Behavioral Ecology Vol. 15 No. 4: 666-672
Behavioral Ecology vol. 15 no. 4 © International Society for Behavioral Ecology 2004; all rights reserved
Substrate type affects caching and pilferage of pine seeds by chipmunks
Department of Biology and the Program in Ecology, Evolution and Conservation Biology, University of Nevada, Reno NV 89557, USA
Address correspondence to J. S. Briggs. E-mail: jenny{at}unr.edu
Received 3 January 2003; revised 16 September 2003; accepted 2 October 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The abiotic environment often influences the ways in which animals interact. By affecting the cues associated with buried seeds, the type of substrate used by seed-caching rodents may alter the relative probabilities of cache pilferage and cache retrieval. We predicted that, after a wildfire, the presence of ash would impair rodents' ability to smell pine seeds on the forest floor. In a laboratory experiment, we compared the foraging success, caching frequency, and cache recovery of chipmunks (six Tamias amoenus and six T. quadrimaculatus) in ash versus sand substrates. Initial results supported our hypothesis: chipmunks found only 2.3% of 108 caches of Jeffrey pine (Pinus jeffreyi) seeds that we buried in ash but found 98% of caches in sand. However, chipmunks made as many or more of their own caches in ash compared with sand (48% for T. amoenus, 73% for T. quadrimaculatus.) When foraging for seeds cached in ash by themselves and by other individuals, they found significantly higher proportions of their own caches (62%) than of caches made by others (25%). However, when foraging in sand, they found high proportions both of their own caches and those of others (86 versus 81%). These results suggest that olfaction is less effective in ash than in sand, that spatial memory enables chipmunks to recover their own caches in ash, and that caching in ash may allow animals to avoid pilferage of stored food. As chipmunks are important dispersers of seeds, changes in their foraging patterns or competitive interactions after fire could significantly affect pine regeneration.
Key words: ash, caching, chipmunks, foraging, Jeffrey pine, olfaction, pilferage, Pinus jeffreyi, spatial memory, scatter hoarding, Tamias.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The storage and recovery of food by animals is influenced by many factors in the physical and social environment (Vander Wall, 1990
). In variable environments, species that store food for later consumption demonstrate several behaviors that appear to optimize the ratio of costs and benefits associated with this strategy (Andersson and Krebs, 1978). First, they select cache sites that permit them to recover a high proportion of stored food (Vander Wall, 1990). Several studies have documented the importance, in cache site selection, of environmental factors such as vegetation structure (Haftorn, 1956; Petit et al., 1989), climate (Brotons, 2000), accessibility during winter (Vander Wall and Hutchins, 1983), and proximity to landmarks (Barkley and Jacobs, 1998; Cheng and Sherry, 1992; Collett et al., 1986; Kamil and Jones, 1997). Consideration of these factors by animals reduces spoilage, increases seasonal access, and facilitates relocation of stored items (Vander Wall and Smith, 1987). Second, animals use diverse behavioral strategies that limit pilferage by members of their own or other species (Hampton and Sherry, 1994). These include active defense (Clarke and Kramer, 1994), optimal spacing of caches (Clarkson et al., 1986; Stapanian and Smith, 1978), selection of inaccessible or cryptic sites (Petit et al., 1989; Preston and Jacobs, 2001), and avoidance of possible competitors when caching (Heinrich and Pepper, 1998; Stone and Baker, 1989).
Animals may also augment their diets, and effectively increase the adaptive value of food hoarding (Andersson and Krebs, 1978), by pilfering from the stores of others whenever possible. Inevitably, some of the strategies animals use to find unstored food and recover their own stored food are the same as those they use to discover and exploit the stores of other animals. Olfaction and directed searching, in which foraging animals focus their explorations in areas commonly used for caching, are common examples (Vander Wall, 1991). Thus, behaviors that minimize an individual's costs of recovering its own food stores (such as caching closer to the nest or selecting memorable and accessible sites) can simultaneously increase the vulnerability of the food to theft by others (Clarke and Kramer, 1994). Successfully balancing these conflicting pressures may be especially critical for scatter-hoarding birds and mammals, which make many small caches throughout their habitat instead of concentrating most or all stored food in a defensible larder. These species use several strategies–olfaction, memory, directed searching, and exploratory digging at preferred sites (Balda et al., 1987; Vander Wall, 1982, 1991)–either singly or in combination when they are looking for food, selecting cache sites, recovering caches, pilfering, and avoiding pilferage.
