inhibits feeding by beaver - Springer Link

Journal of Chemical Ecology, Vol. 20, No. 8, 1994

RED MAPLE (Acer rubrum) INHIBITS FEEDING BY BEAVER (Castor canadensis)

D. M/0LLER-SCHWARZE, * BRUCE A. SCHULTE, LIXING ANNETTE M(ILLER-SCHWARZE, and CHRISTINE

SUN,

M/~ILLER-SCHWARZE

College of Environmental Science and Forestry State University of New York Syracuse, New York 13210 (Received February 1, 1994; accepted March 29, 1994) Abstract--At many beaver (Castor canadensis) sites at Allegany State Park in New York State, red maple (Acer rubrum) is the only or one of the few tree species left standing at the ponds' edges. The relative palatability of red maple (RM) was studied in three ways. (1) At seven beaver sites, the available and utilized trees were recorded and an electivity index (E) computed. Of 15 tree species, RM ranked second or fourth lowest. (2) In experiment I, RM, sugar maple (A. saccharum, SM), and quaking aspen (Populus tremuloides) logs were presented cafeteria style at 10 colonies. RM was the least preferred. (3) Bark of RM was extracted with solvents. Aspen logs were painted (experiment II) or soaked (experiment III) with this RM extract and presented to beaver cafeteria-style, along with aspen and RM controls. This treatment rendered aspen logs less palatable, indicating that a chemical factor had been transferred.

Key Words--Acer rubrum, Adirondacks, aspen, beaver, Castor canadensis, feeding inhibition, Populus tremutoides, red maple.

INTRODUCTION

Beaver (Castor canadensis) at Allegany State Park (ASP) in New York State have clear-cut the areas around many of their ponds. The last remaining single mature deciduous trees are often red maple (Acer rubrum), even though its wood classifies it as a "soft maple" in contrast to sugar maple (A. saccharum), a *To whom correspondence should be addressed. 2021 0098~)33 ]/94/0800-202 [ $07.00/0 © 1994 Plenum Publishing Conporation

2022

MOLLER-SCHWARZE ET AL.

"hard maple." Beaver prefer trees with soft wood. It is possible that red maple (RM) repels herbivores chemically. If so, the antifeedant compound(s) would be candidates for an effective beaver repellent. Ever-increasing beaver populations and damage complaints call for a potent chemical repellent, among other control measures. For these reasons, we undertook a study of food preferences by beaver at ASP. This included a survey of available and selected trees at several beaver sites, cafeteria-style palatability experiments at beaver ponds, chemical extraction of RM, and palatability experiments with RM extracts transferred to logs of the most preferred tree, quaking aspen (Populus tremuloides).

METHODS AND MATERIALS

This study consisted of three steps: (1) Tree preferences by beaver were determined by recording the number of trees available and those utilized. (2) Relative palatability was determined by cafeteria-style choice experiments at beaver ponds. (3) RM bark was extracted with solvents, and the extract transferred to aspen logs. These RM-extract-treated aspen logs were presented to free-living beaver at their ponds in cafeteria-style choice experiments. Relative Utilization of Red Maple by Beaver. The number of intact trees and stumps left by beaver were recorded for each tree species at three sites on December 29 and 30, 1984, and at four different sites on February 19 and 20, 1985. We sampled around the ponds up to 60 m from the water. This is how far from the pond there were signs of beaver foraging. The number of felled trees of each species was related to the number of originally available trees (intact trees plus stumps) by computing the electivity index, E (Jacobs, 1974), for each species j:

Ej = In

(rj) (1 - pj) (pj) (1 - rj)

where r = number of utilized trees of a given species, and p = number of available trees of that species. An E larger than zero represents preference, smaller than zero avoidance. Values near zero mean a species is taken in proportion to its abundance. The significance of E is determined with a chi-square test (Jenkins, 1979). Choice Experiments, Untreated Logs. For experiment I, branches from quaking aspen (Populus tremuloides, QA), sugar maple (Acer saccharum) (SM), and RM were cut into 20-cm-long (average diameter 2.71 + 0.72 cm) sections. These logs were placed at 10 beaver ponds in the evening, 12 at a time (four replications of each species). Preferences for untreated logs were tested twice: at ASP on March 12-15, 1985, and at Cranberry Lake Biological Station (CLBS) in the Adirondacks on July 3-6, 1985. At CLBS, yellow birch (Betula alle-

RED MAPLE INHIBITS BEAVER FEEDING

2023

gheniensis, YB) was added, because it is frequently utilized by beaver there. Equal numbers of logs were used; four for each of the three tree species at ASP, and three for each of the four species at CLBS. The logs were placed at the water's edge near the lodge and arranged 20 cm apart, parallel, in a row in predetermined random sequence. On the following day, the logs were examined for beaver activity. They were classified as removed, fully or partially peeled, sampled, or left intact. Most logs were still at the feeding site, while some were floating nearby or had drifted to the dam. Some were found one or several days later, after beaver had first taken them to the lodge and then discarded the peeled logs. Solvent Extraction of Red Maple. Winter-dormant RM twigs were harvested in ASP from the areas within foraging range of beaver. They were collected and transported to the laboratory in freezing temperatures, and then kept at - 10°C in a walk-in freezer. Still frozen, they were cut into short (20-cm) pieces and shredded in a Wiley mill. The resulting fine chips were placed in 2-liter amber glass jars and soaked in solvent. In the first experiment, five solvents of different polarities were used. These were methanol, ethyl acetate, acetone, ethyl ether, and hexane. The solvent was removed from the chips in a Buechner funnel. For every gram of wood chips, approximately one gram of extract was obtained. In the second experiment, only methanol was used as a solvent. An average log had about 15 g of bark. The original extract was concentrated so that 5 ml contained material from 15 g of bark, the amount to be transferred to each experimental log, as described below. Choice Experiments, Treated Logs. For experiment II, aspen logs painted with RM extract were bioassayed three times at ASP (April 22-26, May 2425, and November 27-30, 1985) and once at CLBS (July 8-9, 1985). As controis, untreated RM and aspen logs and solvent-treated aspen logs were used. Twelve logs per night were placed at each pond. For experiment UI, the three phases took place on April 6-10, 1991, April 7-12, 1992, and April 17-21, 1993. Based on our experience, experiment HI differed from experiment II in three ways: (1) The logs were punctured by rolling them over a nailboard to increase the amount of extract they absorbed. These perforated logs were dried for 2 hr, soaked with concentrated extract or solvent (controls) for 2-3 hr in flat trays, and periodically turned over. Five milliliters of extract was used per log, representing the extract from 15 g bark. (2) After drying for 2-3 hr at ambient temperature, 10 logs each were nailed, evenly spaced 20 cm apart, to a 2-m long pole, so that the beavers removed only logs they were motivated to feed on. In 1991, two such 10-log arrays were placed, tied to two pegs, at each of four ponds as seen in Figure 1. In 1992 and 1993, one array each was placed at 10 sites. (3) Extracts were made from different types and parts of RM trees: In 1991 we used 10 different treatments. These included five different RM extracts: from RM trunk base (beaver area), twigs

2024

MULLER-SCHWARZE ET AL.

Fit3. 1. (A) An array of the 10 treated samples, placed at the edge of an active beaver pond. (B) An array of samples on the day following its placing, after the beaver have made their selections.

2025


TABLE

1.

