Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Snowshoe hare demography and behaviour during a cyclic population low phase Hodges, Karen Elizabeth 1998

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

Item Metadata

Download

Media
831-ubc_1998-271625.pdf [ 9.03MB ]
Metadata
JSON: 831-1.0088796.json
JSON-LD: 831-1.0088796-ld.json
RDF/XML (Pretty): 831-1.0088796-rdf.xml
RDF/JSON: 831-1.0088796-rdf.json
Turtle: 831-1.0088796-turtle.txt
N-Triples: 831-1.0088796-rdf-ntriples.txt
Original Record: 831-1.0088796-source.json
Full Text
831-1.0088796-fulltext.txt
Citation
831-1.0088796.ris

Full Text

SNOWSHOE HARE DEMOGRAPHY AND BEHAVIOUR DURING A CYCLIC POPULATION LOW PHASE by K A R E N ELIZABETH HODGES A.B., Mount Holyoke College, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E DEGREE OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Zoology we accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA JANUARY 1998 © Karen Elizabeth Hodges, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date c* DE-6 (2/88) A B S T R A C T Snowshoe hares, Lepus americanus, undergo a ten-year population cycle. I examined whether hares respond behaviourally to predators, and whether hare behaviour influenced population dynamics. If behaviour reduces hares' condition and fecundity, that could explain why the low phase lasts two to four years. I used a factorial manipulation of food addition and predator reduction in southwestern Yukon to test their effects on hare behaviour and demography. Two areas had food added, one area was fenced to exclude mammalian predators, and one area combined both manipulations. Hares densities were low from 1993 to 1995, and increased after summer 1995. Adult survival was similar throughout. Most deaths were due to predation, despite low densities of predators. Although all treatments had higher hare densities, the manipulations started in the previous increase phase, and densities are probably due to earlier dynamics. Populations increased simultaneously on all sites. Hares on control sites ate more protein and less fibre than did hares elsewhere. These differences resulted from the species and twig sizes they ate. Hares on food addition sites had better overall diets, indicated by lower faecal fibre. On all treatments, hares preferred habitats with little open ground and dense clusters of willow. Hares used the thickest cover when resting, and used more exposed sites while foraging. In summer, they preferred deadfall for immediate cover, but in winter they preferred spruce. Male hares, but not females, had larger home ranges when predators were present. Summer movement rates were also higher on sites with predators. Hares did not respond behaviourally to manipulations of food or predators. Food availability was high and predation risk low: hares may employ many different behavioural ii strategies with similar demographic impacts. There is no support for a behaviourally-mediated explanation for the demography or duration of the low phase. Adult survival and fecundity did not change, which implies that changes in juvenile survival are crucial to population dynamics. Although behaviour does not appear to affect demography during the low phase, it may do so during the increase and especially the decline phases, when food is more limiting and predation more severe. iii T A B L E O F C O N T E N T S ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES viii LIST OF FIGURES x ACKNOWLEDGEMENTS xii CHAPTER ONE: Predator sensitive foraging, sub-lethal effects of behavioural changes, and the snowshoe hare cycle 1 Introduction 1 Description of the snowshoe hare cycle 2 The causes of the cycle 4 The effects of predators on behaviour and physiology 5 Linking behaviour and demography; the Refugium hypothesis 6 My aims, experimental design, and organization in this thesis 7 CHAPTER TWO: Snowshoe hare demography during a cyclic population low 9 Introduction 9 Methods and Materials 12 Study area and treatments 12 Density, population rate of change, and recruitment of snowshoe hares 16 Survival rates of snowshoe hares 16 Winter food availability 17 Food availability per hare 18 Predator numbers in the study area 18 Causes of death of snowshoe hares 19 Condition of snowshoe hares 19 Age of snowshoe hares " 20 Results 20 Density and population rate of change 20 Recruitment and survival of snowshoe hares 25 Food availability 32 Predator numbers and indices of predation risk 32 Causes of death for snowshoe hares 39 Body condition 39 Age structure 45 iv Discussion 45 Theories for dynamics of the low phase 45 Evidence of food limitation? 48 Evidence of predation limitation? 50 Evidence for interacting effects of food and predation 50 Indirect effects of predation 51 Conclusions and directions for further research 52 CHAPTER THREE: Snowshoe hare foraging behaviour during the low phase of the ten-year cycle. 54 Introduction 54 Diet selection by snowshoe hares 56 Experimental design, dietary variables, and predictions 57 Methods and Materials 60 Study area, treatments, and study animals 60 Feeding trials 61 Faecal pellet deposition rate: calibrating a faecal clock 63 Collection of hare faecal pellets for fibre analysis 63 Collection of forage samples 64 Fibre and protein analyses of plants and hares' faecal pellets 64 Snow-tracking snowshoe hares 65 Browse consumption and species preferences of snowshoe hares 66 Estimation of spruce clip biomass 66 Intake rate 67 Strip-barking of willow and spruce 67 Statistical analyses 68 Results 68 Browse weight, protein content, and fibre content 68 Snowshoe hare diets: size and species 73 Snowshoe hare fibre and protein intake 90 Faecal pellet deposition rate as an index of time; hares' intake rates 95 Faecal fibre 95 Discussion 106 Hares' diets during the low phase 106 How did hares choose their diets? 108 Methodological ramifications 111 Can diets during the low phase create population consequences? 112 v CHAPTER FOUR: Habitat use by snowshoe hares during the low phase: a test of predator sensitivity and the 'Refugia' hypothesis. 113 Introduction 113 Methods and Materials 116 Study sites, treatments, and habitat availability 116 Snowshoe hares and their locations 117 Habitat classification: open ground, species clusters, and immediate cover 118 Statistical analysis 119 Results 120 Habitat availability 120 Species clusters and amount of open ground 123 Snowshoe hare habitat use and preferences in summer 123 Habitat at sites where hares were killed by predators 123 Species clusters around snowshoe hares 131 Immediate cover of snowshoe hares 138 Discussion 145 Did snowshoe hares use refugia? 145 What constitutes a refugium? 145 What scale is appropriate for assessing refugia? 148 What habitat characteristics did hares select? ' 149 Cyclic habitat shifts and demographic implications 149 CHAPTER FIVE: The effects of food and predation risk on snowshoe hares' movement patterns. 151 Introduction 151 Methods and Materials 154 Study area, treatments, and study animals 154 Locations and activity of snowshoe hares 155 Summer movement rates 156 Home range estimation 156 Winter movement rates, distances, and time spent at browse sites 157 Dispersal and forays 158 Statistical analyses 158 Results 159 Snowshoe hare home ranges: methodological choices 159 Snowshoe hares'home ranges 159 Summer activity patterns 165 Movement rates 165 Foraging patterns 172 Dispersal and forays 172 vi Discussion 187 Factors influencing movement rates of snowshoe hares 187 High movement rates as an anti-predator strategy 189 Food availability's effect on hares' movements 190 Sex differences in movement patterns 191 Dispersal 192 Scale of movements and potential demographic effects 193 Synthesis and directions for further research 194 CHAPTER SIX. Snowshoe hare behaviour and demography during the low phase: implications for cyclic dynamics 195 Were there behavioural shifts with demographic consequences? 195 Snowshoe hare behaviour during the low phase 196 The strength of the interaction between behaviour and demography 197 The continuing problem of the extended low phase 202 The next generation: research questions for another cycle 202 LITERATURE CITED 204 APPENDICES 220 APPENDIX 1: Calculation of shrub biomass on each grid from belt transects and clip plots. 220 APPENDIX 2: Nutritional quality of hares' major food species: whole twigs and herbs. 222 APPENDIX 3: Incremental changes in nutritional quality: segmented twigs. 223 APPENDIX 4: Hare brow se points. 224 APPENDIX 5: Nutritional quality of hares' diets. 226 APPENDIX 6: Fibre in hares' faecal pellets. 227 APPENDIX 7: Snowshoe hares' proximity to cover species. 228 APPENDIX 8: Snowshoe hare home range sizes. 229 vii LIST OF TABLES Table 2.1. Predicted effects of trophic manipulations on snowshoe hare dynamics during the low phase. 11 Table 2.2. Kaplan-Meier survival estimates for snowshoe hares. 30 Table 2.3. Food availability for snowshoe hares: biomass of birch and willow. 35 Table 2.4. Predator numbers at Kluane during the low and early increase phase of the snowshoe hare cycle. 38 Table 2.5. Causes of deaths of snowshoe hares. 42 Table 2.6. Spring age structure of snowshoe hares during the low phase. 46 Table 2.7. Synopsis of snowshoe hare demography during the cyclic low phase. 49 Table 3.1. Summary of predictions for snowshoe hare diets. 58 Table 3.2. Calculation of dietary indices. 62 Table 3.3. Snowshoe hares' consumption of forbs and incidental woody species. 76 Table 3.4. Dietary preferences of snowshoe hares. 81 Table 3.5. Hares' consumption of spruce and willow bark. 89 Table 3.6. Summary of snowshoe hare diets during the cyclic population low. 109 Table 4.1. Habitat characteristics of the treatment areas. 121 Table 4.2. Relationship between the clustering of hares' main food species and total ground cover. 124 Table 4.3. Summary of habitat use patterns by snowshoe hares. 146 Table 5.1. Predictions for the effects of food supply, predation risk, and density on snowshoe hare movements. 153 Table 5.2 Spearman's coefficients of rank correlation for snowshoe hare home ranges calculated using different methods. 160 viii Table 5.3. Movement rates of snowshoe hares. 170 Table 5.4. Regression analyses of snowshoe hares' movement rates. 173 Table 5.5. Long distance movements of snowshoe hares. 