A long‐standing paradigm in ecology holds that herbivore pressure and thus plant defences increase towards lower latitudes. However, recent work has challenged this prediction where studies have found no relationship or opposite trends where herbivory or plant defences increase at higher latitudes. Here we tested for latitudinal variation in herbivory, chemical defences (phenolic compounds), and nutritional traits (phosphorus and nitrogen) in leaves of a long‐lived tree species, the English oak Quercus robur. We further investigated the underlying climatic and soil factors associated with such variation. Across 38 populations of Q. robur distributed along an 18° latitudinal gradient, covering almost the entire latitudinal and climatic range of this species, we observed strong but divergent latitudinal gradients in leaf herbivory and leaf chemical defences and nutrients. As expected, there was a negative relationship between latitude and leaf herbivory where oak populations from lower latitudes exhibited higher levels of leaf herbivory. However, counter to predictions there was a positive relationship between leaf chemical defences and latitude where populations at higher latitudes were better defended. Similarly, leaf phosphorus and nitrogen increased with latitude. Path analysis indicated a significant (negative) effect of plant chemical defences (condensed tannins) on leaf herbivory, suggesting that the latitudinal gradient in leaf herbivory was driven by an inverse gradient in defensive investment. Leaf nutrients had no independent influence on herbivory. Further, we found significant indirect effects of precipitation and soil porosity on leaf herbivory, which were mediated by plant chemical defences. These findings suggest that abiotic factors shape latitudinal variation in plant defences and that these defences in turn underlie latitudinal variation in leaf herbivory. Overall, this study contributes to a better understanding of latitudinal variation in plant–herbivore interactions by determining the identity and modus operandi of abiotic factors concurrently shaping plant defences and herbivory.

Systematic comparisons of species interactions in urban versus rural environments can improve our understanding of shifts in ecological processes due to urbanization. However, such studies are relatively uncommon and the mechanisms driving urbanization effects on species interactions (e.g. between plants and insect herbivores) remain elusive. Here we investigated the effects of urbanization on leaf herbivory by insect chewers and miners associated with the English oak Quercus robur by sampling trees in rural and urban areas throughout most of the latitudinal distribution of this species. In performing these comparisons, we also controlled for the size of the urban areas (18 cities) and gathered data on CO2 emissions. In addition, we assessed whether urbanization affected leaf chemical defences (phenolic compounds) and nutritional traits (phosphorus and nitrogen), and whether such changes correlated with herbivory levels. Urbanization significantly reduced leaf chewer damage but did not affect leaf miners. In addition, we found that leaves from urban locations had lower levels of chemical defences (condensed and hydrolysable tannins) and higher levels of nutrients (nitrogen and phosphorus) compared to leaves in rural locations. The magnitude of urbanization effects on herbivory and leaf defences was not contingent upon city size. Importantly, while the effects of urbanization on chemical defences were associated with CO2 emissions, changes in leaf chewer damage were not associated with either leaf traits or CO2 levels. These results suggest that effects of urbanization on herbivory occur through mechanisms other than changes in the plant traits measured here. Overall, our simultaneous assessment of insect herbivory, plant traits and abiotic correlates advances our understanding of the main drivers of urbanization effects on plant–herbivore interactions.

Overall, these findings suggest that altitudinal gradients in herbivory and defenses in Q. robur are uncoupled and that elevational variation in herbivory and plant traits responds mainly to abiotic factors.

Summary 1.It is well-known that vascular plants have species-specific effects on soil properties. However, little is known on how individual species forming biocrusts, communities dominated by lichens, mosses and cyanobacteria that are prevalent in many ecosystems world-wide, affect microbial communities and soil variables related to nutrient cycling. 2. We evaluated the relationship of six biocrust-forming lichens (Buellia epipolia, Diploschistes diacapsis, Fulgensia subbracteata, Psora decipiens, Squamarina cartilaginea and Squamarina lentigera) with microbial abundance and multiple variables associated with soil nitrogen (N), carbon (C) and phosphorus (P) cycling and storage. We also evaluated whether the composition of lichen tissues (contents in C, N, P and polyphenols) is related to the C, N, P availability and microbial abundance in soils. Finally, we assessed what lichen species positively and negatively relate to soil fertility compared to bare ground areas without biocrusts. 3. We found contrasted C, N, P availability and soil microbial abundance under the different biocrust-forming lichens. Interestingly, inorganic P and amino acids were the most important factors differentiating lichen microsites. These differences in nutrient availability seem to be related to the C, N and P composition of the lichen tissues. For example, soils under D. diacapsis and P. decipiens, which had the lowest and highest C, N and P contents in their tissues, respectively, had the lowest and highest nutrient availability, respectively. We also found contrasted soil microbes abundance under the different soil lichens. For instance, F. subbracteata and D. diacapsis were negatively related to the abundance of bacteria compared to bare ground areas. 4. Our results support the idea that, as found with vascular plants, biocrust-forming lichens have species-specific effects on soil microbial communities and C, N and P cycling. Thus, continuing considering biocrusts as a unique entity will only add confusion to our knowledge of how they control nutrient availability and microbial abundance in the ecosystems where this key community is prevalent.

Leaf nutrient resorption allows plants to lower their dependence on current soil nutrients, thereby influencing ecosystem-level processes such as litter decomposition and soil nutrient availability. Among different factors controlling nutrient resorption are the availability of plant resources such as nutrients, water or light. The heterogeneous spatial distribution of these resources in natural environments may influence the spatial pattern of nutrient resorption. The spatial variability of leaf N and P resorption efficiency and proficiency and their relationship to the spatial pattern of soil and light resources were evaluated by using descriptive statistics and geostatistics in a Quercus robur L. (pedunculate oak) population. Resorption efficiency and resorption proficiency were significantly higher for P than for N. Levels of N resorption proficiency indicated incomplete resorption in all individuals. However, 80% of individuals exhibited intermediate or complete levels of P resorption. Resorption efficiency and proficiency of leaf N and P showed spatial dependence at the studied scale, with the spatial distribution of P showing higher range (autocorrelation distance) than that of N. The spatial pattern and scale of nutrients in senesced leaves differed from that in green leaves, with senesced leaves having a higher spatial range but lower percentage of variance explained by distance. All soil variables and the light availability index showed spatial dependence at the examined spatial scale. However, only soil water content and extractable-P were significantly correlated with resorption proficiency and efficiency, although these relationships explained a low percentage (<5%) of total variance. The spatial dependence for resorption variables has important implications for sampling design because nearby individuals cannot be considered independent samples. It may also have implications for ecosystem-level processes related to litter quality, which might also exhibit spatial dependence.