Food storability and the foraging behavior of fox squirrels (Sciurus niger).
|Abstract:||Food storage in animals allows foragers to reap when food is plentiful and costs are low and eat when food is scarce and costs are high, thus shifting resources from periods of low value or high availability to periods of high value or low availability. To a caching animal, a food item has two components: its present value for immediate consumption and its future value if stored. We explored some properties of caching in the context of a food's future value using free-living fox squirrels (Sciurus niger) and manipulations of cacheability of supplemental food. We assessed squirrel behavior using giving-up densities (GUDS) of noncacheable food in artificial food patches. Squirrels had higher GUDs in assessment trays when given noncacheable supplemental food than when food was not augmented; when given supplemental food in a highly storable form, squirrels had intermediate GUDs. Thus, future value of food affects the foraging behavior of squirrels through the balancing of present and future needs.|
|Subject:||Fox squirrel (Behavior)|
Kotler, Burt P.
Brown, Joel S.
|Publication:||Name: The American Midland Naturalist Publisher: University of Notre Dame, Department of Biological Sciences Audience: Academic Format: Magazine/Journal Subject: Biological sciences; Earth sciences Copyright: COPYRIGHT 1999 University of Notre Dame, Department of Biological Sciences ISSN: 0003-0031|
|Issue:||Date: July, 1999 Source Volume: 142 Source Issue: 1|
Food caching has evolved numerous times and occurs in diverse taxa (Smith and Reichman, 1986; Vander Wall, 1990). In animals such as spiders, shrikes, shrews and mountain lions (Shull, 1907; Eberhard, 1967; Wemmer, 1969), the time scale for caching and cache recovery involves no more than hours or days (Waite and Grubb, 1988). In chipmunks, nutcrackers and desert rodents (Vorhies and Taylor, 1922; Vander Wall and Balda, 1977; Elliot, 1978), the time scale increases to weeks and months. Haying in pikas (Conner, 1983) and seed maintenance by kangaroo rats (Reichman and Rebar, 1985; Reichman et al., 1986) illustrate elaborate food preparation, packaging and storage. (Smith and Reichman, 1986; Vander Wall, 1990).
Food storing allows animals to reap when food is plentiful and costs are low and eat when food is scarce and costs are high (Smith and Reichman, 1986; Vander Wall, 1990). Hence, the value of a cacheable food item to an animal has two components: present value and future value. By present value we mean the fitness consequences of consuming a food item immediately. By future value we mean the fitness consequences of foregoing consumption in anticipation of needing, recovering and consuming the food item in the near or distant future. Both should affect foraging decisions. Here, we tested for these foraging decision responses by supplementing the diet of free-living fox squirrels with cacheable (hazelnuts in their shells) or noncacheable (hazelnuts without their shells) foods. The squirrel's subsequent patch use behavior will tell us whether fox squirrels can assess and respond to the future value of their founds.
Patch use and the future value of food: theory and predictions. - We used animals' exploitation of depletable food patches to investigate how future value of food affects foraging behavior. We start by giving the conditions for when an optimal forager should leave a resource patch. Then, we use this to explore how effects of current vs. future value of food will alter the manner in which foragers exploit food patches. To maximize fitness, an animal should quit a food patch when its harvest rate in the patch, H, fails to equal the sum of its metabolic, C, predation, P, and missed opportunity, MOC, costs of foraging (Brown 1988, 1992):
H = C + P + MOC (1)
When the animal's harvest rate is a function of remaining food density in the patch (Kotler and Brown, 1990; Brown et al., 1994), the amount of food left behind by the forager in a food patch, the giving-up density (GUD) (Brown, 1988), provides a measure of the forager's quitting harvest rate. In this way, GUDs estimate foraging costs and can be used to evaluate foraging costs and patch use.
Equation (1) is derived by maximizing a fitness generating function that is affected by energy gain, survivorship (both of which are functions of time spent foraging) and the amount of alternative, fitness-enhancing activities performed (a function of time spent performing these activities) subject to a time constraint (Brown, 1988). The resulting optimal patch use strategy is to exploit a resource patch only so long as harvest rates in the patch are greater than the sum of the various foraging costs. Note that the model deals directly with fitness rather than a surrogate for fitness such as energy gain (Krebs et al., 1978; Oaten, 1977; Iwasa et al., 1981; Mitchell, 1990), so relationships such as that between instantaneous energy, gain and lifetime fitness (McNamara, 1982) which otherwise may come into play when examining current behavior and future value are subsumed.
