Food hoarding by Merriam's kangaroo rats: a test of alternative hypotheses

Ecology, Dec, 1995 by Stephen H. Jenkins, Aron Rothstein, Wendy C.H. Green

INTRODUCTION

Food storage is an important adaptation of many animals to temporally variable or unpredictable food supplies (Smith and Reichman 1984, Vander Wall 1990). Vander Wall (1990) thoroughly reviewed the abundant descriptive literature on food hoarding. Despite this rich natural history of hoarding, only a few workers have tested specific hypotheses about the adaptive value of various spatial and temporal patterns of food storage (e.g., Stapanian and Smith 1978, Clarkson et al. 1986, Reichman et al. 1986, Hurly and Robertson 1987, McNamara et al. 1990, Lucas and Walter 1991, Waite and Reeve 1992, 1993, Clarke and Kramer 1994a).

Two major patterns of food storage are larderhoarding, in which all items are stored in one location such as a burrow, and scatterhoarding, in which items are stored in many separate locations (Vander Wall 1990). According to Vander Wall (1990:3), these patterns should not be viewed as a dichotomy but as the ends of a continuum of spatial patterns of food storage. For example, scatterhoards may be individual items, as in the nuts stored by fox squirrels (Sciurus niger; Stapanian and Smith 1978), or small clumps of items, as in the pine seeds stored by Clark's Nutcrackers (Nucifraga columbiana; Vander Wall and Balda 1981) and white-footed mice (Peromyscus leucopus; Abbott and Quink 1970). Furthermore, scatterhoards may be widely spaced throughout an individual's home range or concentrated in particular microhabitats or spatial locations (Vander Wall 1990).

Heteromyid rodents (kangaroo rats and pocket mice) are excellent model subjects for studies of food hoarding. They eat mainly seeds (Reichman 1975), which can be stored for a long time without spoiling. They are well adapted to desert environments (Schmidt-Nielson 1964), where seed production is sporadic and unpredictable. Their external, fur-lined cheek pouches are a striking morphological adaptation for minimizing water loss while carrying seeds to burrows or other storage sites (Vander Wall 1993). Sympatric kangaroo rats sometimes exhibit very different spatial patterns of food storage. For example, banner-tailed kangaroo rats (Dipodomys spectabilis) occupy large mounds in which they larderhoard substantial quantities of seeds (Vorhies and Taylor 1922, Reichman et al. 1985). By contrast, Merriam's kangaroo rats (D. merriami) often scatterhoard small clumps of seeds in numerous shallow holes dug in the soil (Reynolds 1958, Daly et al. 1992). Although competition and resource partitioning have been well studied in desert rodent communities (e.g., Price and Brown 1983, Brown and Munger 1985, Kotler and Brown 1988, Brown and Heske 1990, Heske et al. 1994), no one has explicitly considered the role of food hoarding in the interspecific interactions of kangaroo rats and other desert rodents.

Brown and Heske (1990) argued that kangaroo rats are a keystone guild in the Chihuahuan Desert because their removal caused major changes in both plant and animal components of the community. We suggest that food hoarding plays a central role in the functioning of kangaroo rats as keystone species. Redistribution of seeds through caching may be an important mechanism of competition among rodent species and between rodents and other granivores. Caching also influences dispersal and recruitment of seeds, and therefore distribution and abundance of plants (Price and Jenkins 1986, Vander Wall 1994). Because of its critical role in interactions within and between trophic levels, food hoarding can be considered a "keystone behavior"; as such, it deserves more attention from ecologists.

We designed a laboratory experiment to test several alternative hypotheses about spatial patterns of food storage by Merriam's kangaroo rats. In accordance with Vander Wall's (1990) suggestion that food may be stored in many different dispersion patterns, we considered larderhoarding as well as four distinct ways of distributing scatterhoards. Each of these distribution patterns has characteristic benefits and costs, which are described in the following paragraphs.

1. Larderhoarding. - An obvious advantage of larderhoarded seeds is that they can be used without the extra time and energy costs and risks of predation of going aboveground to dig up scatterhoards. Larderhoarding, however, has at least two potential costs: (1) seeds stored in the relatively warm, moist environment of a burrow may be more likely than scatterhoarded seeds to be attacked by microbes, fungi, or insects (Reichman et al. 1986); (2) a concentrated supply of seeds in a burrow may attract other rodents that steal those seeds (Daly et al. 1992).

2. Scatterhoarding near the burrow. - If an individual could successfully defend an area around its burrow, then making scatterhoards in this area would decrease the likelihood of microbial, fungal, and insect attack compared to larderhoarding, with minimal increases in costs of cache recovery. In fact, some kangaroo rats make numerous scatterhoards directly over their underground burrow systems (Shaw 1934, Hawbecker 1940). Success of this strategy depends on being able to defend the caching area against both conspecifics and heterospecifics. Among kangaroo rats, larger species are often aggressively dominant to smaller species (Congdon 1974, Blaustein and Risser 1976). Thus a smaller kangaroo rat (e.g., D. merriami) probably could not defend an area in which it cached seeds against a larger kangaroo rat (e.g., D. spectabilis). In this situation, a concentration of caches near a burrow would be attractive for stealing by the larger animal. Furthermore, the larger animal might be able to use the burrow entrance as a cue to abundant caches nearby.