Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Herbivore-plant-soil interactions in the boreal forest : selective winter feeding by spruce grouse Mueller, Fritz Paul 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_mueller_fritz.pdf [ 4.51MB ]
JSON: 831-1.0086145.json
JSON-LD: 831-1.0086145-ld.json
RDF/XML (Pretty): 831-1.0086145-rdf.xml
RDF/JSON: 831-1.0086145-rdf.json
Turtle: 831-1.0086145-turtle.txt
N-Triples: 831-1.0086145-rdf-ntriples.txt
Original Record: 831-1.0086145-source.json
Full Text

Full Text

HERBIVORE-PLANT-SOIL INTERACTIONS IN THE BOREAL FOREST:SELECTIVE WINTER FEEDING BY SPRUCE GROUSEbyFRITZ PAUL MUELLERB.Sc.(Honours), The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of ZoologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAFebruary 1993©Fritz P. Mueller, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of  ZOO LO&YThe University of British ColumbiaVancouver, CanadaDate  TeA3 I c I /619 DE-6 (2/88)ABSTRACTThis thesis examines the unusual winter forage selection of spruce grouse(Dendragapus canadensis) in the Yukon. The winter diet of spruce grouseconsists entirely of conifer needles. Spruce grouse feed selectively on individualtrees (feeding trees) of a single species, white spruce (Picea alauca), leavingadjacent trees of similar size and age uneaten. These feeding trees are usedthroughout the winter and some are used repeatedly for several years.Accumulations of droppings on the ground and defoliated branches indicatepreferred trees. Between 0.3 to 5.0 kg dry weight of spruce grouse faecesaccumulate annually under feeding trees and between 25 to 90 % (mean 40 `)/0)of the needles on preferred trees are removed by feeding grouse. Why dospruce grouse feed so selectively?Chapter 2 describes the role of foliar chemistry in the selection of winterforage by spruce grouse. During feeding trials, captive spruce grouse had amarked preference for needles from feeding trees over control trees. Chemicalanalyses of needles also support the hypothesis that needle chemistry accountsfor the winter forage selection of spruce grouse. Concentrations of twomonoterpene antifeedants, camphor and bornyl acetate, and the ratio of resin tonitrogen, were inversely related to grouse forage preferences and may explainthe selection by spruce grouse of individual trees for winter feeding.Feeding trees are exposed during winter to recurring high levels ofherbivory by spruce grouse. This herbivory may affect growth rate, architecture,reproductive output, and chemical defence of selected trees (Chapter 3).Feeding trees have higher lateral (twig) growth rates, longer needles, and longer,more highly branched limbs and more rounded crowns than control trees. Also,cone production is significantly lower in feeding trees than in adjacent controliitrees. The relatively high growth rates, the low reproductive output, and the lowsecondary chemical content of feeding trees suggests a within-tree trade-off inallocation of limited carbon resources.Large amounts of spruce grouse faeces accumulate annually underfeeding trees. Decomposition of these faeces is rapid relative to spruce litter. Inthe nutrient limited boreal forest, spruce grouse faecal inputs under feeding treesmay locally increase soil nutrient availability and nutrient cycling rates (Chapter4). The nitrogen content of grouse faeces may account for the relatively highgrowth rates of feeding trees and may lead to the higher forage quality of feedingtrees compared with control trees.Experimental defoliation and fertilization of white spruce trees suggeststhat the observed differences between feeding trees and adjacent control treesresult, in part, from the effects of selective feeding by spruce grouse. The growthrates of experimentally fertilized trees increased significantly over control treegrowth rates suggesting faecal input is critical to regrowth. The growth rates ofclipped trees did not change in response to simulated grouse herbivory.The combined effects of defoliation and nutrient return by spruce grousemay lead to regrowth that is more palatable than forage on uneaten plants.Spruce grouse feeding may result in patches of highly palatable forage thatattract further feeding, generating a feeding-regrowth feedback loop (Chapter 5).Positive effects on forage palatability and quantity may account for the unusualand prolonged use by spruce grouse of individual trees for winter feeding.Spruce grouse may be farming their food plants.iiiTABLE OF CONTENTSPageABSTRACT^ iiLIST OF FIGURES^ vLIST OF TABLES viACKNOWLEDGEMENTS^ vii, viiiCHAPTER 1. GENERAL INTRODUCTION^ 1CHAPTER 2. SELECTION OF WINTER FEEDING TREES BY SPRUCEGROUSE IN YUKON: EFFECTS OF SECONDARYCOMPOUNDS AND NITROGEN.^ 8Abstract ^ 9Introduction 10Study Area 13Methods 14Results ^ 19Discussion 23CHAPTER 3. SPRUCE GROUSE HERBIVORY: IMPACTS ON GROWTHARCHITECTURE, REPRODUCTION, AND DEFENSE OFWHITE SPRUCE.^ 44Abstract ^ 45Introduction 46Study Area 48Methods 49Results ^ 51Discussion 54CHAPTER 4. FAECAL DECOMPOSITION IN THE BOREAL FOREST:POSITIVE FEEDBACKS BETWEEN SPRUCEGROUSE FORAGING AND NITROGEN CYCLING. ^ 70Abstract ^ 71Introduction 72Study Area 75Methods 76Results ^ 81Discussion 84CHAPTER 5. GENERAL CONCLUSION^  101APPENDIX 1. ^  110REFERENCES CITED^  112ivLIST OF FIGURESPageFig. 1.1 Possible effects of herbivory on plant growth (quantity)and plant forage quality ^  6Fig. 2.1 Map of study area 28Fig. 2.2 Diagram of white spruce branch showing annual growth nodesand typical grouse browse damage^ 30Fig. 2.3 Amounts of three needle types eaten by captive spruce grouseduring feeding trials ^  32Fig. 2.4 Nitrogen content, resin content, and resin-to-nitrogen ratio fortwo age classes of needles from feeding trees and adjacentcontrol trees 34Fig. 2.5 Relationship between nitrogen and resin for two age classes ofneedles from two tree types^ 36Fig. 2.6 Amounts of two monoterpenes, camphor and bornyl acetate, inneedles from feeding trees and adjacent control trees^ 38Fig. 2.7 Nitrogen content and acid-pepsin digestibility of needles fromthree tree types^ 40Fig. 3.1 Growth form of feeding trees and control trees ^ 61Fig. 3.2 Comparison of growth rates (annual twig length and annual ringwidth) of feeding trees and control trees 63Fig. 3.3 Differences in needle length and cone number between feedingtrees and control trees^ 65Fig. 4.1 Nutrient cycling diagram 93Fig. 4.2 Faecal decomposition and mesh bag experiments^ 95Fig. 4.3 Microbial respiration rates of faeces and litter  97Fig. 4.4 Effects of NPK fertilizer on white spruce twig growth^ 99Fig. 5.1 Cause-and effect model of coniferous woody plant chemistrychanges induced by severe defoliation and fertilization ^ 107vLIST OF TABLESPageTable 2.1 ANOVA summary table for nitrogen content, resin content, andresin-to-nitrogen ratio for two age classes of needles fromfeeding trees and adjacent control trees^ 42Table 2.2 Concentration of seven elements in two age classes of needlesfrom feeding trees and adjacent control trees^ 43Table 3.1 Age, height, and circumference of feeding trees and controltrees^ 67Table 3.2 Physical characteristics of experimentally defoliated trees andfertilized trees and adjacent control trees^ 68Table 3.3 Effects of experimental defoliation and fertilization on treetwig growth and needle length^  69Table 4.1 Inputs of spruce grouse faeces under feeding trees ^ 91Table 4.2 Carbon, nitrogen, and C / N ratios of spruce needles,spruce grouse faeces, and soils ^  92Table 5.1 Hypotheses and predictions from model shown in Figure 5.1 ^ 109viACKNOWLEDGEMENTSI particularly thank Tony Sinclair, my supervisor, for believing both in meand in this project and for giving me room to make my own successes andfailures. I also thank David Hik for introducing me to turds and to David Wilcox,for logistical help in the field and in Vancouver, and for encouraging andparticipating in the ballistic pace of the last few years.I am grateful to Charles Krebs, Tony Sinclair, Jamie Smith, Kathy Martin,and other members of the Kluane Project for the opportunity to work with themboth in the field in the Yukon and at UBC in Vancouver.Discussions with John Bryant and Roger Ruess (University of Alaska,Fairbanks), Bob Jefferies (University of Toronto), Tom Whitham (NorthernArizona University), and Erkki Haukioja (University of Turku, Finland) influencedmy thinking a great deal and gave me confidence in my ideas about grouse andtrees.Giles Galzey, Animal Science UBC, assisted me with nutrient analyses, letme use his laboratory, and helped us set up our own forage analysis facilities.Thanks Giles. Canadian Wildlife Service gave us glassware and many otherlaboratory supplies. Ray Anderson, Oceanography UBC, gave us access to hisGLC laboratory and David Williams assisted with the camphor analyses. OzSchmitz and Nic Larter shared the lab and their expertise on acid-pepsindigestions. Bernie Von Spindler, Soil Science UBC, taught me how to use theLECO for carbon analyses. Rudy Boonstra, University of Toronto, braved theSLOWPOKE reactor to analyze spruce needles for elements.David Hik, Karen Hodges, Tony Sinclair, and Jamie Smith made helpfulcomments on drafts of this thesis.I thank Andy Williams, manager of the field station at Kluane Lake, for hishealthy sense of the ridiculous and for saunas.And finally, thanks to everyone at Kluane over the last several years fortheir help in innumerable ways and most of all for their friendship. Thanks, it wasgreat!This research was conducted under Yukon Renewable Resourcesscientific and collecting permit No. 0938. Financial support was provided byNatural Science and Engineering Research Council (NSERC) grants to TonySinclair and Charles Krebs and by NSERC scholarships and Department ofIndian and Northern Affairs grants to myself.Cheers, Fritz.vii"Boyd"viiiCHAPTER 1:GENERAL INTRODUCTION1As prey, plants and animals are different. Plant prey are rarely completelyconsumed or killed when eaten by herbivores; animal prey are nearly alwayskilled when eaten by predators. Plants are modular, consisting of repetitivemulticellular units (Whitham et al. 1991, Haukioja 1991, Schlichting 1986), andhence herbivore impacts on plants are usually sublethal. Plants are alsosedentary. The implications of these obvious differences between plant andanimal prey, although long known to botanists and agriculturalists, are onlyrecently becoming generally appreciated by animal biologists and ecologists(Lindroth 1989, Whitham et al. 1991, Huntly 1991, Bryant et al. 1991). Thepotential for complex direct and indirect interactions between plants and animalsarises because plants live to be eaten many times by herbivores (Dirzo 1984,Bianchi and Jones 1991, Jefferies et al. 1992, Hunter 1992).Herbivores are not passive consumers. They engage in complex andoften dynamic interactions with their food plants and with soils. Throughselective feeding and the direct effects of defoliation and fertilization herbivorescan alter both the chemistry and the production of plants. These changes canaffect soil processes, which in turn affect plants and herbivore forage selection(Pastor 1988, Jefferies 1988).Herbivores, through the effects of defoliation and fertilization, can changeattributes of their plant prey. These changes affect later herbivore-plantinteractions. The consequences of these effects to plants can be either positiveor negative from the herbivore's perspective (Bianchi and Jones 1991, Whithamet al. 1991, Bryant et al. 1992). Such interactions are almost always negativefrom the plant's point of view (Belsky 1986, Mooney et al. 1991).The consequences of herbivory to plant production and palatability aredetermined by the plant species, individual genotype, plant age, season, nutrientstatus of the plant, and the timing and intensity of herbivory (Whitham et al. 1991,2Haukioja 1991, Bryant 1992a and b, Herms and Mattson 1992). Forageproduction (or quantity) often changes, and in some situations increases (e.g.Pellew 1983, McNaughton 1984, Power 1990, Hik et al. 1991), as a consequenceof herbivory (Fig. 1.1(a)).Herbivores can also have a range of effects, positive and negative, onplant forage palatability (or quality) (Maschinski and Whitham 1989, Bryant 1991,Haukioja 1991, Whitham et al. 1991). In some situations, food value (quality) ofplants increases following herbivory (Danell and Huss-Danell 1985, Bryant et al.1991, Irons et al. 1991, Niemela et al. 1991) and in others it decreases (Rhoades1979, Haukioja and Neuvonen 1985, Haukioja 1980, Schultz 1988, Karban andMyers 1989). See Figure 1.1(b).Herbivores usually must compromise between increasing nutrient intakeand avoiding plant toxins. Forage quality (or palatability) for herbivores isdependent upon the ratio of nutrients (e.g. nitrogen, elements, carbohydrates) tofeeding inhibitors and toxins (e.g. plant defense chemicals, lignins, undigestiblefiber). I use the words quality and palatability interchangeably throughout thisthesis to refer to this ratio of nutrients to toxins.Herbivores can directly effect plants through defoliation or throughfertilization with faeces and urine. If the palatability and/or amount of forageincreases as a consequence of herbivory then the potential arises for positivefeedbacks to develop between plants and animals. Herbivory may result inpatches of highly palatable forage that attract further feeding, generating afeeding-regrowth feedback loop (Du Toit et al. 1991, Bryant et al. 1992). The roleand generality of such positive feedback in ecosystem processes and communitydynamics is just beginning to be appreciated (e.g. DeAngelis et al. 1986, Lindroth1989, Gordon and Lindsay 1990, Bianchi and Jones 1991, Robertson 1991,Huntly 1991, Jefferies et al. 1986, 1992).3In this thesis I examine the unusual winter forage selection of sprucegrouse (Dendragapus canadensis)  in Canada's Yukon. The winter diet of sprucegrouse consists of needles from particular individual trees of a single foragespecies, white spruce (Picea glauca). Such extreme specialist feeding isassociated typically with insects and is rare for vertebrates.Spruce grouse and white spruce trees are an ideal system for examininghow plant-animal interactions are mediated through herbivore forage selection,soil microbial processes, and plant secondary chemistry. The importance ofthese processes for ecosystem dynamics, and plant-animal interactions ingeneral, can be studied directly in boreal ecosystems. The boreal forest isrelatively simple, has few dominant herbivores and forage species, andexperimental manipulations of these herbivores and plant species can be carriedout in the field with relative ease.In chapter 2 of this thesis, I examine the role of foliar chemistry in theselection of winter food by spruce grouse. In chapter 3, I compare themorphology and reproductive output of trees eaten by grouse with uneaten trees.By experimentally defoliating and fertilizing white spruce trees I simulated sprucegrouse herbivory to determine the effects of these two manipulations on treegrowth rate and forage palatability. Chapter 4 examines the effects of sprucegrouse faeces on local soil processes and nutrient cycling to determine howspruce grouse faeces may affect white spruce trees and hence spruce grouse.In chapter 5 I conclude with a model showing the positive and negativefeedbacks that may be operating between spruce grouse, white spruce trees,and soils.4Some general questions examined and raised during this study of grouseand trees follow:1. How is herbivore forage selection influenced by defensive chemistry inforage plants?2. What effects do herbivores have on plant growth and forage quality?3. How do herbivore forage selectivity and plant defense chemistry influencesoil microbial processes?4. How do the processes of decomposition and nutrient cycling influenceinteractions between plants and animals?5. What are the effects of herbivores on plant physiology - specifically, withinplant allocation of carbon and nutrients to growth, maintenance, reproduction,and defense?6. What is the role and generality of positive feedback in ecosystem processesand community dynamics?7. How do the direct and indirect effects of herbivores influence the complexand dynamic interactions between herbivores, plants, and soils?5Figure 1.1 This figure shows the possible effects of increasing intensity ofherbivory on (a) plant forage quantity and (b) plant forage quality. The solid linesindicate extremes of a continuum of possible effects of herbivory on plant foragequantity and quality. Figures a and b are not necessarily related. All changes inquantity and quality are viewed as positive or negative from the herbivoreperspective. Part (a) was modified from McNaughton (1983).67CHAPTER 2:SELECTION OF WINTER FEEDING TREES BY SPRUCE GROUSE IN YUKON:EFFECTS OF SECONDARY COMPOUNDS AND NITROGEN.81. ABSTRACTThis chapter examines the role of foliar chemistry in the selection of winterforage by spruce grouse (Dendragapus canadensis) near Kluane Lake,southwest Yukon. In winter spruce grouse feed selectively on individual whitespruce (Picea qlauca) trees, leaving adjacent trees of similar size and ageuneaten. These trees are used throughout the winter and some are utilizedrepeatedly for several years. The winter diet of spruce grouse consists entirely ofconifer needles. Accumulations of droppings on the ground and defoliatedbranches indicate preferred trees. Why do spruce grouse feed so selectivelyduring winter?During feeding trials, captive spruce grouse ate significantly more needlesfrom feeding trees than from either adjacent or random control trees. Thissuggests that selection of individual trees for winter feeding is based on thechemical quality of needles.White spruce needles were analyzed for ether-extractible resin, twomonoterpene antifeedants (camphor and bornyl acetate), total nitrogen, acid-pepsin digestibility, and seven elements (Al, Ca, CI, K, Mg, Mn, and Na). Spruceneedles collected from trees used by spruce grouse contained less ether-extractable resin than needles from similar adjacent trees not used by grouse forfood. Needles from feeding trees also contained significantly smaller amounts oftwo monoterpenes, camphor and bornyl acetate, than needles from adjacentcontrol trees. There were no significant differences in the nitrogen content ofneedles between feeding trees and either adjacent or random control trees,although there was a tendency for nitrogen content to be higher in preferredneedles. Acid-pepsin digestibility of spruce needles was significantly higher intrees eaten by grouse than in random controls not eaten by grouse. No9significant differences in digestibility were found between feeding trees andadjacent controls. No differences were found between trees in theconcentrations of seven elements.These results support the hypothesis that needle chemistry accounts forthe winter forage selection of spruce grouse. Avoidance of antifeedants,specifically camphor and bornyl acetate, and preference for low resin-to-nitrogenratios may explain the unusual selection by spruce grouse of individual trees forwinter