Effects of current vs. future value of food on patch use can be examined by exploring how they should alter the value of the various components of equation (1). We do so by focusing on the cost of predation, P. P is comprised of the product of the risk of predation. [Mu], and the fitness of a forger should it survive to reproduce, F, divided by the marginal value of consuming energy, dF/de (see Brown, 1988, for the derivation).
P = [Mu] [multiplied by] F/(dF/de) (2)
Over short periods of time (e.g., d. or even wk), it is unlikely that [Mu] or F will vary much unless there are large changes in predator numbers and food availability. So changes in foraging behavior that we see during each round of our experiment will most likely come about because of changes in the marginal value of energy. We note that dF/de is the marginal value of consuming energy, now. If the food in a depletable food patch cannot be cached, then the marginal value of energy in equation (2) is just its present value. A similar equation to (2) can also be drawn for missed opportunity cost (Brown, 1988, 1992).
For many animals, the present value of energy changes daily. This occurs because present value depends strongly on the animal's energetic state and its recent consumption of food. An animal whose body (not necessarily its cache) has low energy reserves will have a high present value of energy. In a patch with noncacheable foods, this results in a low cost of predation and/or missed opportunity cost, and a low giving-up density. An animal that is sated and whose body has ample energy reserves will have a low present value of energy and high giving-up densities in patches with perishable food.
Supplementing an animal with a noncacheable vs. a cacheable food should have different effects on the animal's present value of energy, and on the animal's use of other resource patches containing a noncacheable food. When an animal is presented with supplemental food that cannot be cached, it has the option of ignoring it (no effects on its valuation of energy and on how it exploits the patch containing noncacheable food) or consuming it. If it consumes the food, then its energy state increases. This will reduce the forager's present value of energy and result in higher giving-up densities, but it will have negligible effects on its future value of energy. The future value of energy may decline slightly if present consumption of food makes it less likely that food will be required in the future, or the future value of food may increase slightly if consuming food now makes it more likely that the animal will want and be able to take advantage of stored food in the future. In either case, present consumption of food should have a much larger effect on the present than future value of food.
When an animal is presented with supplemental food that can be cached, the animal can ignore it (no effects on the value of energy and its use of the patch containing noncacheable food), consume it (the above effects on present and future value), or cache it for future use. Caching a food item will likely have no or slight effects on the present value of energy and on giving-up densities. (One way in which caching a food item can affect the current value of food is if caching a food item is energetically costly; then, the animal's state may decline and the present value of food will rise somewhat). In contrast, caching a food item should usually affect the future value of food. It can either depress or increase the future value of energy, depending on whether there are diminishing or increasing returns to stored foods, respectively. This can affect subsequent effort put into caching and even overall subsequent foraging effort (Kotler, 1997). For our purposes, caching a food item should have no effect or a small positive effect on the present value of energy.
We can make two predictions regarding an animal's use of food patches containing a noncacheable food:
(1). Supplementing an animal with extra noncacheable food should increase the animal's giving-up density in food patches. Rationale: the extra food consumption depresses the present value of energy. This, in turn, increases the cost of predation and missed opportunity cost and increase GUDs:
[GUD.sub.nc] [greater than] [GUD.sub.0]
where nc = supplemental noncacheable food, and 0 = no supplemental food.
(2). Supplementing an animal with cacheable food should yield a higher or equal giving-up density relative to no additional food, and a lower or equal giving-up density relative to extra noncacheable food. Rationale: the decline in the present value of energy resulting from the food augmentation and its effects on predation costs and missed opportunity costs relates directly to the fraction of the extra food that is stored rather than immediately consumed. The more of the extra food that is stored rather than eaten now, the less will be the effect of the food supplement on the present value of energy. Thus, an augmentation of noncacheable food will lead to a large increase in P and/or MOC; an augmentation of cacheable food will lead to a smaller increase in P and/or MOC, depending on the amount stored rather than eaten:
[GUD.sub.nc] [greater than] [GUD.sub.c] [greater than] [GUD.sub.0]
where c = supplemental cacheable food.