In this article, we focus primarily on the interplay between memory and olfaction in the placement and recovery of caches by chipmunks. We predicted that, in different environmental conditions, chipmunks would use memory and olfaction to different extents, and that shifts in their use of these foraging modes would affect their opportunities for inter- and intra-specific pilferage. Changes in abiotic factors may often cause changes in foraging and caching behavior. For example, some rodents are sensitive to variation in the moisture levels of the substrates in which they forage (Vander Wall, 1993b, 1995, 2000). In dry soils, seeds emit little or no odor, and consequently, yellow-pine chipmunks (Tamias amoenus) and deer mice (Peromyscus maniculatus) had much lower foraging success in dry substrates in both laboratory and field experiments (Vander Wall, 1993b, 2000). However, rates of pilferage increased when either the substrate or the seeds were moistened, causing them to release more odorants (Vander Wall, 2000). Spatial memory seems to play a more important role in cache retrieval in arid environments when olfaction is less effective (Vander Wall, 2000).
Although little is known about the effects of different substrate types on rodents' foraging success and caching strategies, we predicted that a change in the nature of the forest floor would also significantly affect foraging and caching behavior. In semiarid western forests, wildfires convert the forest floor to ash and alter the substrate onto which seeds fall. Not only is the moisture content of ash low, but ash may have a negative effect on the olfactory capacity of animals either by masking the smell of seeds with its own strong odor, or by absorbing organic odorants that would otherwise be released from the seeds (Vander Wall, 2003).
We tested two species of chipmunk, yellow pine (Tamias amoenus) and long-eared (T. quadrimaculatus), captured in a semiarid Jeffrey pine (Pinus jeffreyi) forest, in a laboratory experiment that compared their caching and cache-recovery behavior in ash and sand substrates. Chipmunks are important dispersers and consumers of pine seeds (Tevis, 1953; Vander Wall, 1992). Based on our general hypothesis that seeds buried in ash would be more difficult to locate using olfaction than those buried in sand, we generated several predictions. First, chipmunks will find more seeds buried in sand than in ash. Second, as ash is a less familiar and possibly more caustic substrate than is sand, as well as one in which olfaction may be less effective, chipmunks will cache fewer seeds in ash. Third, as chipmunks are able to use spatial memory to recover their own caches, they will retrieve a significantly higher proportion of their own cached seeds in ash than seeds cached by other animals in ash. However, in sand they will retrieve similar proportions of their own caches and those of others. Finally, as the habitats and diets of the two chipmunk species overlap broadly, we did not expect significant interspecific differences in caching frequency or cache recovery in either substrate.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The experiment was conducted from January–April 2002 in a laboratory at the University of Nevada, Reno, using yellow pine chipmunks and long-eared chipmunks trapped the previous fall in the Carson Range, Washoe County, Nevada. Animals were kept in an animal care facility in individual cages (25 x 25 x 50 cm) on a 12-h light/12-h dark cycle. They were given a diet of mixed wild bird seeds and rodent chow between trials. The experiment comprised five phases: (1) acclimation trials, in which subjects foraged for moistened pine seeds that we had buried in a sand-filled test arena; (2) foraging trials in sand- and ash-filled compartments of the arena; (3) caching trials, in which subjects could bury seeds in sand or ash; (4) retrieval of caches made in phase 3; and (5) foraging by naive subjects for caches made by other animals in sand and ash. All chipmunks were tested in each phase in the same order, and all the trials in one phase were completed before the next phase was initiated. Seventeen individuals were tested in total; six T. amoenus and six T. quadrimaculatus completed all five phases of the experiment.
We conducted experiments in an arena comprised of two wooden enclosures (dimensions, 88 x 232 x 30 cm) connected by a tunnel 7 cm in diameter and approximately 150 cm long. Each enclosure was divided into three compartments, separated by partitions with holes allowing animals to move freely from one compartment to another (Figure 1). The six compartments were filled to a depth of approximately 3 cm with either sand or an ash/sand mixture (75% ash), depending on the phase of the experiment. At the start of each trial, animals were placed in a nest-box connected to the arena to which they could return at any time. Water was available in two of the six compartments (hereafter called boxes), and small rocks and sticks were placed in each box to provide heterogeneity. The design of the arena presented the subjects with patches of each substrate in which they could choose to forage or make caches. The placement of the nest-box, tunnel, water, and partitions effectively "forced" subjects to traverse the different "habitats" (Figure 1) as they might if their natural habitat experienced a low-intensity, patchy, surface fire. Foraging trials lasted 11 h, and caching trials lasted 8 h.
