Water has long been recognized as a key resource for life on Earth. While energy and nutrients are both relevant in understanding the behavioral ecology of species, recent research has shed light on the impact water may have on food webs in arid and semiarid ecosystems. For instance, water-stress could drive animal behavioral strategies for managing both water demand and various other demands organisms face, including the need to avoid predation, maintain a positive energy balance, obtain nutrients, and find mates to reproduce.
Here I present a cafeteria-style field experiment aimed at assessing how variation in the ratio of water to dry biomass (or the hydration state) of cottonwood leaves may affect the consumption behavior of the damp-loving field cricket, G. alogus. I hypothesized that, if net water gain is greater than net water loss incurred during foraging and consumption, crickets would increase consumption of cottonwood leaf material with decreases in leaf hydration to compensate for reduced water intake per gram of leaf tissue consumed. Conversely, if crickets begin to consume leaves of sufficiently low hydration, they may incur water losses greater than the water gained during consumption and therefore may reduce consumption of sufficiently dehydrated leaves. Based on this hypothesized mechanism, I predicted that consumption of cottonwood leaves along a hydration gradient would exhibit a unimodal shape with a peak at intermediate leaf hydration. To test my prediction, I laid 10 wooden door shims, each with 4 leaves, each leaf of 1 of 4 hydration states, in the open along a transect in a Cottonwood-Willow floodplain forest over a single night during the dry pre-monsoon season of the Sonoran desert. I measured total consumption per leaf and tested for differences in average consumption per leaf against leaf hydration using ANOVA. I assessed the support for a linear versus unimodal model of the data using AICc.
Results/Conclusions
There was a significant difference in consumption between the leaf hydration groups (F=3.70, P=0.02). The unimodal model had greater support than a linear model of the consumption data (AICc wt: unimodal=0.69, linear=0.31). The support I found for the influence of pre-formed water on consumption patterns in the field suggests that water-driven behavioral patterns may influence food web dynamics with important consequences for our understanding of ecosystem processes, thus further investigation is well warranted.