feeding.2. INTRODUCTIONHerbivores face several challenges from their food plants. These includevariation in nutritional value and in mechanical and chemical defenses amongplant species, individuals, growth stages, and parts (Spencer, 1988, Provenzaand Balph 1990, Bryant et al. 1991). Many plants have evolved extensive arraysof chemicals (antifeedants) for defense against herbivores (e.g. Rosenthal andJanzen 1979, Crawley 1983, Howe and Westley 1988, Fritz and Simms 1992,Tuomi 1992). Chemical defenses do not provide complete protection becauseherbivores have also evolved anatomical, physiological, and behavioral countersto plant defenses (Lindroth 1988, Distel and Provenza 1990, Provenza et al.1992). A result is that herbivores eat some plants and plant parts more thanothers.Grouse select certain species of plants to eat from those available, andoften prefer individual plants within a single species (e.g. Ellison 1966, Doer et al.1974, Linden 1984, Remington and Braun 1985, Jakubas et al. 1989,Remmington 1990, Andreev 1991, Gjerde 1991). Both nutrient content and1 0chemical defense hypotheses have been proposed to explain this selection.Plant palatability may also explain observed patterns of forage selection byspruce grouse (Dendragapus canadensis) (Ellison, 1976, Hoft et al. 1987).The winter diet of spruce grouse consists entirely of conifer needles(Crichton 1963, Jonkels and Greer 1963, Ellison 1966, Pendergast and Boag1970). In the Yukon, spruce grouse eat needles of certain individual trees of asingle forage species, white spruce (Picea glauca (Moench) Voss). Sprucegrouse in Ontario, northern Michigan, northern Washington, and Alaska also usespecific trees for winter feeding (Ellison 1966, Pendergast and Boag 1971,Gurchinoff and Robison 1972, Hoft et al. 1987). Such extreme specialist feedingis typically associated with insects and is rare for vertebrates.For herbivores that eat woody plants, such as slow growing evergreens,chemical defenses may be especially significant in determining foragepreferences (Bryant et al. 1983, 1988, 1991, Coley et al. 1985). Studies haveindicated that many secondary metabolites produced by conifers are chemicaldefenses and deter herbivore feeding (Bryant and Kuropat 1980, Palo et al.1985, Tahvanainen et al. 1985, Sinclair et al. 1988a, Bryant et al. 1991).The forage preferences of grouse do not correlate consistently with thenutrient quality (energy or nitrogen content) of needles (Bryant and Kuropat1980). Several authors (Pendergast and Boag 1970, Gurchinoff and Robison1972, Bryant and Kuropat 1980, Hoft et al. 1987) suggested that spruce grousemight be avoiding monoterpenes found in conifer needles. Monoterpenes are amajor class of carbon-based plant chemicals, and many have deterrent or toxiceffects on a number of herbivore species (Schwartz et al. 1980, Farentinos et al.1981, Pederson and Welch 1985, Brooks et al. 1987, Sinclair et al. 1988b,Horner et al. 1988). Most studies examining chemical defence hypotheses forspruce grouse forage selection have looked only at resin content of spruce11needles, a crude measure of monoterpene content. Few studies have identifiedspecific compounds with feeding deterrent properties.Two oxygenated monoterpenes, camphor and bornyl acetate, areabundant in white spruce in the Yukon, and one of these chemicals, camphor, isa potent feeding deterrent for snowshoe hares (Sinclair et al. 1988b). Camphor,a monoterpene also found in juniper in Colorado and Utah, reduces microbialactivity in the rumen of deer and thus inhibits digestion (Schwartz et al. 1980).Grouse digestion is microbially mediated (Moss and Hanssen 1980) and mayalso be inhibited by monoterpenes such as camphor and bornyl acetate.In the Yukon, near Kluane Lake, spruce grouse select certain individualwhite spruce (Picea glauca) trees for winter feeding while ignoring adjacentspruce trees of similar physical characteristics. Large accumulations ofdroppings on the ground and heavily defoliated branches indicate preferred trees.Browse damage also indicates some feeding trees are used repeatedly for two ormore years.To explain the use of feeding trees I hypothesized that two oxygenatedmonoterpenes found in white spruce, camphor and bornyl acetate, are feedingdeterrents for spruce grouse and that avoidance of these compounds explainsthe high feeding selectivity of these birds during winter. Alternatively, Ihypothesized that forage selection of spruce grouse is affected by the nutrientcontent of needles; specifically, that feeding trees have both higher nitrogen andhigher digestibility than surrounding trees. A third possibility is that sprucegrouse are balancing nutrient intake and plant defense avoidance.In this chapter I describe the results of chemical analyses testing thepredictions that (i) forage preferences of spruce grouse in the Yukon arenegatively correlated with the concentrations of plant defensive compounds,specifically of camphor and bornyl acetate, and (ii) that forage preferences are12positively correlated with both the nutrient content (nitrogen, minerals) and thedigestibility of needles.3. STUDY AREAThis study was conducted from 1989 to 1992 and is part of a long-termcollaborative study of boreal forest ecology at Kluane, Yukon (Krebs et al. 1992and references therein).Kluane Lake (60° 5' N, 137° 5' W) is located in the southwest Yukon onthe eastern boundary of Kluane National Park. The study area extends 28 km tothe southeast of Kluane Lake along the Alaska Highway and is located within theShakwak Valley, a broad (8-16 km wide) glacial-formed valley which runsnorthwest-southeast along the eastern front-range of the St. Elias Mountains(Fig. 2.1). Elevation is between 900 to 1000 m over most of the study area anddescends to 860 m at Kluane Lake.The climate is semi-arid and characterized by long, cold winters and short,cool summers. The mean annual air temperature at Kluane Lake just to the westof the study area is minus 2.7°C, with July and January means of 12.6°C andminus 21.5°C respectively (Anonymous 1987). The frost free period at HainesJunction, 40 km southeast of the study area, is only 21 days (Webber 1974).Kluane Lake receives a mean annual precipitation (1951-1980) of 223.9 mm withapproximately 70% falling as rain and the remainder as snow (Anonymous,1988). Snow remains on the ground for almost eight months, from the middle ofOctober until the middle of May.The vegetation of the Shakwak Valley is dominated by white spruce (Piceaolauca (Moench) Voss), the only conifer tree species. The vegetation of the area13has been described and mapped by Douglas (1974), Krebs et al. (1986), Dale(1990), and Zbigniewicz and Dale (1992).4. METHODS4.1 Browse damage and tree typesForage preferences of wild spruce grouse were determined from sites ofpast grouse browsing. Spruce grouse characteristically eat only the distal twothirds of individual needles, leaving the base of eaten needles (1-4 mm long)attached to the branch (Fig. 2.2).The age of needles eaten was determined by counting the annual growthnodes along the branches (Fig 2.2). Two general age classes of needles, youngand old, were recognized based on the color of adjacent branches. The color ofbranches adjacent to young needles, between 1 and 5 years, is orange. Thecolor of branches adjacent to needles older than 5 years is grey or black.Three types of white spruce trees were identified: (i) feeding trees (FT), (ii)adjacent control trees (AC), and (iii) random control trees (RC). The first type,feeding trees (FT), were used by spruce grouse for winter feeding and wereidentified in the field by extensively defoliated branches and large accumulationsof droppings on the ground. Feeding trees were highly visible in early May whenneedles damaged during winter feeding turned red prior to falling off. Thesecond type, adjacent control trees (AC), were the nearest trees, usuallybetween 2 to 5 m distant, of similar size and age to the feeding trees, which werenot eaten by grouse. Feeding trees and adjacent control trees were consideredpaired due to similar site characteristics and time of sampling. A third type,14random control trees (RC), were selected randomly on the 40 ha study areasestablished by the Kluane Project.Trees with a trunk circumference, at 1 m height, of less than 25 cm andtrees less than 25 years old were excluded from all trials and analyses since nofeeding trees were found smaller or younger than this (Chapter 3).4.2 Feeding trialsFeeding trials were conducted with captive spruce grouse in the Yukonbetween January and April 1991. Spruce grouse were captured in the field withnoose poles and mist nets and housed in a 9 x 9 x 2.5 m outdoor pen. The penwas covered with 5 x 5 cm nylon mesh and enclosed the lower portions ofseveral white spruce trees so that the habitat was as natural as possible. Allspruce needles on portions of the trees enclosed by the pen were removed sothat the only needles available to grouse were those presented during feedingtrials.During each feeding trial (n = 15) captive spruce grouse were presentedsimultaneously with branches collected from the three tree types describedabove. Only branches from the most intensively browsed feeding trees wereused. Three or four branches of each tree type were collected from the southside, between 2 to 3.5 m height, of each tree. In each trial branches wereweighed and placed randomly on the lower limbs of trees enclosed within thepen. Branches were left for 4 to 6 days to allow feeding and then reweighed.The length of feeding trials was adjusted depending on the number ofbirds in the pen so that the number of "bird days" per trial was approximately thesame (mean = 5 days). Between one and three birds were housed in the pen atany one time. In each trial different grouse were presented branches from15different trees. Most trees sampled for feeding trials were also sampled for laterchemical analyses.4.3 Needle chemistry4.3.1 Sample collectionWhite spruce needles were collected from three tree types (feeding trees,adjacent control trees, and random control trees) and from two age classes ofneedles (young (1 to 5 years) and old (greater than 5 years)) from within eachtree type. White spruce needles were clipped with scissors from 3 to 4 branchestaken from the south side of each tree and frozen for later analyses. The heightof sampled branches was between 1 and 4 meters.All needle samples were collected at Kluane during winter months frommid-November through to early April (1989/90 and 1990/91). Seasonal changesin needle chemistry may be minimal during winter when needles are frozen (Dayet al. 1989).4.3.2 Sample preparationNeedle samples were dried in an oven for 72 hours at 60-65°C. Sampleswere weighed before and after drying to determine dry matter content. Dryneedles were ground in either a Wiley mill or a high speed Brinkman mill to passthrough a 1 mm screen.4.3.3 NitrogenNeedles were analyzed for total nitrogen content using macro-Kjeldahldigestion. Samples were dried and ground and 0.5-1.0 g of each was weighedand digested on a Tecator block digestor at 410°C in sulphuric acid and16hydrogen peroxide using selenium catalyst (Parkinson and Allen, 1975).Nitrogen content of digested samples was then determined colorimetrically on aTechnicon II autoanalyzer (Technicon Autoanalyzer II. 1974. Industrial MethodNo. 321-74A). For each sample two replicates were run and for each replicatetwo autoanalyzer readings were taken. Values are expressed as mean percentof dry matter.4.3.4 Ether extractable resin"Resin" is the ether-extractable product following Sinclair and Smith(1984), Bryant et al. (1983), and Fox and Bryant (1984). Between 3-4 grams offrozen needles were accurately weighed and soaked in 20 ml of ethyl ether for 5days. The ether was decanted into a second weighted jar and allowed toevaporate. The sample was soaked in fresh ether for a total of three washings.The accumulated oil or resin remaining was then weighed and expressed aspercent dry matter. For each forage sample two replicates were run to determinethe mean resin content. The resin weight was determined for two tree types,feeding trees and adjacent control trees, and for two age classes of needles,young and old, within each tree type. Replication for each treatment was equal(n = 17) and the results were analyzed using two-way ANOVA.4.3.5 Monoterpenes (Camphor and Bornyl acetate)Frozen needles (30 g wet weight) were cut into small pieces and placed in200 ml of methanol containing 10 mg of internal standard, borneol, and left for 24hours in a refrigerator at 40C. This methanol extract was filtered through cottonto remove the needles, swirled with 3 g of charcoal for 3 minutes, and thenpassed through filter paper. The resulting methanol extract was roto-evaporated17for 3-4 hours until dry. The concentrated extract obtained was then made up to 3ml with methanol for analysis with gas-liquid chromatography.Gas-liquid chromatography with flame ionization detection was used forquantitative analyses of camphor and borneol acetate. The concentrated extractof needles from 9 feeding trees and 9 non-feeding control trees was analyzedusing a Hewlett Packard 5830A gas chromatograph (GC) with 2 m x 2 mm (I.D.)stainless-steel column packed with 3% Carbowax 20M on Chromosorb W.Helium, delivered at a flow rate of 25 ml/min, was used as the carrier gas. TheGC was temperature programmed to 150°C at 5°C / minute. The injector anddetector temperatures were 2000 C and 250°C, respectively. 1 ul samples of theconcentrated extract with internal standard were injected by hand. A HewlettPackard 18850a computing integrator was used to quantify peak areas. Peakswere identified using relative retention times and data from Sinclair et al. (1988b).Standard calibration curves were prepared for camphor, borneol acetate, andborneol using weighed samples of pure material.4.3.6 Acid-pepsin digestibilityAcid-pepsin digestibility, an index of forage fiber content, was determinedfollowing the procedure of Larter (1992). The acid-pepsin solution was made upwith 2 g of pepsin (Pepsin A, 1:10000 Sigma Chemical Co.) and 8.33 ml of 12 NHCL made up to 1 L with tap water. Two grams of ground needle sample werethen combined with acid-pepsin solution (20 ml) in 20 x 150 mm test tubes andrubber stoppered. The sample was mixed by rotating the tube by hand and thenplaced in a 370C water bath for 48 hours. The tubes were mixed again at 1 hour,6 hours, and 24 hours after being placed in the bath. At 48 hours the test tubeswere removed from the bath. The test tube contents were then vacuum filteredand the remaining particulate matter was dried for 48 hours at 90-100°C. The18percent of dry matter (DM) digested was determined using the following formula:% DM digested = (1 - (remaining particulate matter / 0.2)) x 100%. For eachforage sample two replicates were run to determine mean percent digested. Aone-way ANOVA was used to compare the digestibility of white spruce needlescollected from feeding trees, adjacent control trees, and random control trees.4.3.7 ElementsThe amounts of 7 elements in young (1 year) and old (>1 year) spruceneedles collected during winter 1989/90 were compared between feeding treesand adjacent control trees (n = 17). Spruce samples were analyzed for amountsof aluminum (Al), calcium (Ca), chloride (CI), magnesium (Mg), manganese (Mn),potassium (K), and sodium (Na) by neutron activation analysis using aSLOWPOKE (Safe, Low Power, Kritical Experiment) reactor (Atomic Energy ofCanada Ltd) at the University of Toronto. Samples were irradiated for 5 min at10 kW (5 x 10 11 neutrons cm -2 s-1 ) and standards were prepared and analysedfollowing the same procedures. Corrections were made for isotopic decay basedon the half-life of each isotope. The neutron activation analysis methods usedare the same as described in lacobelli and Jefferies (1991) and Kruger (1971).Data were log transformed prior to ANOVA on SYSTAT.5. RESULTS5.1 Forage preferences of wild spruce grouseSpruce grouse select particular white spruce trees for winter feeding andleave adjacent trees of similar age and size. The percentage of needlesremoved from feeding trees by grouse ranged from 25 to 90 percent, with a mean19defoliation of 40 percent (Chapter 3). On some trees the only available needleswere those that grew during the previous summer.Within feeding trees, spruce grouse browsed selectively upon the youngneedles of growing twigs and avoided feeding on older needles. Examination ofthe percentage of needles eaten on 15 branches from 5 browsed trees indicatedthat young needles on orange twigs (1 to about 5 years) are preferred over oldneedles on black colored twigs (> 5 years) (paired t-test, t 4 = 8.28, P < 0.002).There was no preference by grouse for needles from any particular branch heightor tree aspect.Spruce grouse also appear to avoid feeding in younger trees duringwinter. The youngest feeding tree aged (n = 36) was 30 years old (Chapter 3).There were many white spruce trees younger than this in the study area.5.2 Feeding trialsDuring feeding trials captive spruce grouse ate about twice as manyneedles from feeding tree branches as from either adjacent or random controltrees (one-way ANOVA, F2 ,42 = 21.34, P < 0.001) (Figure 2.3). Captive sprucegrouse ate a mean of 206 g of needles per bird per day.In 1 of 15 feeding trials the random control branches offered were eaten ata faster rate (156 g bird -1 day-1 ) than feeding tree branches (110 g bird -1 day-1 ).With this outlier removed, mean consumption of random control needlesdecreased from 42.1 to 31.8 (g bird -1 day-1 ) and the differences in amount eatenof the 3 tree types increases further.205.3 Foliar Chemistry Analyses5.3.1 NitrogenNeedles from white spruce feeding trees did not have significantly highernitrogen content than needles from adjacent control trees (two-way ANOVA, F1 , 54= 2.715, P = 0.104) (Fig. 2.4 (a); Table 2.1 (a)). There was a tendency fornitrogen to be slightly higher in feeding trees than in adjacent control trees.A second set of spruce needle samples was analyzed to compare thenitrogen content of random control trees with feeding and adjacent control trees(Fig. 2.7 (a)). Differences in nitrogen content between the three tree types werenot significant (one-way ANOVA, F2 ,32 = 1.66, P = 0.206).5.3.2 Ether-extractable resinFeeding trees contained significantly less ether-extractable resin thanadjacent control trees (Fig. 2.4 (b); Table 2.1 (b)). Young needles from feedingtrees contained a mean of 15.3% resin while young needles from adjacentcontrol trees had a mean of 16.7% resin by dry weight.5.3.3 Ratio of Resin to NitrogenThe ratio of resin content to nitrogen content is significantly lower infeeding trees than in adjacent control trees (Fig. 2.4 (c); Table 2.1 (c)). The resinto nitrogen ratio is significantly different between trees but not between ageclasses of needles within trees (Fig. 2.5).5.3.4 Monoterpenes (Camphor and Bornyl acetate)Quantitative chemical analyses of camphor and bornyl acetateconcentrations were made for 9 feeding trees and 9 non-feeding (adjacent)21controls. Needles from white spruce feeding trees contained less camphor andless bornyl acetate than needles from adjacent non-feeding control trees (Figs.2.6 (a and b)). The mean camphor content was 1.49 ± 0.38 a- 1 SE) (mg / g wetwt) for non-feeding control trees and 0.14 ± 0.05 for feeding trees. Thedifference in camphor content between tree types was significant (independent t-test, t 16 = 2.767, P < 0.01, one-tailed). Bornyl acetate concentration was alsosignificantly greater in non-feeding trees (1.96 ± 0.58) than in feeding trees (0.71± 0.12) (t16 = 2.092, P < 0.05, one-tailed test).5.3.5 Forage fiber contentThe acid-pepsin digestibility values for a random sample of white spruceneedles were significantly lower (one-way ANOVA, F2 , 61 = 15.03, P < 0.001), byabout 5 percent, than the digestibility of either feeding tree or adjacent controltree needles (Fig. 2.7 (b)). There was no difference in digestibility of needlesbetween feeding trees and adjacent control trees.5.3.6 ElementsNo significant differences were found between feeding and adjacentcontrol trees in concentrations of seven elements (Al, Ca, CI, Mg, Mn, K, and Na)(Table 2.2). Within trees, there were significant differences (ANOVA; P < 0.05)between young and old needles for CI and K. Young needles have about twicethe concentrations of both CI and K as old needles.5.4 Quality and needle ageBoth nitrogen and ether-extractable resin contents were significantlygreater in young needles than old needles regardless of tree type (Fig. 2.4, Table2.1). Young needles were over 4% higher in resin and about 0.2% higher in22nitrogen than old needles. Two elements, CI and K, were about twice theconcentration in young needles as in old needles (Table 2.2).6. DISCUSSIONThe results support the hypothesis that needle chemistry accounts for thewinter forage selection of spruce grouse. Forage preferences of spruce grousein the Yukon were negatively correlated with the concentrations of plantdefensive compounds, specifically of camphor and bornyl acetate, and werepositively correlated with both the nutrient content (nitrogen, minerals) and thedigestibility of needles.6.1 Feeding trialsFeeding trials tested the ability of spruce grouse to differentiate betweenneedles collected from feeding trees and from uneaten control trees. Themarked preference of captive spruce grouse for needles from feeding trees overneedles from either adjacent or random control trees suggests that sprucegrouse select feeding trees based on chemical characteristics of needles (Ellison1976, Hoft et al. 1987, see also Jakubas and Gullion 1990, 1991). This suggeststhat other hypotheses to explain selection of feeding trees, such as socialorganization and behavior, tree micro-habitat and micro-environment differences,and tree morphology are of secondary importance.How spruce grouse distinguish between different needle types is unknownbut it has been suggested that grouse are able to smell chemical componentssuch as secondary compounds and nitrogen in plant tissue (Hoft et al. 1987,23Jakubas 1989). Post-ingestive cues, such as digestion inhibition, may also affectforage selection by spruce grouse.The tendency of spruce grouse to prefer needles from adjacent controltrees over random control trees suggests that the quality of white spruce asforage is not homogeneous in the region (see also Hockman et al. 1989, Jensen1988, Senn et al. 1992), and that either local environmental gradients in nutrientavailability, water stress, micro-climate, and/or tree genetics influence treechemistry and hence tree selection.Little is known about the distribution and abundance of trees withacceptable chemistry. There may be many suitable unbrowsed trees. The singleoutlying value for the amount of random control tree (RC) needles eaten duringfeeding trials is interesting because it suggests that 1 of 15 (or 7%) of sprucetrees in the boreal forest are potential feeding trees.6.2 Needle chemistryEther-extractable resin content is a crude index of secondary compoundcontent and is affected by plant chlorophylls and secondary metabolites, some ofwhich may be antifeedants for herbivores. Depending on the types and activitiesof secondary compounds present in plants, even slight differences in resincontent may represent considerable differences in palatability.Feeding trees were significantly lower in ether-extractable resin thanadjacent control trees. Although absolute differences in resin content betweentrees are slight, such differences could still explain a large proportion of theselection by spruce grouse for individual trees.The nitrogen content of white spruce needles was low, about 1% dryweight. No significant differences in nitrogen content were found betweenfeeding trees, adjacent control trees and random control trees. Although there24was a tendency for nitrogen to be higher in selected trees (Fig. 2.4 (b)), nitrogenalone does not appear to account for the selection of individual trees.The ratio of resin to nitrogen (Fig. 2.4 (c)) was lower in feeding trees thanin either type of control tree. This suggests that an interaction between resin andnitrogen could affect digestive physiology of spruce grouse and hence mayexplain the selection of feeding trees. Low nitrogen content and high secondarycompound content may reduce forage digestibility since digestion is microbiallymediated (Bryant and Kuropat 1980, Mattson 1980, Palo et al. 1985). Both resincontent and nutritional content may determine which trees are selected forfeeding as has been demonstrated for capercaillie (Tetrao urogallus) inScandinavia (Linden 1984).Within trees spruce grouse prefer to eat young needles rather than oldneedles. However, the resin content of young needles is over 4 percent higherthan the resin content of old needles (Fig. 2.4 (b)). The higher nitrogen contentof young tissues (Fig. 2.4 (a)) may outweigh the costs of eating more highlydefended plant parts.Concentrations of two monoterpenes, camphor and bornyl acetate, wereinversely related to forage preferences shown by the grouse. There was roughly10 times more camphor and about 2.5 times more bornyl acetate in non-feedingcontrol trees as in feeding trees. Since monoterpenes have deterrent or toxiceffects in other herbivores such as snowshoe hares (Sinclair et al. 1988b), it ispossible that camphor and bornyl acetate could at least partly explain the winterselection by spruce grouse of feeding trees.Concentrations of camphor and bornyl acetate are low, less than 6 mg / 30g wet wt. of needles, but during winter even these small differences may beimportant. Monoterpenes, such as camphor, are highly volatile and aromatic,25therefore smell and/or taste may enable grouse to avoid individual plants orspecies high in such compounds (Jakubas 1989).Spruce grouse appear to avoid feeding in young spruce trees duringwinter. Ellison (1976) also noted that spruce grouse avoided young white spruce(< 14 years) in Alaska. This is consistent with a camphor avoidance hypothesis.Sinclair et al. (1988b) found camphor occurred in juvenile (< 2 m high) whitespruce in the Yukon at 4 times the concentration of that in mature spruce (> 5 mhigh). However, this lack of use of young trees by spruce grouse may also berelated to physical characteristics of trees such as height and branch cover.Hoft et al. (1987) also looked at several specific monoterpenes, includingcamphor, as possible chemicals explaining spruce grouse selectivity inWashington. They found no difference in camphor concentration between"activity" trees and adjacent control trees in lodgepole pine (Pinus contorta) andEnglemann spruce (Picea engelmannii). Their results indicate camphor is notpresent in lodgepole pine trees. Camphor was present in Englemann sprucetrees but activity trees were used for roosting and not for feeding (Hoft et al.1987). However, in lodgepole pines, "activity" trees were used for feeding.It appears that spruce grouse select trees with a digestibility higher thanthat of a random sample. Digestibility is inversely related to fiber content (seeLarter 1992). The acid-pepsin digestibility of a random sample of trees is aboutfive percent higher than the digestibility of feeding trees. There was no differencein digestibility between feeding trees and adjacent control trees. Digestibilitydoes not explain why spruce grouse select feeding trees, but digestibility may berelated to site effects which indirectly influence food choice. Among trees withhigher than average digestibility other chemical cues may become important indetermining the final selection of a feeding tree.26There were no differences in amounts of the seven elements between treetypes. None of the elements analyzed explain the selection of feeding trees byspruce grouse. Although there were some differences in element concentrationsbetween young and old needles within trees, it seems unlikely these differenceswould account for within tree preferences of spruce grouse.The results of these chemical analyses are correlations only. Forexample, the lower concentrations of camphor and bornyl acetate in feedingtrees may be a result of spruce grouse feeding and not necessarily a cause ofselection. Grouse feeding activities may change needle chemistry. Intensivebrowsing by grouse may severely reduce the carbon reserves trees can allocateto plant defense chemicals (Chapter 3). The large piles of droppings at the baseof trees may also increase needle palatability due to fertilization (Chapter 4). Theeffects of fertilizer on defensive compounds is generally unknown (Crawley1983), but can result in lower levels of secondary compounds (Coley et al. 1985,Bryant et al. 1991).Spruce grouse feeding preference for specific white spruce trees might bedue to a combination of factors including the levels of certain secondarycompounds and nitrogen. Selection of forage by spruce grouse appears toinvolve a compromise between meeting nutritional requirements and avoidingfoods containing high amounts of fibre and especially secondary metabolites.However, available data suggest that the levels of camphor and bornyl acetateare the most significant differences between preferred and unused trees.Selection of forage by spruce grouse appears to involve a compromise betweenmeeting nutritional requirements and avoiding foods containing high amounts ofsecondary metabolites.27Fig. 2.1. Map of study area. 1. Location of Kluane Lake Research Base and penwhere captive grouse were kept during feeding trials. 2. Intensive study site(Flint Grid). 3. Site of tree clipping and fertilization experiments. 4. Site of faecaldecomposition experiments. 5. Intensive study site (Sulphur Grid). Inset showsposition of study area in southwest Yukon.28I1,\0 Dawson Citysysk...,,N.W.T.2 KM IKLUANELAKEYUKONI Study AreaLL:- _ _ ___t .Nshyfive* 7./.\^ 4,0644,16.4.4 4,■\^5^SULPHURN LAK%,... -.,••••■■ •••■■•.... ••%.WhitehorseB.C.-.....,LAKE....,.. /99114 ;,.-,.4446* ---.4.040 \4 #t)tiii,j.. \.. _ ....."••• -..-._N.:_......, \y.■••• ''.Fig. 2.2. Diagram of a white spruce branch showing annual growth nodes andcharacteristic browse damage of spruce grouse.3031Fig. 2.3. Amounts of 3 different needle types eaten per day by captive sprucegrouse during feeding trials (n = 15). White spruce needles from feeding trees(FT), adjacent control trees (AC), and random control trees (RC) were presentedsimultaneously. Amounts eaten are means ± 1 SE.3233Fig. 2.4. The nitrogen content (% dry weight) (a), ether-extractable resin content(b), and the resin to nitrogen ratio (c) for two age classes of needles (young andold) from feeding trees (FT) and adjacent control trees (AC). Shaded barsindicate old needles. Error bars are + 1 SE. Statistical differences werecompared with three 2-way ANOVA (see ANOVA summary tables, Table 2.1 (a,b, and c)).340.6FT AC8FT ACFT^ACb.^18Tree type35Fig. 2.5. The relationship between ether-extractable resin content and nitrogencontent for two age classes of needles from feeding trees (FT) and adjacentcontrol trees (FT). (0 = young, FT; A = old, FT; 0 = young, AC; and * = old, AC).Error bars are ± 1 SE.3637Fig. 2.6. Camphor content (mg / 30 g wet wt) (a) and Bornyl acetate content (b)of feeding trees (FT) and adjacent control trees (AC). Values are means forindividual trees.3839Fig. 2.7. Nitrogen content (% dry weight) (a) and acid-pepsin digestibility (% drymatter digested) (b) of white spruce needles collected from feeding trees (FT),adjacent control trees (AC), and random control trees (RC). Error bars are ± 1SE.40Tree type41Table 2.1. Summary tables of results of 2-way ANOVA of (a) nitrogen content (%dry matter), (b) amount of ether-extractable resin (% dry matter), and (c) the ratioof resin to nitrogen content for two tree types (feeding and adjacent control) andtwo age classes of needles (young and old) (see Fig. 2.1 (a, b, and c)). Maineffects are tree type (TREE) and needle age (AGE).a. nitrogen content.TREATMENT SUM-OF-SQUARES DF MEAN-SQUARE F-RATIO PTREE 0.076 1 0.076 2.715 0.104AGE 0.873 1 0.873 31.280 0.000TREE*AGE 0.000 1 0.000 0.002 0.963ERROR 1.786 64 0.028b. resin content.TREATMENT SUM-OF-SQUARES DF MEAN-SQUARE F-RATIO PTREE 43.271 1 43.271 7.899 0.007AGE 281.808 1 281.808 51.446 0.000TREE*AGE 0.965 1 0.965 0.176 0.676ERROR 350.578 64 5.478c. ratio of resin to nitrogen.TREATMENT SUM-OF-SQUARES DF MEAN-SQUARE F-RATIO PTREE 236.593 1 236.593 13.625 0.000AGE 8.844 1 8.844 0.509 0.478TREE*AGE 7.322 1 7.322 0.422 0.518ERROR 1111.354 64 17.36542Table 2.2. Mean concentrations (ppm) of seven elements for two age classes ofneedles from two tree types, feeding trees and adjacent control trees (n = 17),(±1 S.E.). (ANOVA; P < 0.05).ElementFeeding treeYoung^OldAdjacentYoungcontrolOld PAl 310 664 245 357 0.500(106) (407) (56) (70)Ca 8524 14917 6610 12779 0.150(2282) (5375) (1218) (3289)Cl 568 229 344 166 0.017(215) (116) (73) (47)K 292 102 215 95 0.001(87) (39) (38) (25)Mg 62 45 46 31 0.229(20) (21) (10) (9)Mn 381 649 318 538 0.106(63) (151) (33) (98)Na 31 45 27 33 0.997(10) (25) (8) (13)43CHAPTER 3:SPRUCE GROUSE HERBIVORY: IMPACTS ON GROWTH, ARCHITECTURE,REPRODUCTION, AND DEFENSE OF WHITE SPRUCE.441. ABSTRACTSelective feeding by spruce grouse (Dendragapus canadensis) may affectgrowth rate, architecture, reproductive output, and chemical defence of preferredtrees. Individual white spruce (Picea glauca) trees in southwestern Yukon areexposed during winter to recurring high levels of defoliation by spruce grouse.Spruce grouse select particular spruce trees for winter feeding, leaving adjacentspruce trees of similar age and size uneaten. Between 25-90 % (mean 40 %) ofthe needles on preferred (feeding) trees were removed by feeding grouse.Tree morphology differed between feeding trees and uneaten controltrees. Twig length on feeding trees was significantly greater than on control treesindicating that feeding trees have relatively high lateral growth rates. Needlelength on feeding trees was significantly greater than needle length on controltrees. Feeding trees were similar in age but greater in height than either adjacentor random control trees. Feeding trees also tended to have longer, more highlybranched limbs and more rounded crowns than uneaten control trees. Coneproduction was significantly lower in feeding trees than in adjacent control trees.Growth rates of experimentally fertilized trees increased significantly overcontrol tree growth rates. However, the growth rates of defoliated trees did notchange in response to simulated grouse herbivory. Experimental defoliation andfertilization of trees suggests that the observed differences between feeding treesand adjacent control trees result, in part, from the effects of selective feeding byspruce grouse. Needle quality may also be positively affected by defoliation andfertilization.The relatively high growth rates, the low reproductive output, and the lowsecondary chemical content (Chapter 2) of feeding trees suggests a within treetrade-off in allocation of limited carbon resources.452. INTRODUCTIONPlants have limited resources to support their physiological processes.Hence all requirements cannot be met simultaneously and trade-offs occurbetween growth, maintenance, storage, reproduction, and defense fromherbivores. Consequently, there is often a strong inverse relationship betweenthe allocation of resources to growth and non-growth processes (Coley et al.1985, Bazzaz et al. 1987, Briggs and Shultz 1990, Chapin 1991, Herms andMattson 1992).Secondary chemical synthesis competes with growth and reproduction forcommon resources (Fagerstrom 1989, Herms and Mattson 1992) During periodsof rapid growth, secondary chemical metabolism may be nutrient and/or energylimited. This trade-off has ecological consequences for plant resourcepartitioning and allocation patterns in different environments. Plants must growfast enough to compete with neighbors, and yet still maintain physiologicaladaptations and defenses required for survival in the presence of herbivores andpathogens (Herms and Mattson 1992).Most northern conifer species produce carbon-based compounds thatdeter herbivores (Bryant and Kuropat 1980, Palo et al. 1985, Tahvanainen et al.1985, Sinclair et al. 1988b, Bryant et al. 1991). These secondary compounds arecarbon-based rather than nitrogen-based because nitrogen is a scarce resourcefor boreal plants (Coley et al. 1985). Both low nitrogen content and highsecondary compound content in conifer tissues reduce digestibility sinceherbivore digestion is microbially mediated (Rhoades and Cates 1976, Bryantand Kuropat 1980, Mattson 1980, Palo et al. 1985).The resource availability hypothesis (Bryant et al. 1983, Coley et al. 1985,Coley 1988, Herms and Mattson 1992) has been typically used to explain46variation in the degree of herbivory and the effectiveness of plant defensesamong plant species. In resource limited environments, such as the borealforest, slow growing plants with large investments in antiherbivore defenseshould be favored by natural selection over fast growing plants with necessarilylower investment in plant defense. White spruce trees in the boreal forest atKluane, Yukon are inherently slow growing and invest heavily in plant defenses(Sinclair et al. 1988b, Smith et al. 1988).White spruce is the only conifer species present at Kluane, Yukon(Douglas 1974). Individual white spruce trees vary in growth rate as well as inthe degree of browsing by spruce grouse (Chapter 2). The highly selectivefeeding on individual white spruce trees by spruce grouse, makes this a suitablesystem for examining predictions of the resource availability and carbon/nutrientbalance hypotheses within a single species of plant.The resource availability and carbon/nutrient balance hypotheses predictthat there will be a negative correlation between the growth rate and the level ofdefense of individual trees. Trees growing on more infertile soils may be slowergrowing and, in turn, have higher levels of carbon-based chemical defensessince carbon demands for growth are lower. Hence, the prediction is that sprucegrouse will feed on more rapidly growing, less defended white