|
Because chipmunks cannot easily detect dry seeds buried in dry substrates (Vander Wall, 1993b, 1998), we increased the moisture content of the Jeffrey pine seeds used in acclimation and foraging trials by placing seeds in a hydration chamber (100% RH) for 11 h before we buried them. The mean moisture content of the seeds (±1 SD) after hydration was 14.8 ± 1.9%, well above the threshold moisture level of around 5% at which chipmunks have been shown to detect Jeffrey pine seeds in dry substrates (Vander Wall, 1998
, 2003). We tested whether seeds retained a comparable level of moisture in both substrates throughout the foraging trials by burying sample caches in the arena, then removing two caches from each substrate at hourly intervals. After weighing, oven-drying, and reweighing the seeds, we found that seeds lost moisture at a similar rate in the two substrates. There was a sharp decline during the first hour, followed by a gradual decrease in moisture over time (mean water loss over 9 h = 7.4 ± 1.1% in ash and 6.4 ± 1.1% in sand). The moisture content of seeds in both substrates did not fall below the 5% threshold and was similar after 9 h (6.6% in ash, 6.3% in sand). Thus, moisture levels of seeds probably did not change enough to affect subjects' foraging success in either substrate. The design of the arena allowed us to record the removal and caching of seeds at the end of each trial. The bottom of each box in the arena consisted of a 6-mm wire mesh screen, stretched over a removable plywood floor. When the floor was removed, the sand or ash in the box drained through the screens into receptacles situated below. Any cached or undiscovered seeds on the screens could be counted and mapped.
Phase 1: acclimation
Before each subject was tested, we deprived it of food for 12 h to increase its motivation to forage during the trial. It was then allowed approximately 11 h to explore the arena and search for buried seeds. During this phase, each of the six boxes in the arena contained sand in which we had made three caches of two Jeffrey pine seeds at a depth of 1.5 cm. This depth is typical of chipmunk caches in the field (Vander Wall, 1993a). We also placed a single seed on the sand surface in each box to indicate to the animal that food was present in the arena.
Caches were buried in sites chosen arbitrarily (i.e. scattered relatively evenly) within each box. To minimize the possibility of accidental discovery, caches were spaced at least 30 cm apart and at least 10 cm from the edges of the box or from rocks and sticks. Previous observations of subjects in laboratory conditions have shown that they often dig near walls and objects when exploring a new, restricted environment (Vander Wall, personal observation). After making the caches, we leveled and brushed the surface of the substrate to remove traces of disturbance.
At the end of each foraging trial, the subject was removed from the arena, all cache sites were checked, and the presence or absence of the original caches was recorded. If a subject had not found any of the 36 buried seeds, it was tested a second time in exactly the same way after an interval of 1 week. If it failed to find any buried seeds in the second trial, it was dropped from the experiment.
Phase 2: foraging for caches in ash and sand
In foraging trials, the sand substrate in boxes 1, 3, and 5 was replaced with the ash/sand mixture (Figure 1b). Cache size and depth were the same as in phase 1; however, the locations of caches were different. Placement of the rocks and sticks was also altered, and no seeds were placed on the surface of the substrates. In all other respects, trials were conducted as in phase 1.
Phase 3: caching seeds in ash and sand
During caching trials, the substrate in each box was the same as in the previous phase (alternately ash and sand), but no buried seeds were present. Instead, a glass dish containing 12 Jeffrey pine seeds was placed on the substrate surface in the center of each box (72 total seeds). Animals were not deprived of food before these trials. Each animal was left in the test arena until it had removed all of these seeds or until 8 h had elapsed, whichever occurred first. We drained the sand and ash out of the boxes and recorded the number and size of all caches made in each box.
All animals completed one caching trial in early February and a second trial in early March. The two sets of trials were conducted in the same manner, but in the March trials we introduced new rocks and sticks, altered their placement in the boxes relative to the previous arrangement, and taped pictures to the compartment walls. These changes were intended to reduce the chance that subjects would search in vain for caches they had made in the first trial. Two distinct sets of objects (rocks, sticks, and pictures) were used in the March trials: half of the subjects were arbitrarily assigned to be tested when the arena contained object set A and half when it contained object set B. The results of both the February and March trials were used in the analyses of caching behavior. However, only the caches made by each subject in the March trial were reestablished in phases 4 and 5 of the experiment.
Phase 4: retrieval of caches in ash and sand by knowledgeable subjects
Phase 4 was initiated 1 week after completion of phase 3. We reestablished the object set (A or B) that had been present for each subject in its second caching trial. Each animal was deprived of food for 10–12 h and then allowed 11 h to forage for the caches it had made in that trial. These caches had been replaced in the arena in their original locations, but they differed from the original caches in several respects. To standardize the conditions experienced by each animal, each cache consisted of two hydrated seeds buried 1.5 cm deep. If an animal had made many (more than 25) caches in a box or placed caches very close to each other, some of these caches were not reestablished in an effort to standardize the numbers of caches available to each subject. Caches were eliminated if they were closer than 5 cm to another cache, but at least two-thirds of each animal's original caches in each box were retained to reduce the chance that the animal would stop foraging if it discovered that many of its caches were gone. At the end of each trial, all cache sites were checked, and the presence or absence of the seeds was recorded.
