A Field Test To Optimal Hoarding Theory In Small Scatter-hoarding Rodents

Food hoarding is an important adaptive strategy adapted by food-hoarding animals to overcome the periods of food scarcity. There are two patterns of food hoarding:larder hoarding and scatter hoarding. For many hoarders, the risk of food pilferage by other animals might make great loss of caches and subsequently endanger their survival during the period of food scarcity and reproductive capacity during the coming breeding season. In order to maximize a net reward,hoarders must consider trade-offs between the costs and rewards associated with hoarding and make fine-scale behavioural decisions when hoarding, including tactics of increasing harvest, finding suitable sites for hoarding, decreasing cache loss and enhancing cache retrieval etc.Under natural conditions, hoarding behaviors of hoarders might have been shaped to an optimal level at which hoarders may maximize net rewards from food hoarding, weighted by the probability of cache survival and retrieval. At here, we hypothesized that, at community level, scatter hoarders prefer to hoard seeds under optimal conditions including in suitable microhabitats and at an optimum dispersal distance, density and cache depth. Under these optimal conditions, maximal net rewards would be achieved by maximizing cache survival and minimizing the energy costs associated with caching. Unambiguous tests of this optimal hoarding model are rare.In this study, we tested the optimal hoarding hypothesis at community level in small rodents that hoarders hoard seeds in optimum microhabitats, and at an optimal density and depth to maximize net rewards under relatively stable environments. We tracked artificially released Himalayan hazelnut, Corylus ferox, seeds in subtropical forest. We measured the microhabitat, cache density and cache depth of rodent-made caches to explore what conditions are favoured by small rodents for hoarding. We then tracked a series of man-made caches in different microhabitats, densities and depths while holding one trait constant (either microhabitat, density or depth). We predicted that if caches within the rodent-favoured microhabitats survived longer than others (indicated by a constant high survival rate), the optimal cache microhabitat hypothesis would be supported; if cache survival reaches a constant high level at the rodent-favoured cache density and/or cache depth, the optimal cache density and/or cache depth hypothesis would be supported. This study was conducted during April to November2011in the Badagongshan National Natural Reserve, Sanzhi County, Hunan China.We found that small scatter rodents preferred to bury seeds at a2-3m apart,2-3cm depth and a micro-habitat with shrub. Caches within the rodent favored levels of cache density (2-3m), cache depth (2-3cm) have a constant high level of survival rate, supporting our hypothesis. However, cases in the rodent favored habitats (under and at the edge of shrubs) disappeared most quickly, which is inconsistent with our prediction.These findings suggest that scatter hoarding rodents tend to hoard food items at optimal cache density and cache depth to gain net reward at community level, and prefer to hoard in safer micro-habitats to avoid the risks of predation. This is the first study to focus on multiple factors related to optimal seed hoarding at community in small rodents in a subtropical environment.