Causality of the relationship between grasses and herbivores and methodological issues
When interpreting our results, it needs to be borne in mind that the positive relation between herbivore species richness and grass species richness can act in both directions. First, herbivores may support grass richness by acting as a dominance control mechanism, suppressing potentially dominant species, and increasing microsite heterogeneity (e.g., Chaneton and Facelli 1991, Ritchie and Olff 1999, Jacobs and Naiman 2008, La Plante and Souza 2018). This is further supported by the positive effect of herbivore species richness manifest on basalts and crests, where the most intense competition among grass species can be assumed because this habitat/bedrock combination is where grasses reach the highest dominance; this also holds for their relative dominance with regards to other species (Hejda et al. 2022). Further, crests host fewer grazers than both seasonal and especially perennial rivers, and this may also contribute to their positive effect on grass species richness – presuming a unimodal relation between grazing intensity and grass richness (e.g. Frank 2005), the herbivores-grasses relationship can be expected to be positive under the conditions of relatively low herbivore abundances. Lastly, we believe that the presented causality of the relationships also applies to perennial and seasonal rivers due to abundant herbivores, such as elephants (MacFadyen 2019), that spend more time by the rivers, and their grazing selectivity is likely to decrease with grazing intensity (Cornelissen and Vulink 2015).
Alternatively, the richness of grasses can affect the richness of herbivores by providing a more diverse food supply, following the logic of the predator-prey relationship (e.g., Kallay and Cohen 2008, Malard et al. 2020). It is impossible to solve this dilemma using comparative data, and even exclosure experiments tend to give ambiguous results (Chikorowondo et al. 2017, Li et al. 2017, Fenetahun et al. 2021). Moreover, it is likely that both mechanisms with opposite directions are in play with differing importance depending on the specific environmental settings. It is probably more important in crests, where grazing selectivity is expected to be higher under low grazing intensities. In general, we suggest that the relationships between grasses and herbivores may work in both directions, but in environments with negative relationships, the effects of herbivores on vegetation prevail.