Phase 5: foraging for cached seeds in ash and sand by naive subjects
For this final phase, the visual environment of the arena was changed again. All subjects previously tested when the arena contained object set A were now exposed to object set B, and vice versa. The alterations of the landscape were again intended to suggest to animals that they should not "expect" their recent caches to be present in the boxes. In each trial, the arena in fact contained a set of caches in locations unfamiliar to the individual being tested. For all subjects, the total number of caches, and the relative proportions of caches in ash and sand, were the same as those present for that animal in phase 4. For example, a subject that had searched for five of its own caches in sand and 12 in ash in phase 4 was presented with five caches in different sites in sand, and 12 caches in different sites in ash.
For each subject, the new set of cache locations was randomly selected from the pooled set of all caches made by the individuals of the same species that had cached when the opposite object set was in place in phase 4. Thus, in contrast to the arbitrary cache sites we used in phase 2, we now buried seeds in actual cache sites previously chosen by other animals but unknown to each naive forager. However, to minimize the number of caches that naive animals could discover by using random search or previous memory rather than olfaction, the following exceptions to the randomly generated cache distributions were made: (1) caches were not placed less than 5 cm from sites previously used by the subject being tested; (2) no more than one cache per box was made near a corner, at the edge of a box, or at the edge of a rock or stick, because these sites had regularly been excavated by many subjects in previous phases; and (3) caches made in corners (common for T. quadrimaculatus) were moved 5 cm away from those corners. At the end of each trial, all cache sites were checked, and the presence and absence of seeds were recorded. Any subject (there were three) that had retrieved one cache or less during the trial was retested a week later in exactly the same way, after being deprived of food for 24 h instead of 12 h.
Analyses
T tests were used to compare the number of caches found by T. amoenus and T. quadrimaculatus in phase 1, and the size of caches made by T. amoenus and T. quadrimaculatus in phase 3. Chi-squared tests were used to determine whether each species' caching frequency in the two substrates (ash and sand) differed from the expected 50:50 distribution. For all other analyses, we used repeated-measures ANOVA. First, to compare foraging success in phases 1 and 2, we compared the number of caches found in boxes 1, 3, and 5 with the number found in boxes 2, 4, and 6 in phase 1 (when all boxes contained sand) versus phase 2 (when boxes 1, 3, and 5 contained ash). Species and box set (1, 3, and 5 or 2, 4, and 6) were the between-subjects factors, and phase of experiment (1 or 2) was the within-subjects factor.
We performed repeated-measures ANOVAs on the results of phase 3 to examine the effects of species, substrate, trial, and box on (1) mean number of caches made and (2) mean number of seeds cached. Finally, we assessed the effects of experience on foraging success by comparing the proportion (arc-sine transformed) of caches found by individuals when they were either knowledgeable or naïve of cache locations, in phases 4 and 5, respectively. Species and substrate (ash or sand) were the between-subject factors, and experience was the within-subjects factor.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Phase 1
Nine of the 12 subjects found 80–100% of the caches of moistened seeds buried in sand (Figure 2a). The other three subjects found 30% of the 18 available caches. The mean number of caches found by T. amoenus (12.3) was not significantly different from the mean number found by T. quadrimaculatus (15.4; t test, t = 1.13, df = 9, p =.29).
|
Phase 2
The success rate of animals foraging in boxes 1, 3, and 5, which contained ash (Figure 2b), was much lower than their success rate in boxes 2, 4, and 6 (sand). Overall, subjects in phase 2 only found three of the 108 caches buried in ash (2.8%), compared with 98% (106 of 108) of those in sand. Observations of footprints showed that all 12 animals had explored 80–100% of the surface of all boxes. Repeated-measures ANOVA identified a significant decrease in the mean number of caches that individuals found in phase 2 versus phase 1 (F1,10 = 165.97, p <.0001). The number of caches found in the two substrates was also significantly different (F1,10 = 14.15, p =.004), and there was a significant interaction between substrate type and phase of experiment (F1,10 = 417.65, p <.0001), which showed that the number of caches found in boxes 1, 3, and 5 differed between phase 1 and 2 (when the sand was replaced with ash), but the number found in boxes 2, 4, and 6 (sand in both phases) did not differ. Foraging success of the two species was not significantly different (F1,10 = 2.62, p =.14).