184 Table 5.6. Summary of snowshoe hare movements during a cyclic population low. 188 ix LIST O F FIGURES Figure 2.1. Study sites in the Shakwak trench near Kluane Lake, Yukon. 14 Figure 2.2. Spring snowshoe hare densities through a population cycle near Kluane Lake, 1986-1996. 21 Figure 2.3. Annual rate of increase of snowshoe hares. 23 Figure 2.4. Recruitment rate of snowshoe hares. 26 Figure 2.5. Annual survival rates of radio-collared snowshoe hares. 28 Figure 2.6. Relationship of recruitment and survival to the population rate of increase. 33 Figure 2.7. Woody browse available per hare at end of winter. 36 Figure 2.8. Predator and hare densities. 40 Figure 2.9. Body condition of female snowshoe hares in spring. 43 Figure 3.1. Dry weight of hares' main winter food species. 69 Figure 3.2. Fibre content of browse. 71 Figure 3.3. Protein content of browse. 74 Figure 3.4. Species composition of snowshoe hares' diets. 78 Figure 3.5. The proportion of spruce eaten as spruce clips. 82 Figure 3.6. Willow eaten by hares: the proportion of willow twigs of juvenile growth form. 84 Figure 3.7. Mean diameter at point of browse for hares' main browse species. 87 Figure 3.8. Dietary fibre of snowshoe hares. 91 Figure 3.9. Dietary protein of snowshoe hares. 93 Figure 3.10. Faecal pellet deposition rate of snowshoe hares. 96 Figure 3.11. Intake rates of snowshoe hares. 98 x Figure 3.12. The relationship of faecal fibre to dietary protein and dietary fibre. 100 Figure 3.13. Faecal fibre of snowshoe hares. 102 Figure 4.1. Hares'use of open habitat. 125 Figure 4.2. Preference of hares for closed habitat. 127 Figure 4.3. Habitat at snowshoe hare mortality sites. 129 Figure 4.4. Snowshoe hares' summer use of dense birch, willow, spruce, and deadfall. 132 Figure 4.5. Snowshoe hares' summer preferences for dense birch, willow, spruce, and deadfall. 135 Figure 4.6. Hares' microhabitat choices: use of thick cover. 139 Figure 4.7. Snowshoe hares'summer use of cover species. 141 Figure 4.8. Snowshoe hares' winter use of cover species. 143 Figure 5.1. Snowshoe hare home ranges: comparison of data sources. 161 Figure 5.2. Snowshoe hare home ranges and core utilization. 163 Figure 5.3. Snowshoe hare home ranges in summer. 166 Figure 5.4. Snowshoe hare activity patterns in summer. 168 Figure 5.5. Hares'movement rates as functions of their density. 175 Figure 5.6. The relationship between home range size and movement rates. 178 Figure 5.7. Time snowshoe hares spent at winter browse sites. 180 Figure 5.8. Distance snowshoe hares travelled between browse sites in winter. 182 Figure 5.9. Dispersal distances of snowshoe hares. • 185 Figure 6.1. A conceptual model of the role of behaviour in hare demography. 199 xi A C K N O W L E D G E M E N T S My debt is great. During this degree, I have had the great good fortune to learn with and from a number of wise people who have been unstintingly gracious with their ideas, time, and patience. This thesis—and more importantly, my thinking and knowledge—have benefitted greatly. In the field, I was blessed by the help of Kim Allcock, Christie Spence, Liz Gillis, Andrea Kalousek, and Sylvia Wood. I learned many field techniques from Cathy and Frank Doyle, Mark O'Donoghue, Craig and Sue Olsen, and David Hik. Mark O'Donoghue also helped me with database structures and programming. Without this group of people, my data set and my life would be impoverished. Tamie Hucal and Mike Blower, Andy and Carole Williams, and Craig and Sue Olsen all enhanced the quality of my life at the Arctic Institute, and kindly offered me visiting privileges in Whitehorse. I have been able to present data on predators via the efforts of Frank Doyle, Mark O'Donoghue, Liz Hofer, Peter Upton, and the other predator trackers. Similarly, food availabilities were derived from the effort of many CSP personnel; Wes Hochachka was instrumental in the analyses of the shrub data. Teresa Chu and Jan Jekielek allowed me to use some of their data on snowshoe hare locations. I am indebted to CSP personnel for twig cutting, hare trapping, data entry, and the general maintenance that is required for fences and food addition treatments. I quite simply could have come nowhere close by my own efforts. The labwork necessary to this thesis was largely undertaken by Tina Raudzus, with the able assistance of Christie Spence and several others. Gary Shorthouse, Gilles Galzy, and Peter Garnett were instrumental for the analyses of protein. Greg Sharam gave me spruce samples. When I initiated this research, I received much useful advice from David Hik, Fritz Mueller, and Andrea Byrom, on the logistics of working in the Yukon. As I have written up this thesis, I have commiserated with John Pritchard, Lisa Thompson, Durrell Kapan and Lance-Barrett-Lennard on the vagaries of time. Alice Kenney gave me the map of the study area. My thinking, writing, statistics, organization, and clarity of expression have been enhanced by the thoughtful comments of Carol Stefan, Tim Karels, Liz Gillis, Jennifer Ruesink, Kim Allcock, Mark O'Donoghue, Shawna Pelech, Dave Board, Frank Doyle, Karl Vernes, Diane Srivastava, and Chris Fonnesbeck. This thesis is much the richer for their considered responses. As I wrote this thesis, my housemates Liz Gillis, Jennifer Ruesink, Jocylyn McDowell and Tara Smith cheered, encouraged, sustained, and generally nurtured me. I am certain that my progress would have been neither so swift nor so pleasant without them. In a similar vein, my office-mate, Kathy Heise, kept her desk clear for the critical month when paper was everywhere. I have greatly appreciated the time, attention, and patience my research committee has bestowed upon me; I know myself the wiser for responding to their comments. My kindest thanks are therefore owing to Charley Krebs, Lee Gass, Carl Walters, Sally Otto, and Colin Clark. Dennis Chitty was also continuously interested in and supportive of my work. Tony Sinclair, my supervisor, unfailingly focussed my thinking and helped me to see how to address the questions I wanted to with the available data and techniques. I am also indebted to him for allowing me a breadth of experience during this degree, with time in Australia and discussions about biodiversity. My education has been greatly enhanced by the various opportunities and discussions in which he has encouraged me to take part. xii My church has been a source of sustenance as well. I am especially grateful to Karen Buschert and Geoff Vanderkooy, Hannelore Schowalter, Andre Pekovich, Corinne Elliot, and Helmut Lemke. My greatest debt is owed to my family, who consented to have me live far away, who enjoyed my innumerable pictures of mountains and hares, who knew when to ask salient questions and when to refrain from asking, and who have reminded me often that there are other disciplines and activities. Much of my sanity and perspective and courage have been derived from you. I was financially supported by a National Science Foundation Graduate Research Fellowship. Research grants from the Natural Resources and Engineering Research Council to Tony Sinclair and Charley Krebs provided research funding. Northern Studies Training Program grants helped to house my field assistants. A short acknowledgements section is not an competent forum for expressing to these people the depth of my appreciation for what they have done for me. I have hopes for three things: that somehow the thanks behind the words comes clear to them, that somehow I may justify the effort they have made on my behalf, and that somehow I may have the opportunity to repay to them some of the kindness and goodness I have experienced. To my friends— Thanks, thanks, more thanks would not suffice to pay The debt I owe. For much I know, I know As fruit of your spent time; how then to say My debt—that 'tis too large, my deep felt woe. Despite my debt, ye yet again must lend— Ye must give speech! to my heart a full tongue, That, profligate, I may make a true end And hymn your praise, 'til all your valour's sung. Ah friends! Then would I speak of wisdom gained, Of mood restored, and laughter-lightened task, Of heart's ease and soul's help, and that ye deigned To offer, oft before I thought to ask. To say all this, prove I still the suitor To beg the words; ye are yet my tutor. xiii C H A P T E R 1 P R E D A T O R SENSITIVE F O R A G I N G , S U B - L E T H A L E F F E C T S O F B E H A V I O U R A L C H A N G E S , AND T H E SNOWSHOE H A R E C Y C L E INTRODUCTION This thesis is predicated upon two theories: the first, that predators can influence the behaviour of their prey, and the second, that behavioural shifts can cause profound demographic changes. In this thesis, I test whether these ideas apply to snowshoe hares (Lepus americanus). Snowshoe hares undergo periodic fluctuations in abundance approximately every ten years (Green and Evans 1940, Elton and Nicholson 1942, Keith 1963, Finerty 1980). During these changes in numbers, hares also change their behaviour (Keith 1966, Wolff 1980, Hik 1994, 1995), and numerous people have attempted to determine the causal relationships between the various behavioural and demographic changes (Wolff 1980, 1981, Boutin 1984a, Hik 1994, 1995, Krebs 1996, Boonstra etal. 1998a, 1998b). The low phase of the cycle has received much of this theoretical attention, since it can last two to four years despite the high potential fecundity of hares (Cary and Keith 1979, C.I. Stefan ms. in prep). Direct effects of food supply and predation are not able to account for the extended duration of this phase (Krebs 1996, Boonstra et al. 1998b), and people therefore find a behavioural explanation attractive. Behavioural patterns at low hare densities could prolong the low phase through a cascade of sub-lethal effects on nutrition, body condition, and fecundity (Boonstra and Singleton 1993, Hik 1994, 1995, Boonstra et al. 1998b). To date, few studies 1 have examined either the demographics or the behaviour of hares during the cyclic low phase; in this thesis, I attempt to fill some of these knowledge gaps. Description of the snowshoe hare cycle Each population cycle lasts 8 to 11 years, and the cycle is broadly synchronous across northern North America (Finerty 1980, Smith 1983, Sinclair et al. 1993, Sinclair and Gosline 1997). Peak densities can exceed low densities by more than 100-fold; peak densities in fall are approximately 5-10 hares per hectare, and low densities are typically below 0.1 hares