The Latitudinal Biotic Interaction Hypothesis (LBIH) states that the intensity of biotic interactions increases from high to low latitudes. This hypothesis, which may partly explain latitudinal gradients in biodiversity, remains hotly debated, largely due to variable outcomes of published studies. We used meta-analysis to identify the scope of the LBIH in terrestrial ecosystems. For this purpose, we explored the sources of variation in the strength of latitudinal changes in herbivory, carnivory, and parasitism (119 publications) and compared these gradients with gradients in the diversity of the respective groups of animals (102 publications). Overall, both herbivory and carnivory decreased towards the poles, while parasitism increased. The latitudinal gradient in herbivory and carnivory was threefold stronger above 50–60º than at lower latitudes and was significant due to interactions involving ectothermic consumers, studies using standardized prey (i.e. prey lacking local anti-predator adaptations) and studies aimed at testing LBIH. The poleward decrease in biodiversity did not differ between ectothermic and endothermic animals or among climate zones and was four-fold stronger than decrease in herbivory and carnivory. The discovered differences between the gradients in biotic interactions and biodiversity suggest that these two global macroecological patterns are likely shaped by different factors.

All data were extracted from published studies that fit the following criteria: i) the data were collected from natural ecosystems; ii) the data were collected from at least two study sites located at least 2 degrees of latitude apart; and iii) the magnitude of the effect was available from the study or could be calculated from the data presented in the publication or provided by authors.

We identified suitable studies based on earlier reviews of latitudinal changes in biotic interactions, and we searched for additional publications in the ISI Web of Science using the keywords ‘latitud*’, ‘geographic’, ‘biotic interactions’, ‘herbivor*’, ‘predat*’, ‘carnivor*’ and ‘parasit*’. The search was completed in December 2020. We included only studies that contained direct quantitative estimates of the intensity of herbivory (the percentage of plant biomass or leaf area lost to herbivores or the proportion of damaged leaves), carnivory (the mortality of prey or the predator attack rates) or parasitism (prevalence; i.e. percentage of infected hosts). We rejected studies where the interaction intensity was deduced from the abundance of herbivores/predators and where interactions could have been affected by human activities, such as herbivory by semi-domesticated reindeer or by large animals whose numbers have considerably decreased over the past decades due to human-induced changes in vegetation.

The vast majority of studies exploring latitudinal patterns in biotic interactions employed sites located at about the same (usually low) elevations. If elevation varied greatly within a study, we removed sites with deviating elevations from effect size calculations to prevent their undue influence on latitudinal pattern. When a study reported data collected over several years, we selected the year with the highest number of sites or with the longest latitudinal gradient. However, if data from different years were pooled by the authors, then we used the pooled data. When a study reported data collected from multiple localities from two or three climate zones, we divided these data by climate zones and then by continents. Entire global gradients were excluded from the analyses when our focus was on gradients either located within one climate zone or crossing a border between the two adjacent zones.

The larger part of the data on latitudinal changes in biodiversity was directly extracted from the published databases (Hillebrand, 2004; Kinlock et al., 2018), resulting in 166 effect sizes (ES, hereafter) for taxonomic groups involving herbivorous, carnivorous and omnivorous animals. However, studies conducted at high latitudes were underrepresented in these databases compared with our database on biotic interactions. This prompted us to search for additional studies from subarctic and arctic regions in the Web of Science and in elibrary.ru (published in English and Russian, respectively) using the keywords ‘latitud*’, ‘diversity’, ‘polar’, ‘arctic’, ‘tundra’ and ‘boreal’.

All explanations and references are provided in the README file.