spruce.Are feeding trees different than adjacent and random control trees ingrowth rates and morphology? The resource availability hypothesis (Bryant et al.1983, Coley et al. 1985, Herms and Mattson 1992) predicts that spruce grousewill feed in fast growing, poorly defended white spruce and that slow growingplants will be highly defended and avoided by spruce grouse. A secondhypothesis, the herbivore optimization (or plant over-compensation) hypothesis(McNaughton 1979, 1983, Jefferies 1988, Hik et al. 1991), also predicts thatwhite spruce trees, perhaps through the effects of fertilization from faecal inputs,47should increase growth in response to herbivory. Alternatively, defoliation"stress" could decrease growth rates of feeding trees relative to control trees. Inthe first part of this chapter I compare morphology of feeding trees and controlsto test predictions of the above hypotheses.In the second part of this chapter two experimental manipulations of whitespruce trees were conducted to examine the hypothesis that observeddifferences in morphology between feeding trees and control trees result from theeffects of selective and intensive feeding by spruce grouse. Experimentaldefoliation of individual trees simulated the effects of grouse defoliation, whilefertilization experiments simulated the effects of accumulating faeces at the baseof selected trees.In this chapter I examine the following questions: (1) How do growth ratesand reproductive output of feeding trees compare with those of control trees?, (2)What are the impacts of grouse herbivory, through the combined effects ofdefoliation and fertilization, on growth and reproduction of white spruce trees?,(3) Are white spruce trees able to overcompensate for leaf area lost to feedinggrouse during previous winters?, and (4) Are spruce grouse selecting trees thatare already different in growth rates, morphology, and needle chemistry or aredifferences in these attributes arising through the feeding activities of sprucegrouse?3. STUDY AREAThis study was conducted from 1989 to 1992 near Kluane Lake, Yukon.Tree measurements were taken throughout the larger study area (Fig. 2.1) duringsummers. Both tree defoliation and fertilization experiments were conducted in awhite spruce stand at the northwest end of the study area (site 3, Fig. 2.1).484. METHODS4.1 Types of treesThree types of white spruce trees were identified and measured for growthand morphology characteristics: (i) feeding trees (FT), (ii) adjacent control trees(AC), and (iii) random control trees (RC). These tree types are described inChapter 2.4.2 MorphologyAge, trunk circumference, height, needle length, needle density, and conenumber were measured for each tree type. Trees were cored with an incrementborer and the total number of annual rings were counted to determine age.These tree cores were also used to measure growth rates (see below). Trunkcircumference was measured at a height of approximately 1 m. Tree height wasestimated by visual triangulation with known horizontal distances. Mean needlelength was determined as follows. Twenty needles from the most distal twig oneach of twenty branches, between 1 and 3 meters height, were measured withcalipers to the nearest 0.1 mm. Mean needle density was determined bycounting the number of needles found on a three centimeter length of one-yearold twig from four different branches at 1 to 3 meters height. The total number ofcones on each tree was counted using 7 x 24 binoculars. Photographs offeeding trees and control trees were used to compare tree growth forms.4.3 Growth ratesBoth annual twig length and annual ring width were measured to estimatethe growth rates of trees. Twenty twigs per tree were measured with calipers todetermine mean annual twig growth, a measure of lateral growth rate. The most49distal twig was measured from each of twenty branches on a tree between 1 and3 meters in height.Mean annual ring width for the last ten years of growth was alsodetermined for each of the three tree types. Annual rings were measured to thenearest tenth of a millimeter with calipers. One core was taken with an incrementborer from the south side of each tree, between 0.5 m and 1.0 m height, tominimize within tree variation in annual ring width.4.4 Percent defoliationIn May and June, 1990 and 1991, the proportion of all needles removedfrom feeding trees by spruce grouse was estimated visually for 36 feeding trees.I tried to score trees with an accuracy of ± 5%.4.5 Defoliation and fertilization experimentsDefoliation and fertilization experiments were conducted at site 3 (Fig.2.1). Experimental trees, between 2.5 m and 6.5 m in height, were randomlyselected and the nearest neighboring tree within this height range was used as acontrol.Twenty trees were experimentally defoliated with scissors to simulatedefoliation by feeding spruce grouse. An equal number of adjacent control treeswere also chosen. During winter (February to April) from 25 to 40 percent (mean30 percent) of all needles were removed at one time from each tree during twoconsecutive years (1990 and 1991). Cumulative defoliation of each tree after twoyears of clipping was about 40 percent. This level of defoliation wasapproximately the same as the mean percentage of needles removed by grousefrom feeding trees. Approximately equal proportions of needles were removedfrom all branches on each experimental tree. While trees were being defoliated50care was taken not to remove buds or to damage bark and I tried to mimic theeffects of grouse feeding as closely as possible.Fertilizer was applied to eighteen trees. In early spring (June) of 1991,250 g of chemical fertilizer (7 N: 2 P: 1 K) was dissolved in 4 L of water andapplied to the base of each experimental tree. Eighteen control trees received 4L of water without fertilizer. This fertilizer application rate was approximately thesame as the estimated input of nitrogen to the base of feeding trees thoughspruce grouse faeces (Chapter 4).Experimental and control trees were measured in late winter of 1990 (pre-treatment), 1991 and 1992 (post-treatment) to determine twig growth rate andneedle length. Twig length and needle length of both types of experimental treeand of control trees were determined as above. Twig and needle length of bothexperiment and control trees were measured prior to treatment as temporalcontrols. Repeated measures ANOVA were used to compare differences ingrowth rate and needle length of experimental and control trees.5. RESULTS5.1.1 Morphology: age, height, and circumferenceThe mean age of 99 trees was 68 ± 5 years (SE). There were nosignificant differences in age between feeding trees, adjacent control trees, andrandom control trees (Table 3.1).Despite general similarity in age, feeding trees were significantly greater inheight than either adjacent control trees or random control trees (Table 3.1).Tree circumference also differed significantly between the three tree types (Table3.1). The minimum age, height, and circumference of feeding trees was 30years, 5 m, and 28 cm, respectively.515.1.2 Growth formDefoliated trees tended to have different branch and crown architecturefrom undefoliated trees (Fig. 3.1). Limbs on heavily used trees appear to be bothlonger and more highly branched than limbs on either adjacent or random controltrees. The crowns of feeding trees were more rounded than the crowns ofcontrol trees.5.1.3 Growth ratesMeasurement of both twig lengths and annual growth rings indicate thattrees eaten by grouse had more rapid lateral growth rates than uneaten controltrees.The mean twig length of feeding trees (4.95 ± 0.24 cm (1 SE), n = 41) wassignificantly greater than the twig length of either adjacent control trees (4.41 ±0.39 cm (1 SE), n = 18) or of random control trees (3.62 ± 0.21 cm (1 SE), n =50). These differences in twig length are statistically significant (one-wayANOVA, F = 8.77, P = 0.000) See Figure 3.2 (a).The mean width of annual growth rings over ten years was greater forfeeding trees (0.160 ± 0.015 cm (1 SE), n = 40) than for random control trees(0.137 ± 0.009 cm (1 SE), n = 49). There was no significant difference betweenthe annual ring widths of feeding trees and adjacent control trees (one-wayANOVA, F = 1.07, P = 0.348) (Figure 3.2 (b)).5.1.4 Needle length and densityFeeding trees had longer needles and lower densities of needles thaneither adjacent control or random control trees. Needles were longer on feedingtrees than on either adjacent control trees or on random control trees (one-way52ANOVA, F = 4.66, P = 0.012) (see Figure 3.3 (a)). Needle density (mean ± 1 SE)on feeding trees was slightly lower on feeding trees (54.4 needles per 3 cm ± 1.2,n = 38) than on either adjacent control trees (55.5 needles per 3 cm ± 2.0, n =16) or on random control trees (59.5 needles per 3 cm ± 1.1, n = 45) (one-wayANOVA, F = 5.13, P = 0.008).5.1.5 Cone productionFeeding trees produced fewer cones than control trees (see Figure 3.3(b)). In 1992, a mast cone year, the total number of mature cones (mean ± 1 SE)on feeding trees averaged 72 ± 17 cones per tree for feeding trees in comparisonto 165 + 32 cones per tree for adjacent controls (paired-t = 2.72, tr,-v.05(1)22 =p < 0.01).In both 1990 and 1991 there was complete failure of cone production onfeeding trees, although this was not quantified. In 1990 and 1991 coneproduction was also low in non-feeding trees sampled at Kluane (35.4 and 29.7cones per tree, respectively), (Kluane Boreal Forest Ecosystem Project, NSERCReport, February 1992).5.2 Grouse defoliation of feeding treesThe visual estimates of the percentage of needles removed from feedingtrees (n = 36) by spruce grouse ranged from 25 to 90 percent, with a meanpercent defoliation of 40 ± 2 (SE).5.3 Tree defoliation and fertilization experimentsAll trees selected for experiments and controls were between 2.5 m and6.5 m in height and all were less than 50 years old. There were no significant1.717,53differences (two-sample t-test) between experimental and control trees in age,height, or circumference (Table 3.2).5.3.1 Effects on twig growthThe growth rates of experimentally fertilized trees were significantly higherthan control tree growth rates (Table 3.3). Repeated measures ANOVA on twiglength indicated significant fertilizer effects (F 1 , 34 = 4.244, P = 0.047), and yeareffects (F 1 , 34 = 18.11, P < 0.001), but no significant interaction between year andfertilizer treatment. Also, see Chapter 4, Fig. 4.4, for the results of a seperatefertilization experiment on twig growth.The growth rates of defoliated trees did not change in response tosimulated grouse herbivory (Table 3.3).5.3.2 Effects on needle growthFertilizer had little effect on needle length (Table 3.3) (P > 0.6 for fertilizerand year effects, repeated measures ANOVA).Needle length also did not differ significantly between experimentallydefoliated trees and control trees (Table 3.3). A repeated measures ANOVAindicated no significant defoliation effects or year effects and no interactionbetween year and treatment.6. DISCUSSION6.1 Growth rateFeeding trees grew more rapidly than uneaten control trees. Feedingtrees had high lateral growth rates but did not differ in rates of woody materialaccumulation. This is consistent with both limited resource availability and plant54over-compensation in response to herbivory. Both hypotheses predict thatgrowth rates of feeding trees will be higher than for control trees. Within apopulation the fastest growing plants are often the least resistant to herbivores(Bryant et al. 1983, 1992b).Figure 3.2 (a) indicates that there is variation in growth rates of individualwhite spruce trees which may reflect underlying variation in soil quality or geneticdifferences. The clumped distribution of feeding trees on a 40 hectare study area(Sulphur, Fig. 2.1) of relatively continuous habitat is consistent with either ofthese possibilities (see Appendix I).Twig growth on experimentally fertilized trees was greater than forunfertilized control trees. Twig growth of experimentally defoliated trees was notsignificantly different from that of undefoliated control trees. This suggests thatfertilization with spruce grouse faeces is the mechanism accounting for the higherrelative growth rates of feeding trees.Nutrient availability, soil moisture, plant competition light availability, timingof herbivory, type of tissue loss may all interact to determine the degree of plantcompensatory response to herbivory (Whitham et al. 1991). Slow growingconifers such as white spruce are though to have little potential for compensatoryresponse regardless of ecological conditions (Whitham et al. 1991, Bryant et al.1991). Growth of these species may be limited by resource availability.6.2 Growth formDefoliated trees show different architecture than undefoliated trees.Feeding trees appear to have longer, more highly branched, limbs and morerounded crowns than control trees (Fig. 3.1). The different growth form offeeding trees may result from repeated exposure to herbivory by grouse over55several years. Grouse feeding may change tree growth form either (i) througheffects on tree growth rates or (ii) by changing apical dominance of trees.The observed growth form of feeding trees may result from the destructionof terminal shoots by feeding spruce grouse over several years. Shoot mortalityeliminates the normal patterns of apical dominance that determine treearchitecture (Kozlowski 1971, Whitham and Mopper 1985) and may stimulate theproduction and growth of lateral buds.Removal of the terminal vegetative buds on twigs by herbivores can alsostimulate development of latent buds (Owens and Molder 1984, Whitham andMopper 1985). With this inhibition removed, latent buds may form numerousepicormic shoots i.e. new shoots that develop on older twigs. This could lead tothe more highly branched structure of feeding trees. It would also result in anincrease in the surface area of needles available for feeding by spruce grouse.6.3 Cone productionCone production was significantly lower in feeding trees than in adjacentcontrol trees in 1992, a mast year. Feeding trees had an average of 44% fewercones than adjacent control trees. In two non-mast years feeding trees hadvirtually no cones.The lower reproductive output of feeding trees in 1991 may have resultedfrom the activity of feeding grouse: (i) defoliated trees may reallocate limitedresources to growth rather than to non-growth processes such as reproduction,(ii) defoliation may change dominance and stimulate growth of lateral shoots overterminal shoots, and (iii) feeding spruce grouse may eat developing seed-conebuds during winter.White spruce is monoecious, producing both male pollen-cones andfemale seed-cones on the same tree. Pollen and seed cones are usually borne56on separate shoots. Female cones are produced by the long terminal shootsnear the top of the tree while male pollen-cones are produced by the shorterlateral shoots at the sides and base of the tree (Owen and Molder, 1984).Herbivory may stimulate production of lateral male-bearing shoots relative toterminal female bearing shoots and indirectly cause white spruce to halt femalecone production and become functionally male plants (Whitham and Mopper1985, Bryant et al. 1991). It would be interesting to see if cone production isreversed over time if spruce grouse are experimentally excluded from usingfeeding trees.Spruce grouse may directly alter the reproductive output of feeding treesby eating developing cone buds during the winter proceeding cone development.Cone buds are high in nitrogen and carbohydrates and may be a preferred foodof spruce grouse. Examining grouse crops for cone buds, comparing the numberof developing cone buds on feeding trees and control trees, and observingbehavior of spruce grouse would allow one to examine this hypothesis.6.4 Defoliation and fertilization experimentsSimulated defoliation had no effects on either twig length or on needlelength of white spruce. Experimental fertilization increased twig length and therewas a slight tendency for fertilizer to increase needle length (Table 3.3).Fertilization rather than defoliation appears to account for the relatively highgrowth rates of feeding trees.All defoliation and fertilization experiments were conducted with youngwhite spruce trees that were less than half the mean age of feeding trees. I wasunable to experimentally defoliate larger trees for logistic reasons. How whitespruce trees of different ages respond to fertilization and to defoliation isunknown.57Trees were defoliated with scissors. Defoliation damage resulting fromscissors appeared similar to the activities of spruce grouse but see Baldwin(1990) for a discussion of potential problems with herbivory simulations.Timing may also be important in determining the response of white sprucetrees to fertilization and to defoliation. Trees were experimentally defoliatedduring winter in a single period of one to two days. Spruce grouse return to eat infeeding trees throughout winter. Trees were fertilized in spring, perhaps justslightly after the main flush of soluble nutrients from spruce grouse faeces(Chapter 4).6.5 Forage qualityA number of lines of evidence suggest that forage palatability (or quality)of white spruce will increase as a consequence of defoliation by spruce grouse.Spruce grouse return to the same feeding trees for several years (Ellison1976, Hoft et al. 1987) Traditional and intensive use of feeding trees does notpromote effective inducible defenses (e.g. Schultz 1988, Karban and Myers1989) in winter defoliated white spruce. There may be physiological constraintson plant chemical resistance. Forage quality of conifers can increase inresponse to both fertilization and defoliation. This and absence of induceddefense may explain why spruce grouse can use the same trees for severalyearsHerbivory often increases forage quality of plants. Scots pine (Pinus sylvestris) lacks an effective inducible defence against its insect herbivores.Defoliation of Scots pine leads to a relative carbon deficiency for the plant andreduces the accumulation of carbon-based substances in the foliage (Niemela etal. 1991). In Sweden, browsing by moose (Alces alces) on birch (Betula spp.)can also induce regrowth that is more palatable than forage on unbrowsed plants58(Danell and Huss-Danell 1985). Shrubs browsed by black-tailed prairie dogsshow similar responses (Coppock et al. 1983).Plant stress, for example being eaten, can also increase plant suitability toinsect herbivores (Mattson and Addy 1975, Rhoades 1979, White 1984, Williamsand Myers 1984, Jones and Coleman 1991, Watt et al. 1991). Rhoades (1979)postulated that as food, stressed plants have a net positive effect on herbivorefitness. White (1984) found that plant stress results in increased concentrationsof mobile nitrogen in leaves, particularly amino acids.6.6 Trade-offsThe relatively high growth rate, the low reproductive output, and the lowsecondary chemical content (chapter 2) of feeding trees suggests a trade-off inallocation of limited carbon resources.In