Phase 3
Repeated-measures ANOVA did not identify a significant overall difference between the number of caches made in trial 1 and trial 2 (F1,10 = 0.22, p =.647), nor were the interactions between trial and substrate or species significant, so data from trials 1 and 2 (Figure 3a,b) were pooled for analyses. Also, no significant box effect was identified when box was used as a blocking factor (F5,50 = 1.31, p =.27), although an interaction between box and trial was nearly significant (F5,50 = 2.36, p =.053).
|
The mean number of caches placed in ash was significantly greater than the mean number placed in sand for all individuals combined (F5,50 = 7.35, p =.02). Ten of the 12 subjects (four T. amoenus and all T. quadrimaculatus) placed most of their caches in ash substrate (Figure 2a). The mean number of caches made by T. amoenus (33.8 ± 26.7 caches in ash and 36.8 ± 17.9 in sand) was significantly greater than that made by T. quadrimaculatus (18.7 ± 12.25 in ash and 6.8 ± 7.7 in sand; F1,10 = 12.30, p =.006). ANOVA identified no difference in the two species' use of ash and sand substrates (species x substrate interaction: F5,50 = 0.34, p =.88). However, chi-squared analyses showed that for T. quadrimaculatus, the distribution of caches between the two substrates was significantly different (
2 = 32.95, df = 1, p <.001), whereas the distribution was not significantly different for T. amoenus (2 = 0.77, df = 1, p >.25). The analysis using number of seeds cached (rather than number of caches) as the response variable showed that T. quadrimaculatus and T. amoenus cached similar numbers of seeds overall (344 versus 343, respectively, in trial 1; 297 versus 350 in trial 2; F1,10 = 0.67, p =.43). This suggests that the larger size of caches made by T. quadrimaculatus was responsible for the significant difference between species in mean number of caches made (Figure 3a,b). In fact, the mean cache size was significantly greater for T. quadrimaculatus (6.1 ± 2.1 seeds) than for T. amoenus (1.5 ± 0.2 seeds; t = 5.52, df = 10, p <.001). There was a significant effect of substrate on number of seeds cached (F1,10 = 14.77, p =.003). This ANOVA identified a significant effect of box on number of seeds cached (F5,50 = 4.87, p =.001) as well as an interaction between box and species (F5,50 = 3.02, p =.02). Also, although the main effect of trial was not significant (F1,10 = 2.78, p =.13), there was a significant interaction between trial and substrate (F1,10 = 8.30, p =.02), showing a change in number of seeds placed in the two substrates in the two trials.
Phases 4 and 5
A significant experience x substrate interaction (F1,10 = 14.63, p =.003) revealed that knowledge of cache sites increased foraging success in the ash substrate only. In sand, knowledgeable versus naive subjects of both species found similar mean proportions of caches (81% versus 72% for T. quadrimaculatus, 91% versus 90% for T. amoenus) (Figure 4). Thus, the analysis revealed no main effects of experience (F1,10 = 3.33, p =.10) or species (F1,10 = 2.99, p =.11). However, the main effect of substrate was highly significant (F1,10 = 61.82, p <.0001): subjects found a lower proportion of caches in ash compared to sand (Figure 4). The interaction between substrate and species was also highly significant (F1,10 = 27.55, p =.0004), reflecting the greater overall success rate achieved by T. quadrimaculatus versus T. amoenus when foraging in ash (Figure 4). Knowledgeable T. quadrimaculatus retrieved more (91% on average) of their own caches made in ash than did knowledgeable T. amoenus (33%). The difference between the two species in mean percentage of caches found in ash by naïve animals was also great (37% versus 13%).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results of phases 1 and 2 showed that, as predicted, chipmunks found significantly fewer caches in ash compared with sand, and that T. amoenus and T. quadrimaculatus did not differ in foraging success during these phases. However, the results of caching trials in phase 3 caused us to reject our hypothesis that animals would prefer to make caches in sand. Ten of the 12 subjects placed most of their caches (from 56–95%) in ash. Thus, we rejected the possibility that the results of phase 2 could be explained by animals' avoidance of ash. Rather, the high frequency of caching in ash, combined with the low foraging success in ash, suggested that we might have identified an unexpected process.
In designing phases 4 and 5, we sought to distinguish between four alternative explanations for the results of phases 2 and 3. First, animals might not have foraged actively in the ash boxes because previous experience in the acclimation trials had exposed them only to seeds buried in sand. This possibility could be discounted in phases 4 and 5, because in these trials subjects foraged for their own caches, many of which they had made in ash. Although there was considerable individual variation, all subjects made some caches in ash (more than six for 11 of 12 subjects). Thus, if a "training effect" had influenced the results of phase 2, it was no longer relevant in phases 4 and 5.
Second, animals might not have foraged actively in ash boxes because they were satiated after consuming the seeds they had found in sand. This seemed unlikely, because in phase 2, on average, animals stopped finding caches before they had consumed the same number of seeds as in the earlier trial. A third possibility was that animals did not find seeds in ash because the caches were made by humans at arbitrary locations, not in sites where chipmunks would actually cache and forage. This may have affected the results, but in phases 4 and 5 we avoided this issue by presenting the animals with caches in sites that they or others had actually used.