To test the two predictions, we chose fox squirrels (Sciurus niger) living under natural conditions on the edges of oak-maple forests at The Morton Arboretum, Lisle, Illinois. Fox squirrels store nuts and acorns for future use during the winter months (Brown and Yeage, 1945; Jacobs, 1989; Vander Wall, 1990). In all of the experiments presented here, we used artificial food patches provisioned with sunflower seeds as assessment trays to measure the squirrels' giving-up densities, perceptions of foraging costs, and changes in present value of food. The fox squirrels do not cache the sunflower seeds. Hence, our food patches satisfy the criteria of being depletable, in accordance with equation (1), and containing a noncacheable food, in accordance with the effects on present value of energy. To examine whether squirrels distinguish between current value and future value of food items, we presented squirrels with supplemental food that differed in cacheability (hazelnuts with or without shells) - and therefore in potential future value - at the same time we measured squirrel patch use behavior in the assessment trays filled with sunflower seeds. To verify, that squirrels treat hazel nuts with and without shells differently, we provided piles of each and observed fox squirrels as they handled the food. We presented either hazelnuts in their shells or nuts with their shells removed five at a time and measured the fate of the nuts. We presented squirrels with a total of 55 nuts. Of these, squirrels cached all but one of the intact nuts, and ate all but three of the hazelnuts without shells. Thus, the squirrels altered their caching behavior in response to the different types of augmentation.
We used three different sites, each separated by over 1 km and each containing three stations separated by more than 50 m. At each station, we presented squirrels with food supplements of hazelnuts placed in small plastic trays (28 X 14 X 6 cm) placed midway between the station's two food patches (assessment trays). The two assessment trays at a station were placed in different microhabitats, bush and open (to allow comparison with other work on the foraging behavior of squirrels at the same sites; Brown et al., 1992). At the parkland site, we placed the bush tray at the base of a tree and open tray 5 to 10 m away; at the two meadow sites, we placed the bush tray 5 m inside the woods, and the open tray 5 m into the meadow. Augmentation of hazelnuts, each of 50 g (not including the mass of the shells), came in either a highly storable form (in the shell) or a perishable form (shell removed). We also had a control treatment of no augmentation. We rotated the augmentation treatments daily among sets of three stations according to a Latin square design, to assess the foraging behavior of the squirrels under the various treatments of extra food. We measured GUDs as follows. The food patches consisted of plastic trays (55 x 28 x 6 cm) filled with 4 L of dry, sifted bank sand into which we thoroughly mixed 9 g of sunflower seeds. Fox squirrels readily dig through the sand for the seeds; with sufficient effort, all depths of the tray are accessible to the squirrels. We set up trays shortly after dawn and collected the remaining seeds in the late afternoon. This gave squirrels ample opportunity to complete their foraging in each tray. We then sifted the sand of each tray to remove the remaining seeds and returned the seeds to the laboratory to be cleaned of debris and weighted to obtain the GUD. We log-transformed the data prior to analysis.
The GUDs in our assessment food patches under the various treatments allow us to test the two predictions. In the absence of food augmentation (control), [GUD.sub.0], provides a baseline measurement of the squirrels' foraging costs and present value of energy. In the presence of extra hazelnuts without shells (less storable), [GUD.sub.nc] tells us how improving the energetic state of the squirrels affects foraging costs through alterations in the present value of energy. In the presence of extra hazelnuts in their shells (more storable), [GUD.sub.0] tells us if squirrels appropriately distinguish between present and future value when caching food.
We have previously used this system of food patches and sites on other studies of diet selection (Brown and Morgan, 1995), patch assessment (Schmidt and Brown, 1996), and effects of microhabitat and supplemental food on the predation and missed opportunity, costs of foraging (Brown et al., 1992; see also Bowers et al., 1993). Typically, one to three different individuals use the food patches of a station, a few individuals will move among stations of a site, and individuals do not move among sites during the course of experiments. Daily switching of experimental treatments did not influence the numbers or dispositions of foragers at a station. At the time of the study, removal of supplemental hazelnuts and food from patches was exclusively by fox squirrels. Observations showed that squirrels removed the entire augmentation each day. Squirrels consumed shelled hazelnuts at the station or removed them to the base or limbs of nearby trees for consumption; unshelled hazelnuts were most frequently cached and commonly, but less frequently, shelled and consumed. While trying not interfere with or disturb the squirrels, we made sufficient observations to verify the assumptions of the experiment.