nutrient deficient environments such as the boreal forest, both carbon(carbohydrate) and nutrients (particularly nitrogen and phosphorus) may belimiting for plant growth. Fertilizer experiments indicate that white spruce trees atKluane are nutrient limited. Therefore, feeding trees could also be carbonlimited. This could occur, first, through defoliation by spruce grouse which wouldreduce the total leaf surface of feeding trees. If a large enough proportion ofneedles are removed from feeding trees it is possible that net photosynthesis rateof the trees would become limiting. Second, slow growing conifers, like whitespruce, store most nitrogen and carbohydrates in their leaves and below-groundstorage is minimal (Bryant et al. 1983, Herms and Mattson 1992). Therefore,extensive defoliation of white spruce may reduce both carbon production andcarbon storage and may change the carbon/nutrient balance of the entire plant.If carbon becomes limiting trees may be forced to increase photosyntheticarea in order to increase the tree's ability to produce additional photosynthate.59Heavily defoliated trees may be forced to allocate remaining carbon resources toregrowth rather than to reproduction and defense.The regrowth of leaves following defoliation requires substantial quantitiesof nutrients and carbohydrates (Herms and Mattson 1992). Carbohydratedemands of rapid regrowth may reduce carbon-based secondary metabolitesynthesis. This may result in patches of highly palatable forage that attractfurther feeding, generating a feeding-regrowth feedback loop (see also Du Toit etal. 1990, Bryant et al. 1992a and b). Such trees may be considered analogous tograzing lawns (e.g. McNaughton 1984).60Figure 3.1. Photos of white spruce trees showing growth form of (a) feedingtrees, and (b) control trees.61Figure 3.2. Comparison of growth rate differences between feeding trees (FT),adjacent control trees (AC), and random control trees (RC) as measured by (a)annual twig length and (b) annual ring width. Error bars are ± 1 SE.63Tree type64Figure 3.3. Differences in (a) needle length and (b) cone number betweenfeeding trees (FT), adjacent control trees (AC), and random control trees (RC).Error bars are ± 1 SE.65Tree type66Table 3.1. Comparison of age, height, and circumference between feeding trees(FT), adjacent control trees (AC), and random control trees (RC) (Mean ± 1 SE).F and P statistics of one-way ANOVA are also presented.FT AC RC Fage 72.0+8.5 90.7+15.5 43.6+1.3 F = 2.85 P=0.063(years) -n=3-6 n=T4 n=49height 9.7+0.6 8.4+0.6 7.5+3.8 F = 4.16 P=0.018(m) n=39 n=i5 n=49Circumfer 77.6+6.7 92.3+10.5 57.1+5.3 F = 5.88 P=0.004(cm) n=-36 n=15 n=5067Table 3.2. Physical characteristics of experimentally defoliated and fertilizedtrees and adjacent control trees (mean ± 1 SE). Experiment and control treesare compared with a two-sample t-test (significance level p < 0.05).Defoliated^Control^t-values^Prob.Age (years)^30.5 +1.2 30.2 + 1.5Height (m)^3.8 + 0.1^4.1 + 0.2Circumfer (cm) 22.2 + 1.0 23.7 + 1.0_t38= 0.18 p = 0.86t38= 1.67 p = 0.10t38= 0.99 p = 0.33Fertilized Control t-values Prob.Age^(years) 35.8 + 2.5 32.4 +^1.7 t 34= 1.13 p = 0.27Height^(m) 4.9 +^0.3 4.9 +^0.3 t 28= 0.19 p = 0.86Circumfer^(cm) 28.9 +^1.5 29.1 +^1.9 t 32= 0.06 p = 0.9568Table 3.3. Results of experimental defoliation and fertilization on twig length andneedle length of white spruce trees. F and P statistics of repeated measuresANOVA are given in the text. Values are means with 1 SE given in parentheses.All lengths are in centimeters.a. Fertilization experiment (n = 18 trees in each cell)twig length^needle lengthyear^fertilized^control^fertilized^control1991 4.36 5.30 1.37 1.37(0.32) (0.29) (0.02) (0.03)1992 5.39 6.04 1.38 1.35(0.26)^(0.37)^(0.03)^(0.03)b. Defoliation experiment (n = 20 trees in each cell)twig length^needle lengthyear defoliated control defoliated control1990 5.85 5.26 1.40 1.43(0.26) (0.40) (0.04) (0.04)1991 6.62 5.90 1.46 1.42(0.27) (0.35) (0.04) (0.04)1992 7.07 6.89 1.40 1.37(0.38) (0.40) (0.04) (0.04)69CHAPTER 4:FAECAL DECOMPOSITION IN THE BOREAL FOREST: POSITIVEFEEDBACKS BETWEEN SPRUCE GROUSE FORAGING ANDNITROGEN CYCLING.701. ABSTRACTFew studies have looked at decomposition processes and nutrient cycling inboreal ecosystems. This chapter examines some direct and indirect interactionsbetween selective winter herbivory by spruce grouse, individual white sprucetrees, and local soil processes. Between 0.3 to 5.0 kg dry weight (dwt) of sprucegrouse faeces accumulate annually under feeding trees. Annual faecal nitrogeninputs to the base of feeding trees range from 3 to 51 g N tree -1 or from 0.7 to3.7 g N m -2 . Carbon / nitrogen ratios of spruce needles, grouse faeces, and soilsare high (48, 48, and between 25 to 50, respectively). Decomposition of sprucegrouse faeces was estimated using mesh bag experiments. From April toSeptember of the first year faeces lost > 50 % mass, > 40% total N, and almostall soluble N (NH4+). Leaching of NH4+ coincided with spring snow-melt. Duringexperimental incubations, microbial respiration rates, a measure ofdecomposition rate, were two to three times higher for spruce grouse faeces thanfor spruce litter (61 vs. 26 mg CO2 / g dwt / 14 days at 21°C, mean values duringthe first summer). Respiration rates of three year old faeces were comparable tothose of one year old spruce litter. Respiration rates were also positively relatedto water content of faeces. During incubations there was net nitrogenmineralization in spruce grouse faeces of +1.01 and +3.94 (mg NH4+ / g dwt / 14days at 21°C) in June and August 1991, respectively. Fertilization experimentsindicate that white spruce trees are strongly nitrogen limited. Twig growth ofexperimentally fertilized trees increased significantly, 14 and 25 %, relative tounfertilized trees. In the nutrient limited boreal forest, spruce grouse faecalinputs under feeding trees may locally increase soil nutrient availability andnutrient cycling rates. The nitrogen input from spruce grouse faeces may be amechanism leading to the relatively high growth rates of feeding trees and to the71higher forage quality of feeding trees compared with control trees (see Chapters2 and 3). The combined effects of defoliation and nutrient return on the nutritionand growth responses of white spruce trees and on soil processes may ultimatelyfeedback positively upon spruce grouse forage selection and account for theprolonged use of individual trees for winter feeding.2. INTRODUCTIONHerbivores play an important role in nutrient cycling in both terrestrial(Zlotin and Khodashova 1980, Botkin et al. 1981, Floate 1981, Coppock et al.1983, Cargill and Jefferies 1984, McNaughton 1984, McNaughton et al. 1988,Pastor et al. 1988, Day and Detling 1990, Moore 1992) and aquatic (Kitchell et al.1979, McDonald 1985, Carpenter 1986, Power 1990, Setala 1991) communitiesthrough mobilization and redistribution of nutrients. Productivity in terrestrialecosystems depends upon the amounts of nutrients available and on the rates ofnutrient cycling among plants, litter, soil, and animal biomass. Herbivory maydirectly and indirectly increase rates of nutrient cycling and mineralization ofelements (Ruess and McNaughton 1987, Jefferies 1988, Pastor et al. 1988,Ruess et al. 1989, Holland and Detling 1990). These effects depend on thetiming and amount of herbivory, the distribution and amount of faecal material,and the ability of the vegetation to utilize these faecal nutrients. The organicnitrogen content of animal faeces increases plant growth and changes plantforage quality in terrestrial ecosystems where nitrogen limits primary productivity(e.g. Ruess and McNaughton 1987, Hik et al. 1991). Soil processes, in turn,affect the supply and quality of browse and the rate at which plants recover frombrowsing (Richards 1987). The combined effects of defoliation and nutrient72return by herbivores ultimately may feed back in a positive manner uponherbivore forage selection and population growth (McNaughton 1984, Day andDetling 1990).Studies of the effects of faecal decomposition on nutrient cycling rateshave been conducted in tropical grasslands of the Serengeti (Ruess andMcNaughton 1987, McNaughton et al. 1988), temperate prairie grasslands(Schimel et al. 1986, Whicker and Detling 1988, Day and Detling 1990), andarctic coastal salt marshes (Jefferies 1988, Ruess et al. 1989). In each study,grazed sites exhibited increased nutrient availability, nutrient cycling rates, andforage quality compared to ungrazed sites. Increased rates of flow of nutrientsinto sediments via herbivore faeces appears to be an important mechanism formaintaining high rates of nutrient cycling in the grazing food web on sites of lownutrient availability. Deposition of faeces by herbivores may directly influence thenitrogen content of plant tissues, secondary compounds, and hence herbivorefeeding rates.Selective foraging by herbivores can cause the rate of nutrient cycling toincrease resulting in positive feedbacks. This is especially true of nitrogenmineralization of nitrogen limits net primary production in many terrestrialecosystems (Sprent 1987), and the nitrogen content of plant tissues maydetermine herbivore feeding rates (Mattson 1980, Robbins 1983), as well as thekinds and amounts of chemical defenses (e.g. Karban and Myers 1989).Feeding preferences, tissue chemistry, and litter decay rates are correlated.Forage digestibility for herbivores and litter decomposition are bothmicrobially mediated and determined by nutrients, structural carbohydrates,lignins, and secondary compounds of plant tissues (Bryant et al. 1991, Pastorand Naiman 1992). The chemical quality of plant material partly determineswhich plants are eaten (Bryant and Kuropat 1980) as well as their fate once they73enter the soil (White 1986, Horner et al. 1988). Production of defensivecompounds also appears to be inversely correlated with growth rates, nutrientuptake rates, and retention time of leaves (Coley et al. 1985, Coley 1988, Hermsand Mattson 1992).Northern boreal forest ecosystems are characterized by large seasonalfluctuations in day length, a short growing season, low soil temperatures, lownitrogen availability, and slow rates of nutrient cycling (Flanagan and Van Cleve1983, Van Cleve et al. 1991). The effects of herbivores on soil nutrientavailability and nutrient cycling may be particularly pronounced when nutrientavailability is low. In boreal ecosystems, the availability of nitrogen is animportant factor limiting production (Nadelhoffer et al. 1984, Flanagan and VanCleve 1983, Van Cleve et al. 1991) and the rate at which elements are releasedis largely governed by litter decomposition rate (Berg 1988, Pastor et al. 1988).This process is mediated by soil micro-organisms (Smith and Paul 1990).However, where herbivores consume large amounts of above-ground biomass,rates of nutrient cycling may be influenced by both changes in litter quality andthe deposition of faeces (Pastor et al. 1988, Bryant et al. 1991, Pastor andNaiman 1992).Few studies have looked at decomposition processes and nutrient cyclingin boreal ecosystems. In this chapter I examine some of the direct and indirecteffects of selective winter herbivory by an avian herbivore (spruce grouse,Dendragapus canadensis) on individual white spruce trees (Picea  glauca) and onlocal soil processes.In winter, spruce grouse in Yukon eat only conifer needles and feedheavily upon particular individual white spruce trees. These feeding trees areused throughout the winter and some are utilized repeatedly for several years.Each winter large numbers of spruce grouse faeces accumulate on the ground74under preferred trees. Both forage quality of needles (Chapter 2) and twiggrowth rates (Chapter 3) are higher on feeding trees than on either adjacent orrandom control trees. Needles from feeding trees have both significantly lowerratios of resin to nitrogen and lower concentrations of two monoterpene plantdefenses, camphor and bornyl acetate, than needles from control trees (Chapter2). Intensive and selective herbivory by spruce grouse affects growth rate, plantarchitecture, reproductive output, and chemical defence of selected trees(Chapter 3).I propose that selective foraging by spruce grouse also has local effectson soil nutrient availability, nutrient cycling rates, and the quantity and chemicalquality of litter returned to soils under feeding trees (Fig. 4.1).To test mechanisms by which spruce grouse may be accelerating nutrientcycling rates I examine the following questions: (1) What amount of grousefaeces are deposited annually under feeding trees?, (2) What are thedecomposition rates of grouse faeces and spruce litter?, (3) Are nutrients fromdecomposing spruce grouse faeces available for use by spruce trees?, and, if so,(4) What effects do such nutrient inputs have on spruce tree growth and quality?3. STUDY AREAThe study site was located in the Shakwak Trench near Kluane Lake,Yukon (Fig. 2.1, Site 4). See Chapter 2 for descriptions of the climate andvegetation of the region.Soils in the study area are typically eutric brunisols (Zbigniewicz and Dale199N). Both the parent material and the soil of the region are poor in essentialmineral nutrients (Zbigniewicz and Dale 199N, Douglas and Knapik 1974). InJune 1988, mean organic matter in the soil near Site 4 was 11.4 + 2.7 % and the75amounts of NH4+ and NO3" were 11.7 + 1.9 ppm and 0 ppm, respectively(Zbigniewicz and Dale 199N). Soil ammonium (NH 4+) concentrations in theShakwak Valley range from 3 - 49 ppm and soil nitrate (NO 3- ) concentrationsrange from 0 - 5 ppm (Zbigniewicz and Dale 199N).4. METHODS4.1 Grouse faeces under feeding treesSpruce grouse faeces were counted under feeding trees in June of 1990and 1991. At this time of year it is possible to distinguish new faeces (less thanone year old) from older faeces. New faeces are less weathered and lighter incolor than old faeces. Only grouse faeces less than one year in age werecounted to estimate annual input rates of faeces to the base of feeding trees.The density of spruce grouse faeces under 36 feeding trees wasestimated by placing a 0.25 m x 0.25 m quadrat directly on the ground, between30 to 60 cm from the trunk of each feeding tree, and all faeces within eachquadrat were counted. Faeces were counted in eight quadrats systematicallyplaced quadrates for each feeding tree; two replicates at each of the north, east,west, and south sides. The mean density of faeces in these eight replicates wasused to estimate the density of grouse faeces under each feeding tree.The area under each feeding tree covered with spruce grouse faeces wasestimated by laying a tape measure on the ground around the perimeter of thearea covered with grouse faeces. These circumference measurements wereused to estimate the surface area covered with faeces. I assumed that a circleapproximated the shape of the area covered with grouse faeces. Theseestimates of faecal density and of area covered were used to calculate the totalnumber of grouse faeces accumulating underneath each feeding tree per year.76Grouse have two kinds of faeces, intestinal and caecal (Moss and Hansen1980). Intestinal faeces form the bulk of faeces found under feeding trees. Theyhave high dry matter content and are regular in shape and size. The second kindof faeces, caecal faeces, are fluid, have a distinct caecumy odor, and are voidedfrom the caecum once or twice a day. Caeca! faeces have two to three times thenitrogen content of intestinal faeces and make up a small percentage of totalfaecal output. Caecal droppings make up about 12 percent (dry matter) of thetotal faecal output of red grouse in Scotland (Moss et al. 1990).Caecal faeces were not included in any of the analyses that followbecause of their infrequency and because of the difficulty in counting individualfaeces. Therefore estimates of both faecal accumulation rate and of nitrogeninput rate are conservative lower limits.4.2 Decomposition of faeces4.2.1 Grouse faeces in mesh bagsDecomposition of grouse faeces was measured in summer 1991. Fresh,frozen grouse faeces were collected from under feeding trees and from snowburrow sites in February, March, and April 1991. Collected faeces were mixed,while frozen, so that quality of subsamples was uniform. Care was taken to avoidfragmenting individual faeces. Approximately 9-11 g of fresh faeces wereweighed, put in labelled fiberglass mesh bags (10 cm x 10 cm, 1 mm mesh size),and placed on the study area prior to snow melt. Samples were laid directly onthe soil at Site 4 on 24 April. Subsequently bags were collected on 15 May (n =20), 6 June (n = 20), 11 June (n = 5), 24 June (n = 6), 8 July (n = 20), 8 August (n= 20), and 19 September (n = 15). Mesh bags were weighed immediately afterremoval from the field. Faeces were oven-dried at 60 0 C for 72 hours and77reweighed to determine water content and mass loss.. Sub-samples of grousefaeces were analyzed for total nitrogen, soluble nitrogen (NH 4+, NO3- ), andphosphate (PO4-3) on a Techicon autoanalyzer (see below).4.2.2 Incubations of grouse faeces and spruce litterMicrobial respiration and net nitrogen mineralization rates were measuredduring summer 1991 and compared between three ages of grouse faeces (one,two, and three years) and with spruce litter collected from under control sprucetrees. Approximately 15 g fresh weight of sample (faeces or litter) were placedon a moistened sponge at the bottom of a 500 ml Mason jar, together with a 20ml scintillation vial containing 15 ml 1 N NaOH to trap respired 00 2 . Sampleswere then placed in an insulated box and incubated for 14 days at constanttemperature (21°C). All incubations were kept in the dark and three jars withoutfaeces served as controls. NaOH traps were changed after 1 and 7 days andtotal respiration was calculated as the cumulative total of the three traps. Theamount of CO2 respired was determined by titrating diluted NaOH (1:6) with 0.05N HCI. The volume (ml) of titrant used to change the pH from 8.3 to 3.7 wasused to determine the amount of 002 respired according to the following formula:mg CO2 = ml titrant x 0.6 x 6 (Jenkinson and Ladd 1981). Soluble ammoniumwas measured initially and after 14 days and the difference was taken as netnitrogen mineralization or immobilization.4.3 SoilsSoil samples were collected from under feeding trees and adjacent controltrees (n = 18 from each type) during June 1992. One block (3 cm x 3 cm x 3 cm)of soil was collected from the south side of each tree, about 0.5 m from the trunk,from a depth of about 5 cm. Soils were dried in an oven at 60°C to constant78weight and then passed through a 1 mm mesh screen prior to analyses for totalcarbon and total nitrogen.4.4 Chemical analyses of faeces, soils, and spruce needles4.4.1 NitrogenFaeces, soils, and needles (see collection methods for spruce needles inChapter 2) were analyzed for total nitrogen content using macro-Kjeldahldigestion. Samples were dried and ground and 0.5-1.0 g of each was weighedand digested on a Tecator block digestor at 4100C in sulphuric acid andhydrogen peroxide using selenium catalyst (Parkinson and Allen 1975).Nitrogen content of digested samples was then determined colorimetrically on aTechnicon II autoanalyzer (Technicon Autoanalyzer II. 1974. Industrial MethodNo. 