A fourth explanation for our results seemed the most likely: animals could not smell the seeds we had buried in ash, but when caching they treated the ash compartments as relatively secure locations for their own caches, which they could relocate using spatial memory. All the predictions that we generated from this hypothesis were supported by the results of phases 4 and 5. Animals foraging in ash found significantly more of their own caches than caches of which they had no knowledge, suggesting that memory was important in relocation. In sand, their success finding known and unknown caches of moistened pine seeds did not differ significantly, presumably because olfaction could be effectively used in these conditions and memory did not contribute significantly to successful retrieval (Vander Wall, 2000). In addition, a significantly higher proportion of caches overall was found in sand compared with ash, indicating that seeds placed in ash are harder to detect by both knowledgeable and naive foragers, and may be less vulnerable to pilferage. Inspection of the ash boxes after trials indicated that although many naive animals had searched extensively in certain areas (for example, along the edges, corners, and objects in these boxes) they had but often failed to locate seeds 1–5 cm away from their excavations. Taken together, these results support the hypothesis that memory is the key factor allowing animals to retrieve seeds buried in a substrate that masks the olfactory signal of the seeds, thereby making ash a more secure location for the placement of caches.
The results of phase 5 led us to reject the hypothesis that the two species would have similar foraging success. T. quadrimaculatus found a higher proportion of both known and unknown caches in ash, and individuals often appeared to direct their searches (Vander Wall, 1991), whether successful or not, to the areas in which many caches were made by this species: near objects, corners, and walls of boxes. However, the conclusion that T. quadrimaculatus was significantly more successful overall may be erroneous, as the results reflect the unusual behavior of two individuals. During their second foraging trials, both subjects systematically excavated all six boxes and found 92% and 67%, respectively, of the caches present in the ash. This behavior resembled exploratory digging (Vander Wall, 1982), did not appear to involve the use of olfaction and was very different from that of the other animals in the study as well as from their own behavior in earlier phases.
As subjects did not place all of their caches in ash in phase 3 or retrieve all of their ash caches in phase 4 when they had knowledge of the cache sites, we can infer that there may be a cost associated with retrieval in ash. Studies of caching and retrieval in captive Clark's nutcrackers led Bednekoff and Balda (1997) to speculate that the costs of different choices affect the performance levels of the birds in different situations. Initially, nutcrackers retrieved caches most accurately early in trials and were less accurate in later retrieval attempts (Kamil and Balda, 1990). However, when costs of retrieval were imposed by experimenters, the nutcrackers made fewer errors (Bednekoff and Balda, 1997). For our subjects, olfaction is probably the most effective foraging technique under certain conditions. Presumably, less precise memory of the location of a cache is required in sand, because olfaction can be used to assist memory. Bednekoff and Balda's (1997) results suggest that if we increased the cost of caching in sand (e.g., by increasing pilferage there), animals might place higher proportions of caches in ash even though those caches are more difficult to retrieve. Alternatively, if we allowed animals to look only for the seeds they had buried in ash, their accuracy of retrieval in this substrate should increase. Tests of these possibilities could predict the response to pilferage in field situations. We anticipate that the frequency of pilferage itself would be lower in ash but realize that a direct test is necessary to confirm this. Cachers may leave visual and olfactory cues at cache sites that were eliminated by the methods in the present study but might assist potential pilferers to find the buried seeds.
Studies of pilferage in black-capped chickadees (Baker and Anderson, 1995; Hampton and Sherry, 1994) have shown the birds stop using cache sites that are regularly robbed. Merriam's kangaroo rats (Dipodmys merriami) responded to pilferage by increasing the degree to which they larder-hoarded, rather than scatter-hoarded, food (Preston and Jacobs, 2001). Emery and Clayton (2001) demonstrated that laboratory-raised western scrub jays (Aphelocoma californica) altered their caching behavior in response to their own experience of pilfering from others plus their own memory of being watched while caching. Although we tested individuals one at a time, we used adult animals trapped in the wild that presumably had considerable experience caching and pilfering seeds. Thus, although we did not directly test the effects of social context on caching behavior, our subjects' actions may reflect an "awareness" of the competitive environment in the field.