SAMPLES

TESTED

DURING

Controls Aspen, intact (Ai) Aspen, punctured (Ap) Aspen, EtOH soaked (Ae) Red maple intact (Ri) Red maple punctured (Rp) Red maple, EtOH soaked (Re) Scots pine, intact (Si) Extracts tested Extract of RM (juvenile regrowth) on aspen (Arj) Extract of RM base on aspen (Arb) Extract of RM mature twigs on aspen (Art) Extract of aspen twigs on Aspen (Aat) Extract of RM juvenile regrowth on RM (Rrj) Extract of aspen on RM (Ra)

1991, 1992, A N D

1993 ~

1991

1992

1993

x x x x x

X

X

X

X

X

X

X

X

X

X

X

X

x x x x

"Carrier logs: A = aspen; R = red maple; S = Scots pine. Extract treatments: a = aspen; b = tree base; e - EtOH; i = intact; j = juvenile regmwth (after cutting by beaver); p = perforated; r = red maple; t = twigs from mature tree. o f mature trees from two b e a v e r colonies, and one upland site b e y o n d the foraging range o f beaver, and j u v e n i l e regrowth after b e a v e r cutting. T h e r e m a i n i n g five samples were controls, including intact aspen, bark-punctured aspen, e t h a n o l - s o a k e d aspen, intact R M , and punctured R M . In 1992 we extracted R M tree base bark, j u v e n i l e regrowth after cutting by beaver, and twigs from mature trees at b e a v e r sites. R M logs treated with aspen twigs extract were added as a new sample to test w h e t h e r palatability o f an undesirable species can be increased by a c h e m i c a l factor from a preferred species. Because the different R M extracts did not h a v e different effects in 1991 and 1992, only o n e type o f R M extract (juvenile regrowth, b e a v e r site) was used in 1993 and applied to b o t h aspen and R M . A s p e n extract was also applied to R M and aspen. T h e r e m a i n i n g six samples were controls: aspen and R M , each intact, perforated, or E t O H - s o a k e d . T h e aspen logs for 1991 a v e r a g e d 2.71 + 0.72 c m in d i a m e t e r (N = 80) and for 1993, 2.58 c m (162 c m 2 surface, N = 30) and the R M logs in 1993, 2.73 c m ( 1 7 1 . 4 c m 2, N = 30). F o r c o m p a r i s o n , on one aspen stand, saplings not used by b e a v e r averaged 0.89 + 2.58 (SE) c m (range 0 . 4 - 1 . 8 cm, N = 12) in diameter, while those that were beaver-cut averaged 1.95 + 1.88 c m (range 0 . 9 - 3 . 1 cm; N = 12). In a n o t h e r study, b e a v e r first cut experimental saplings ranging b e t w e e n 3.1 and 4 c m diameter, d e p e n d i n g on distance (Fryxell and Doucet, 1993). B e a v e r that had m o v e d into a new area preferred a s p e n o f 1 . 5 - 7 . 5 c m d i a m e t e r (Basey et al., 1988).

2026

M O L L E R - S C H W A R Z E ET AL.

Statistically, each 10-1og array constituted a block. Because samples were either eaten or avoided (including untouched and sampled), and rarely halfway peeled, nonparametric tests were used. Lost logs were not included in the analysis, even though almost all that were found later had been peeled. Multiple samples were compared by Cochran Q test, and samples were compared pairwise by the McNemar's test. RESULTS

Relative Utilization of Red Maple. The results of two surveys of available and utilized trees are presented as electivity indices in Figures 2 and 3. In the first survey (three beaver sites) RM ranks fourth lowest in preference. Only hawthorn (Crataegus sp.), white pine (Pinus strobus), and Scots pine (P. sylvestris) rank still lower (Figure 2). In the second survey, RM ranks second lowest, between Norway spruce (Picea abies) and Scots pine (Figure 3). In general, the rankings agree with food preferences known for beaver from other studies (e.g., Fryxell and Doucet, 1993). Small sample sizes--sometimes only Elecflvity I n d e x -5

-4 |

-3 I

-2 i

AH" AS" JU" MA" HP" YBBC" BE" WHMUSM" RM"

HAWP" SP"

-I I

0

t |

2 I

12(**)

7(..) 49(**) 36 (**) 9 (ns)

29 t") 16 (*) 9 (n.) 44(n.)