per hectare (Keith 1990). Populations take several years to build from low densities to peak densities, but the population decline is usually rapid. The numerical changes are accompanied by changes in reproduction, dispersal, and survival. Hares can have up to four litters in a summer, with 2-10 leverets per litter. The number of litters per year and the number of leverets per litter both display cyclic variation, with the highest natality rate approximately three years before the highest population densities (Cary and Keith 1979, Keith 1990, C.I. Stefan ms. in prep). Dispersal appears to be highest during the peak and early decline phases (Windberg and Keith 1976a, Boutin et al. 1985), but no more than -28% of 'losses' from a tagged population were due to dispersal, so dispersal cannot account for the decline (Boutin etal. 1985). Survival rates of both adult and juvenile hares are lowest during the decline phase, and highest during the low and early increase phases (Trostel et al. 1987, Keith 1990). Most deaths are due to predation (Boutin etal. 1986, Keith 1990, Krebs etal. 1995). Juvenile survival rates are typically lower than adult survival rates (Keith 1990), but whereas many estimates of adult survival are from radio-telemetry, juvenile survival has typically been estimated from trapping 2 data, and only one study has examined juvenile survival via radio-telemetry (Gillis 1997). Since juveniles also disperse frequently (Boutin et al. 1985, Gillis 1997), estimating their survival by trapping almost certainly provides a severe underestimate. Juvenile survival strongly affects cyclic dynamics (Keith 1990, Krebs etal. 1995, Krebs 1996, E.A. Gillis, D. Haydon, C.I. Stefan and C.J. Krebs, ms. in prep), so better estimates of their survival would be useful. Survival of leverets may also fluctuate through the cycle (Keith 1990, O'Donoghue and Krebs 1992, C.I. Stefan ms. in prep), but variation in their survival affects cyclic dynamics less than does juvenile survival (E.A. Gillis, D. Haydon, C.I. Stefan and C.J. Krebs, ms. in prep). The predators of snowshoe hares also undergo a ten-year cycle, but typically with a 1-3 year lag after the hares (Keith et al. 1977, Keith 1990, Boutin et al. 1995). Great horned owls (Adamcik and Keith 1978, Adamcik et al. 1978, Rohner 1996), goshawks (Doyle and Smith 1994), lynx (Brand et al. 1976, Brand and Keith 1979, O'Donoghue et al. 1997), and coyotes (Nellis and Keith 1976, Todd etal. 1981, Todd and Keith 1983, O'Donoghue etal. 1997) all undergo numerical changes through the cycle that result from changes in dispersal, survival, and reproduction. Some prey switching occurs, too, so predators' functional responses change through the cycle (Keith 1990, O'Donoghue 1997). .Over-all, hares are at greatest risk of predation during the decline phase, and at least risk during the late low phase and early increase phase (Hik 1994, Boutin etal. 1995). Hares' food species also exhibit cyclic patterns. When hare densities are high, the plants suffer heavy browsing; extensive regrowth then occurs during the low phase (Smith et al. 1988, Keith 1990). Food limitation may occur at peak densities (Pease et al. 1979, Keith 1983, Keith 1990), but other studies have not been able to confirm food shortage (Smith et al. 1988, Krebs et 3 al. 1995). In addition, many of snowshoe hares' food plants contain secondary metabolites which act as anti-feedants (Bryant and Kuropat 1980, Bryant et al. 1992, Schmitz et al. 1992), and concentrations of these metabolites vary during the hare cycle as plants regrow (Bryant 1981a, Fox and Bryant 1984, Sinclair et al. 1988). The causes of the cycle Any explanation for the cycle needs to provide an understanding of how and why both survival and fecundity change through the cycle. There are four main possibilities: the cycle could be intrinsic to the hares, or it could be due to hare-food interactions, hare-predator interactions, or both hare-food and hare-predator interactions (Finerty 1980, Keith 1990, Krebs et al. 1992). Intrinsic changes in fecundity and juvenile survival could lead to the cycle, with changes in adult survival due to delayed responses on the part of the predators (Chitty 1967, Krebs 1978b). A hare-food interaction could limit hares' nutrition during the decline and low phases, via a flush of secondary metabolites resulting from heavy browsing at the peak; again, survival changes could result either from a predatory time lag or from changes in hares' behaviour as they seek less-defended plants (Bryant 1981, Fox and Bryant 1984). Hare-predator interactions could clearly cause the survival changes, but fecundity changes are more difficult to explain under this hypothesis (Trostel etal. 1987). By far the most empirical support is for a joint hare-food and hare-predator mechanism for the cycle (Keith 1990, Krebs et al. 1995, 1996). Manipulations of food and predation have indicated that neither by itself is adequate to alter the dynamics of the cycle, although both increase the densities of hares throughout the cycle. Joint reduction of predation and addition of 4 food has come closest to interrupting the cycle, indicating that their interactive effects on hares are crucial to the cycle (Krebs et al. 1992, 1995, 1996). The low phase of the cycle remains enigmatic despite the strong evidence for a three trophic level interaction as the cause of the cycle (Krebs et al. 1995, Boonstra et al. 1998b). During the low phase, predation is low and food is ample, and yet densities do not immediately increase. Behavioural limitation could maintain the low phase, if either natality or survival of juveniles were negatively affected by behavioural changes of hares (Hik 1994, Krebs et al. 1995, Boonstra et al. 1998b). The effects of predators on behaviour and physiology Predators can affect the movements, activity, habitat selection, and diet selection of small mammals (Lima and Dill 1990). Many small mammals utilize denser cover when predators are nearby or when the amount of light increases the risk of predation (Kotler 1984b, Brown 1988, Brown etal. 1988, Dickman et al. 1991, Kotler et al. 1991, 1992, Longland and Price 1991, Kotler 1992, Lagos et al. 1995). Similarly, many small mammals reduce the amount of time they are active, reduce their movements, and reduce their home ranges when predation risk is high (Brown etal. 1988, Desy et al. 1990, Jedrzejewski and Jedrzejewska 1990, Kotler et al. 1991, Saarikko 1992, Hughes and Ward 1993, Jedrzejewski etal. 1993, Fenn and Macdonald 1995). Fewer studies have explicitly considered dietary shifts as a response to risk, but there is some evidence that animals eat nutritionally poorer diets when risk is high (Abrahams and Dill 1989, Godin 1990, Sih and Moore 1990, Anholt and Werner 1995). Snowshoe hares also attempt to reduce the predation risk they experience. Hares exposed to high predation risk during a decline phase used denser habitats, ate foods that were more common, and reduced their movements (Hik 1994). Similarly, hares avoided clearings more on bright moon-lit nights than on dark nights (Gilbert and Boutin 1991). Hares are physiologically affected by high predation risk; they are often lighter (Hik 1994, K.E. Hodges, E.A. Gillis and C.I. Stefan, ms. in prep), and hormonal challenges indicate that hares exposed to predators had higher free Cortisol levels and were less able to respond to stress than were hares protected from mammalian predators (Boonstra and Singleton 1993, Boonstra et al. 1998a). Linking behaviour and demography; the Refugium hypothesis Individual behaviour can affect demography in a number of ways. Any change in behaviour that affects fecundity, dispersal, or mortality will by definition affect the demography of the population. Theoretically, the effects of changes in behaviour on demography can be substantive (McNamara and Houston 1982, Ives and Dobson 1987, McNamara 1990, Abrams 1993). Empirically, changes in habitat use can affect demography directly, through habitat-specific survival rates (Kotler et al. 1991, Longland and Price 1991, Dickman et al. 1991, Dickman 1992, Stuart-Smith and Boutin 1995), or indirectly, by affecting the foraging rates or nutritional quality of the diet, which in turn can affect the starvation rate, fecundity, and probability of dispersal (McNamara and Houston 1987, Lima and Dill 1990). Changes in activity and movement rates operate similarly. For snowshoe hares the linkage between behaviour and demography has been formalized into the refugium hypothesis (Wolff 1980, 1981, Hik 1994, 1995), in which it is postulated that hares utilize dense, protective habitats during the low phase of the cycle, then use increasingly open, risky habitats as their numbers increase. The argument states that when hares are in refugia, their safety increases, but at the cost of reduced nutrition, which leads to poorer body condition 6 and reduced fecundity. The low phase would thus be maintained by low reproduction or poor survival of offspring that suffer from deleterious maternal effects (Boonstra et al. 1998b). This hypothesis has partial support: hares do suffer higher mortality in more open areas (Sievert and Keith 1985, Hik 1994, Murray etal. 1994, Rohner and Krebs 1996, Cox etal. 1997), they appear to shift the habitats they use during the cycle (Keith 1966, Wolff 1980, 1981, Hik 1994), their diets fluctuate through the cycle (Pease et al. 1979, Smith et al. 1988, Keith 1990, Hik 1994), and food shortage affects hares' condition and fecundity (Windberg and Keith 1976b, Cary and Keith 1979, Vaughan and Keith 1981). Additionally, hares which are exposed to predation risk suffer deleterious effects of stress, either from repeated encounters with predators or from the putative nutritional and habitat shifts, and these stress effects may reduce their fecundity or the viability of their offspring (Boonstra and Singleton 1993, Boonstra et al. 1998a). My aims, experimental design, and organization in this thesis In this thesis, I am primarily concerned with establishing how predation risk and food supply affect hares' behaviour. Secondarily, I describe the demographic patterns that occur during the low phase, and attempt to link the behavioural shifts to demographic consequences. I conducted this research as part of the Kluane Boreal Forest Ecosystem Project, and therefore used the large-scale manipulations of food addition, predator reduction, and food addition+predator reduction that were established in 1987-1988, during the previous population increase (Krebs et al. 1995). A coyote was inside the fence of the food addition+predator reduction treatment from November 1995 through January 1996. Because it killed many of the hares, food addition was stopped in January. I report demographic data from this treatment for the spring of 1996, but I did not collect any behavioural data after November 1995. 