Changes in the dynamics of seed caching and consumption may affect pine regeneration after fire. Our results suggest that after a fire, chipmunks' ability to find seeds in an ashy substrate is significantly reduced. The most likely explanations for this effect are that odorant molecules bind to ash or that the odor of ash masks the odor of seeds (Vander Wall, 2003). Thus, although seeds may readily be buried in ash, either by abiotic processes or by caching, animals may not recover or consume them to the same degree that they do in unburned forest. However, they may prefer to cache the seeds that they obtain elsewhere in burned substrates, if pilferage in these areas is reduced. Fires expose areas of mineral soil that represent favorable microsites for the germination and survival of seeds of several conifer species, including Jeffrey pine (Kauffman and Martin, 1998; Kilgore, 1973), and convert organic matter to patches of deep ash on the forest floor, which contain higher levels of many nutrients in a readily accessible form (Blank and Zamudio, 1998; Johnson et al., 1998). Thus, the decisions of individual chipmunks to cache, pilfer, or forage for seeds in the mosaic of different substrates present after a fire may affect the success of establishment and the subsequent rate of growth of seedlings.
Other studies have also suggested that the behaviors and interactions of seed-caching animals may have important implications for plant succession. Blue jays (Cyanocitta cristata) that cached nuts of several oak species in a woodland/prairie landscape in Iowa (Johnson et al., 1997) exhibited a significant preference for cache sites in burned areas after a fire. Johnson et al. (1997) speculated that decreases in populations of seed predators in burned areas served to protect jays' stores there. In a prairie-forest ecotone in Kansas, fox squirrels (Sciurus niger) caching walnuts (Juglans nigra) also appear to balance the risks of predation and pilferage such that, in poor mast years, their activities facilitate the dispersal of walnut trees onto the prairie (Stapanian and Smith, 1986). As the plants that rodents and corvids disperse (e.g., oaks, walnuts, and pines) are often considered to be late successional species, the caching decisions of individual animals can serve to hasten the process of plant succession.
|
ACKNOWLEDGEMENTS |
|---|
We thank Jennifer Armstrong for her assistance in the laboratory and Maurie Beck and Michael Sears for their help with statistical analyses. The design of the experiment was improved by discussion with members of the Biology Department and Program in Ecology, Evolution and Conservation Biology at U.N.R., in particular Maurie Beck, Jennifer Hollander, Ted Thayer, and Michael Sears. Funding from the Nevada Biodiversity Initiative and National Science Foundation Grant DEB-9708155 contributed to the present study. We thank the Whittell Forest for permission to trap subjects in Little Valley. We appreciate the comments of David Westneat, Stephen Jenkins, Rich Briggs, Kellie Kuhn, and an anonymous reviewer on early versions of the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Andersson M, Krebs J, 1978. On the evolution of hoarding behaviour. Anim Behav 26:707-711.[CrossRef]
Baker MC, Anderson P, 1995. Once-pilfered sites not avoided by black-capped chickadees. Anim Behav 49:771-785.
Balda RP, 1980. Recovery of cached seeds by a captive Nucifraga caryocatactes. Z Tierpsychol 52:331-346.
Balda RP, Bunch KG, Kamil AC, Sherry DF, Tomback DF, 1987. Cache site memory in birds. In: Foraging behavior (Kamil AC, Krebs JR, Pulliam HR, ed). New York: Plenum Press.
Barkley CL, Jacobs LF, 1998. Visual environment and delay affect cache retrieval accuracy in a food-storing rodent. Anim Learning Behav 26:439-447.
Bednekoff PA, Balda RP, 1997. Clark's nutcracker spatial memory: many errors might not be due to forgetting. Anim Behav 54:691-698.[CrossRef][Web of Science][Medline]
Blank RR, Zamudio DC, 1998. The influence of wildfire on aqueous-extractable soil solutes in forested and wet meadow ecosystems along the eastern front of the Sierra-Nevada range, California. Int J Wildl Fire 8:79-85.
Brotons L, 2000. Individual food-hoarding decisions in a nonterritorial coal tit population: the role of social context. Anim Behav 60:395-402.[CrossRef][Web of Science][Medline]
Cheng K, Sherry DF, 1992. Landmark-based spatial memory in birds (Paris atricapillus and Columba livia): the use of edges and distance to represent spatial positions. J Comp Psych 106:331-341.[CrossRef]
Clarke MF, Kramer DL, 1994. The placement, recovery and loss of scatter hoards by eastern chipmunks, Tamias striatus. Behav Ecol 5:353-361.
Clarkson K, Eden SF, Sutherland WJ, Houston AJ., 1986. Density dependence and magpie food hoarding. J Anim Ecol 55:111-121.
Collett TS, Cartwright BA, Smith BA, 1986. Landmark learning and visuo-spatial memory in gerbils. J Comp Physiol A 158:835-851.[CrossRef][Medline]
Emery NJ, Clayton NS, 2001. Effects of experience and social context on prospective caching strategies by scrub jays. Nature 414:443-446.[CrossRef][Medline]
Haftorn S, 1956. Contributions to the food biology of tits especially about storing of surplus food. Part II. The coal-tit (Parus aterL.). Det Kongelige Norske Videnskabers Selskabs Skrifter 2:1-52.