Tree

Species

Sl (~1 21 (**) I'-:~:r:~::-'r: 84 (**)

I:':';':+:+:':':':q!:!:!ir:!;':] I:'~:-;-:':-:-:':'!

4 (**) 32 (**)

FIG. 2. Utilization of 15 tree species by three beaver colonies at Allegany State Park, New York, fall 1984, expressed by the electivity index E. Bars on right: preferred trees; bars on left: trees selected against. Significance levels: * = 0.05; ** = 0.01 (X2 test). Species abbreviations: AH: ash (Fraxinus sp.); AS: aspen (Populus tremuloides); JU: Juneberry (Amelanchier sp.); MA: maple (Acer sp.); HP: hop hornbeam (Ostrya virginiana); YB: yellow brich (Betula lutea); BC: black cherry (Prunus serotina); BE: beech (Fagus grandifolia); WH: witchhazel (Hamamelis virginiana); MU: musclewood (Carpinus caroliniana); SM: sugar maple (Acer saccharum); RM: red maple (Acer rubrum); HA: hawthorne (Crataegus sp.); WP: white pine (Pinus strobus); SP: Scots pine (Pinus

sylvestris).

2027


Elec//vity Index -5

-4

I

-3

I

-2

I

-I

i

0

i

I

i

2

i

3

I

AP" JU" WIBC" AS" 43 (**) Iq;:::;:;:::::::;:;:l BE" 151 (**) [;::::::::zl MU" T r e e HP" Species 8 ~ (*') SM" 7S (**) RP" t::;.'::::::::;:l 3 (us) HE" ::::::::::::::::::::::::: 6 (**) YBNS" ~:~:r:~:~:~:~.'!:r:~:~',:.:~:~:~:~!~il 22 (**) RMSP" :::::::::::::':':::::::::::::::::::::::::::::::':::::::sn(**)

I (us) 1 Ca.) 1 (w) 43 (**) 7 (**) [;5:;:;5:-:~:5:;:;:1

FIG. 3. Utilization of 15 tree species by four beaver colonies at Allegany State Park, New York, winter 1984/85. Species Abbreviations same as Figure 2 with these new species: AP: apple (Pyrus malus); WI: willow (Salix sp.); RP: red pine (Pinus resinosa); HE: eastern hemlock (Tsuga canadensis); NS: Norway spruce (Picea abies). one tree or stump of one species--result in nonsignificant rankings. In Figure 3, for instance, the significant preferences series starts with black cherry (fourth bar). Experiment I: Choice Experiments, Untreated Logs. In the first choice test at ASP, beaver consumed 94.7% of QA, 76.3% of SM, and 46.3% of RM. The preference is significant (X 2 = 45.9; df = 6; P < 0.001). The second choice test, at CLBS, took place in July when beaver take little woody vegetation. Of the four tree species offered, QA was taken most often, but only 38.9% of the logs were utilized (YB: 13.9%; SM: 11.1%). RM was never taken. The preference is significant (X 2 = 21.68; df = 6; P < 0.01). Experiment II: Pilot Tests with Extract Painted Logs. First, three pilot tests were performed: In the first test, in April 1985, the painting of aspen logs with RM had no effect on consumption. Aspen, solvent-painted aspen, and RM-painted aspen were eaten equally often (80.6%, 86.8%, and 82.6%, respectively). Only intact RM was consumed less (58.3%). In the second test, in May 1985, aspen logs painted with RM extract were consumed as often (55%) as control logs painted with solvent (57.5%) or intact aspen (60%). Intact RM, however, was taken significantly less than the other three samples (X 2 = 9.269; df = 2; P < 0.01). Thus, the painting had no effect. For the third test, in November 1985, the percentages eaten were 42.5%, 40%, 36.6%, and 22.5% respectively, for aspen, solvent control, RM-treated aspen, and intact RM. Still, no significant differences were found. Likewise, in the test at CLBS in July 1985, painting logs did not reduce their palatability.

2028

MULLER-SCHWARZE ET AL.

