A meta-analysis of 126 published studies shows that exposure to artificial light at night induced strong responses for physiological measures, daily activity patterns and lifehistory traits.

Artificial light has transformed the nighttime environment of large areas of the earth, with 88% of Europe and almost 50% of the United States experiencing light-polluted night skies [1]. The consequences for ecosystems range from exposure to high light intensities in the vicinity of direct light sources to the very widespread but lower lighting levels further away [2]. While it is known that species exhibit a range of physiological and behavioral responses to artificial nighttime lighting [e.g., 3-5], there is a need to gain a mechanistic understanding of whole ecological community impacts [6, 7], especially to different light intensities. Using a mesocosm field experiment with insect communities, we determined the impact of intensities of artificial light ranging from 0.1 to 100 lux on different trophic levels and interactions between species. Strikingly, we found the strongest impact at low levels of artificial lighting (0.1 to 5 lux), which led to a 1.8 times overall reduction in aphid densities. Mechanistically, artificial light at night increased the efficiency of parasitoid wasps in attacking aphids, with twice the parasitism rate under low light levels compared to unlit controls. However, at higher light levels, parasitoid wasps spent longer away from the aphid host plants, diminishing this increased efficiency. Therefore, aphids reached higher densities under increased light intensity as compared to low levels of lighting, where they were limited by higher parasitoid efficiency. Our study highlights the importance of different intensities of artificial light in driving the strength of species interactions and ecosystem functions.

SignificanceInitial species loss can lead to follow-on extinctions because of the interconnectedness of species in ecosystems. The loss of biodiversity through human impact can change the structure of ecological communities, which, models predict, can affect the likelihood of secondary extinctions. Uniquely, we provide an empirical demonstration of this with a plant-insect food web experiment by focusing on a particular secondary extinction mechanism. This mechanistic approach leads to specific, empirically testable hypotheses that can inform the development of predictive models on the consequences of changes in ecological community structure.

Animals often engage in mutualistic associations with microorganisms that protect them from predation, parasitism or pathogen infection. Studies of these interactions in insects have mostly focussed on the direct effects of symbiont infection on natural enemies without studying community‐wide effects. Here, we explore the effect of a defensive symbiont on population dynamics and species extinctions in an experimental community composed of three aphid species and their associated specialist parasitoids. We found that introducing a bacterial symbiont with a protective (but not a non‐protective) phenotype into one aphid species led to it being able to escape from its natural enemy and increase in density. This changed the relative density of the three aphid species which resulted in the extinction of the two other parasitoid species. Our results show that defensive symbionts can cause extinction cascades in experimental communities and so may play a significant role in the stability of consumer‐herbivore communities in the field.

Species extinction rates due to human activities are high, and initial extinctions can trigger cascades of secondary extinctions, leading to further erosion of biodiversity. A potential major mechanism for secondary extinction cascades is provided by the long-standing theory that the diversity of consumer species is maintained due to the positive indirect effects that these species have on each other by reducing competition among their respective resource species. This means that the loss of one carnivore species could lead to competitive exclusion at the prey trophic level, leading to extinctions of further carnivore species. Evidence for these effects is difficult to obtain due to many confounding factors in natural systems, but extinction cascades that could be due to this mechanism have been demonstrated in simplified laboratory microcosms. We established complex insect food webs in replicated field mesocosms and found that the overharvesting of one parasitoid wasp species caused increased extinction rates of other parasitoid species, compared to controls, but only when we manipulated the spatial distribution of herbivore species such that the potential for interspecific competition at this level was high. This provides clear evidence for horizontal extinction cascades at high trophic levels due to the proposed mechanism. Our results demonstrate that the loss of carnivores can have widespread effects on other species at the same trophic level due to indirect population-dynamic effects that are rarely considered in this context.