7 There are two potential problems associated with these manipulations. First, the original aims of the Kluane Boreal Forest Ecosystem Project were to determine the demographic changes associated with these manipulations, and indeed every manipulation resulted in higher densities than occurred on control areas (Boutin et al. 1995, Krebs et al. 1996). Therefore, when I examine behavioural differences, it is always possible that both increased density and the manipulation affect the hares. This problem is partially obviated by working over three years; if density affects hares' behaviour, then each treatment should differ from itself as densities change. Second, due to the large scale and high cost of the manipulations, the predator exclosure and food addition + predator exclosure treatments were not replicated. I discount the possibility that observed differences in behaviour and demography are due to site effects: I used anywhere from two to seven control sites to ensure I sampled the range of natural variation, both in sites and in hares' behaviours. If behaviour of hares on these unreplicated manipulations still differed from the behaviour of hares on multiple control sites, despite the range of natural variation among control sites, it is likely that treatment effects rather than site effects are responsible for the differences. In chapter 2,1 describe the demographics of snowshoe hares during the low and early increase phases. In chapters 3-5,1 examine the effects of food addition and predator reduction on hares' diets, habitat selection, and movements, respectively. In the final chapter, I refer back to the framework established in this chapter, and ask whether snowshoe hares' behaviour contributes to their demography and to the prolonged duration of the low phase. 8 C H A P T E R 2 SNOWSHOE H A R E D E M O G R A P H Y DURING A C Y C L I C P O P U L A T I O N L O W INTRODUCTION For decades, ecologists have known that snowshoe hares (Lepus americanus) exhibit an 8 to 11 year cycle in abundance, with hare densities changing 20-200 fold (Green and Evans 1940, Elton and Nicholson 1942, Keith and Windberg 1978, Finerty 1980, Keith et al. 1984). The hare cycle is broadly synchronous across the boreal and northern temperate forests of North America (Smith 1983, Sinclair et al. 1993, Sinclair and Gosline 1997), and is associated with predictable changes in reproduction and survival (Keith 1990). The underlying mechanisms causing these changes in survival and reproduction have been much debated. To form a cycle of the appropriate duration, there must be a density-dependent lag of 1-3 years in one or both of these factors. The low phase of the cycle often lasts two to four years, with apparent lags in vegetative regrowth and predator numbers, indicating that the dynamics during the low phase may be crucial to the generation of the cycle. The main hypotheses proposed to explain the dynamics of the cycle involve trophic interactions: hare-food, hare-predator, or food-hare-predator. Changes in plant quantity or quality, resulting from heavy browsing during hare population peaks, may create the necessary time lag as plants regrow lost biomass or adjust their chemical defenses (Pease et al. 1979, Bryant 1981, Fox and Bryant 1984). Changes in predation pressure may be sufficient to cause the cycle as predators respond functionally and numerically (Trostel et al. 1987, Krebs et al. 1992, O'Donoghue et al. 1997). There are two forms of the food-hare-predator hypothesis: the Keith 9 hypothesis states that food shortage in the winter of peak densities initiates the decline phase, then predation continues the high mortality that characterizes the decline phase (Keith 1974, 1990). Alternatively, recent evidence from the Yukon suggests that both food supply and predation are integral to the population dynamics of hares at all stages of the cycle (Krebs et al. 1995, 1996). Despite these potential explanations, hare population dynamics during the two to four year phase of low numbers are still poorly understood (Boonstra and Singleton 1993, Hik 1994, 1995, Krebs et al. 1995, Boonstra et al. 1998a, 1998b), partly because previous studies have focussed on the transition from peak densities to low densities (Keith and Windberg 1978, Wolff 1980, 1981, Keith 1990, Hik 1995). Furthermore, food is apparently abundant during the low phase of the cycle, and predator numbers low, both of which should promote high hare survival; fecundity is also high during this phase (Smith et al. 1988, Keith 1990, Boutin et al. 1995, Krebs et al. 1995, Stefan 1998). Direct effects of food limitation or predation are therefore seemingly insufficient to explain the length of the low phase. If predators cause indirect effects on the nutrition, body condition, and reproduction of hares during the low phase, these indirect effects might be sufficiently large to create the extended duration of this phase (Wolff 1980, 1981, Boonstra and Singleton 1993, Hik 1994, 1995, Krebs 1996, Boonstra etal. 1998a, 1998b). In this chapter, I test these hypotheses experimentally for the low phase of the cycle (see also Table 2.1). If the cyclic low is extended because of a hare-food interaction, then addition of high quality food should cause an earlier increase than that observed for hare populations without supplemental food. This earlier increase could be due to increased fecundity resulting from better nutrition, or to higher survival if starvation or predation deaths were reduced by the abundance of food [see McNamara and Houston (1987) and Sinclair and Arcese (1995) for the difficulty in 10 Table 2.1. Predicted effects of trophic manipulations on snowshoe hare dynamics during the low phase. Statements are relative to the control populations. Food limitation could be either an absolute shortage of available biomass, or a shortage of food low in secondary compounds. The predation hypothesis refers to mortality effects alone, or to mortality and indirect effects of risk on hares' behaviour and physiology. Statements in bold indicate the most likely mechanism for numerical change, while those in regular type might change in the direction indicated or might not differ from control populations. For the interaction hypothesis, the food addition+predator reduction treatment should have greater impact on hare dynamics than either single factor manipulation, as indicated by the + signs. Manipulation Food Limitation Predation Interaction increase earlier density higher -growth rate higher -survival better -recruitment higher -condition better -mean age lower increase same density same -growth rate same -survival same -recruitment same -condition same -mean age same increase earlier density higher -growth rate higher -survival better -recruitment higher -condition better -mean age lower predator reduction increase same density same -growth rate same -survival same -recruitment same -condition same -mean age same increase earlier density higher -growth rate higher -survival better -recruitment higher -condition better -mean age higher increase earlier density same -growth rate higher -survival better -recruitment higher -condition better -mean age higher food addition + predator reduction increase earlier density higher -growth rate higher -survival better -recruitment higher -condition better -mean age lower increase earlier density higher -growth rate higher -survival better -recruitment higher -condition better -mean age higher increase earlier+ density higher+ -growth rate higher+ -survival better+ -recruitment higher+ -condition better+ -no age prediction 11 separating starvation and predation deaths]. If, instead, a hare-predator interaction causes the dynamics of the low phase, then hare populations protected from predators should have a shorter low phase than their control counterparts. The mechanism here is liable to be increased survival, although fecundity might also increase; predators may suppress reproduction in Clethrionomys and Microtus voles (Ylonen 1989, 1994, Ylonen and Ronkainen 1994, Norrdahl and Korpimaki 1995). Finally, if dynamics during the low phase of the cycle are due to an interaction of all three trophic levels, then the joint manipulation of predators and food availability should have greater effects on fecundity, survival, density, and duration of the low phase than does manipulation of either single factor. For the low phase of the ten-year cycle, my purposes are therefore 1) to describe the demographic parameters of the hares; 2) to quantify food availability and predation pressure; 3) to test experimentally the trophic level hypotheses for this phase by manipulating food and predators; and 4) to examine the indirect effects of predators on hares' body condition, age structure, and recruitment. I used control, food addition, predator reduction, and food addition+predator reduction manipulations established by the Kluane Boreal Forest Ecosystem Project by summer 1988 (Krebs etal. 1992, 1996) to address these questions. M E T H O D S AND M A T E R I A L S Study area and treatments In this research, I focus on 1993-1996, thus encompassing the low phase of the ten-year cycle. This research is part of a ten-year (1986-1996) examination of community organization conducted by members of the Kluane Boreal Forest Ecosystem Project (Krebs et al. 1992) in the southwestern Yukon near Kluane Lake (60°57'N, 138° 12'W), in a broad glacial valley about 900 12 m above sea level. The 350 km2 study area is predominantly composed of white spruce (Picea glauca) forest on the valley floor, but extends up the slopes into the alpine tundra of the Ruby Range to the northeast and the Kluane Range to the southwest. The valley floor has occasional pockets of aspen (Populus tremuloides) and poplar (P. balsamifera), multiple small lakes, and several swampy areas. Willow (Salix glauca, S. alaxensis) and bog birch (Betula glandulosa) are common both in the swamps and as understory in the spruce forests. I used seven 34 ha treatment areas which the Kluane Boreal Forest Ecosystem Project established before or during 1988 (Figure 2.1, Krebs et al. 1992, 1995). Three areas were unmanipulated controls. Two areas (food) were provisioned ad lib. with commercial rabbit chow (minimum 16% crude protein) spread along four 570 m cutlines. One area (fence) was contained within a 1 km2 electric and chicken wire fence which excluded the main terrestrial predators, lynx (Lynx canadensis) and coyote (Canis latrans). Ten ha of the trapping area inside the fence was covered with monofilament line strung through the trees to deter goshawks (Accipiter gentilis), Harlan's hawks (Buteo jamaicensis), and great horned owls (Bubo virginianus). The final treatment (food+fence) was a combination of food addition and fence, but had no monofilament. Snowshoe hares were trapped on each of these areas every March and October for population estimates. Sporadic trapping occurred at other times to maintain the number of radio-collared hares for estimates of survival. Eighty-six Tomahawk live traps (Tomahawk Live Trap Co., Tomahawk WI) were spaced in four trapping lines across the 34 ha grids; the effective trapping area was 60 ha. Traps were baited in the evening with alfalfa (and rabbit chow on the food addition areas), and checked the following morning at first light. Each hare was identified with a Monel #3 eartag (National Band and Tag Co., Newport KY). For every capture, we 13 Figure 2.1. Study sites in the Shakwak trench near K l u a n e L a k e , Y u k o n . L a k e s and ponds are represented in grey shading. The A l a s k a H i g h w a y roughly bisects the study area, and the marked areas are the treatments on wh ich snowshoe hare populations were studied. The boxes wi th in the fenced areas indicate the location of the trapping grids wi th respect to the fences. Terrestrial predators were studied throughout the area; raptors were studied intensively f rom east o f the Predator Exc losu re + F o o d treatment area back to K l u a n e L a k e , and wi th less intensity through the rest o f the study area. In this thesis, I discuss results f rom controls 1-3, both food grids, and both predator exclosures. 14 15 recorded eartag, location, sex, reproductive condition (males: abdominal or scrotal testes; females: lactating, pregnant, or not pregnant), length of right hind foot, and weight. Some hares were radio-collared with 40 g radio-collars (Lotek, Newmarket ONT) equipped with mortality censors that doubled the pulse rate when the radio-collar did not move for four hours. Density, population rate of change, and recruitment of snowshoe hares Hare densities were estimated every March from 3 to 7 nights of trapping per experimental grid. Population sizes for each treatment grid were derived as the average of jackknife estimates from program CAPTURE (Otis et al. 1978, Boulanger and Krebs 1994) and Jolly-Seber estimates (Seber 1982). The annual recruitment rates were estimated as the geometric mean (averaged from spring to spring) of Jolly-Seber dilution rates (Krebs 1989). The annual finite rate of increase was calculated as N t + 1/N t. Survival rates of snowshoe hares Adult snowshoe hare survival rates were estimated based on radio-collared hares. Each hare's mortality date was defined as the day after it was last known to be alive, unless the state of the corpse or sign at the death site indicated a more recent death. Survival rates were calculated by program Pollock (C. J. Krebs unpublished), which uses the non-parametric Kaplan-Meier estimator (Kaplan and Meier 1958, Pollock etal. 1989a, 1989b, Krebs 1989). Kaplan-Meier estimation allows for the staggered entry of radio-collared hares, and allows information to be included for hares even if their final fates were unknown or their deaths were due to human causes. Such censoring of information maximises the use of data; other methods would excise all information about such hares, whereas the Kaplan-Meier procedure excludes only information about the time or manner of death. 16 I censored records for hares for which time of death could not be determined to within two weeks and for hares that had incomplete records due to human-caused deaths (trap death, road kill), radio-collar failure, or loss of radio-signal due to dispersal. For these cases, the censoring date was the last date on which the hare was known to be alive. I assume that most loss of radio-signals was not due to deaths of the hares and destruction of the radio-collars. Hares that originated on the fence and food+fence treatments but died outside of the fences were censored on the last date for which I knew them to be inside the fences. If hares survived beyond the time interval under examination, their censoring date was the last day of the interval. Winter food availability In May-June of 1995, 5 to 10 transects were conducted on each treatment area to estimate shrub biomass. Along each 90-180 m transect, all willow and birch bushes within a 2 m wide belt were measured for number of stems >5 cm, basal circumference, and maximum height. The biomass of >5 mm twigs along each transect was calculated from predictive regression equations using these variables with constants fitted from data from 76 birch bushes and 104 willow bushes that were measured, cut, and weighed (Appendix 1; all regressions had r2 values of >0.90; C.J. Krebs, A.R.E. Sinclair, and W. Hochachka, unpublished data). Bushes were sampled from all of the treatment areas for which the belt transects were conducted, to ensure the regressions can be applied to all areas. The biomass per transect was converted into estimates of kg dry weight/hectare (Appendix 1; A.R.E. Sinclair and C.J. Krebs, unpublished data). These estimates are expected to be constant from year to year since most changes in shrub biomass occur for twigs <5 mm diameter (A.R.E. Sinclair and C.J. Krebs, unpublished data). 17 The proportion of the standing biomass composed of small twigs (<5 mm diameter) was estimated annually. Every May, bushes were cut down on -50 1 m2 plots on two control areas, one food area, and the food+fence area. For each shrub species, biomass of >5 mm twigs and <5 mm twigs was recorded. In 1994 and 1995, 20 individual bushes each of birch and willow were cut down on the other three sites. Annual small twig biomass was calculated using the estimate of >5 mm standing biomass and the ratio of <5 mm twigs to total biomass (see Appendix 1). Food availability per hare The above calculations provide estimates of the standing amount of food left in early May, before leaf-out and summer growth. As such, values represent yearly minima of available small twigs. The amount of browse available per hare at the end of winter was estimated by dividing the available kg/ha of small twigs (willow and birch combined) by the March density of hares. Predator numbers in the study area Predator densities were studied by other researchers affiliated with the Kluane Boreal Forest Ecosystem Project. Numbers of great horned owls, goshawks, and Harlan's hawks were assessed in the spring and summer, using playback techniques, sightings, and records of owl hootings (F.I. Doyle unpublished data, Rohner and Doyle 1992, Doyle and Smith 1994, Boutin et al. 1995). Raptor censuses were conducted intensively within a 100 km2 core study area. Lynx and coyote numbers were determined through the course of each winter (November-March), by a combination of trapping and radio-collaring, track transects, and records of coyote howling (Boutin etal. 1995, O'Donoghue etal. 1997). 18 Causes of death of snowshoe hares Hares' radio-collar frequencies were monitored a minimum of three times a week from the Alaska Highway, using a large antenna mounted on a vehicle. When mortality signals were heard, Project personnel located the radio-collar by using handheld receivers and yagi or H antennas (Telonics, Mesa AZ). At each site of death, information on predator tracks, feathers, scat, whitewash, hare remains, and manner of eating were collected. We used these clues to identify the cause of death (C. Doyle and C.J. Krebs, unpublished protocol). Corpses found intact were collected for necropsy. I used conservative criteria for determining the cause of death. If all of the evidence present at a site indicated a particular predator species, I assigned the kill to that species. If, however, most of the evidence indicated a particular species but there were inconsistencies or the evidence was scanty, I classified the death as 'unknown predator'. Deaths were classified as 'non-predation' if there were no signs of predation (usually, non-depredated hares were found intact). Similarly, if the evidence for predation rather than non-predation was poor, I assigned the death to the conservative category of 'unknown death'. I assume that the distribution of causes for 'unknown deaths' is similar to that for known deaths, especially since most 'unknown deaths' were due to a delay in finding the radio-collar and hence a reduction in observable evidence. Condition of snowshoe hares Condition of female hares in the spring was estimated as observed weight/predicted weight, where predicted weight was derived from an equation relating weight to skeletal size: weight (g) = 1.97 + right hind foot length (mm)148. The parameters were derived from 1249 captures of female hares from control sites in the springs of 1989-1996 (K.E. Hodges, E.A. Gillis, 19 C L Stefan, ms. in prep). The index indicates relative condition, with a value of one indicating the hare is in average condition relative to the female population at large (O'Donoghue and Krebs 1992, Krebs and Singleton 1993). Age of snowshoe hares Snowshoe hares can have up to four litters in a summer, and parturition dates are synchronized between females. Therefore juvenile hares trapped in summer could be assigned to one of the four litter groups, based on weight and right hind foot length. Juveniles caught in the fall were often adult weight, so all hares caught for the first time in the fall were classified as juveniles of unknown litter unless they could be identified as third or fourth litter. We added a year to all hares' ages on January 1 st. RESULTS Density and population rate of change Snowshoe hare populations were generally at their lowest densities in the springs of 1993 and 1994 (Figure 2.2). Hare populations on all sites declined from 1992 to 1993. From 1993 to 1994, two control populations and the food+fence population declined but populations on the other sites increased. All populations increased annually thereafter, with the exception of a decline on the fenced site from 1995 to 1996. The annual rate of increase in 1995-1996 was much higher than in the preceding three years (Figure 2.3). Because spring densities were lowest from 1993-1995, and rates of increase either below one or generally close to it, these years will be considered as the low phase. I will consider that the increase phase started in summer 1995. With the exception of the fence treatment in 1996, all manipulations increased density, but no manipulated population increased earlier than the control populations. From 1993-1995, the 20 Figure 2.2. Spr ing snowshoe hare densities through a populat ion cyc le near K l u a n e L a k e , 1986-1996. This study commenced in M a r c h 1993, and continued unti l A p r i l 1996, encompassing the low and early increase phases o f the cyc le . 