We ran seven replicates of the Latin square as follows. Period 1: 6, 7, 8 April 1994; period 2: 9, 19, 20 April 1994; period 3: 21, 22, 23 July 1994; period 4: 26, 28, 29 July 1994; period 5: 2, 5, 6 August 1994; period 6: 22, 24, 25 November 1994; period 7: 9, 10, 11 December 1994. In all cases, seed trays were set out after dawn and collected before dusk.
We used a partially hierarchical ANOVA to analyze the data (Brownlee, 1965). Day, site and treatment (supplemental noncacheable food, supplemental cacheable food, and no supplemental food) were the variables of the Latin square experimental design. Station was a variable nested within site, and days were nested within period (3 d each), different temporal replicates of the Latin square. The temporal replicates of the Latin square are not "true" replicates in the sense of representing new sets of stations and squirrels. However, they do increase the power for testing the predictions at our nine stations that represent in total from 20-25 different squirrel individuals. Microhabitat (bush vs. open) was a variable fully crossed with all other independent variables.
The augmentation treatments influenced the GUDs of fox squirrels in assessment trays ([ILLUSTRATION FOR FIGURE 1 OMITTED]; Table 1). Also, GUDs were lower in the bush than the open microhabitat (reflecting greater safety near cover) and varied across time periods (reflecting especially low predation costs in the summer when cover was greatest) and sites (Table 1). As predicted, GUDs were lowest in the absence of extra hazelnuts and highest when the augmentation consisted of non-cacheable hazelnuts (shells removed) (prediction 1). Extra non-cacheable food increased foraging costs and giving-up densities through a reduction in the present value of food. The GUD under the augmentation of cacheable hazelnuts (shells intact) was intermediate between the other treatments (prediction 2). The similarity of this GUD to those either of the previous two treatments determines how much of the value of hazelnuts in their shells is future vs. present value, respectively. It appears that nearly half (44%) of the value of the hazelnuts in their shells is future value. The large drop in GUD in the assessment tray when the augmented food is highly cacheable compared to when it is not precludes the effect being due solely to the cacheable augmentation being worth less because of the need to spend time and energy opening the hazelnuts.
An animal's foraging decisions are influenced in several ways when its food is cacheable and has future value. Normally, as a forager gathers food, its internal state is altered as its energy reserves increase; it becomes less "hungry." In terms of foraging theory, its marginal value of energy declines, and hence, its cost of predation (for a given level of mortality risk) increases (Brown, 1988, 1992). Eventually, cost of predation will increase to the point that the forager is no longer profiting from further food gathering, and the animal retreats to shelter. However, when a food item of high future value is gathered for later consumption, it has less effect on the forager's current state and on the forager's subsequent fired gathering activities. In the case of the fox squirrels in the experiments reported here, augmenting with a highly storable food (hazelnuts in their shells) reduced the fox squirrel's marginal value of energy less than augmenting with a more perishable food (hazelnuts with shells removed). Perishability of previously harvested food influences the subsequent patch use behavior of the squirrels. When we gave squirrels hazelnuts with shells, squirrels responded by foraging assessment trays containing nonstorable food more intensively and quitting them at lower GUDs than when we gave the squirrels hazelnuts without shells.
Another consideration concerns possible seasonal changes in the future value of stored food (Barry, 1976; Thompson and Thompson, 1980). For instance, hazelnuts may have little future value in the late spring when food is plentiful and anything cached for the following winter must remain hidden and intact for many months. At this time, foraging decisions should be shaped mostly by the present value of cacheable food items. By autumn, however, the same food item may be worth much more for its future value. Interestingly, we found no interaction effects between the different periods and the supplemental food treatments, suggesting that the future value of stored food remained unchanged, perhaps due to the need to survive the entire year to be able to reproduce. This is in contrast to the effect of time period itself, where GUDs (and therefore foraging costs; see equation 1) were lowest during the summer months when energetic foraging costs, C, were presumedly lowest.
