Introduction
Increases in species’ abundances and diversity towards the equator are well documented (Dobzhansky, 1950; Willig et al. , 2003), and may be associated with a heightened intensity of biotic interactions at lower latitudes (i.e. , the “biotic interactions hypothesis”; Mittelbach et al. , 2007; Schemske et al. , 2009). Evidence for stronger biotic interactions in the tropics has been found for many mutualisms and antagonisms (Schemske et al. , 2009; Zvereva & Kozlov, 2021); however, evidence for latitudinal gradients in herbivory, one of the most ecologically significant biotic interactions, remains mixed (Anstett et al. , 2016; Zhang et al. , 2016).
Herbivory is particularly important because it involves a large proportion of the biodiversity and energy flows in terrestrial ecosystems (Becerra, 2007; Agrawal et al. , 2012). The classical latitudinal herbivory hypothesis (LHH) suggests that plant losses to herbivory decrease with latitude (Coley & Aide, 1991; Johnson & Rasmann, 2011). Although this hypothesis has been popular in the literature, its generality has been called into question in the last ten years because latitudinal trends in herbivory tend to be highly variable in both strength and direction (Moles et al. , 2011; Moles & Ollerton, 2016).
Some biases among LHH studies may partly explain the inconsistency in patterns reported thus far. The LHH was initially formulated as a community-level hypothesis, in which herbivory was expected to be greater in tropical versus temperate communities (Coley & Aide, 1991; Coley & Barone, 1996). However, many recent studies have focused on herbivory gradients within specific clades or even single species (i.e. , species found along a latitudinal gradient), rather than community-wide herbivory (see review in Anstett et al. , 2016). Such non-random, species-level measurements of herbivory are likely to bias conclusions regarding both absolute levels of herbivory and spatial patterns in community-wide herbivory (Zvereva & Kozlov, 2019). Furthermore, plant species may be unevenly distributed among communities (varying in abundance), and thus the results of LHH studies may be affected by species-specific factors, as well as limited by species’ ranges (Anstett et al. , 2016). Therefore, direct measurements of community-wide herbivory are better suited to studies of the LHH, as originally proposed, and essential for understanding the consequences of herbivory for biodiversity and ecosystem functioning (e.g. , nutrient cycling, productivity and trophic transfer rates) (Anstett et al. , 2016; Zvereva et al. , 2020).
Abiotic and biotic correlates of latitude can potentially affect community-wide herbivory through both intraspecific and interspecific pathways. At the intraspecific level, abiotic and biotic factors can drive variability in herbivory by directly influencing the abundance of insect herbivores (Maron et al. , 2014) and/or by indirectly altering plant susceptibility to herbivore damage (Moreira et al. , 2018; Loughnan & Williams, 2019). For example, plant populations growing under more benign climates may suffer greater herbivore damage, as longer growing seasons increase the number of insect generations and warmer winters result in lower insect mortality (Sakata et al. , 2017). In addition, plants growing in high resource environments may also invest less in defense traits (Hahn & Maron, 2016) and are therefore more susceptible to herbivore damage.
Apart from causing intraspecific variability in herbivory, latitudinal abiotic and biotic correlates are also predicted to alter the composition and diversity of natural plant assemblages and to modify the abundance of individual plant species (i.e. , community evenness) (Yang et al. , 2011; Ma et al. , 2017), ultimately affecting community-wide herbivory. These “species turnover effects” result from species’ differences in susceptibility to herbivory and also levels of adaptation to abiotic environmental conditions (Agrawal, 2007). For example, the resource availability hypothesis posits that lower-resource environments can select plant species with slow growth rates, but enhanced herbivore resistance (Coley et al. , 1985; Hahn & Maron, 2016). Individual species differences lead to habitat-driven turnover in plant community composition, potentially resulting in divergent patterns of community-wide herbivory over space. Overall, abiotic and biotic latitudinal correlates can affect community-wide herbivory through both intraspecific and/or interspecific pathways. However, the relative importance of these two pathways remains largely uninvestigated.
To test for latitudinal patterns in community-wide herbivory and to evaluate the relative roles of intraspecific herbivory variability versus species’ turnover effects, we conducted a survey of both population- and community-wide herbivory along a 1,500-km latitudinal gradient in the grasslands of the Eastern Qinghai-Tibetan Plateau. Grasslands are the dominant ecosystem on the Qinghai-Tibetan Plateau, a diverse region with highly variable abiotic conditions (Yao, 2019); grassland communities in this region vary dramatically in composition, productivity and species richness, and plant defense traits have been shown to vary along environmental gradients within the region (Xiaoet al. , 2021). In this study, we asked: 1) Does community-wide herbivory decrease with latitude and, if so, what is the relative importance of intraspecific herbivory variability versus species’ turnover in driving this latitudinal pattern? 2) What abiotic (i.e. , climatic and edaphic factors) and biotic (i.e. , plant community factors) factors drive latitudinal variation in population- and community-wide herbivory, and how are these drivers related to intraspecific herbivory variability and/or species’ turnover effects?