Hampton RR, Sherry DF, 1994. The effects of cache loss on choice of cache sites in black-capped chickadees. Behav Ecol 5:44-50.
Heinrich B, Pepper JW, 1998. Influence of competitors on caching behavior in the common raven, Corvus corax. Anim Behav 56:1083-1090.[CrossRef][Web of Science][Medline]
Johnson DW, Susfalk RB, Dahlgren RA, Klopatek JM, 1998. Fire is more important than water for nitrogen fluxes in semi-arid forests. Env Sci Pol 1:79-86.[CrossRef]
Johnson WC, Adkisson CS, Crow TR, Dixon MD, 1997. Nut caching by blue jays (Cyanocitta cristata L.): implications for tree demography. Am Midl Nat 138:357-370.[CrossRef]
Kamil AC, Balda RP, 1990. Differential memory for cache sites by Clark's nutcracker (Nucifraga columbiana). J Exp Psych: Anim Behav Proc 16:162-168.[CrossRef]
Kamil AC, Jones JE, 1997. The seed-storing corvid Clark's nutcracker learns geometric information about landmarks. Nature 390:276-279.[CrossRef]
Kauffman JB, Martin RE, 1988. Fire behavior, fuel consumption, and forest-floor changes following prescribed understory fires in Sierra Nevada mixed conifer forests. Can J For Res: (??)455–462.
Kilgore BM, 1973. The ecological role of fire in Sierran conifer forests. Quat Res 3:496-513.[CrossRef]
Petit DR, Petit LJ, Petit KE, 1989. Winter caching ecology of deciduous woodland birds and adaptations for protection of stored food. Condor 91:766-776.
Preston SD, Jacobs LF, 2001. Conspecific pilferage but not presence affects Merriam's kangaroo rat cache strategy. Behav Ecol 12:517-523.
Stapanian MA, Smith CC, 1978. A model for seed scatterhoarding: coevolution of fox squirrels and black walnuts. Ecology 59:884-896.[CrossRef]
Stapanian MA, Smith CC, 1986. How fox squirrels influence the invasion of prairies by nut-bearing trees. J Mammal 67:326-332.[CrossRef]
Stone ER, Baker MC, 1989. The effects of conspecifics on food caching by black-capped chickadees. Condor 91:886-890.
Tevis L, 1953. Stomach contents of chipmunks and mantled squirrels in northeastern California. J Mammal 34:316-324.[CrossRef]
Vander Wall SB, 1982. An experimental analysis of cache recovery in Clark's nutcracker. Anim Behav 30:84-94.[CrossRef]
Vander Wall SB, 1990. Food hoarding in animals. Chicago: University Chicago Press.
Vander Wall SB, 1991. Mechanisms of cache recovery by yellow pine chipmunks. Anim Behav 41:851-863.[CrossRef]
Vander Wall SB, 1992. The role of animals in dispersing a "wind-dispersed" pine. Ecology 73:614-621.[CrossRef]
Vander Wall SB, 1993a. Cache site selection by chipmunks (Tamias spp.) and its influence on the effectiveness of seed dispersal in Jeffrey pine (Pinus jeffreyi). Oecologia 96:246-252.[CrossRef]
Vander Wall SB, 1993b. Seed water content and the vulnerability of buried seed to foraging rodents. Am Midl Nat 129:272-281.[CrossRef]
Vander Wall SB, 1995. Influence of substrate water on the ability of rodents to find buried seeds. J Mammal 76:851-856.[CrossRef]
Vander Wall SB, 1998. Foraging success of granivorous rodents: effects of variation in seed and soil water on olfaction. Ecology 79:233-241.[CrossRef]
Vander Wall SB, 2000. The influence of environmental conditions on cache recovery and cache pilferage by yellow pine chipmunks (Tamias amoenus) and deer mice (Peromyscus maniculatus). Behav Ecol 11:544-549.
Vander Wall SB, 2003. How rodents smell buried seeds: a model based on the behavior of pesticides in soil. J Mammal 84:1089–1099.
Vander Wall SB, Hutchins HE, 1983. Dependence of Clark's nutcracker, Nucifraga columbiana, on conifer seeds during the post-fledging period. Can Field-Nat 57:208-214.
Vander Wall SB, Joyner JW, 1998. Recaching of Jeffrey pine (Pinus jeffreyi) seeds by yellow pine chipmunks (Tamias amoenus): potential effects on plant reproductive success. Can J Zool 76:154-162.[CrossRef]
Vander Wall SB, Smith KG, 1987. Cache-protecting behavior of food-hoarding animals. In: Foraging behavior (Kamil AC, Krebs JR, Pulliam HR, eds). New York: Plenum Press.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||