21 v - CN CO O O "5 cr o c o c o o o o TD O O <> o - I * CN TJ O O O C 0 cu o c + TJ O O * 9 + 22 Figure 2.3. Annual rate of increase of snowshoe hares. These rates were calculated as N t + 1 / N t , from spring to spring. Values lower than one indicate a declining population. 23 24 food addition site had densities ~4.2x higher than control areas, the fence manipulation had densities ~2.5x higher, and the food+fence manipulation had densities ~10.9x higher. In 1992-1993 and 1993-1994, annual rates of increase were similar for all treatments. In 1994-1995, two control populations had higher rates of increase than elsewhere, and in 1995-1996 populations on the fence and food+fence treatments had lower rates of increase than did the food and control populations. Recruitment and survival of snowshoe hares Recruitment rates generally increased on control sites during the low phase of the cycle (Figure 2.4), but two control populations had their recruitment rates reduced during the first year of population increase (1995-1996). On both food areas, recruitment was lowest in 1994-1995. The fence and food+fence manipulations had little change in recruitment rates. Control populations generally had higher recruitment rates than did the manipulated populations, but in all years some of the control populations were similar to the food populations. Both the food and control populations had higher recruitment rates than the fence or food+fence populations. Annual survival was generally higher on the fence and food+fence treatments than on the control and food areas (Figure 2.5). Hares on control sites had annual survival rates of 17-29%, with the highest survival in the year of early increase; hares on the fenced sites had annual survival rates about double that of the hares on control sites. During 1995-1996, the hares on food+fence had low survival, comparable to the control hares; a coyote was inside the fence during the winter. Survival rates were similar within each year, with no obvious seasonal effects (Table 2.2). Hares on the fence and food+fence treatments had higher survival than hares on control sites from 25 Figure 2.4. Recruitment rate of snowshoe hares. The values are geometric means of Jolly-Seber dilution rates, calculated from spring to spring. The values are N a c t u a l/N p r e d i c t e d , with N p r e d i c t e d derived from the estimation of what the population would have been had no animals been added to it (through immigration or births). 26 (d}EJ uojjnijp jeqes-Aiiop) luewijruosj 27 Figure 2.5. A n n u a l survival rates o f adult radio-collared snowshoe hares. Rates are K a p l a n -M e i e r estimates calculated from 1 A p r i l to 31 M a r c h ; the first bar for each treatment is 1993-1994, then 1994-1995, and then 1995-1996. Bars are mean surv iva l , and lines are 9 5 % confidence intervals derived f rom Greenwood's standard error (Po l lock et al. 1989b). Su rv iva l rates are significantly different from each other i f the confidence intervals do not overlap. U s u a l l y at least 10 hares were radio-col lared per treatment at any g iven t ime. 28 L N N N N N N N N N N N N S J S K c o O a ) o o r ^ c D L O " ^ r c o c \ i o o o o o o o o 9JBJ |BA|AjnS lenuue 29 Table 2.2. Kaplan-Meier survival estimates for snowshoe hares. Estimates are for 30-day survival within the delineated periods. Confidence limits were derived from Greenwood's standard error (Pollock et al. 1989b), and survival estimates are significantly different from each other if their confidence limits do not overlap. Sample sizes refer to independent collaring occasions, rather than to individual snowshoe hares (i.e. hares could be re-collared and would then be counted twice). n 30-day survival 95% C L May 1 - August 31 control 1993 13 0.87 0.73 - 0.96 control 1994 66 0.91 0.85 - 0.96 control 1995 52 0.91 0.86 - 0.95 food 1993 15 0.83 0.68 - 0.93 food 1994 13 0.95 0.87 - 1.00 food 1995 37 0.91 0.84 - 0.96 fence 1993 7 1.00 1.00- 1.00 fence 1994 13 0.94 0.86- 1.00 fence 1995 12 1.00 1.00- 1.00 food+fence 1993 45 0.87 0.78 - 0.94 food+fence 1994 31 0.96 0.91 - 1.00 food+fence 1995 23 0.95 0.90- 1.00 Sept. 1 - Dec. 31 control 1993 22 0.84 0.73 - 0.92 control 1994 51 0.82 0.75 - 0.88 control 1995 ' 45 0.90 0.85 - 0.95 food 1993 32 0.80 0.70 - 0.88 food 1994 23 0.94 0.89 - 0.99 food 1995 34 0.92 0.86 - 0.97 fence 1993 15 0.98 0.94- 1.00 fence 1994 20 0.96 0.90 - 1.00 fence 1995 17 0.95 0.87- 1.00 food+fence 1993 43 0.97 0.94- 1.00 food+fence 1994 31 0.96 0.92- 1.00 food+fence 1995 29 0.80* 0.67 - 0.88 30 Table 2.2 continued. Kaplan-Meier survival estimates for snowshoe hares. n 30-day survival 95% C L January 1 - A p r i l 30 control 1994 38 0.89 0.79-0.97 control 1995 53 0.88 0.80-0.94 control 1996 57 0.93 0.87 -0.97 food 1994 28 0.88 0.79-0.95 food 1995 31 0.89 0.80-0.95 food 1996 23 0.99 0.97- 1.00 fence 1994 17 0.97 0.91 - 1.00 fence 1995 15 0.96 0.91 - 1.00 fence 1996 15 0.79 0.58-0.90 food+fence 1994 35 0.96 0.93 - 1.00 food+fence 1995 25 0.97 0.92- 1.00 food+fence 1996 26 0.84* 0.72 - 0.92 *A coyote was inside the fence from at least November 9, 1995 through January 5, 1996. 31 September through December in the first two years, and fenced hares survived slightly better than control hares in two of the three summer periods. The population rate of increase was more related to the recruitment rate than to annual survival (Figure 2.6). Recruitment rate (measured by the Jolly-Seber dilution rate) was a good predictor of the rate of population increase (r2=0.61, F, 10=18.1, p<0.01). Adult snowshoe hare survival and population rate of increase were not related during the low phase (r2=0.035, F, 10=0.36, p<0.56). Food availability Small twigs were plentiful during the low phase in hare numbers (Table 2.3), with willow consistently more abundant than birch. There was no clear pattern of increase or decrease in willow and birch biomass through the years of the population low. At the end of every winter, the amount of willow and birch remaining could have fed the existing hare population for another entire winter (Figure 2.7). Although food per hare was lowest on the food+fence treatment, there was still ample food left there to support the existing population for another winter, especially since the calculation of food per hare is an underestimate, as it does not include other food plants or growth of plants in summer. Predator numbers and indices of predation risk The summer adult raptor populations remained similar through the low phase, but lynx and coyote numbers showed more interannual variation (Table 2.4). There were 1-2 pairs of goshawks in each summer, 9-10 pairs of Harlan's hawks, and 10-11 pairs of great horned owls. Both lynx and coyote numbers were lowest in the final winter of the low phase (5-6 coyotes, 8-9 lynx) and both increased slightly in numbers by the first winter of snowshoe hare increase. 32 \ Figure 2.6. Relationship of recruitment and survival to the population rate of increase. Each point represents one treatment in one year (from spring to spring). A. Recruitment and rate of increase. Recruitment rate values are N a c t u a l/N p r e d i c t e d , with N p r e d i c t e d derived from the estimation of what the population would have been had no animals been added to it (through immigration or births), y = 0.205 + 0.202x. B . Annual adult survival and rate of increase, y = 2.29 - 1.21x. 33 34 Table 2.3. Food availability for snowshoe hares: biomass of birch and willow. Values are kilograms/hectare (dry weight), mean (95% CI). Large twigs are >5 mm diameter and small twigs are <5 mm diameter. Values were derived from estimates of standing biomass and the ratio of small twigs to total biomass (see Appendix 1). Biomass of large twigs is similar from year to year (A.R.E. Sinclair and C.J. Krebs, unpublished data). of Large Twigs 1993 1994 1995 1996 Birch control 1 2 2 1 (0.2-3) 0 0 control 2 71 32 (28-36) 33 (29-37) 22 (6-47) 55 (37-80) control 3 7 ?? 26(15-55) 17(12-27) ?? food 1 371 ?? 220 (143-324) 217 (148-307) ?? food 2 394 ?? 76 (46-135) 74 (46-135) ?? fence 620 ?? 258 (147-406) 605 (357-1022) ?? food+fence 240 45 (43-46) 84 (82-87) 112 (77-157) 110(82-143) Willow control 1 1811 627 (460-820) 419(406-431) 441 (317-579) 420 (255-614) control 2 2094 434 (353-522) 376 (358-395) 430 (343-522) 366 (310-425) control 3 238 ?? 499 (329-814) 118 (76-171) ?? food 1 1322 ?? 729 (443-1125) 732 (496-1038) ?? food 2 394 ?? 336 (218-511) 693 (448-1139) ?? fence 1919 ?? 462 (292-660) 800 (549-1108) ?? food + fence 2591 556 (537-577) 411 (400-422) 618 (465-788) 547 (396-699) 35 Figure 2.7. Woody browse available per hare at end of winter. The four treatment lines indicate availability, as kg browse of <5 mm twigs per hare. Biomass was assessed in May of each year, and hare densities in March. The winter requirement line indicates that -54 kg browse are necessary per hare for an entire winter, based on a 180 day winter and the estimate of Pease et al. (1979) that a hare needs to eat 300 g (fresh weight) daily. 