Scatterhoarding by territorial red squirrels: a test of the optimal density model

A simple null model for scatterhoarding is presented, which demonstrates that prediction 1 is not exclusive to the Optimal Density Model and that "optimal densities" were not the primary goal of the caching animal, but rather the result of a positive relationship between food value and investment in caching.

Abstract

We observed a high degree of scatterhoarding in a population of red squirrels and tested two predictions of the Optimal Density Model (ODM): (1) large food items will be cached at a greater distance from their source than small items; and (2) caches will be uniformly distributed about their source. Caching experiments supported prediction 1. Red squirrels carried large food items farther than small items before caching them. Prediction 2 was not supported; caches were distributed nonuniformly about their source both within and among caching bouts. We present a simple null model for scatterhoarding, which demonstrates that prediction 1 is not exclusive to the Optimal Density Model. Analyses of our cone-caching data and published data suggested that "optimal densities" were not the primary goal of the caching animal, but rather the result of a positive relationship between food value and investment in caching (carrying distance).