321-74A). For each sample two replicates were run and for each replicatetwo autoanalyzer readings were taken. Values are expressed on a dry weightbasis.Soluble nutrients were extracted from fresh and aged faeces with 30 ml of2 N KCI. Extracts were filtered, frozen, and transported to the University ofBritish Columbia for analysis for NH 4+, NO3- , and PO43 using a Technicon IIautoanalyzer.4.4.2 CarbonThe amounts of total carbon present in soils, faeces, and needles weredetermined using a LECO Gasometric Carbon Analyzer (Model No. 572-200).About 0.05 g of dry, ground sample was accurately weighed and placed in aceramic crucible with tin and iron accelerator chips. CO2 produced duringcombustion of the sample was absorbed by KOH solution and the volumetric79difference gave the percent carbon (% C) according to the following formula: %C = LECO reading x correction factor x (1 / sample weight). The correction factoraccounts for temperature and atmospheric pressure changes between sampleruns. A blank and a known standard were run at the start of each analysisperiod.4.5 Tree growthThe growth response of spruce trees to artificial fertilizer application wasmeasured. Three study areas were fertilized by aerial spraying with chemicalfertilizer at the beginning of the growing season (May to June) of years 1987,1988, and 1989 (Krebs et al. 1992, Kluane Boreal Ecosystem Project, NSERCReport 1992). No fertilizer application was made in 1986. In 1987 nitrogenfertilizer (ammonium nitrate) was applied at 25 g / m 2 . In 1988 NPK fertilizer wasapplied at 17.5 g N / m2 , 5 g P / m2 , and 2.5 g K / m2 . NPK fertilizer was used in1989 at half the application rate of 1988.In 1990 annual twig lengths were measured for the previous four years ofgrowth from 25 randomly selected trees on each of three fertilized study areasand three control study areas. Twig length was measured with calipers to thenearest millimeter for growth that occurred in years 1986, 1987, 1988, and 1989.Three branches were measured on each tree. Relative twig growth, as apercentage of initial growth, was determined by comparing growth after fertilizerapplication with growth that occurred in 1986, prior to fertilizer application.Results were analyzed with a repeated measures ANOVA.805. RESULTS5.1 Grouse faeces under feeding treesThe maximum direct faecal nitrogen input to the base of feeding trees andthe measurements and calculations used to obtain these estimates are given inTable 4.1. Annual faecal nitrogen inputs range from 3 to 51 g N per tree or from0.7 to 3.7 g N per square meter.Faeces and needles accumulate under the same tree they originate from.Observations of wild grouse indicate that transportation of material by sprucegrouse from other areas and trees is minimal.5.2 Chemical content of needles, faeces, and soil.The initial total nitrogen content of grouse faeces (mg N / g dwt) for 12samples was 10.249 ± 0.573 (1 SE). The initial soluble nitrogen content (mgNH41- / g dwt) of spruce grouse faeces was 7.6 ± 1.8 (1 SE), n = 4 samples.Nitrate (NO -3) and phosphate (PO43) were both at low concentration and notdetectable in fresh spruce grouse faeces.The initial total nitrogen and total carbon content of spruce grouse faeces,fresh spruce needles, and of soils are given in Table 4.2. C/N ratios were alsodetermined because they determine net mineralization rates of organic nitrogen(Lunz and Chandler 1946, Richards 1987).There were no significant differences in either nitrogen or carbon contentbetween the three types of spruce needles, so these are also expressed as meanvalues. Spruce needles were also found to be high in secondary compounds(see Chapter 2); fresh spruce needles are about 14% crude resin. Both cruderesin and two monoterpenes, bornyl acetate and camphor, were found to be at81significantly higher concentrations in needles from trees not eaten by sprucegrouse. No spruce litter was analyzed for chemical quality.The carbon content of fresh grouse faeces was slightly (3.3%) butsignificantly lower than for fresh spruce needles. There were no significantdifferences in nitrogen or in C/N ratios for any of the three needle types. Thechemical composition of grouse faeces closely resembled those of needles.The nitrogen and carbon contents of soils collected at 5 cm depth frombeneath feeding trees and adjacent control trees were also compared. Thecarbon content of these soils was about 25 percent, approximately half thecarbon content of fresh needles and faeces. The organic matter content ofcollected soils was about 42 percent (organic matter content = % C x 1.72). Thechemical analyses of soil total nitrogen content were not completed but probablyrange from 0.5 to 1.0 percent dry weight (Zbigniewicz and Dale, 199N, Hik andMueller 1993). Thus C/N ratios of soils are likely between 25 and 50.5.3 Decomposition of grouse faeces5.3.1 Mesh bag experimentsMesh bag experiments indicated that from April to September, 1991 therewas > 50 % mass loss, > 40% total N loss, and nearly complete loss of soluble Nfrom decomposing spruce grouse faeces (Fig. 4.2).Mass loss from spruce grouse faeces followed a typical negativeexponential decay (Fig. 4.2 (a)). Only 49 percent of the initial mass of faecesremained at the end of September. The total nitrogen content of remainingfaeces declined about 40 percent. Total nitrogen was lost at a fairly constant ratefrom the beginning of April until the end of September (Fig 4.2 (b)). Solublenitrogen (NH 4+) loss from spruce grouse faeces was rapid and almost complete82by the end of May (Fig. 4.2 (c)). Leaching of soluble nitrogen from faecescoincided with the spring snow melt. After the spring snow-melt, water content offaeces was largely dependent upon the frequency and duration of rainfall. Thechanges in percent water content of grouse faeces were therefore large, rangingfrom 14% to 78% over the course of the summer.5.3.2 Incubation experimentsRates of microbial respiration and nitrogen mineralization were determinedin experimental incubations of spruce grouse faeces and of spruce litter.Microbial respiration rate is a measure of microbial activity and thus an index ofdecomposition rate (Richards 1987).The respiration rates (mg CO2 / g dry weight / 14 days at 21°C) of differentaged spruce grouse faeces and of spruce litter are shown in Figure 4.2 (a).Respiration rates of spruce grouse faeces were about three times higher thanrespiration rates of spruce litter. Respiration rates were unstable but declined asspruce grouse faeces aged. After three years the respiration rates of grousefaeces were comparable to respiration rates of spruce litter. The high variabilityin respiration rates of faeces in year one is probably related to changes in watercontent of faeces associated with variable rainfall events. Respiration rates offaeces are positively correlated with the water content of faeces when collectedfrom the field (Fig. 4.3 (b)).During incubations there was net nitrogen mineralization in spruce grousefaeces. The mineralization rates during incubations in 1991 are +1.01 and +3.94(mg NH441 g dwt / 14 days at 21°C) during June and August, respectively.835.4 Fertilizer experiment and tree growthFigure 4.4 shows the relative twig growth of experimentally fertilized whitespruce trees and unfertilized control trees. There were no differences in 1987(year of fertilization). One year after fertilization, spruce trees respondedstrongly, increasing lateral twig growth by 14 % in 1988 and by 25 % in 1989relative to prefertilization growth rates.6. DISCUSSION6.1 InputsLarge quantities of spruce grouse faeces are deposited under feedingtrees and annual faecal nitrogen inputs range from 3 to 51 g N per tree or from0.7 to 3.7 g N m -2. These are maximum direct inputs since not all nitrogen isreleased from decomposing faeces in the first year and indirect inputs such asmicrobial seeding and enhanced decomposition of adjacent litter are notconsidered.Estimates of maximum direct nitrogen input are conservative for thefollowing reasons. Firstly, caecal faeces were not included in any of thecalculations. Caecal faeces of red grouse in Scotland (Moss et al. 1990)comprised 12 percent of the total dry matter output of faeces and the nitrogencontent of caecal faeces was 2.2 times higher than the nitrogen content ofintestinal faeces. If these values are similar for spruce grouse thenapproximately 23 percent of the annual faecal nitrogen input is unaccounted forin the estimates given in Table 4.1. Secondly, indirect effects on nutrient inputsto soils are not included in these estimates.Very few or none of the faeces under feeding trees are from other areas ortrees. Faeces and needles accumulate under the same tree they originate from84and transportation of material by spruce grouse from other areas and treesappears minimal.6.2 DecompositionDecomposition of spruce grouse faeces is rapid compared withdecomposition of spruce litter. Microbial respiration rates are 2 to 3 times higherin grouse faeces than in spruce needle litter (Fig. 4.3 (a)). Initial decompositionof spruce grouse faeces is also rapid compared to the faeces of other herbivores(hare and moose) in the boreal forest (Hik and Mueller 1993). This is probablydue to the relatively high surface to volume ratio of grouse faeces.Mass loss rates of white spruce needles are low compared to rates foundfor spruce grouse faeces. In mesh bag experiments in southern Alberta, Tayloret al. (1991) found that white spruce needles lost 22% of mass in the first year,and only 53% after four years. Needle quality (N = 47.4%, C = 60%) wascomparable to this study although, because of latitude, the number of daysduring which decomposition can occur in the Yukon is considerably fewer than insouthern Alberta. Spruce grouse faeces appear to lose mass over four timesfaster than spruce needles.6.3 Chemical qualityWhen climate and site factors such as soil type are constant,decomposition rates are regulated primarily by the chemical composition andphysical structure of litter (Swift et al. 1979, Taylor et al. 1989). The initialnitrogen content or the ratio of carbon to nitrogen are often considered to becritical to decomposition. Many studies have tried to predict from C/N ratioswhen net nitrogen mineralization will occur (e.g. Edmonds 1980, Berg andEkbohm 1983, Stohlgren 1988). Lunz and Chandler (1946) suggested that85nitrogen mineralization in forests should occur at C/N ratios between 20/1 and30/1. Edmonds (1980) also found critical C/N ratios in the 20/1 to 30/1 range forconifer needles. The C/N ratio of white spruce needles was about 48 whichsuggests net nitrogen immobilization.The C/N ratio of the soil samples is greater than 25:1 and the totalnitrogen content of the sediments is less than 1.0% by weight. Harmsen and VanSchewen (1955) and Campbell (1978) report that the generally accepted valuesfor equilibrium between net rates of immobilization and mineralization of nitrogenare C/N ratios of 20-25:1 and a soil nitrogen content of 1.5 to 2 %. Althoughthere is a large range of variability in the critical percentages of nitrogen and inC/N ratios at which net immobilization gives way to net mineralization (Hayes1986), the low nitrogen values and high C/N ratios indicate that netimmobilization should predominate in the sediments for much of the summer.The high C/N ratios of spruce needles and soils collected from belowfeeding trees suggest net immobilization of nitrogen. Conversely, netmineralization of soluble nitrogen was observed during incubations of freshspruce grouse faeces. It is likely that net mineralization predominates early andlate in the season (Nadelhoffer et al. 1984, Bazely and Jefferies 1989) perhaps ata time when nitrogen is most required by spruce trees.It is unknown how much of the nitrogen from decomposing faecesbecomes available in soil for use by spruce trees. Studies with radioactive 15 Nare required to determine conclusively if nitrogen from faeces is incorporated intogrowing white spruce. The rapid decomposition rates of faeces, the netmineralization of faeces during incubations, and the increased growth of sprucetrees when experimentally fertilized are all consistent with the idea that faecalnutrients increase soil nutrient availability and that these nutrients areincorporated into growing spruce trees.86White spruce is a poor quality food for boreal herbivores (Chapter 2,Bryant and Kuropat 1980, Sinclair et al. 1988a, Sinclair et al. 1988a and b): it isalso a low quality litter (Flanagan and Van Cleve 1983, Moore 1984, Pastor et al.1988). The same chemical properties that reduce white spruce food qualityreduce its value for soil microbes. These properties include low concentrations ofnitrogen and high concentrations of toxic secondary metabolites, lignins, waxes,and cutins that are indigestible by herbivores (Robbins 1983, Bryant et al. 1991)and slow to decompose (Swift et al. 1979).Most northern conifer species produce carbon-based compounds thatdeter herbivores (e.g. Bryant and Kuropat 1980, Palo et al. 1985, Tahvanainen etal. 1985, Sinclair et al. 1988a and b, Bryant et al. 1991). These secondarycompounds are carbon-based rather than nitrogen-based because nitrogen is ascarce resource to boreal plants (Coley et al. 1985). Both low nitrogen contentand high secondary compound content reduce digestibility (Feeney 1980,Rhoades and Cates 1976, Bryant and Kuropat 1980, Mattson 1980, Palo et al.1985) and also make litter difficult to decompose (Melillo et al. 1982, Flanaganand Van Cleve 1983, White 1986, Homer et al. 1988) since both herbivoredigestion and decomposition are microbially mediated. Therefore, the carbonand nitrogen cycles of boreal forests are tightly linked by positive and negativefeedback loops between decomposers, plants, and herbivores because nitrogenavailability controls net carbon fixation but the types of carbon compoundsproduced control both nitrogen availability and browsing intensity.Spruce litter is of lower quality than fresh spruce needles. Spruce treesremove nitrogen but little of the secondary compounds (carbon-based) fromneedles before they fall off (Chapin et al. 1986). Grouse eat fresh needles andtherefore return relatively high quality faeces to soil, whereas natural litter is oflow quality.87The quality of leaf litter from feeding trees differs from litter from controltrees (Chapter 2) and probably affects relative decomposition rates. Thedecomposition of needles and faeces from feeding trees is probably elevatedrelative to control trees because of the higher forage, and hence litter, quality offeeding trees.Monoterpenes reduce the quality of white spruce needles as food forherbivores and reduce its quality as litter. The relatively low concentrations ofresin and of monoterpenes (Chapter 2) in needles from feeding trees increasespotential nutrient cycling rates above that of surrounding trees. The possibility ofgrouse herbivory further increasing feeding tree quality as forage and as litterleads to further positive feedbacks between forage selection and soil processes.6.4 Tree growthFertilization experiments demonstrate that white spruce trees are nutrientlimited (Fig. 4.4). Nitrogen is generally considered to be the most importantlimiting nutrient to tree growth in boreal forest soils (Van Cleve et al. 1983, 1991,Chapin et al. 1986). Based on the growth response of white spruce to fertilizer(Fig. 4.4) production of this species appears to be nitrogen limited in the studyarea.Production of heavily defoliated white spruce (feeding trees) may besustained during the summer months by the input of soluble nitrogen, largely asammonium ions, from faeces. Nutrient inputs from grouse faeces may accountfor relatively high growth rates of feeding trees and may also lead to the higherforage quality of feeding trees compared with controls.The effect of nitrogen fertilization on the chemical content of spruceneedles is not known although it has been suggested that fertilization willincrease foliar nitrogen concentrations (Shaver and Chapin 1980, Bryant et al.881991, Bjorkman et al. 1991) and decrease foliar concentrations of carbon-basedsecondary compounds such as monoterpenes (Muzika et al. 1989, Bryant et al.1992b, Rogers et al. 1992).6.5 Indirect effectsThere are a number of indirect effects, mediated through spruce grouse,which probably act to increase local decomposition rates, substrate quality, andultimately nutrient cycling rates below feeding trees. First, spruce grouse may beseeding faeces with microbes and thus increasing decomposition rates. Microbeseeded faeces may increase decomposition rates of adjacent litter above that oflitter not adjacent to faeces (e.g. Ruess and McNaughton 1987). Second,nutrient release from rapidly decaying litter types could stimulate thedecomposition of adjacent recalcitrant litter types. Transportation of nutrientsamong litters of different quality may result in more rapid and efficient utilizationof litter substrates by soil microbes. Third, perhaps accumulations of grousefaeces, which retain relatively large amounts of water, increase the moisturecontent of soils beneath feeding trees and hence indirectly further increasedecomposition rates of adjacent needle litter and of the grouse faecesthemselves. There may also be indirect effects through spruce grouse forageselection. Spruce grouse select trees with low ratios of resin to nitrogen. Forageof high quality decomposes more rapidly than low quality forage. Finally, sprucegrouse may also accelerate the return of large quantities of dead needles (endsof browsed needles which dry out, turn red, and fall off) to the base of feedingtrees.Results of this study are consistent with the hypothesis that selectiveforaging by spruce grouse changes the quantity and chemical quality of litterreturned to the soil, thus changing local soil nutrient availability and nutrient89cycling rates. Defoliation-mediated changes in foliar chemistry may increasedecomposition rates of faeces and litter.Nutrient inputs in the form of grouse faeces may account for unusuallyhigh growth rates of feeding trees and may also lead to the higher forage qualityof feeding trees compared with controls. Increased rates of cycling of nutrients(nitrogen) through spruce grouse (Fig. 4.1) may account for high growth rates offeeding trees and may explain why spruce grouse use feeding trees for manyyears in succession. Perhaps spruce grouse feed in individual trees to increasethe amount of faeces which collect at the base of feeding trees, therebyincreasing positive feedback between grouse forage selection and forage quality.Additions of grouse faeces and increased rates of nutrient cycling may offset ordelay the potentially negative impacts of heavy defoliation by spruce grouse onspruce trees.90Table 4.1. Spruce grouse faeces under feeding trees (n = 26). Estimate ofmaximum direct faecal nitrogen input.Mean SE Range n1. Density^(faeces/quadrat) 79.4 7.1 31.5-173.3 252.