Experiment IlL Choice Tests with Extract-Soaked Logs. The second series with treated logs at ASP covered the three spring seasons 1991-1993. In 1991, a total of 320 logs were placed. Eight arrays of 10 logs each were never touched and therefore not counted. Of the remaining 240 logs, 193 were recovered. Another 47 had been carried away and remained missing. (Most missing logs found later were peeled.) Beaver consumed 80.0%, 60.0%, and 32.0%, respectively, of the three key samples (N = 25 for each)--intact aspen (Ai), aspen treated with extract from juvenile regrowth of red maple at a beaver site (Arj), and intact red maple (Ri). These differences were significant (Cochran Q test: Q = 14.53; P < 0.001, one-tailed). Samples Ai and Arj, compared by the McNemar test, were not significantly different, due to a small sample size. The three extracts of red maple (r) twig and tree base did not differ in their effects. Therefore, they were combined into the Ar category (aspen logs treated with red maple extract). For a larger sample size, all 1991 RM extract samples (Ar; N = 104) were compared first to all three types of control aspen (Ac; N = 75) including intact, punctured, and soaked in EtOH, then to all red maple controls (Rc; N = 48) including intact and punctured (Rp), independent of the original blocks. Ac were consumed most often (73.6%), and Rc least often (37.5%). Ar were intermediate (57.7%). They were significantly more avoided than Ac (X 2 = 4.02; P < 0.05, two-tailed), and eaten significantly more often than Rc (X 2 = 4.58; P < 0.05; two-tailed). In 1992, the same treatments were repeated, and an aspen extract (Aa) added to test whether a "super-aspen" could be created. First, the four treatments with expected extreme results were compared by Cochran Q test. These were Aa, Ap, Ar, and Rp (N = 35 for each treatment). The differences in this multiple comparison were highly significant (X 2 = 13.47; df = 3; P < 0.001, two-tailed). In a matched two-sample comparison, Ar was avoided more than Ap (McNemar test, X 2 = 3.273; P < 0.05; one-tailed). Ar was eaten less often than Ai, but not significantly so. The beaver did not treat Aa as super-aspen; they consumed Aa and Ai equally often (X 2 = 0.8; NS). In 1993, the only type of RM extract tested on aspen was from juvenile regrowth (stump sprouts) at a beaver site, henceforth labeled Ar. As expected, Ri, Ar, and Ai differed significantly (Cochran Q test, Q = 31.4; P < 0.001; N = 33 for each treatment). The beavers' responses to the three key samples Aa, Ai, and Ar also differed significantly (Cochran Q test: Q = 13.76; df = 2; P < 0.001, two-tailed). Aspen extract was used on aspen (Aa) and RM (Ra). Treating RM with aspen extract (Ra) raised acceptance of RM significantly (Q = 14.6; P < 0.0005), compared with Ri and Rr. Pairwise comparisons by McNemar's test showed that beaver avoided Ar more than Aa (X 2 = 2.77; P < 0.05), while Ar was consumed slightly less than Ai (X 2 = 2.5; P = 0.06). They accepted Ra more often than Rr (X 2 = 7.11; P < 0.005). In other words,

2029


treatment by RM or aspen extract resulted in clear palatability differences for both aspen and RM logs. Ri and Rr were never eaten (except for one Ri log), so that treatment with RM extract (i.e., changing Ri into Rr) could not make RM any more unpalatable than Ri already was. The results for 1993 are summarized in Figure 4. Trends over Time. In 1991, the total number of logs utilized increased from 36.8% on the first day to 77.6% on the fourth, and in 1992 from 46% on the first, to 77.8% on the last day. Neither trend was significant (Spearman rank correlation coefficient). In 1993, no such trend was observed. Colony Differences. The beaver colonies differed significantly in their responses. Some colonies never touched the experimental logs, others only on certain days. Eight of 40 arrays (20%) were not touched in 1991, 15 of 50 (30%) in 1992, and 12 of 50 (24%) in 1993. These were all excluded from the analysis on the assumption that beaver may not have encountered the samples or were not even present at the site when the samples were presented. For those colonies that consumed at least some of the logs, multiple range analysis revealed three significantly different groups. Colonies that consumed more logs were also less selective. In 1992 the three colonies (sites 50, 61, and 58) that consumed the least number of logs consumed the red maple controls Ri and Rp near zero levels (0.55 and 0.01%, respectively). By contrast, the three colonies that consumed the most (sites 49, 62, and 81B), used the two RM controls Ri and Rp at averages of 32.3 and 30.9%, respectively. In 1993, the three colonies that

