Exotic plants accumulate and share herbivores yet dominate communities via rapid growth

Herbivores may facilitate or impede exotic plant invasion, depending on their direct and indirect interactions with exotic plants relative to co-occurring natives. However, previous studies investigating direct effects have mostly used pairwise native-exotic comparisons with few enemies, reached conflicting conclusions, and largely overlooked indirect interactions such as apparent competition. Here, we ask whether native and exotic plants differ in their interactions with invertebrate herbivores. We manipulate and measure plant-herbivore and plant-soil biota interactions in 160 experimental mesocosm communities to test several invasion hypotheses. We find that compared with natives, exotic plants support higher herbivore diversity and biomass, and experience larger proportional biomass reductions from herbivory, regardless of whether specialist soil biota are present. Yet, exotics consistently dominate community biomass, likely due to their fast growth rates rather than strong potential to exert apparent competition on neighbors. We conclude that polyphagous invertebrate herbivores are unlikely to play significant direct or indirect roles in mediating plant invasions, especially for fast-growing exotic plants.


Supplementary Tables
Supplementary Table 1

Herbivore cage design
Herbivore cages (Supplementary Figure 13)  proportion of adult weevils are likely to be parasitized by a biological control agent (Microctonus aethiopoides, Hymenoptera: Braconidae), they were maintained for three weeks to purge parasitized individuals, before four weevils (two of each sex, determined by examining the shape of the posterior ventrite following Bright (1994) 36 ) were added to the center of each mesocosm. Because direct feeding of adults was difficult to observe, their presence was determined by characteristic match-head size notches on clover leaf margins.
Moreover, because clover root weevil and lucerne weevil larvae feed belowground, sampling could only be conducted during the final mesocosm harvest by thoroughly searching through plant roots and the mesocosm soil as it was homogenized. Clover root weevil larvae were encountered in only a single mesocosm (host confirmed as red clover, Trifolium pratense, using RFLP analysis) and no lucerne weevil larvae were found. These two weevils are oligophagous, with narrow host plant ranges. Clover root weevil has been recorded feeding Greasy cutworm, Agrotis ipsilon (Hufnagel): Greasy cutworm is a cosmopolitan caterpillar that self-colonized five mesocosms in low abundance. It was removed from -Herbivore mesocosms when encountered and left when found in +Herbivore mesocosms. Based on the Plant-SyNZ Database, greasy cutworms have been recorded feeding on 15 host plant species in New Zealand, and were recorded on 3 plant species in the mesocosms.
Common forest looper, Pseudocoremia suavis Butler: Common forest looper is a native moth species that self-colonized five mesocosms. It was removed from -Herbivore mesocosms when encountered and left when found in +Herbivore mesocosms. Based on the Plant-SyNZ Database, common forest looper have been recorded feeding on 69 host plant species in New Zealand, and were recorded on 10 plant species in the mesocosms.

Leafhopper, Anzygina zealandica (Myers):
Anzygina zealandica is a common native leafhopper that self-colonized several mesocosms in high abundance and was subsequently identified using Knight (1976) 38  Pea aphid, Acyrthosiphon pisum Harris: Pea aphids self-colonized seven mesocosms where they fed on three exotic legume species. This herbivore species was removed from -Herbivore mesocosms when encountered and left when found in +Herbivore mesocosms. Plants were surveyed by systematically examining plant tissue for aphid colonies. If less than 500 aphids were found (as was always the case for this species), they were counted as accurately as possible. Based on the Plant-SyNZ Database, pea aphids have been recorded feeding on 11 host plant species in New Zealand, and were recorded on 3 plant species in the mesocosms.
Cherry-oat aphid, Rhopalosiphum padi (L.): Cherry-oat aphids self-colonized several mesocosms, with a small number of plants experiencing severe outbreaks. Thus, we located a source population on Yorkshire fog-grass (Holcus lanatus) adjacent to the experimental site, which was used to infest previously uncolonized +Herbivore mesocosms with five alates (winged adults) on 27 October 2017.
This herbivore species was also actively removed from -Herbivore mesocosms when 78 encountered. Plants were surveyed by systematically examining plant tissue for aphid colonies. If less than 500 aphids were found, these were counted as accurately as possible.  Polymerase chain reaction (PCR) protocol:

Trait-based analyses
To investigate potential mechanisms underlying invertebrate herbivore presence, diversity, biomass, and damage to plants, we explored whether variation in these response  43 . Insect reproduction status represents a trait outcome that allowed us to distinguish between aggregative vs. population responses of herbivores. We predicted that plant-herbivore interactions would be strongest on nitrogen-fixing and mycorrhizal plants, grasses and forbs, plants with high SLA and biomass, and for herbivore species that reproduced in the mescososms. To test these predictions, we modelled each response variable (transformations and model error distributions were the same as in Supplementary Table 26) as a function of the trait main effects and their pairwise interactions with plant provenance, and with the random effects of plant species, herbivore species (for models of herbivore presence and biomass only), and mesocosm nested within plant community.
We found that the probability of an invertebrate herbivore species occurring on a given plant was 2.3 times higher on plants that do not fix nitrogen than on nitrogen-fixers (F = 15.29, P = 86 0.002; Supplementary Table 19), and increased with total plant biomass (F = 24.95, P = 0.00003; Supplementary Table 19). Moreover, the probability of a herbivore species occurring on a given plant also depended on the interaction between plant provenance and insect reproduction status in the mesocosms (F = 6.00, P = 0.015; Supplementary Fig. 6, Supplementary