At least three things determine the future value of a food item: its nutrient and energy content, its likelihood of consumption in the future, and its likely value to the animal in the future (Andersson and Krebs, 1978; Moreno et al., 1981; Kagel et al., 1996). The larger the food item's content of energy or nutrients, the larger will be its future value (Frank, 1991; Whisaw and Tomie. 1989). The higher the probability of consuming the food item in the future, the higher also will be its future value (Kagel et al., 1986). This aspect of future value depends on the survivorship of the forager, the storability of the food (Shaw, 1936; Post and Reichman, 1991), and the likelihood of recovering the food item (Baker et al., 1988). We note that anything that affects the likelihood of recovering a food item in the future, be it storability or likelihood of the food item being pilfered by a neighbor, will have the same effect on future value. In addition, the future value of a cacheable food item increases with the likelihood of a high and unmet marginal value of energy at some time in the future. The likelihood of future consumption and the likely marginal value of energy at the time of consumption determine what the forager will earn (in terms of per capita growth rate) if it saves the food item for future consumption (Andersson and Krebs, 1978). Similar considerations also hold for animals that store food as fat for hibernation. Frank (1991) showed that Beldings ground squirrels (Spermophilus beldingi) eat plants high in fatty acids in the autumn because the fatty acids are necessary for breaking hibernation, even though these plants are not appropriate for the autumn diet.
Caching animals may modify food items in order to alter the foods' future value. Fox squirrels typically remove the endosperm from acorns of oak species that germinate in the autumn shortly after they are released from the parent tree (Fox, 1982; Smallwood and Peters, 1986). This increases the storability of the acorns and increases their future value. A consequence of this is that squirrels seldom provide seed dispersal for these oak species. Acorn species which do not germinate until the spring are cached intact, and squirrels commonly act as agents of seed dispersal. Kangaroo rats (Dipodomys spp.) will manipulate seeds in their caches by handling the seeds and even urinating on them in order to reduce spoilage from fungi (Reichman and Rebar, 1985; Reichman et al., 1986).
The storability of a food item can greatly affect how a forager treats the item both in terms of collection and subsequent use. Reichman and colleagues have examined this issue in detail, particularly with regards to caching behavior of eastern woodrats (Neotoma floridana). Reichman (1988) examined how woodrats handled laboratory chow when it was presented along with acorns or grapes. When presented with acorns, laboratory chow was ignored by woodrats; the storability of acorns meant that it was superior both for immediate consumption and for future use. When presented with grapes, laboratory chow was stored by woodrats, and grapes were consumed. Here, the utility of an item varied with its storability and whether it was to be eaten or stored, Post and Reichman (1991) conducted a similar experiment with woodrats involving the effect of distance from burrow and presence of competitors. Again, food storability affected the foragers' behavior, with perishable foods being consumed and storable foods being collected for future consumption. Also, foraging activity close to the burrow was more likely to be devoted to caching, while the presence of competitors meant that woodrats collected perishable food first. In a similar experiment, white-footed mice (Peromyscus leucopus) in the presence of a competitor cached dry, storable food first and only cached less storable food later (Sanchez and Reichman, 1987). Finally, Gendron and Reichman (1995) demonstrated theoretically that food abundance affects the best caching strategy in regards to the collection of storable vs. perishable foods. When food is scarce, animals using a strategy where they choose foods according to future value and eat from their cache according to current value do best: when food is abundant, a "first in, first out" strategy does best, where perishable foods are gathered at first, and less perishable foods later in the season as the cache fills. These works serve to demonstrate that food storability affects current foraging behavior and future cache use.
This report provides evidence that foraging decisions are affected by more than just the maximization of the instantaneous rate of energy acquisition. The future value of a food item can shape current foraging decisions, including patch use, total foraging activity, and food selectivity (Reichman, 1988).
Acknowledgments. - We are grateful to Christopher Dunn and the staff of the Morton Arboretum who facilitated this research. We would also like to thank Ken Schmidt, James Thorson and Robert Morgan for help with field work. Zvika Abramsky, Leon Blaustein, Marc Mangel, Anthony Oldfield, James Thorson and David Ward provided useful comments on the manuscript.
ANDERSSON, M. AND J. KREBS. 1978. On the evolution of hoarding behavior. Anim. Behav., 26:707-711.
