Multiple studies have shown that predacious arthropods attracted by extrafloral nectaries (EFNs) attack and remove herbivores, benefitting the plant. In interior Alaska, EFN frequency in quaking aspen (Populus tremuloides Michx.) is highly variable within and among ramets and clones. In aspen, EFNs vary in number from 0 to 6 per leaf with 0 or 2 EFNs most common. Past studies have shown that leaves with EFNs sustain less damage by a major herbivore, the aspen leaf miner (Phyllocnistis populiella Chambers; hereafter ALM), than those lacking EFNs, despite receiving similar egg loads. Similar reductions in ALM damage correlate with increasing shoot EFN frequency, although it is unknown if reductions are the result of additive leaf level or non-additive shoot level effects. Using experimental and observational approaches, we investigated a) whether the leaf- and shoot-level benefits of EFNs extend to ramets and clones with high EFN frequency, b) how patterns of EFN expression influence predator abundance and predation rates, and c) how these relationships influence aspen growth rates. Experiments conducted at nine sites in interior Alaska excluded crawling arthropods, including common ALM predators, from ramets. Observational data was collected from 13 sites. Sites varied in EFN frequency, predator, and herbivore abundance.

Results/Conclusions

Across sites, exclusion of crawling arthropods led to a significant increase in mining damage relative to controls at the leaf, shoot, and ramet levels, and the magnitude of this effect was positively correlated with EFN frequency. Damage on leaves without EFNs was inversely related to shoot-level EFN frequency on control but not exclusion ramets, suggesting that beneficial effects of predators “spill over” to leaves lacking EFNs. Leaves with EFNs had lower ALM damage than leaves without EFNs across treatments, suggesting that EFNs may correlate with other plant defenses. Across ramets and sites, mining damage decreased with increasing EFN frequency. Experimental and EFN-mediated reductions in mining were chiefly explained by differential predation on ALM eggs. Egg predation was significantly greater on control than exclusion ramets and when predators were present, EFN frequency correlated positively with egg predation rates at leaf, shoot, ramet, and site levels. Site-level population estimates of predacious mites and ants correlated negatively with average ALM damage. While experimental exclusion did not affect ramet growth rates, natural variation in ramet EFN frequency was positively correlated with growth.