University of Tennessee, Knoxville, Tennessee, United States
Plants often produce structural and chemical defenses in response to herbivory, which can increase their resistance to herbivory in the short term. However, inducible defenses are typically energetically or ecologically costly and can change the functional or numerical response of herbivores – and therefore affect long-term dynamics of plant and herbivore populations. Theory suggests that defense onset and decay rates and the relationship between consumer density and defense strength are key in determining the consequences of inducible defenses for consumer-resource dynamics. Importantly, if multiple traits are used in defense, then differences in their timing, cost of production, longevity, and efficacy against herbivores could confer sustained resistance. However, we have few empirical examples of how herbivore induced defenses change through time across multiple defense traits, or how herbivore density affects defense expression at the level of genetically variable plant populations. Our goal is to investigate the effect of herbivore density on the timing of expression of multiple defense traits and how these traits contribute to resistance across time.We established 70 experimental populations of upland cotton, Gossypium hirsutum, in a field common garden. Each population consisted of ten plants from a pool of 45 wild genotypes enclosed in a mesh cage. We manipulated the density (none, 6 (low), 12 (medium), 25 (high) per plant) of Spodoptera exigua caterpillars in each population, and caterpillars moved freely among plants within the cage. At five timepoints following caterpillar introduction, we measured resistance against subsequent herbivory via bioassays with naïve caterpillars, leaf structural defenses (i.e., trichomes, terpenoid glands), and chemical defenses via HPLC. From 4-12 days following caterpillar introduction, mean induced resistance increased with herbivore density but did not vary through time. Compared with the control treatment, plants with high herbivore density reduced the relative growth rate of naïve caterpillars to below zero and leaf area consumed by 83%. After 21 days (coincident with caterpillar pupation), resistance returned to control levels. Structural leaf traits varied in their timing of peak induction: trichome densities were greatest (2.0x minimum level) at four days, and leaf terpenoid gland densities were greatest (1.3x minimum level) at seven days.Offsets in the timing of peak induction among defensive traits may be critical in the maintenance of high resistance through time (i.e., across herbivore development) while minimizing production of costly defensive traits. Moderate induction rates and herbivore density-dependent defense strength may jointly stabilize plant and herbivore population dynamics.