BAKER, M. C., E. STONE, A. E. BAKER, R. J. SHEEDEN, P. SKILLICORN AND M. D. MANTYCH. 1988. Evidence against observational learning in storage and recovery of seeds by black-capped chickadees. Auk, 105:492-497.
BARRY, W. J. 1976. Environmental effects of food hoarding in deermice (Peromyscus). J. Mammal., 57: 731-746.
BOWERS, M. A., J. JEFFERSON AND M. G. KUEBER, 1993. Variation in giving-up densities of foraging chipmunks (Tamias striatus) and squirrels (Sciurus carolinensis). Oikos, 66:229-236.
BROWN, J. S. 1988. Patch use as an indicator of habitat preference, predation risk, and competition. Behav. Ecol. Sociobiol., 22:37-47.
-----. 1989. Desert rodent community structure: a test of four mechanisms of coexistence, Ecol. Monogr., 59:1-22.
-----. 1992. Patch use under predation risk: 1. Models and predictions. Ann. Zool. Fenn., 29:301-309.
-----, B. P. KOTLER AND T. VALONE. 1994. Foraging under predation: a comparison of energetic and predation costs in rodent communities of the Negev and Sonoran deserts. Aust. J. Zool., 42: 435-448.
-----, R. A. MORGAN AND B. D. DOW. 1992. Patch use under predation risk: II. A test with fox squirrels, Sciurus niger. Ann. Zool. Fenn., 29:311-318.
----- AND -----. 1995. Effects of foraging behavior and spatial spatial on diet selectivity: a test with fox squirrels. Oikos. 74:122-136.
BROWN. L. G. AND L. E. YEAGER. 1945. Fox squirrels and gray squirrels in Illinois. Bull. Ill. Nat. Hist. Surv., 23:449-536.
BROWNLEE, K. A. 1965. Statistical theory and methodology in science and engineering, 2nd ed. Wiley, New York, New York.
CONNER, D. A. 1983. Seasonal changes in activity patterns and the adaptive value of haying in picas (Ochotona princeps). Can. J. Zool., 61:411-416.
EBERHARD, W. 1967. Attack behavior of diguetid spiders and the origin of prey wrapping in spiders. Psyche, 74:173-181.
ELLIOT, L. 1978. Social behavior and foraging ecology of the eastern chipmunk (Tamias striatus) in the Adirondack Mountains. Smithson. Contrib. Zool., 265:1-107.
FOX, J. E. 1982. Adaptation of gray squirrel behavior to autumn germination of white oak acorns. Evolution, 36:800-809.
FRANK, C. L. 1991. Adaptations tot hibernation in the deposit of fats of a ground squirrel (Spermophilus beldingi). Can. J. Zool., 69:2707-2711.
GENDRON, R. P., AND O. J. REICHMAN. 1995. Food perishability and inventory management: a comparison of three caching strategies. Am. Nat., 145:948-968.
JACOBS, L. 1989. Cache economy of the gray squirrel. Nat. Hist., 98:40-47.
IWASA, Y. M., M. HIGASHI AND N. YAMAMORA, 1981. Prey distribution as a factor determining the choice of optimal foraging strategy. Am. Nat., 117:710-723.
KAGEL, J. H., L. GREEN AND T. CARACO. 1986. When foragers discount the future: constraint or adaptation. Anim. Behav., 34:271-283.
KOTLER, B. P. 1997. Patch use by gerbils in a risky environment: manipulating food and satiety to test four models. Oikos, 78:274-282.
----- AND J. S. BROWN. 1990. Harvest rates of two species of gerbilline rodents. J. Mammal., 71:591-596.
KREBS, J. R., A. KACELNIK AND P. TAYLOR. 1978. Test of optimal sampling by foraging great tits. Nature, 275:27-31.
MCNAMARA, J. M. 1982. Optimal patch use in a stochastic environment. Theor. Popul. Biol., 21:269-288.
MITCHELL, W. A. 1990. All optimal control theory of diet selection: the effects of resource depletion and exploitative competition. Oikos, 58:16-2.
MORENO, J., A. LUNDBERG AND A. CARLSON. 1981. Hoarding of individual nuthatches, Sitta europaea and marsh tits Parus palustris. Holarctic Ecol., 4:263-269.
OATEN, A. 1977. Optimal foraging in patches: a case for stochasticity. Theor. Popul. Biol., 12:263-285.