^Input rate(faeces/m2 /year) 1270 114 504-2772 253. Circumference^(m)^ofarea covered by faeces 8.9 0.5 6.0-18.0 264. Area^(m2 /tree)^covered byfaeces 6.8 0.9 2.9-25.8 265. Input rate(faeces/tree/year) 9076 1695 2256-37853 256. Weight of individualfaeces^(g dwt/turd)0.131 0.004 0.06-0.62 1767. Weight of faeces(g dwt/m2 /year) 166 15 66-363 258. Weight of faeces(g dwt/tree/year) 1189 222 296-4959 259. Faecal nitrogen(mg N/g dwt faeces) 10.25 0.57 8.43-15.13 1210. Faecal N input(mg N/m2 /year) 1705 153 677-3722 2511. Faecal N input(mg N/tree/year) 12187 2275 3029-50827 2591Table 4.2. Mean total carbon, total nitrogen, and C/N ratios of fresh spruceneedles, spruce grouse faeces, and soils compared between feeding trees andcontrol trees. Values are means with 1 SE and sample size in parentheses. (FT)= feeding tree, (AC) = adjacent control tree, and (RC) = random control tree.(Soil nitrogen analyses were not completed).SubstrateTotal C(%)Total N(%)C/NratioSpruce needles (FT) 48.5 1.05 46.4(0.6,10) (0.03,10) (1.2,10)Spruce needles (AC) 47.2 0.98 48.2(0.6,10) (0.02,10) (1.1,10)Spruce needles (RC) 48.4 0.98 49.5(0.2,10) (0.05,10) (2.4,10)Spruce needles (mean) 48.0 1.01 47.8(0.3,30) (0.02,26) (0.8,26)Spruce grouse faeces 44.7 1.02 47.7(0.3,8) (0.06,12) (1.7,8)Soil^(FT) 26.9(2.3,18)Soil^(AC) 22.6(2.6,19)Soil^(mean) 24.7(1.8,37)92Fig. 4.1. Nitrogen cycle diagram showing major linkages between spruce grouse,white spruce trees, and soils.9394Fig. 4.2. Decomposition of spruce grouse faeces. Results of mesh bagexperiments between April and September 1991: (a) percent dry matterremaining, (b) total nitrogen (mg N / g dwt) remaining, and (c) soluble nitrogen(mg NH4+ / g dwt) remaining in decomposing spruce grouse faeces. Means ± 1S.E. shown.95Date96Fig. 4.3. Microbial respiration rates (mg CO2 / g dwt / 14 days at 21°C) duringexperimental incubations of spruce grouse faeces and white spruce litter. (a)Microbial respiration rates during incubations of three age classes of sprucegrouse faeces and of spruce litter. Spruce litter is indicated by the dashed line.(b) Microbial respiration rates of grouse faeces in relation to water content.9798Fig. 4.4. Relative twig growth (%) of experimentally fertilized white spruce treesand control trees from three experimentally fertilized areas and three controlareas. Twig growth is relative to prefertilization growth rates in 1986. Shadedbars indicate growth of trees in three fertilized areas; unshaded bars indicatecontrol areas.99CHAPTER 5:GENERAL CONCLUSION101In this chapter I conclude by presenting and describing a model of positiveand negative feedbacks between spruce grouse, spruce trees, and soils. Thismodel may answer the two central questions of this thesis: (i) Why do sprucegrouse feed so selectively during winter? and (ii) Why do spruce grouse continueto feed in the same individual trees throughout winter for several years insuccession?In chapter 2 I examined the role of foliar chemistry in the selection ofwinter forage by spruce grouse. The main points of this chapter are as follows:a. Spruce grouse fed very selectively on individual plants of a single foragespecies, white spruce. Young tissues were preferred within selected plants.b. During feeding trials captive spruce grouse had a marked preference forneedles from feeding trees over control trees. This suggests that tree selectionwas based on needle chemistry.c. The ratio of resin to nitrogen was significantly lower in feeding trees than incontrol trees. Most of this difference was due to the significantly lower resincontent of feeding trees, although there was also a tendency for nitrogen to behigher in feeding trees than in control trees.d. Concentrations of two monoterpenes, camphor and bornyl acetate, wereinversely related to forage preferences by spruce grouse in Yukon. Thissuggested that camphor and bornyl acetate content influence selection of feedingtrees.e. The chemical content of needles was different between feeding trees andcontrol trees. These differences in forage quality may explain the unusualselection by spruce grouse of individual trees for winter feeding.102In the third chapter I examined differences in growth rate, growth form,and reproductive output between feeding trees and control trees. I also tried todetermine experimentally if these differences between trees result from theeffects of feeding spruce grouse. I found that:a. Spruce grouse ate from 25 to 90% (mean 40%) of all the needles onfeeding trees.b. Trees eaten by grouse differed in shape from uneaten control trees.Feeding trees had higher lateral growth rates and longer needles than controltrees. Cone production of feeding trees was lower than in control treesc. Defoliation and fertilization experiments suggest that observed differencesbetween feeding trees and control trees result from the effects of selectivefeeding by spruce grouse.d. Feeding trees grew more rapidly and allocated less carbon to defense andreproduction than did more slowly-growing control trees.In chapter 4 I examined the role of spruce grouse faeces in this herbivore-plant-soil interaction. Some of the main results of this chapter are:a. Annual inputs of spruce grouse faeces under feeding trees were large.Direct faecal nitrogen inputs ranged from 3 to 51 g N tree-1 .b. Decomposition of grouse faeces was relatively rapid. Mesh bagexperiments indicated that during the first summer there was > 50 % mass loss,> 40% total N loss, and almost complete loss of soluble N from decomposingspruce grouse faeces.c. Microbial respiration rates, a measure of decomposition rate, for grousefaeces were higher than for spruce litter and also depended upon water contentof faeces.103d. Experimental fertilization of white spruce trees indicated that tree growthwas nitrogen limited.e. Nutrient inputs from grouse faeces may account for the relatively highgrowth rates of feeding trees and may also lead to the higher forage quality offeeding trees compared with controls.f. Spruce grouse faecal inputs under feeding trees may locally increase soilnutrient availability and nutrient cycling rates.A MODEL OF POSITIVE FEEDBACKS BETWEEN SPRUCE GROUSEFORAGING, WHITE SPRUCE TREES, AND SOILS.Spruce grouse may induce favorable changes in their host plants, whitespruce trees, through defoliation and fertilization of individual plants (Fig. 5.1).The figure shows the effects of severe defoliation by spruce grouse and offertilization through grouse faeces on the growth and foliar chemistry of whitespruce trees and hence on litter quality. Defoliation and fertilization by sprucegrouse may induce regrowth that is more palatable than forage on uneatenplants. Positive effects on forage palatability induce further feeding. This mayresult in patches of highly palatable forage that attract further feeding, generatinga feeding-regrowth feedback loop such as described by Bryant (1992a and b)and by Du Toit et al. (1990).In Table 5.1, I summarize the main hypotheses and predictions that arisefrom this model. Aspects of these hypotheses were tested in this thesis andindicate that spruce grouse may be farming.104FUTURE STUDIESSpruce grouse and white spruce trees at Kluane Lake are an ideal systemfor studying the role of animal behavior in ecosystem processes and communitydynamics. The boreal forest is relatively simple, has few dominant herbivoresand forage species, and experimental manipulations of these herbivores andplant species can be carried out in the field with relative ease.At Kluane there is only one species of conifer, white spruce. Sprucegrouse feed entirely on the needles of white spruce during the long winters andare extremely selective in their choice of individual plants for feeding. Sprucegrouse browse damage is distinctive and can be confused with no otherherbivore present at Kluane. From defoliation damage it is possible to accuratelytell which trees have been eaten by grouse and which trees have not. Within treeforage preferences and the age of needles eaten throughout a winter can also bedetermined. Trees are sedentary and long lived. Once feeding trees are found,it is possible to predict sites of future spruce grouse use. It is also easy tomeasure growth rates and reproductive output of white spruce. Spruce grouseare cyclically abundant and once found are easy to observe at close proximitywith minimal disturbance. The behavior of spruce grouse during winter is stilllargely undescribed. This system of grouse and trees is ideal for testing some ofthe ideas of optimal foraging and may also be good for developing ideas aboutthe possible effects of positive and negative feedbacks between herbivores,plants, and soils on animal population dynamics. Finally, the highly localizedforaging behavior of spruce grouse and the relative simplicity of the boreal forestmakes feedbacks between herbivores, plants, and soils more obvious than inmany ecosystems.105This thesis has raised many questions some of which may have generalimplications for understanding ecosystem processes and perhaps communitydynamics. I encourage anyone interested to begin working on this system.Further studies of this system will, I think, be rewarding for some time to come. Iwill try to help anyone starting a project on grouse and trees by freely sharing mydata and my understanding of spruce grouse natural history.106Fig. 5.1 Cause-and-effect model of coniferous woody plant chemistry changesinduced by severe defoliation and fertilization. Spruce grouse change the growthand palatability of white spruce needles and hence litter quality through effects offertilization by faeces and severe defoliation. Positive effects on palatabilityinduce further feeding. Arrows from boxes indicate positive effects unlessmarked (-). Small arrows within boxes indicate directions of change. Adaptedfrom Bryant et al. 1992 and Du Toit et al. 1990.107108TABLE 5.1 HYPOTHESES AND PREDICTIONS FROM MODEL SHOWN IN FIGURE 5.1HYPOTHESIS 1:^FEEDING TREES ARE OF INHERENTLY DIFFERENT PALATABILITY THAN CONTROLTREES.HYPOTHESIS 2:^SPRUCE GROUSE FEEDING INCREASES NEEDLE PALATABILITY OF FEEDINGTREES, THIS FEEDS BACK POSITIVELY TO FACILITATE FURTHER FEEDING.THERE ARE TWO WAYS THIS CAN HAPPEN:HYPOTHESIS 2A:^DEFOLIATION EFFECTS.HYPOTHESIS 2B:^FERTILIZER EFFECTS FROM FAECES.PREDICTIONS:^FORAGE QUANTITY^ FORAGE QUALITY(GROWTH) NITROGEN^SECONDARY COMPOUNDSdefoliation fertilization defoliation fertilization defoliation fertilizationH1 - + 0 0 0 0H 2A + + + 0 - 0H 2B 0 + 0 + 0(where + = increase, o = no change, and - = decrease)APPENDIX 1One 600 m x 600 m study area (Sulphur, see Fig. 2.1) was systematically andintensively searched in May 1990 for all feeding trees that were in use during theproceeding winter. Habitat on Sulphur is generally continuous white spruceforest (Dale 1990). The distribution of feeding tree locations is shown in thefigure that follows. Feeding tree distribution was clumped. Sulphur is used fairlyintensively by spruce grouse compared to other areas in the Shakwak Valley(Fig. 2.1).110111REFERENCES CITEDAndreev, A.V. 1991. Winter habitat segregation in the sexually dimorphic black-billed capercaillie Tetrao urogalloides. Ornis Scandinavica 22: 287-291.Anonymous. 1987. Climate of Alberta with data for Yukon and NorthwestTerritories. Report 1987. Tables of temperature, precipitation and sunshine.The atmospheric environment service, Environment Canada. p. 91.Baldwin, I.T. 1990. Herbivory simulations in ecological research. Trends inEcology and Evolution 5: 91-93.Bazely, D.R. and Jefferies R.L. 1989. Lesser snow geese and the nitrogeneconomy of a grazed salt marsh. Journal of Ecology 77: 24-34.Bazzaz, F.A., N.R. Chiariello, P.D. Coley, and L.F. Pitelka. 1987. Allocatingresources to reproduction and defense. Bioscience 37: 58-67.Belsky, A.J. 1986. Does herbivory benefit plants? A review of the evidence.American Naturalist 127: 870-892.Berg, B. 1988. Dynamics of nitrogen ( 15N) in decomposing Scots pine (Pinussylvestris) needle litter. Long-term decomposition in a Scots pine forest. VI.Canadian Journal of Botany 66: 1539-1546.Berg, B. and G. Ekbohm. 1983. Nitrogen immobilization in decomposingneedles at variable carbon:nitrogen ratios. Ecology 64: 63-67.Bianchi, T.S., C.G. Jones.^1991.^Density-dependent positive feedbacksbetween consumers and their resources. In: Comparative analyses ofecosystems: patterns, mechanisms, and theories. Edited by J.J. Cole, G.M.Lovett, and S.E.G. Findlay. Springer-Verlag, New York. pp. 331-340.Bjorkman, C., S. Larsson, and R. Gref. 1991. Effects of nitrogen fertilization onpine needle chemistry and sawfly performance. Oecologia(Berlin) 86: 202-209.Botkin, D.B., J.M. Melillo, and L.S.Y. Wu. 1981. How ecosystem processes arelinked to large mammal population dynamics. In: Dynamics of LargeMammal Populations. Edited by C.W. Fowler and T.D. Smith. John Wiley andSons, New York. pp. 373-387.Briggs, M.A. and J.C. Schultz. 1990. Chemical defense production in Lotuscorniculatus L. II. Trade-offs among growth, reproduction, and defense.Oecologia (Berlin) 83: 32-37.Brooks, J.E., J.H. Borden, H.D. Pierce. 1987. Foliar and cortical monoterpenesin Sitka spruce: potential indicators of resistance to the white pine weevil,Pissodes strobi Peck (Coleoptera: Curculionidae). Canadian Journal ofForest Research 17: 740-745.112Bryant, J.P., and P.J. Kuropat. 1980. Selection of winter forage by subarcticbrowsing vertebrates: the role of plant chemistry. Annual Review of Ecologyand Systematics 11: 261-285.Bryant, J.P., F.S. Chapin, Ill, and D.R. Klein. 1983. Carbon/nutrient balance ofboreal plants in relation to vertebrate herbivory. Oikos 40: 357-368.Bryant, J.P., J. Tuomi, and P. Niemala. 1988. Environmental constraint ofconstitutive and long-term inducible defenses in woody plants. In: Chemicalmediation of coevolution. American Institute of Biological Sciences. pp. 367-389.Bryant, J.P., P.B. Provenza, J. Pastor, P.B. Reichart, T.P. Clausen, and J.T. duToit. 1991. Interactions between woody plants and browsing mammalsmediated by secondary metabolites. Annual Review of Ecology andSystematics 22: 431-446.Bryant, J.P., K. Danell, F. Provenza, P.B. Reichardt, T.A. Clausen. 1992a.Effects of mammal browsing on the chemistry of deciduous woody plants. In:Phytochemical induction by herbivores. Edited by D.W. Tallamy and M.J.Raup. John Wiley & Sons, New York. pp. 135-154. In press.Bryant, J.P., P.J. Kuropat, P.B. Reichart, and T.P. Clausen. 1992b. Controlsover the allocation of resources by woody plants to chemical antiherbivoredefense. In press.Campbell, C.A. 1978. Soil organic carbon, nitrogen and fertility. In: Soil organicmatter. Edited by M. Sehnitzer and S.U. Khan. Elsevier, Amsterdam. pp.173-271.Cargill, S.M. and R.L. Jefferies. 1984. The effects of grazing by lesser snowgeese on the vegetation of a sub-arctic salt marsh. Journal of AppliedEcology 21: 669-686.Carpenter, R.C. 1986. Partitioning herbivory and its effects on coral reef algalcommunities. Ecological Monographs 56: 345-363.Chapin, F.S. III. 1991. Integrated responses of plants to stress. Bioscience 41:29-37.Chapin, F.S. III., P.M. Vitousek, and K. Van Cleve. 1986. The nature of nutrientlimitation in plant communities. American Naturalist 127: 48-58.Coley, P.D. 1988. Effects of plant growth rate and leaf lifetime on the amountand type of antiherbivore defense. Oecologia (Berlin) 74: 531-536.Coley, P.D., J.P. Bryant, and F.S. Chapin, III. 1985. Resource availability andplant antiherbivore defense. Science 230: 895-899.Coppock, D.L., J.K. Detling, J.E. Ellis, and M.I. Dyer. 1983. Plant-herbivoreinteractions in a mixed-grass prairie. I. Effects of black-tailed prairie dogs onintraseasonal aboveground plant biomass and nutrient dynamics and plantspecies diversity. Oecologia (Berlin) 56: 1-9.113Crawley, M.J. 1983. Herbivory: the dynamics of animal-plant interactions.University of California Press, Berkeley.Crichton, V. 1963. Autumn and winter foods of spruce grouse in central Ontario.Journal of Wildlife Management 27: 597.Dale, M.R.T. 1990. Two-dimensional analysis of spacial pattern in vegetation forsite comparison. Canadian Journal of Botany 68: 149-158.Danell, K. and K. Huss-Danell. 1985. Feeding by insects and hares on birchesearlier affected by moose browsing. Oikos 44: 75-81.Day, T.A. and J.K. Detling. 1990. Grassland patch dynamics and herbivoregrazing preference following urine deposition. Ecology 71 : 180-188.Day, T.A., E.H. DeLucia, and W.K. Smith. 1989. Influence of cold soil and snowcover on photosynthesis and leaf conductance in two Rocky Mountainconifers. Oecologia (Berlin) 80: 546-552.DeAngelis, D.L., W. Post, and C.C. Travis. 1986. Positive feedbacks in naturalsystems. Springer-Verlag, Berlin.Dirzo, R. 1984. Herbivory: a phylocentric overview. In: Perspectives on plantpopulation ecology. Edited by R. Dirzo and J. Srukhan. Sinauer AssociatesInc., Sunderland, Massachusetts. pp. 141-165.Distel, R.A., and Provenza, F. D. 1990. Experience early in life affects voluntaryintake of blackbrush by goats. Journal of Chemical Ecology 17:180-188.Doerr, P.D., L.B. Keith, D.H. Rusch, and C.A. Fisher. 1974. Characteristics ofwinter feeding aggregations of ruffed grouse in Alberta. Journal of WildlifeManagement 38: 601-615.Douglas, G.W. 1974. Montane zone vegetation of the Alsek River region,southwest Yukon. Canadian Journal of Botany 52: 2505-2532.Douglas, G.W. and Knapik, L.J. 1974. Montane zone soil characteristics inKluane National Park. Canadian Wildlife Service, Dept. of the Environment.p.65.Du Toit, J.T., J.P. Bryant, and K. Frisby. 1990. Regrowth and palatability ofAcacia shoots following pruning by African savanna browsers. Ecology 71:149-154.Edmonds, R.L. 1980. Litter decomposition and nutrient release in Douglas fir,red alder, western hemlock, and Pacific silver fir ecosystems in westernWashington. Canadian Journal of Forest Research 14: 395-400.Ellison, L. 1966. Seasonal food and chemical analysis of winter diet of Alaskanspruce grouse. Journal of Wildlife Management 30: 729-735.Ellison, L. 1976. Winter food selection by Alaskan spruce grouse. Journal ofWildlife Management 46: 205-213.114Fagerstrom, T. 1989. Anti-herbivory chemical defense in plants: a note on theconcept of cost. American Naturalist 133: 281-287.Farentinos, R.C., P.J. Capretta, R.E. Kepner, and V.M. Littlefield.^1981.Selective herbivory in tassel-eared squirrels: role of monoterpenes inponderosa pines chosen as feeding trees. Science 213: 1273-1275.Feeny, P. 1980. Biochemical coevolution between plants and their insectherbivores. In: Coevolution of plants and animals, Revised edition. Edited byL.E. Gilbert and P.H. Raven. University of Austin Press, Austin, Texas.Flanagan, P.W. and K. Van Cleeve. 1983. Nutrient cycling in relation todecomposition of organic matter quality in taiga ecosystems. CanadianJournal of Forest Research 13: 795-817.Floate, M.J.S. 1981. Effects of grazing by large herbivores on nitrogen cycling inagricultural ecosystems. In: Terrestrial nitrogen cycles: processes,ecosystem strategies and management impacts. Edited by F.E. Clark and T.Rosswall. Ecological Bulletins (Stockholm) 33: 585-601.Fox, J.F. and J.P. Bryant. 