Current theory predicts that the intensity of biotic interactions, and particularly herbivory, decreases with increasing latitude and elevation. However, recent studies have revealed substantial variation in both the latitudinal and elevational patterns of herbivory. This variation is often attributed to differences in study design and type of data collected by different researchers. Here, we used a standardised sampling protocol along elevational gradients in six mountain ranges, located at different latitudes within temperate Eurasia, to uncover the sources of variation in elevational patterns in insect herbivory on woody plant leaves. We discovered the considerable variation in elevational patterns among different mountain ranges; nevertheless, herbivory generally decreased with increasing elevation at both the community-wide and individual plant species levels. This decrease was mostly due to openly living defoliators, whereas no significant association was detected between herbivory and elevation among insects living within plant tissues (i.e. miners and gallers). The elevational decrease in herbivory was significant for deciduous plants but not for evergreen plants, and for low-stature plants but not for tall plants. The community-wide herbivory increased with increases in both specific leaf area and leaf size. The strength of the negative correlation between herbivory and elevation increased from lower to higher latitudes. We conclude that elevational gradients in herbivory demonstrate considerable variation, and that this variation is mostly associated with herbivore feeding habit, some plant traits and latitude of the mountain range.

Sampling protocol

The sampling was conducted in the second half of the growth season, when the majority of insect herbivores had completed their feeding, but well before the start of leaf fall at any of the study sites. At each site, we collected a total of 15‒25 branches (greater numbers corresponded to more diverse plant communities) from common woody plant species. We selected branches for sampling while standing at a distance of 5‒10 m away (at this distance, the level of herbivory on the selected branch could not be assessed visually) to avoid unconscious selection bias. We disregarded a few branches that were grazed by vertebrate herbivores, which usually remove the entire shoots or greater parts of adjacent leaves; this ‘coarse’ damage is easily distinguished from a ‘fine’ and relatively minor damage imposed on individual leaves by invertebrate herbivores. All samples were collected and processed by the same persons (MVK and VZ) who had no a priori knowledge of the herbivory levels in the visited regions and/or on the collected plant species.

In the Alps and in Cairngorms, we first selected up to five woody plant species that dominated the community (i.e. that comprised the greater part of foliar biomass), and we then haphazardly collected one branch (or group of branches) from each of five mature individuals of each selected species (on a first-found, first-sampled basis). In the other mountain ranges, we did not select the species to be collected. Instead, we collected one branch from each of 15‒25 individuals of woody plants irrespectively of their identity using the following procedure. Each of two collectors walked along a straight line, making stops at approximately 10 m intervals and collecting a branch from a species which had the greatest foliar biomass (as estimated visually) within a 2 × 10 m plot behind the collector. This change was introduced to better reflect herbivory in species-rich plant communities and to avoid possible subjectivity in the selection of plant species to be collected. All other details of the protocol were the same across all regions, and we assumed that under both these protocols the numbers of sampled individuals of each species were roughly proportional to the contributions of these species to the community-wide foliar biomass. Importantly, the sampling protocol was always identical within each gradient, and thus a minor difference outlined above is unlikely to influence the results of meta-analysis based on correlations between herbivory and elevation within individual gradients.

Measurements of herbivory

In the laboratory, each leaf was carefully examined for the presence of damage imposed by chewing, galling and mining invertebrates. In conifers, we classified current-year needles that were missing from the shoot as having been consumed by insects, because undamaged needles of evergreen conifers generally persist for several years. We assigned each leaf/needle (leaf hereafter) to one of the damage classes according to the percentage of the leaf area that was consumed or otherwise damaged by insects, as follows: 0 (intact leaves), 0.01–1%, 1–5%, 5–25%, 25–50%, 50–75% and 75–100%. The last class included the petioles of fully consumed leaves and missing needles. We calculated the proportion of foliage area lost to insects (i.e. the leaf herbivory level) by multiplying the numbers of leaves in each damage class by the respective median values of the damaged leaf area (i.e. 0 for intact leaves, 0.5% for the damage class 0.01–1%, etc.); the obtained values were then summed for all damage classes and divided by the total number of leaves (including undamaged ones) in a sample.

Measurements of leaf traits

We measured leaf traits from each branch collected from Altai, Avachinskij, Caucasus and Troodos, but only from one of several conspecific branches collected from the same site in the Alps and in Cairngorms. Soon after branch collection, when the plant leaves were still turgid, we sampled 1‒125 undamaged leaves (depending on leaf size; the median value was 8 leaves) and weighed them to the nearest 1 mg. We then punched out 5‒12 discs (4.5, 8 or 12 mm in diameter, depending on leaf size) from those leaves, dried the discs and the remaining parts of leaves for 24 h at 105ºC and then weighed them to the nearest 0.1 mg. When the leaves were too small to allow punching of leaf discs, we press-dried the leaves and measured their total area from their photographs using Adobe Photoshop 2020. This measured area was then divided by 0.902 to correct for leaf shrinkage during drying. This correction factor was obtained by averaging the ratio between dry and fresh areas of 60 leaf discs collected from leaves of 20 plant species from our study regions. We did not measure SLA in needle-bearing plants. The dry weight of an individual leaf was calculated by dividing the weight of the punched leaves plus discs by the number of leaves. The water content in leaf tissues was calculated by dividing the difference between the wet and dry weight of a leaf by its wet weight. The SLA was calculated as the total area of the discs or leaves (mm²) divided by their dry weight (mg). Leaf area (mm²) was calculated by multiplying dry leaf weight (mg) by SLA (mm² mg^-1).

The data and metadata are saved as plain text (.txt) files.