36 c o o T J O o CD O c CO 0 o c a + T J O O m » 4 c CD e 0 '13 cr 0 >— 0 < Xr CD CD CD un CD CD CD CD CO CD CD CM CD CD O O O CD O O O LO O O O o o o CO o o o CN o o o 9 j e i | jed S6;MI news 6>| 37 Table 2.4. Predator numbers at Kluane during the low and early increase phases of the snowshoe hare cycle. Raptor numbers represent pairs of birds in 100 km2, and mammal numbers represent individuals in 350 km2. The data are from F.I. Doyle (unpublished), Boutin et al. 1995, and O'Donoghue et al. 1997. Raptor numbers were assessed in summer, and mammal numbers in winter. SUMMER 1992 1993 1994 1995 1996 Great Horned Owl 14 10 11 10 10 {Bubo virginianus) Goshawk 1 2 2 1 1 (Accipiter gentilis) Harlan's Hawk 9 10 10 10 10 (Buteo jamaicensis) WINTER 91-92 92-93 93-94 94-95 95-96 Coyote 17 9 5 6 7 (Canis latrans) • Lynx 28 15 9 8 12 {Lynx canadensis) 38 Predator densities were therefore low but fairly consistent throughout the cyclic low phase, despite the large change in hare densities (Figure 2.8). Causes of death for snowshoe hares Predation accounted for 75-100% of all deaths of adult hares on control sites (Table 2.5). The other treatments had similar proportions of predation deaths, except for the food+fence treatment, which had slightly elevated numbers of non-predation deaths. Lynx and coyotes were responsible for most of the predation deaths (-65%) on the food and control areas. Goshawks and owls accounted for only 9-21% of predation deaths on these sites (the remainder of predation deaths were due to mustelids or were unidentifiable). On the fence and food+fence sites, raptors caused 51-89% of the predation deaths. In the winter of 1995-1996 a coyote inside the food+fence manipulation was responsible for 37% of total predation mortality on this site during the three year study period and 46% to 96% of the predation during that winter (the exact amount is unclear because 50% of the kills there that winter could not be identified to predator species). Body condition Female snowshoe hares on the unmanipulated control sites were in below average condition in the first year of the low and the first year of the increase (Figure 2.9). Female hares on all treatments had higher mean condition than females on control sites, and hares on the food and fence treatment were in better condition than hares on the food addition sites (two-way ANOVA, treatment F 3 318=15.0, p<0.001, followed by Tukey post-hoc tests). Annual effects were statistically indetectable (two-way ANOVA, year F 3 318=0.90, N.S.), but the interaction of 39 Figure 2.8. Predator and hare densities. Circles represent the density of lynx + coyotes, and squares represent the density of goshawks + Harlan's hawks + great horned owls. The dotted line indicates the average density of hares on three control sites. 40 CD -*—» O >. O o _ + o c Q-to to Qi CD . C 4» 4 • 41 CO CD bfl rt E "> D ai O T3 D< CJ ai S rt ~<3 5" CO oi -8 c o ai o .5 rt CO rt S co o o rt s ai ^ .2 •a o T3 > tt) CO <J -fi. J3 o Ci o o 1— P H -a Oi < cfc CO ^ "§ ai - O >-> co r j ai w T3 T3 3 C O « c 5 o ^ F rt1 " ^*-> - r t e ^ ^ •>.•«» ai ai ai to G -fi ai ai o &1 -fi Ci T3 3 Si ai co ai 2 -fi S H ai ai C Ml +3 C -fi 5 w 2 co -rt s .2 -fi "J2 l i 3 5 ai co co 3 £ rt 3 ° ai .n 8 0 co "rt 3 rt rt ai U CO rt « CD 3 - f i rt ti_ H o CO « C I) ai ai ai & ai -5 <D ^ CD ^- —1 co U 60 3 2 "53 0 rt S 3 7 3 - CD g =3 T3 - ° 12 2 li T3 C fi "2. „ ,C rt c 2 a) o 5 ^ rt " C -rt X) a s ^ a. £ .2 § M> ^ rt c ai ^ T3 rt S ^ ai rt > rt P b & ai O — ai •4—* rt C "So o ai CO rt ai C rt "3 1 -j» CO C -r rt c S #o rt Oi u CM rt T3 e o c c 3 cu c o c c 3 rt co cu 0 0 0 O 0 0 0 0 0 0 0 cn 0 O 0 0 0 0 m d ON 00 0 0 0 O 0 0 0 /—) 0 1—1 0 00 IT-) —• '—1 •—• —' —1 '—1 00 •—1 00 O N 00 p 1 0 6 C N O N r^ ; O N 00 O N in  0 C N O C O r- 00 V D 00 00 O N 00 N O r- Os 0 0 00 00 N O C O ^ in 00 1—* */S ON (N CN O O C N in 00 <— d ~ ~ C N —1 CO CO 1/0 >0 ON 00 ON O ~ t-H CO CN C N — ^ o CN NO CN CO CN IT) N O ON ON O N ON O N ON " S co m S O N O N ON Q O N O N ON ^ O N vn C N C N O m -1 0 0 0 1/0 CO NO CO —' — — CO C N o 0 co C N o in C N co N O 00 N O —< C N C N co 00 C N >n CO — — •3- m N O O N O N O N O N O N O N 'O CO T t g O N ON _ O O N ON O N — — ~ i >n O N O CO >n C N C N o o 0 C N O O O O — 0 0 O N O* O 0 q —; p in O N d 0 CN C O o in 0 0 ^ 0 0 m m C N o o o o o C N O O O O 1- m \D O N O N O N O N O N O N CU U B co T f r O N O N O N cv ON O N O N — — ^ ai o c ai —; N O co N O O N CO CO CO r- r~ 00 N O 0 d C N CO -sf ON NO CO NO CN 0 CO CO CO C N C N NO N O C N 06 co in C N H m N O O v i —1 C N O O N O O N O O 3 O CU u c CU ta + o o ta Tt- >n O N O N O N O N CO 1^" ON ON ON ON CO 00 ON CN CO ai o NO C ON ~^ — <u ON 3 — O 4 2 Figure 2.9. Body condition of female snowshoe hares in spring. Values are mean ± SE. The condition index is observed weight/predicted weight, with predicted weight derived from the relationship of skeletal size to weight (K.E. Hodges, E.A. Gillis, and C.I. Stefan, ms. in prep). Values are relative, and 1 indicates average condition. Means represent 2-92 hares (x = 20.8). 43 CD CD CD H CD CD CD H LO CD CO H i LO CD CO TD O O LL CD CD CO CD CD CD O c CD LL + "O o o CD CD CO CD CD H CD CD CD CO CD CD LO CD CD H i LO CD CD O O CD CD CO CD CD CD O C CD H CD CD CO CD CD CD O OO O CM r - O CD O OO O xapin uoiiipuoo Apoq 44 treatment and year was significant (F9318=4.38, p<0.001). The interaction term may be due to three anomalous years: control hares were in poorer condition in 1993 than in later years, and in 1996 fed hares were in better condition and hares on the food+fence site in poorer condition than in previous years. Age structure On control sites, yearling hares composed 83.3-89.3% of the breeding population during the low of the cycle; in 1996, the first year of increase, yearlings composed 89.6% of the population (Table 2.6). On all four treatments, the proportion of yearlings increased and mean age decreased through the three years of the low. For both the fence and food+fence treatments, yearlings composed a smaller fraction of the breeding population than on the control sites, and both mean age and maximum age were higher than for control populations. DISCUSSION Theories for dynamics of the low phase The competing trophic models for dynamics of the snowshoe hare cycle are that food alone is limiting (food limitation), that predation alone is limiting (predation limitation), or that the two interact-continually or sequentially—to limit hare populations. Previous research has not resolved the issue, since studies have found conflicting evidence and have examined different phases of the cycle. Evidence from the low phase of the cycle is particularly scanty. The interaction of both factors has become the favorite, since the bulk of evidence suggests both food and predation contribute to cyclic dynamics (Keith and Windberg 1978, Keith 1990, Krebs et al. 1995, 1996). Despite the evidence in favour of this position, two questions remain: whether the interaction is sequential or continuous, and whether predation affects hare demographics primarily 45 Table 2.6. Spring age structure of snowshoe hares during the low phase. All hares were aged a year on January 1, so a population composed entirely of the previous summer's young would have a mean age of 1.0. Percent young is the percentage of the population aged one. hares mean age: percent age of caught years ± SE young oldest hares Control 1993 12 1.17 ± 0 11 83 3 2 1994 14 1.21 ± 0 16 85 7 3 1995 28 1.11 ± 0 06 89 3 2 1996 106 1.11 ± 0 03 89 6 3 Food 1993 20 1.55 ± 0 21 70 0 4 1994 26 1.35 ± 0 19 84 6 5 1995 53 1.08 ± 0 05 94 3 3 1996 213 1.15 ± 0 03 87 3 4 Fence 1993 10 1.70 ± 0 30 60 0 3 1994 15 1.53 ± 0 27 73 3 4 1995 24 1.33 ± 0 18 79 2 5 1996 5 1.60 ± 0 25 40 0 2 Food+Fence 1993 52 1.87 ± 0 12 38 5 4 1994 39 1.72 ± 0 17 64 1 4 1995 36 1.36 ± 0 13 75 0 4 1996 71 1.42 ± 0 11 74 7 6 46 as a mortality agent (the predation interaction hypothesis) or also through indirect effects on condition and fecundity (the risk interaction hypothesis). Keith's sequential food-predation hypothesis states that predation is limiting but food adequate during the low phase (Keith et al. 1977, Keith and Windberg 1978). Keith and Windberg (1978) suggested that the increase phase did not begin until predation mortality declined, as a result of the numerical lag in predator populations (Keith et al. 1977). Although Keith suggested that malnutrition (due to limited food supply) was liable to increase both starvation and predation deaths (Keith et al. 1977, Pease et al. 1979, Keith 1990), during low phases hares appear not to be malnourished (Keith and Windberg 1978, Keith 1990). Keith's arguments suggest that during the low phase predation limits hare numbers predominantly through its effects as a mortality agent. During the low phase, Keith's arguments become identical to the predation interaction hypothesis. In contrast, predation may also affect prey indirectly through the stress of chases or predator avoidance, and limitation of foraging time or locations. Behavioural trade-offs between safety and food are common in small mammals (e.g. Brown et al. 1988, Desy et al. 1990, Lima and Dill 1990, Lagos et al. 1995) and have been shown for hares during other phases of the cycle (Gilbert and Boutin 1991, Hik 1994, 1995, Rohner and Krebs 1996). According to this argument, the low phase can be explained as a result of behavioural trade-offs by hares, which can affect body condition, fecundity, and both starvation and predation deaths (Boonstra and Singleton 1993, Hik 1994, 1995, Krebs 1996, Boonstra et al. 1998b). For the risk interaction hypothesis, food and predation are both essential to population dynamics since they determine the behavioural trade-offs. 47 Table 2.7 summarizes the effect of phase of cycle and treatment on the demographic parameters which are discussed below. Evidence of food limitation? The evidence presented in this chapter contradicts the idea that winter food was limiting to hares during the low phase. Food density was high, and food availability per hare was much higher than food needed per hare. Even if there were changes in the secondary compound content of these species as a response to heavy browsing from hares at peak densities (Fox and Bryant 1984, Sinclair et al. 1988b), hares during the low had many times the biomass that they needed and presumably could eat less defended plants. There were no starvation deaths during the low phase; the few non-predation deaths appeared to result from cold temperatures or old age, and necropsies failed to find evidence of starvation (C.J. Krebs, unpublished