Ecosystem engineers, organisms that modify the physical environment, are generally thought to increase diversity by facilitating species that benefit from engineered habitats. Recent theoretical work, however, suggests that ecosystem engineering could initiate cascades of trophic interactions that shape community structure in unexpected ways, potentially having negative indirect effects on abundance and diversity in components of the community that do not directly interact with the habitat modifications. We tested the indirect effects of a gall‐forming wasp on arthropod communities in surrounding unmodified foliage. We experimentally removed all senesced galls from entire trees during winter and sampled the arthropod community on foliage after budburst. Gall removal resulted in 59% greater herbivore density, 26% greater herbivore richness, and 27% greater arthropod density five weeks after budburst. Gall removal also reduced the differences in community composition among trees (i.e., reduced beta diversity), even when accounting for differences in richness. The community inside galls during winter and through the growing season was dominated by jumping spiders (Salticidae; 0.87 ± 0.12 spiders per gall). We suggest that senesced galls provided habitat for spiders, which suppressed herbivorous arthropods and increased beta diversity by facilitating assembly of unusual arthropod communities. Our results demonstrate that the effects of habitat modification by ecosystem engineers can extend beyond merely providing habitat for specialists; the effects can propagate far enough to influence the structure of communities that do not directly interact with habitat modifications.

The red-eared slider turtle (Trachemys scripta elegans; RES) is often considered one of the world’s most invasive species. Results from laboratory and mesocosm experiments suggest that introduced RES outcompete native turtles for key ecological resources, but such experiments can overestimate the strength of competition. We report on the first field experiment with a wild turtle community, involving introduced RES and a declining native species of conservation concern, the western pond turtle (Emys marmorata; WPT). Using a before/after experimental design, we show that after removing most of an introduced RES population, the remaining RES dramatically shifted their spatial basking distribution in a manner consistent with strong intraspecific competition. WPT also altered their spatial basking distribution after the RES removal, but in ways inconsistent with strong interspecific competition. However, we documented reduced levels of WPT basking post-removal, which may reflect a behavioral shift attributable to the lower density of the turtle community. WPT body condition also increased after we removed RES, consistent with either indirect or direct competition between WPT and RES and providing the first evidence that RES can compete with a native turtle in the wild. We conclude that the negative impacts on WPT basking by RES in natural contexts are more limited than suggested by experiments with captive turtles, although wild WPT do appear to compete for food with introduced RES. Our results highlight the importance of manipulative field experiments when studying biological invasions, and the potential value of RES removal as a management strategy for WPT.

1. For arboreal lizards, the ability to cling or adhere to the substrate is critical for locomotion during prey capture, predator escape, thermoregulation and social interactions. Thus, selection on traits related to clinging is likely strong.2. Correlations between morphology, performance and habitat use have been documented in arboreal lizards, providing a framework for using functional traits to predict habitat use in the field.3. We tested the hypothesis that clinging performance predicts habitat use in an actively assembling community of introduced lizards in Hawaiʻi comprised of anoles (Anolis carolinensis, A. sagrei) and day geckos (Phelsuma laticauda).4. We measured morphological traits (toepad area and lamellae number) and tested clinging performance on two artificial and eight natural substrates in the laboratory. We measured habitat use in 10 m × 10 m outdoor enclosures where habitat availability was controlled and the lizard species assemblage was manipulated to reflect all species combinations. The enclosure experiment generated more than 9,000 habitat use observations from 360 lizards. 5. Morphological traits that predict performance in Anolis were not predictive in Phelsuma, indicating that direct measures of performance are necessary for comparisons between the genera. Measuring clinging performance on multiple substrates provided key insights intopatterns of habitat use. While all three species performed best on an artificial smooth substrate (acrylic), performance on natural substrates predicted which texture (rough vs. smooth) was most often used by each species. 7. Performance predicted perch height use: species with the greatest clinging performance (A. carolinensis and P. laticauda) across substrates perched twice as high as A. sagrei. 8. We did not observe habitat shifts in the height or texture of perches used by any species in response to experimental manipulation of the lizard species assemblage. 9. Our results highlight the inextricable link between ecology, morphology and performance, the importance of measuring functional traits in ecologically relevant ways, and the potential for resource partitioning to be influenced by differences in the ability to attach to different substrates.

Trophic morphology affects resource acquisition; therefore, species differences in such traits may be informative for inferring resource use overlap and potential species interactions. In lizards, head size and shape determine the size and hardness of prey that can be consumed. Lizards with large differences in head morphology are expected to overlap less in prey use than lizards with more similar traits. Stable isotopes are increasingly being used to describe diet, yet how traditional functional traits affect isotopic diet is often not clear a priori. We measured head size, head shape, 𝛿15N, and 𝛿13C under controlled resource availability in an enclosure experiment using introduced lizards in Hawaiʻi to test whether functional traits predict isotopic diet. Brown anoles Anolis sagrei had the tallest and narrowest heads, the highest values of 𝛿13C, and the lowest values of 𝛿15N. Gold dust day geckos Phelsuma laticauda had the shortest and widest heads, the lowest values of 𝛿13C, and the highest values of 𝛿15N. Green anoles Anolis carolinensis were intermediate in both diet and morphology. As a result of isotopic diet overlap, green anoles have reduced competitor‐free resource space in the presence of both of the other lizard species. Head shape was the best predictor of diet and the only trait that explained variation within as well as among species. Head size was sexually dimorphic, and therefore the weaker diet correlations with this trait may be explained by sexual selection. Breadth in morphospace did not correlate with isotopic diet breadth, nor did the amount of overlap in morphospace predict the amount of overlap in isotopic diet space. While lizards were able to locally depress prey in experimental enclosures, no shifts in diet were detected in response to the presence of heterospecifics. The generality of head shape in predicting isotopic diet, and whether it does so independent of habitat use, warrants additional study. Head shape provides a potentially fruitful avenue for trait‐based approaches to studying ecology and evolution in lizards. Read the free Plain Language Summary for this article on the Journal blog.