90-

m

80-

7o

TT T

TTJ-: 1 1 11

50

o

2O

o

T ~T o A. Rr

Rp

Ri

Re

Ra

Red Maple L o g s

Ap

Ar Aa

Ai

Ae

Aspen Logs

Treatment

FIG. 4. Consumption of 10 differently treated samples during April 1993. For abbreviations see text.

2030

MOLLER-SCHWARZE ET AL. 100 100-

1 g

iI~

a"

75-

~o-

,¢ 0Rr

Ri

Rn

Ar

Ai

An

Sample

FIG. 5. Average percent of bark area removed from treated logs at the three least consuming (stippled bars) and three most consuming (black bars) colonies in 1993. consumed the least (sites 42, 82, and 150L) only took aspen controls (Figure 5). Of these, super-aspen (Aa) was utilized most (59%). However, intact aspen (Ai) averaged only 33.3%. The three colonies that consumed the most (sites 53, 61, and 150U), increased their consumption of aspen controls and superaspen, Ai, even to a 100% average, but their choices also extended to other samples: they accepted Ar (66.7%), and even some RM samples, such as Ra (11.2%) and Re (4.7%). Both probably were more palatable because of their specific modification by aspen extract or EtOH, respectively. The latter may have extracted some deterring compounds. DISCUSSION

Red maple was clearly avoided by beaver in our study area. It may rank differently in other areas. For instance, we noted that beaver utilized red maple to a greater degree in the surroundings of Cranberry Lake Biological Station in the Adirondacks in New York State. The mature red maple trees still standing at otherwise clear-cut beaver sites at ASP most likely grew up during the beaverless time between the near-extinction of beaver in the 19th century and their reappearance at ASP in 1937 (Shadle and Austin, 1939). Logging further favored the RM by eliminating competition from other trees. In other studies, beaver have utilized red maple least of all tree species (Shadle and Austin, 1939; Henry and Bookhout, 1970) or consumed it according to its local abundance (Belovsky, 1984; Johnston and Naiman, 1990). Beaver


2031

have not been reported to prefer red maple. In some studies, the different species of Acer have not been distinguished (Shadle et al., 1943; Jenkins, 1975, 1978, 1979, 1980, 1981). The work by Shadle and Austin (1939) is particularly pertinent. They counted trees cut by the first beaver that moved into Allegany State Park. From their arrival in the summer of 1937 to November 1938, the beaver cut 226 trees. Only one was a red maple, leaving open how much of it the beaver actually consumed. The authors state " . . . the maples and birches, although present, were not much used . . . . " The abundance of the various tree species were not reported in that study. Silver maple (Acer saccharinum), a soft maple like RM, was selected against by a beaver population in Wisconsin to the point that it remained the dominant tree at lower elevations near beaver ponds (Barnes and Dibble, 1988). It is well known to foresters that red maple is less afflicted by herbivorous insects and less browsed by deer than is sugar maple (N. Richards, personal communication). We assume that phenolics render the red maple less palatable to beaver. Most Acer species contain considerable amounts of condensed and hydrolyzable tannins, usually in a reciprocal relationship (Bate-Smith, 1977). In the field, a simple test with FeCI3 solution placed on a cross section of a green branch shows a difference between red maple and sugar maple. Red maple turns a deep purple, sugar maple a pale green. Different growth forms are expected to contain different levels of secondary plant compounds. However, the relationships are not always clear. For instance, quaking aspen juvenile-type regrowth after cutting by beaver is consumed less by beaver, and yet it contains lower levels of salicin and tremuloidin than adult-form sprouts (Basey et al., 1990). A marsupial, the brush-tailed possum, Trichosurus vulpecula, fed more on different clones of poplars (Populus sp.) having lower levels of salicin and its derivatives (Edwards, 1978). Colonies differed considerably in their feeding levels and selectivity. Some never touched the samples, others fed on some samples, while still others fed on most samples with little discrimination (Figure 5). Two colonies even preferred Scots pine (Pinus silvestris) to aspen samples. They had been feeding on Scots pine before and during the time of the experiment. Other authors (e.g., Jenkins, 1975, 1981) have also noted that tree preferences by beaver can vary between colonies and even within colonies, both spatially and over time. It appears that the aspen logs remained attractive to the beaver to varying degrees despite the modification by the extracts. Therefore, for future experiments, we recommend not only higher concentrations of extracts, but also carder logs from trees of medium preference, such as black cherry (Prunus serotina), sugar maple, or musclewood (Ostrya sp.) in areas similar to ASP. These neutral trees could be rendered more palatable by aspen extract and less palatable by RM extract. Eventually, of course, repellents will have to be tested on the

