to and expression of multiple defensive traits.