Human activity is rapidly increasing the radiance and geographic extent of artificial light at night (ALAN) leading to alterations in the development, behavior, and physiological state of many organisms. A limited number of community-scale studies investigating the effects of ALAN have allowed for spatial aggregation through positive phototaxis, the commonly observed phenomenon of arthropod movement toward light. We performed an open field study (without restricted arthropod access) to determine the effects of ALAN on local arthropod community composition, plant traits, and local herbivory and predation rates. We found strong positive phototaxis in 10 orders of arthropods, with increased (159% higher) overall arthropod abundance under ALAN compared to unlit controls. The arthropod community under ALAN was more diverse and contained a higher proportion of predaceous arthropods (15% vs 8%). Predation of immobilized flies occurred 3.6 times faster under ALAN; this effect was not observed during the day. Contrary to expectations, we also observed a 6% increase in herbivory under ALAN. Our results highlight the importance of open experimental field studies in determining community-level effects of ALAN.

Plant tolerance of herbivory, i.e., the ability to recover after damage, is an important component of how plants cope with herbivores. Tolerance has long been hypothesized to be constrained evolutionarily by plant resistance to herbivores, traits that allow plants to cope with stressful growing conditions, and traits that influence the timing of damage in relation to reproduction. Variation in tolerance and resistance can be caused by differences in the identity of the plant (e.g., genotype, species, clade) and by the context of the herbivore threat (e.g., identity of the herbivore, type of damage it causes, abiotic conditions in which the plant is growing). To date, the vast majority of studies have explored trade-offs with tolerance within species. Here, we test hypotheses of constraints on tolerance using comparative approaches in a clade of mustards, emphasizing the variety of contexts in which damage is realistically tolerated. We estimated tolerance to leaf damage, tolerance to apical clipping at the bolting stage - simulating browsing -, and resistance to a specialist and generalist lepidopteran herbivore for a group of native mustards, grown in field soils unique to each population and in a common potting soil. Resistance to herbivores was soil dependent, while surprisingly, tolerance was not. Phylogenetic signal in resistance to specialist and generalist lepidopteran herbivores was present, but only when plants were grown in field soils. Tolerance had low phylogenetic signal. Tolerance to leaf damage was unrelated to tolerance to simulated browse. We found no evidence for a resistance-tolerance trade-off, and some evidence for a soil-dependent positive correlation between tolerance and resistance to both herbivores. Drought-tolerant species had poorer ability to tolerate browse damage, and earlier flowering species tended to be less tolerant to leaf damage. Our results suggest that tolerance trades off with traits that allow mostly annual, monocarpic Streptanthus (s.l.) to persist in drought-prone conditions but is largely unrelated to resistance to herbivores. Our study highlights a need for a new framework for tolerance to herbivory that explicitly acknowledges that the relationship among tolerance, resistance, and traits that ameliorate abiotic stress.