2032


specific tree species that are to be protected. This includes ornamental, fruit, and forest trees. Beaver take logs into the water and peel them there. They also pulled whole arrays of our experimental logs attached to a pole into the water. RM extract may leach out when the logs are thus immersed. We also observed occasionally that beaver felled a red maple into their pond, but did not consume the branches at all or only after a few days' soaking in the water. The winter food caches in the water may also permit leaching out of compounds that inhibit feeding. Could such "food processing" be a regular strategy of beaver? If so, in future experiments, presoaked logs should be more acceptable to beaver. Beaver use smell (Doucet et al., 1994) and possibly taste and other senses to select food. The frequent sampling, i.e., biting into experimental logs and naturally growing trees, suggests that the sense of taste is important. Learning by feedback from ill effects of unpalatable plants is also likely. A taste experience may condition an odor aversion so that subsequently an odor suffices as a signal. In black-tailed deer (Odocoileus hemionus columbianus), compounds derived from red pepper failed to inhibit feeding, suggesting that the trigeminal nerve may not play a role (Sullivan et al., 1985). It is possible that red maple has developed chemical defenses in response to herbivory by beaver. Beaver and red maple have coexisted since the Tertiary, ample time for coevolution to have Occurred. Furthermore, red maple is insectpollinated. This facilitates faster selection (than wind pollination) in the limited areas along streams and around lakes where beaver dwell. To test this hypothesis, the levels of secondary plant compounds in upland populations of red maple far removed from beaver should be compared with those of red maples along streams. More extensive experiments than our 1991 test are needed to establish palatability differences between upland red maples and those in beaver-populated stream valleys. Red maple logs can also be made more attractive to beaver by treating them with extract of aspen, a preferred food. This fact, together with its opposite, the feeding inhibition by red maple extracts, shows that chemical factors, and not merely texture or other properties of bark or wood, play a role. Our findings can be applied in beaver management to reduce wildlife damage. We propose that extracts and compounds from less palatable tree species should be potent beaver repellents that can be applied to individual specimen trees. To keep herbivores out of entire areas, " f e a r " repellents instead of negative odors or taste compounds are recommended (Sullivan et al., 1985). Recent field tests showed that predator odors applied to aspen sticks reduced feeding by beaver (Engelhart and Miiller-Schwarze, 1994). A repellent has to prevent cutting bark and felling trees, not just feeding, as in our experiments. As beaverproof tree cover in larger areas and for longer time-frames, we suggest planting the less vulnerable red maple near beaver sites, along with conifers such as pine

RED MAPLE INHIBITS BEAVER FEEDING and spruce species. Conversely,

2033 where desirable, one could possibly render

s o m e tree s p e c i e s m o r e p a l a t a b l e to c a p t i v e o r f r e e - r a n g i n g b e a v e r b y t r e a t i n g them with extracts from aspen.