POST, D. AND O. J. REICHMAN, 1991. Effects of food perishability, distance, and competitors on caching behavior by eastern woodrats. J. Mammal., 72:513-517.
REICHMAN, O. J. 1988. Caching behavior by eastern woodrats, Neotoma floridana, in relation to food perishability. Anim. Behav., 36:1525-1532.
----- AND C. REBAR. 1985. Seed preferences by desert rodents based on levels of moldiness. Anim. Behav., 33:726-729.
-----. A. FATTAEY AND K. FATTAEY. 1986. Management of sterile and mouldy seeds by a desert rodent. Anita. Behav., 34:221-225.
SANCHEZ, J. C. AND O. J. REICHMAN. 1987. The effects of conspecifics on caching behavior of Peromyscus leucopus. J. Mammal., 68:695-697.
SCHMIDT, K. A. AND J. S. BROWN. 1996. Patch assessment in fox squirrels: The rule of resource density, patch size, and patch boundaries. Am. Nat., 147:360-380.
SHAW, W. T. 1936. Moisture and its relation to the cone-storing habit of the western pine squirrel. J. Mammal., 17:337-349.
SHULL, A. F. 1907. Habits of the short-tailed shrew, Blarina brevicauda. Am. Nat., 41:495-522.
SMALLWOOD, P. AND W. D. PETERS. 1986. Gray squirrel food preferences: The effects of tannin and fat concentrations. Ecology, 67:168-174.
SMITH, C. C. AND O. J. REICHMAN. 1984. The evolution of food caching by birds and mammals. Annu. Rev. Ecol. Syst., 15:329-351.
STEPHENS. D. W. AND J. R. KREBS. 1986. Foraging theory. Princeton University Press, Princeton, New Jersey.
THOMPSON, D. C. AND P. S. THOMPSON. 1980. Food storage and caching behavior of urban gray squirrels. Can. J. Zool., 58:701-710.
TILMAN, D. 1982. Resource competition and community structure. Princeton University Press, Princeton, New Jersey.
VANDER WALL, S. B. 1990. Food hoarding in animals, University of Chicago Press, Chicago, Illinois.
----- AND R. P. BALDA. 1977. Coadaptations of the Clark's nutcracker and the pinion pine for efficient seed harvest and dispersal. Ecol. Monogr., 47:89-111.
VORHIES, C. T. AND W. P. TAYLOR. 1922. Life history of the kangaroo rat, Dipodomys spectabilis spectabilis Merriam. U.S. Dep. Agric. Bull., 1091:1-40.
WAITE, T. A. AND T. C. GRUBB. JR. 1988. Diurnal caching rhythms in captive white-breasted nuthatches (Sitta carolinensis). Ornis Stand., 19:68-70.
WEMMER, C. 1969. Impaling behavior of the loggerhead shrike, Lanius ludovicianus. Z. Tierpsychol., 26:208-224.
WHISAW, I. Q. AND J. A. TOMIE. 1989. Food-pellet size modifies the hoarding behavior of foraging rats. Psychobiol., 17:93-101.
TABLE 1. - Analysis of variance of giving-up densities (log-transformed) in food patches of squirrels foraging on sunflower seeds. The contrasts of control versus cacheable and cacheable versus noncacheable are a priori, one-tailed tests Variable SS DF MS F Group variables Treatment 15.451 2 7.725 9.78(***) Control vs. Cacheable 4.344 1 4.344 5.50(**) Cacheable vs. Noncacheable 3.572 1 3.572 4.52(*) Site 110.133 2 55.067 38.45(***) Period 48.584 6 8.097 2.92(*) Nested variables Station 8.592 6 1.432 1.81 Day 38.090 14 2.769 3.51(***) Fully crossed variable Microhabitat 52.064 1 52.064 65.90(***) Interaction effects Treatment x Microhabitat 0.024 2 0.012 0.02 Treatment x Period 11.192 12 0.933 1.18 Station x Microhabitat 23.689 8 2.961 3.75(***) Day x Microhabitat 55.915 20 2.796 3.54(***) Error 221.921 281 0.790 * P [less than] 0.05; ** P [less than]: 0.01; *** P [less than] 0.001
|Gale Copyright:||Copyright 1999 Gale, Cengage Learning. All rights reserved.|