| Abstract: | 1. Insect herbivores choose their food according to its protein to carbohydrate ratio, but the reasons why different species have contrasting intake targets remain unclear. According to the growth rate hypothesis, P‐rich insects have higher growth rates. It is therefore expected that P‐rich insects will prefer plants that are high in protein (hence in nitrogen, N) and phosphorus (P). 2. To test this hypothesis, the ecological stoichiometry and the interaction network frameworks were combined. The food preferences of 24 plant species by 23 grasshopper species were determined, and the N and P contents of both trophic levels were measured. The weighted mean P and N contents of the consumed plants, which represent the grasshoppers' feeding niche, were highly correlated, indicating that the grasshoppers' diets are spread along a single functional niche axis. The links between the stoichiometry of the plants and their consumers were then tested with the fourth corner analysis, a multivariate technique combining the plant traits, the insect traits and the interaction network. 3. In line with the earlier hypothesis, P‐rich grasshoppers consumed plants high in N and P, probably because their growth rate is higher. These findings therefore introduce a mechanism that accounts for interspecific differences in diet preference. They also contribute to an understanding of how herbivore communities might respond to P and N limitation in ecosystems, and how complex interaction networks can influence biogeochemical cycles of N and P. |