Acknowledgments--We thank the officers and staff of Allegany State Park under Regional Directors Hugh Dunne, the late Dann Colvin, and recently, James M. Rich for their support over the years. Thanks go to Mr. Rong Tang for help in preparing the extracts, Dr. Stephen V. Stehman for statistical advice, Dr. Francis X. Webster for chemical advice, and Dr. Stephen Teale for critically reading the manuscript. SUNY College of Environmental Science and Forestry provided a field camp at Allegany State Park.

REFERENCES

BARNES, W.J., and DIBBLE, E. 1988. The effects of beaver in riverbank forest succession. Can, J. Bot. 66:40-44. BASEY, J.M., JENKINS, S.H., and BUSHER, P.E. 1988. Optimal central-place foraging by beavers: Tree-size selection in relation to defensive chemicals of quaking aspen. Oecologia 76:278282. BASEY, M.J., JENKINS, S.H., and MILLER, G.C. 1990. Food selection by beavers in relation to inducible defenses of Populus tremutoides. Oikos 59:57-62. BATE-SMITH, E.C. 1977. Astringent tannins of Acer spp.. Phytochemistry 16:1421-1426. BELOVSKY, G.E. 1984. Summer diet optimization by beaver. Am. Midl. Nat. 111:209-222. DOUCET, C.M., WALTON,R.A., and FRYXELL,J.M. 1994. Perceptual cues used by beavers foraging on woody plants. Anirn. Behav. 47:1482-1484. EDWARDS, W.R.N. 1978. Effect of salicin content on palatability of Populus foliage to opossum ( Trichosurus vulpecula ). N. Z. J. Sci. 21:103-106. ENGELHART, A., and MOLLER-SCHWARZE,D. 1994. Responses of beaver (Castor canadensis Kuhl) to predator odors. FRYXELL, J.M., and DOUCET, C.M. 1993. Diet choice and the functional response of beavers. Ecology 74:1297-1306. HENRY, D.B., and BOOKHOUT,T.A. 1970. Utilization of woody plants by beavers in northeastern Ohio. Ohio J. Sci. 70:123-127. JACOBS, J. 1974. Quantitative measurement of food selection: A modification of the forage ratio and Ivlev's electivity index. Oecologia 14:413-417. JENKINS, S.H. 1975. Food selection by beaver: A multidimensional contingency table analysis. Oecologia 21 : 157-173. JENKINS, S.H. 1978. Food selection by beaver: Sampling behavior. Brevora (Harvard Mus. Comp. Zool.) 447:1-6. JENKINS, S.H. 1979. Seasonal and year-to-year differences in food selection by beavers. Oecologia 44:112-116. JENKINS, S.H. 1980. A size-distance relation in food selection by beavers. Ecology 61:740-746. JENKINS, S.H. 1981. Problems, progress and prospects in studies of food selection by beavers. Proceedings International Furbearer Symposium, pp. 559-579. JOHNSTON, C.A., and NAIMAN, R.J. 1990. Browse selection by beaver: Effects on riparian forest composition. Can. J. For. Res. 20:1036-1043.

2034


SHADLE, A.R., and AUSTIN, T.S. 1939. Fifteen months of beaver work at Allegany State Park, NY. J. Mammal. 20:299-303. SHADLE, A.R., NAUTH, A.M., GESE, E.C., and AUSTIN, T.S. 1943. Comparison of tree cuttings of six beaver colonies in Allegany State Park, New York. J. Mammal. 24:32-39. SULLIVAN, T.P., NORDSTROM, L.O., and SULLIVAN, D.S. 1985. The use of predator odors as repellents to reduce feeding damage by herbivores. II. Black-tailed deer (Odocoileus hemionus columbianus). J. Chem. Ecol. 11:921-935.