Abstract
BACKGROUND: By sensing environmental cues indicative of pathogens or herbivores, plants can "prime" appropriate defenses and deploy faster, stronger responses to subsequent attack. Such priming presumably entails costs-else the primed state should be constitutively expressed-yet those costs remain poorly documented, in part due to a lack of studies conducted under realistic ecological conditions. We explored how defence priming in goldenrod (Solidago altissima) influenced growth and reproduction under semi-natural field conditions by manipulating exposure to priming cues (volatile emissions of a specialist herbivore, Eurosta solidaginis), competition between neighbouring plants, and herbivory (via insecticide application). RESULTS: We found that primed plants grew faster than unprimed plants, but produced fewer rhizomes, suggesting reduced capacity for clonal reproduction. Unexpectedly, this effect was apparent only in the absence of insecticide, prompting a follow-up experiment that revealed direct effects of the pesticide esfenvalerate on plant growth (contrary to previous reports from goldenrod). Meanwhile, even in the absence of pesticide, priming had little effect on herbivore damage levels, likely because herbivores susceptible to the primed defences were rare or absent due to seasonality. CONCLUSIONS: Reduced clonal reproduction in primed plants suggest that priming can entail significant costs for plants. These costs, however, may only become apparent when priming cues fail to provide accurate information about prevailing threats, as was the case in this study. Additionally, our insecticide data indicate that pesticides or their carrier compounds can subtly, but significantly, affect plant physiology and may interact with plant defences.