1984. Instability of the snowshoe hare and woodyplant interaction. Oecologia (Berlin) 63: 128-135.Fritz, R.S. and Simms, E.L. 1992. Plant resistance to herbivores and pathogens:Ecology, evolution and genetics. University of Chicago Press.Gjerde, I. 1991. Cues in winter habitat selection by Capercaillie. II. Experimentalevidence. Ornis Scandinavica 22: 205-212.Gordon, I.J. and W.K. Lindsay. 1990. Could mammalian herbivores "manage"their resources? Oikos 59: 270-280.Gurchinoff, S. and W.L. Robison. 1972. Chemical characteristics of jackpineneedles selected by feeding spruce grouse. Journal of Wildlife Management36: 80-87.Harmsen, G.W, and van Schreven, D.A. 1955. The mineralization of organicnitrogen in the soil. Advances in Agronomy 7: 299-398.Haukioja, E. 1980. On the role of plant defenses in the fluctuation of herbivorepopulations. Oikos 35: 202-213.Haukioja, E. 1991b. Cyclic fluctuations in density: interactions between adefoliator and its host tree. Acta Oecologica 12: 77-88.Haukioja, E. 1991. The influence of grazing on the evolution, morphology andphysiology of plants as modular organisms. Phil. Trans. R. oc. Lond. 333:241-247.Hayes, R.J. 1986. Mineral nitrogen in the plant-soil system. Academic Press,New York.Herms, D.A. and W.J. Mattson. 1992. The dilemma of plants: to grow or defend.The Quarterly Review of Biology 67: 283-335.115Hik, D.S. and F.P. Mueller. 1993. Decomposition of herbivore faecal materialand nutrient cycling in the northern boreal forest, Yukon. Functional Ecology.In process.Hik, D.S., H.A. Sadul, and R.L. Jefferies. 1991. Effects of the timing of multiplegrazing by geese on net above-ground primary production of swards ofPuccinellia phryganodes. Journal of Ecology 79: 715-730.Hockman, J.N., J.A. Burger, and D.W. Smith. 1989. Spatial and temporalvariability of foliar nutrient levels in fraser fir Christmas trees. Forest Science35: 632-639.Hoft, R.S., J.T. Ratti, and R. Croteau. 1987. Experimental analysis of winterfood selection by spruce grouse. Journal of Wildlife Management 51: 159-167.Holland, E.A. and J.K. Detling. 1990. Plant responses to herbivory andbelowground nitrogen cycling. Ecology 71: 1040-1049.Horner, J.D., J.R. Gosz, and R.G. Cates. 1988. The role of carbon-based plantsecondary metabolites in decomposition in terrestrial ecosystems. AmericanNaturalist 132: 869-883.Howe, H.F., and L.C. Westley. 1988. Ecological relationships of plants andanimals. Oxford University Press, Oxford.Hunter, M.D. 1992. Interactions within herbivore communities mediated by thehost plant: the keystone herbivore concept. In: Effects of resource distributionon animal-plant interactions. Edited by M.D. Hunter, T. Ohgushi, and P.W.Price. Academic Press, San Diego, Calif. pp. 287-325.Huntly, N. 1991. Herbivores and the dynamics of communities and ecosystems.Annual Review of Ecology and Systematics 22: 477-503.lacobelli, A. and R.L. Jefferies. 1991. Inverse salinity gradients in coastalmarshes and the death of stands of Salix: The effects of grubbing by geese.Journal of Ecology 79: 61-73.Irons III., J.G., J.P. Bryant, and M.W. Oswood. 1991. Effects of moose browsingon decomposition rates of birch litter in a subarctic stream. Canadian Journalof Fisheries and Aquatic Sciences 48: 422-444.Jakubas, W.A.^1989. Ruffed grouse feeding behavior and ecology: itsrelationship to the chemical composition of quaking aspen flower buds. Ph.D.thesis, University of Minnesota.Jakubas, W.A. and G.W. Gullion. 1990. Coniferyl benzoate in quaking aspen: aruffed grouse feeding deterrent. Journal of Chemical Ecology 16: 1077-1087.Jakubas, W.A. and G.W. Gullion. 1991. Use of quaking aspen flower buds byruffed grouse: its relationship to grouse densities and bud chemicalcomposition. Condor 93: 473-485.116Jakubas, W.A., G.W. Gullion, and T.P. Clausen. 1989. Ruffed grouse feedingbehavior and its relationship to secondary metabolites of quaking aspenflower buds. Journal of Chemical Ecology 15: 1899-1917.Jefferies, R.L. 1988. Pattern and process in arctic coastal vegetation inresponse to foraging by lesser snow geese. In: Plant form and vegetationstructure, adaptation, plasticity and relation to herbivory. Edited by M.J.A.Werger. SPB Academic Publishing, The Hague. pp. 281-300.Jefferies, R.L., D.R. Bazely, and S.M. Cargill. 1986. Effects of grazing on tundravegetation - a positive feedback model. In: Rangelands: Resource undersiege. Proceedings of the Second International Rangeland Conference.Edited by P.J. Joss, P.W. Lynch, and O.B. Williams. Australian Academy ofScience, Canberra. p. 50.Jefferies, R.L., J. Svoboda, G. Henry, M. Raillard, and R. Ruess. 1992. Tundragrazing systems and climatic change. In: Arctic ecosystems in a changingclimate; an ecophysiological perspective. Edited by F.S. Chapin III., R.L.Jefferies, J.F. Reynolds, G.R. Shaver, and J. Svoboda. Academic Press, SanDiego. pp. 391-412.Jenkinson D.S. and J.N. Ladd. 1981. Microbial biomass in soil: measurementand turnover. In: Soil biochemistry, Vol. 5. Edited by E.A. Paul and J.N. Ladd.Decker, New York. pp. 415-471.Jensen, T.S. 1988. Variability of Norway spruce (Picea abies L.) needles;performance of spruce sawflies (Gilpinia hercyniae Htg.) Oecologia (Berlin)77: 313-320.Jones, C.G. and J.S. Coleman. 1991. Plant stress and insect herbivory: towardan integrated perspective. In: Response of plants to multiple stresses. Editedby H.A. Mooney, W.E. Winner, and E.J. Pell. Academic Press. San Diego.pp. 249-280.Jonkels, C.J. and K.R. Greer. 1963. Fall food habits of spruce grouse innorthwestern Montana. Journal of Wildlife Management 27: 593-596.Karban, R. and J.H. Myers. 1989. Induced plant responses to herbivory. AnnualReview of Ecology and Systematics 20: 331-348.Kitchell, J.F., R.V. O'Neill, D. Webb, G.W. Gallepp, S.M. Bartell, J.F. Koonce, andB.S. Ausmus. 1979. Consumer regulation of nutrient cycling. Bioscience 29:28-34.Kozlowski, T.T. 1971. Growth and development of trees. Academic Press, NewYork.Krebs, C.J., B.S. Gilbert, S. Boutin, A.R.E. Sinclair, and J.N.M. Smith. 1986.Population biology of snowshoe hares. I. Demography of food supplementedpopulations in the southern Yukon, 1976-84. Journal of Animal Ecology 55:963-982.117Krebs, C.J., R. Boonstra, S. Boutin, M. Dale, S. Hannon, K. Martin, A.R.E.Sinclair, J.N.M. Smith, and R. Turkington. 1992. What drives the snowshoehare cycle in Canada's Yukon? In: Wildlife 2001: Populations. Edited by D.R.McCullough and R.H. Barrett. Elsevier Applied Science, London and NewYork. pp. 886-896.Krebs, C.J., R. Boonstra, S. Boutin, M.R.T. Dale, S. Hannon, A.R.E. Sinclair,J.N.M. Smith. 1988. Kluane boreal forest ecosystem project. A collaborativespecial project, N.S.E.R.C. Annual Report. June 1988. p. 25.Kruger, P. 1971. Principles of activation analysis. Wiley-Interscience, New York.Larter, N.C. 1992. Forage fiber analyses: a comparison of two techniques.Australian Journal of Wildlife Research 19: 289-93.Linden, H. 1984. The role of energy and resin contents in the selective feedingof pine needles by capercaillie. Ann. Zool. Fennici 21: 435-439.Lindroth, R.L. 1988. Adaptations of mammalian herbivores to plant chemicaldefenses. In Chemical Mediation of Coevolution, ed. K. C. Spencer, pp. 415-45. New York: Academic Press.Lindroth, R.L. 1989. Mammalian herbivore-plant interactions. In: Plant-animalinteractions. Edited by W.G. Abrahamson. McGraw-hill Book Company, NewYork. pp. 163-204.Lunz, H.J., and R.F. Chandler. 1946. Forest soils. John Wiley and Sons, NewYork.Maschinski, J. and T.G. Whitham. 1989. The continuum of plant reponses toherbivory: The influence of plant association, nutrient availability, and timing.American Naturalist 134: 1-19.Mattson, W.J. 1980. Herbivory in relation to plant nitrogen content. AnnualReview of Ecology and Systematics 11: 119-161.Mattson, W.J. and N.D. Addy. 1975. Phytophagous insects as regulators offorest primary production. Science 190: 515-522.McDonald, M.E. 1985. Growth enhancement of a grazed phytoplanktonivorusfish and growth enhancement of other grazed algae. Oecologia (Berlin) 67:132-136.McNaughton, S.J. 1979. Grazing as an optimization process: grass-ungulaterelationships in the Serengeti. American Naturalist 113: 691-703.McNaughton, S.J. 1983. Compensatory plant growth as a response toherbivory. Oikos 40: 329-336.McNaughton, S.J. 1984. Grazing lawns: animals in herds, plant form, andcoevolution. American Naturalist 124: 863-886.McNaughton, S.J., R.W. Ruess, and S.W. Seagle. 1988. Large mammals andprocess dynamics in African ecosystems. Bioscience 38: 794-800.118Melillo, J.M., J.D. Aber, and J.F. Muratore. 1982. Nitrogen and lignin control ofhardwood leaf litter decomposition dynamics. Ecology 63: 621-626.Mooney, H.A., W.E. Winner, and E.J. Pell (editors) 1991. Response of plants tomultiple stresses. Academic Press. San Diego.Moore, P.D. 1992. Pastoral plumbing. Nature 357: 644.Moore, T.R. 1984. Litter decomposition in a subarctic spruce-lichen woodland,eastern Canada. Ecology 65: 299-308.Moss, R. and I. Hansen. 1980. Grouse Nutrition. Nutrition Abstracts andReviews, Series B 50: 555-567.Moss, R., I.B. Trenholm, A. Watson, and R. Parr. 1990. Plant growth andnitrogen metabolism of red grouse Lagopus lagopus scoticus in spring. OrnisScandinavica 21: 115-121.Muzika, R.M., K.S. Pregitzer, and J.W. Hanover. 1989. Changes in terpeneproduction following nitrogen fertilization of grand fir (Abies grandis (Doug.)Lindl.) seedlings. Oecologia (Berlin) 80: 485-489.Nadelhoffer, K.J., J.D. Aber, and J.M. Melillo. 1984. Seasonal patterns ofammonium and nitrate uptake in nine temperate forest ecosystems. Plantand Soil 80: 321-335.Niemela, P., J. Tuomi, and T. Lojander. 1991. Defoliation of the scots pine andperformance of diprionid sawflies. Journal of Animal Ecology 60: 683-692.Owens, J.N. and M. Molder. 1984. The reproductive cycle of interior spruce.B.C. Ministry of Forests, Victoria, B.C.Palo, R.T., K. Sunnerheim, and 0. Theander. 1985. Seasonal variation ofphenols, crude protein and cell wall content of birch (Betula pendula Roth.) inrelation to ruminant in vitro digestibility. Oecologia (Berlin) 65: 314-318.Parkinson, J.A. and S.E. Allen. 1975. A wet oxidation procedure suitable for thedetermination of nitrogen and mineral nutrients in biological material.Commun. Soil Sci. Plant Anal. 6: 1-11.Pastor, J. and R.J. Naiman. 1992. Selective foraging and ecosystem processesin boreal forests. American Naturalist 139: 690-705.Pastor, J., R.J. Naiman, B. Dewey, and P. McInnes. 1988. Moose, microbes,and the boreal forest. Bioscience 38: 770-776.Paul, E.A. 1989. Soils as components and controllers of ecosystem processes.In: Toward a more exact ecology. Edited by P.J. Grubb and J.B. Whittaker.Blackwell Scientific Publications, Oxford. pp. 353-374.Pederson, J.C. and B.L. Welch. 1985. Comparison of ponderosa pines as feedand non-feed trees for abert squirrels. Journal of Chemical Ecology 11: 149-157.119Pellew, R.A. 1983. The giraffe and its food resource in the Serengeti. I.Composition, biomass, and production of available browse. African Journal ofEcology 21: 241-267.Pendergast, B.A., and D.A. Boag. 1970. Seasonal changes in diet of sprucegrouse in central Alberta. Journal of Wildlife Management 34: 605-611.Pendergast, B.A., and D.A. Boag. 1971. Nutritional aspects of spruce grouse incentral Alberta. Condor 73: 437-443.Power, M.E. 1990. Resource enhancement by indirect effects of grazers:armored catfish, algae, and sediment. Ecology 71: 897-904.Provenza, F.D., and Balph, D.F. 1990. Applicability of five diet-selection modelsto various foraging challenges ruminants encounter. In: BehavioralMechanisms of Food Selection. Edited by R. N. Hughes, Vol. 20, NATO ASISeries G: Ecological Sciences, Heildelberg:Springer-Verlag. pp. 423-59.Provenza, F.D., J.A. Pfister, and C.D. Cheney. 1992. Mechanisms of learning indiet selection with reference to phytotoxicosis in herbivores. Journal of RangeManagement 45: 36-45.Remington, T.E. 1990. Food selection and nutritional ecology of blue grouseduring winter. Ph.D. thesis, University of Wisconsin-Madison.Remington, T.E., and C.E. Braun. 1985. Sage grouse food selection in winter,North Park, Colorado. Journal of Wildlife Management 49: 1055-1061.Rhoades, D.F. 1979. Evolution of plant chemical defense against herbivores. In:Herbivores: their interaction with secondary plant metabolites. Edited by G.A.Rosenthal and D.H. Janzen. Academic Press, New York. pp. 3-54.Rhoades, D.F. and R.G. Cates. 1976. Toward a general theory of plantanti herbivore chemistry. Recent Advances in Phytochememistry 10: 168-213.Richards, B.N. 1987. The microbiology of terrestrial ecosystems. LongmanScientific and Technical, England.Robbins, C.T. 1983. Wildlife feeding and nutrition. Academic Press, New York.Robertson, D.S. 1991. Feedback theory and Darwinian evolution. Journal ofTheoretical Biology 152: 469-484.Rogers, A.R., D. Williams, A.R.E. Sinclair, T.P. Sullivan, and R.J. Andersen.1992. Does nursery breeding reduce anti-herbivore defenses of whitespruce? Further support for the resource availability hypothesis. CanadianJournal of Botany In press.Rosenthal, G.A. and D.H. Janzen. 1979. (eds.) Herbivores: their interaction withsecondary plant metabolites. Academic Press, New York.Ruess, R.W. and S.J. McNaughton. 1987. Grazing and the dynamics of nutrientand energy regulated microbial processes in the Serengeti grasslands.Oikos 49: 101-110.120Ruess, R.W., D.S. Hik, and R.L. Jefferies. 1989. The role of lesser snow geeseas nitrogen processors in a sub-arctic salt marsh. Oecologia (Berlin) 79: 23-29.Schimel, D.S., W.J. Parton, F.J. Adamsen, R.G. Woodmansee, R.L. Senft. 1986.The role of cattle in the volatile loss of nitrogen from a shortgrass steppe.Biogeochemistry 2: 39-52.Schlichting, C.D. 1986. The evolution of phenotypic plasticity in plants. AnnualReview of Ecology and Systematics 17: 667-693.Schultz, J.C. 1988. Many factors influence the evolution of herbivore diets, butplant chemistry is central. Ecology 69: 896-897.Schwartz, C.C., J.G. Nagy, W.L. Regelin. 1980. Juniper oil yield, terpenoidconcentration, and antimicrobial effects on deer. Journal of WildlifeManagement 44: 107-113.Senn, J., S. Hanhimaki, and E. Haukioja. 1992. Among-tree variation in leafphenology and morphology and its correlation with insect performance in themountain birch. Oikos 63: 215-222.Setala, H. and I. Makela. 1991. Galerucella nymphaeae (Col., Chrysomedidae)grazing increases Nuphar leaf production and affects carbon and nitrogendynamics in ponds. Oecologia (Berlin) 86: 170-176.Shaver, G.R. and Chapin, F.S. III. 1980. Response to fertilization by variousplant growth forms in an Alaskan tundra: nutrient accumulation and growth.Ecology 61: 662-675.Sinclair, A.R.E. and J.N.M. Smith. 1984. Do plant secondary compoundsdetermine feeding preferences of snowshoe hare? Oecologia (Berlin) 61:403-410.Sinclair, A.R.E., C.J. Krebs, J.N.M. Smith, and S. Boutin. 1988a. Populationbiology of snowshoe hares. III. Nutrition, plant secondary compounds andfood limitation. Journal of Animal Ecology 57: 787-806.Sinclair, A.R.E., M.K. Jogia, and R.J. Andersen. 1988b. Camphor from juvenilewhite spruce as an antifeedant for snowshoe hares. Journal of ChemicalEcology 14: 1505-1513.Smith, J.L. and E.A. Paul. 1990. The significance of soil microbial biomassestimations. In: Soil Biochemistry, Vol. 6. Edited by J.M. Bollag and G.Stotzky. Marcel Dekker, New York. pp. 357-396.Smith, J.N.M., C.J. Krebs, A.R.E. Sinclair, R. Boonstra. 1988. Populationbiology of snowshoe hares. II. Interactions with winter food plants. Journal ofAnimal Ecology 57: 269-286.Spencer, K.C., ed. 1988. Chemical mediation of coevolution. Academic Press.Sprent, J.I. 1987. The ecology of the nitrogen cycle. Cambridge UniversityPress, Cambridge.121Stohlgren, T.J. 1988. Litter dynamics in two Sierra mixed conifer forests. II.Nutrient release in decomposing leaf litter. Canadian Journal of ForestResearch 18: 1136-1144.Swift, M.J., O.W. Heal, J.M. Anderson. 1979. Decomposition in terrestrialecosystems. Blackwell, Oxford.Tahvanainen, J., E. Helle, R. Julkenen-Titto, and A. Lavola. 1985. Phenoliccompounds of willow bark as deterrents against feeding by mountain hare.Oecologia (Berlin) 65: 319-323.Taylor, B.R., C.E. Prescott, W.J. Parsons, and D. Parkinson. 1991. Substratecontrol of litter decomposition in four Rocky Mountain coniferous forests.Canadian Journal of Botany 69: 2242-2250.Taylor, B.R., D. Parkinson, and W.F.J. Parsons. 1989. Nitrogen and lignincontent as predictors of litter decay rates: a microcosm test. Ecology 70: 97-104.Tuomi, J. 1992. Toward integration of plant defence theories. Trends in Ecologyand Evolution 7: 365-367.Van Cleve, K., F.S. Chapin III, C.T. Dyrness, and L.A. Viereck. 1991. Elementcycling in taiga forests: state-factor control. Bioscience 41: 78-88.Van Cleve, K., L. Oliver, R. Schlentner, L.A. Viereck, C.T. Dryness. 1983.Productivity and nutrient cycling in taiga forest ecosystems. CanadianJournal of Forest Research 13: 747-766.Waring, R.H. 1991. Responses of evergreen trees to multiple stresses. In:Response of plants to multiple stresses. Edited by H.A. Mooney, W.E.Winner, and E.J. Pell. Academic Press, San Diego. pp. 371-390.Watt, A.D., S.R. Leather, and G.I. Forrest. 1991. The effect of previousdefoliation of pole-stage lodgepole pine on plant chemistry, and on thegrowth and survival of pine beauty moth (Panolis flammea) larvae. Oecologia(Berlin) 86: 31-35.Webber, B.L. 1974. The climate of Kluane National Park. Project Report No. 16,Atmospheric Environment Service, Environment Canada. p. 300.Whicker, A.D. and J.K. Detling. 1988. Ecological consequences of prairie dogdisturbances. Bioscience 38: 778-779.White, C.S. 1986 Volatile and water-soluble inhibitors of nitrogen mineralizationand nitrification in a ponderosa pine ecosystem. Biol. Fertil. Soils 2: 97-104.White, T.C.R. 1984. The abundance of invertebrate herbivores in relation to theavailability of nitrogen in stressed food plants. Oecologia (Berlin) 63: 90-105.Whitham, T. G. and S. Mopper.^1985. Chronic herbivory: impacts onarchitecture and sex expression of pinyon pine. Science 228: 1089-1091.122Whitham, T.G., J. Maschinski, K.C. Larson, and K.N. Paige. 1991. Plantresponses to herbivory: the continuum from negative to positive andunderlying physiological mechanisms. In: Plant-animal interactions:evolutionary ecology in tropical and temperate regions. Edited by P.W. Price,T.M. Lewinsohn, G.W. Fernandes, and W.W. Benson. John Wiley and Sons,Inc.Williams, K.S. and J.H. Myers. 1984. Previous herbivore attack of red alder mayimprove food quality for fall webworm larvae. Oecologia (Berlin) 63: 166-170.Zbigniewicz, M.W. and M.R.T. Dale. 199N. Studies of boreal forest vegetation inthe Kluane Lake, Yukon area. I. Plant composition and environmental factors.Unpublished manuscript.Zlotin, R.I. and K.S. Khodashova. 1980. The role of animals in biological cyclingof forest-steppe ecosystems. Dowden, Hutchinson and Ross Inc.,Pennsylvania.123


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items