Complex consequences of herbivory and interplant cues in three annual plants

Abstract

Information exchange (or signaling) between plants following herbivore damage has recently been shown to affect plant responses to herbivory in relatively simple natural systems. In a large, manipulative field study using three annual plant species (Achyrachaena mollis, Lupinus nanus, and Sinapis arvensis), we tested whether experimental damage to a neighboring conspecific affected a plant's lifetime fitness and interactions with herbivores. By manipulating relatedness between plants, we assessed whether genetic relatedness of neighboring individuals influenced the outcome of having a damaged neighbor. Additionally, in laboratory feeding assays, we assessed whether damage to a neighboring plant specifically affected palatability to a generalist herbivore and, for S. arvensis, a specialist herbivore. Our study suggested a high level of contingency in the outcomes of plant signaling. For example, in the field, damaging a neighbor resulted in greater herbivory to A. mollis, but only when the damaged neighbor was a close relative. Similarly, in laboratory trials, the palatability of S. arvensis to a generalist herbivore increased after the plant was exposed to a damaged neighbor, while palatability to a specialist herbivore decreased. Across all species, damage to a neighbor resulted in decreased lifetime fitness, but only if neighbors were closely related. These results suggest that the outcomes of plant signaling within multi-species neighborhoods may be far more context-specific than has been previously shown. In particular, our study shows that herbivore interactions and signaling between plants are contingent on the genetic relationship between neighboring plants. Many factors affect the outcomes of plant signaling, and studies that clarify these factors will be necessary in order to assess the role of plant information exchange about herbivory in natural systems.

Figure 1. The effect of neighbor wounding and relatedness (maternal siblings or unrelated) on subsequent herbivore damage (log-transformed percent leaf damage) to conspecific neighbors of three experimental plant species in the field.

Least Square Mean +/− SE.

Figure 2. The effect of neighbor wounding and relatedness (maternal siblings or unrelated) on lifetime seed production (log-transformed) of three experimental plant species in the field.

Least Square Mean +/− SE.

Figure 3. The effect of neighbor wounding and relatedness (maternal siblings or unrelated) on developmental phenology of conspecific neighbors of three experimental plant species in the field.

Filled bars are the number of bolting plants, and empty bars are the number of flowering plants in each treatment two weeks after the wounding of neighboring plants (May 9, 2011). The total number of plants in each treatment is shown on each bar.

Figure 4. The effect of neighbor wounding and relatedness (maternal siblings or unrelated) on the percent leaf area (log-transformed) of conspecific neighbors of three potted experimental plant species consumed by generalist Spodoptera caterpillars (as an indicator of leaf palatability to generalists).

The potted plant was exposed to a damaged or undamaged neighbor in the field for 2 days. At this point the plant was moved indoors, and a feeding trial with a neonate Spodoptera caterpillar was initiated. Least Square Mean +/− SE.

Figure 5. For the crucifer Sinapis arvensis, the effect of neighbor-wounding and relatedness (maternal siblings or unrelated) on various measures of leaf palatability to a specialist herbivore (Pieris).

The potted plant was exposed to a damaged or undamaged neighbor in the field for 2 days. At this point the plant was moved indoors, and a no-choice feeding trial with a neonate Pieris caterpillar was initiated. Graphs show a) the percent leaf area removed by Pieris (log-transformed), and b) Pieris caterpillar mass (log-transformed) at the end of the feeding trial. Bars indicate least square mean +/− SE.

Figure 6. A summary of experimental results.

For each response variable, the bar represents the log response ratio where the conspecific neighbor (emitter) was experimentally wounded (W) versus unwounded (UW). A positive bar indicates that the response variable of the receiver plant was greater when the emitting plant was wounded than when it was unwounded; a negative bar indicates the converse. The effect of neighbor wounding is shown both related and unrelated pairs of each of the three plant species in this study. For example, in related pairs of A. mollis, wounding to a neighbor increased the leaf removal that a focal plant suffered compared to focal plants with unwounded neighbors. Palatability to a specialist herbivore (Pieris) could only be assessed for one plant species (S. arvensis) and was not tested (NT) for other plant species.

Figure 7. A conceptual diagram of potential ecological interactions.

a) Interactions predicted by plant signaling theory, and b) a more generalized scheme showing possible outcomes of damage to neighboring plants in a complex community setting.