data). This pattern of abundant food affirms previous research at Kluane, which has not demonstrated food shortage for hares (Krebs et al. 1986a, 1986b, Sinclair etal. 1988b). Keith (1990, Pease et al. 1979) has demonstrated food shortage during peak winters, and his experimental creation of food-short populations (Vaughan and Keith 1981) indicates that when hares are malnourished, deaths due to starvation are common and that hares have low body weights and lose large amounts of weight over winter. These patterns were not apparent in the hares at Kluane (C.J. Krebs unpublished data), confirming that food was not limiting during the low phase. Food addition did, however, result in higher population densities and better mean body condition, but the lack of reproductive or survival changes on the food 48 Table 2.7. Synopsis of snowshoe hare demography during the cyclic low phase. Population rate of increase, survival, and recruitment were calculated from spring to spring, 1993-1994, 1994-1995, and 1995-1996. Density, condition, and mean age were calculated for March of each year. A . Treatment effects. Values are relative to control populations, for the three years of the low phase combined. The final column indicates which trophic level hypothesis was supported. The + signs indicate more effect than on the single factor treatments. food fence food+fence hypothesis supported density higher higher higher+ interaction increase rate similar similar similar — survival similar higher higher predation recruitment similar-lower lower lower — condition better better better+ interaction mean age similar higher higher predation B . Comparison with other phases of the cycle. Values are from control sites and are relative to the general pattern of the low phase. Survival and increase rate data from Krebs et al. 1995, 1996, unpublished; condition and age data from K.E. Hodges, E.A. Gillis, and C. Stefan, unpublished. Recruitment indices varied too much in each year and between sites to offer a clear or consistent pattern. increase peak decline increase rate higher lower lower survival similar lower lower condition poorer better poorer mean age similar higher higher 49 addition sites indicates that food limitation by itself is not a viable explanation for the duration of the cyclic low phase. Evidence of predation limitation? There were few predators during the low phase of the cycle, and little reproduction of predators (O'Donoghue et al. 1997). In contrast to Keith et al.'s (1977) data, which showed marked changes in predator density in each year of the low phase, at Kluane predator densities declined and then remained low for several years. Hares on control sites had an annual mortality rate of 71-83%. Despite the scarcity of predators at Kluane, more than 3/4 of these deaths were caused by predation. On control and food sites, most of the predation was by lynx and coyotes. On the fence and food+fence sites, most of the predation was by raptors, but the increased proportion of avian kills did not completely compensate for the lack of mammalian kills. The survival data support the predation hypothesis, since the reduction of predators resulted in higher annual survival rates, but food addition did not. Evidence for interacting effects of food and predation The higher densities on all treatments relative to the controls may indicate that dynamics were affected by a food-predation interaction, since density changes potentially result from differential recruitment or survival. Immigration is unlikely to be of large enough magnitude to create the observed density differences (Boutin et al. 1985, Krebs et al. 1995). But these treatment sites were not established de novo at the start of the low phase; the treatments began during the previous increase phase. All treatments have had higher populations than controls since at least 1990 (until the fence population declined below control levels in 1996). It could be that higher treatment populations during the low phase are the result of higher populations during 50 the peak and decline phases, rather than the result of differing demographic rates during the low phase. But even if that were the case, densities at those times support a food-predation interaction hypothesis (Krebs et al. 1995, 1996). The better body condition and older age structures of the treatment populations also support the interaction hypothesis. Condition was better on both single factor manipulations, and best on the joint manipulation. Mean age was higher on just the fence and food+fence treatments, but hares lived to older ages on all treatment sites than on control sites. Although older hares tend to be in better condition than yearlings (Keith 1990, K.E. Hodges, E.A. Gillis, C.I. Stefan ms. in prep) the observed treatment patterns are not a result of the older populations: when mean ages were similar, female condition was still better in manipulated populations. The same patterns also occur when just condition of older adults is examined (K.E. Hodges, E.A. Gillis, C.I. Stefan, ms. in prep). If predation risk alone affected age and condition, the food treatments should not have shown the patterns they did. Indirect effects of predation Predation risk affected condition and age structure. Hares on all treatments had better condition and higher mean age than hares on control sites. As additional confirmation of this idea, the food+fence treatment showed a reduction in condition—to the level of the control hares—in 1996. In November 1995, a coyote got inside the fence; in January 1996, food addition was halted because the coyote was still inside and experimental integrity was already effectively destroyed. Snowshoe hare condition decreased rapidly after these changes. Changes in condition and age structure did not appear to lead to increased recruitment during this phase of the cycle. The recruitment indices are based on trapping data, and thus it is 51 impossible to distinguish the effects of fecundity, juvenile survival, and dispersal on the estimates. Fecundity may be affected by the condition changes; studies from other phases of the cycle indicate that over-winter weight loss influences fecundity, but age does not (Keith and Windberg 1978, Cary and Keith 1979, Keith 1990). C.I. Stefan (unpublished data) found no link between body weight and fecundity for the first two litters at Kluane during the early increase phase, but there was a positive effect of weight on the size of the third litter. Thus, based on the evidence available so far, it is impossible to determine whether age and condition changes are simply side effects of reduced predation risk, or whether they form an integral part in changes of fecundity that lead to numeric changes. Conclusions and directions for further research The evidence from the cyclic low phase at Kluane refutes the food limitation hypothesis for hare population dynamics. Food was not limiting, and the reactions of the fenced populations should not have occurred if the food hypothesis were true. The predation hypothesis and Keith's sequential food-predation hypothesis (which equals the predation hypothesis during the low phase) are also refuted. Population responses on the food addition areas should not have occurred under these hypotheses. The data are consistent with the risk interaction hypothesis. The observed treatment effects on density, condition, age structure, and possibly survival support this hypothesis. The differing densities might reflect the dynamics that occurred earlier in the cycle, or could reflect dynamics that occur during the low phase of the cycle. In either case, the fact that all treatments showed higher densities than the control sites indicates that food and predation interact to affect 52 hares' dynamics. The changes in condition and age structure show that predation does have indirect as well as direct effects on hare populations. Several issues remain unresolved. Fecundity data and juvenile survival data are scarce for this phase of the cycle. Fecundity is apparently high (Cary and Keith 1979, Stefan 1998) and Gillis (1997) found that first and second litter juveniles survived as well as adults during the early increase phase, but that third and fourth litter juveniles survived less well. Knowledge of both fecundity and juvenile survival would be helpful in understanding the dynamics of this phase, especially since there is some evidence that four litters are born during the low phase (Cary and Keith 1979, Stefan 1998) and since juvenile survival may be crucial to hare population dynamics (Keith and Windberg 1978, Krebs etal. 1986b, Keith 1990, Gillis 1997, E.A. Gillis, D. Haydon, C.I. Stefan, and C.J. Krebs ms. in prep). Although the evidence supports the risk interaction hypothesis, the food+fence manipulation did not have a shorter low phase than the control areas had. This inability to change the timing of the cycle could be due to the small (1 km2) size of the fenced area and even smaller food addition area, the ability of hares to disperse out of the relatively high density food+fence area, or an insufficient change in food availability or predation risk (since raptors still had unrestricted access). Additionally, the behavioural mechanisms underlying the indirect effects on condition and age structure are not clear. Changes in age structure could be simply a reflection of reduced predation rates, but it is also possible that behavioural changes affect the susceptibility of hares to mortality. Potential behavioural changes include diet, habitat, and movement; these possibilities will be addressed in subsequent chapters. 53 C H A P T E R 3 SNOWSHOE H A R E F O R A G I N G B E H A V I O U R DURING T H E L O W P H A S E O F T H E T E N - Y E A R C Y C L E INTRODUCTION Predators can cause prey to change their behaviour (Lima and Dill 1990), and some behavioural shifts have demographic effects (Hik 1995, Sinclair and Arcese 1995). Although previous research has been able to show that fecundity and survival of individuals is affected by predator sensitive behaviour, it is not entirely clear from field research whether population dynamics as a whole can be affected by predator-prey interactions mediated by behavioural shifts. Behavioural effects on population dynamics are possible theoretically (Ives and Dobson 1987, McNamara and Houston 1987, McNamara 1990, Abrams 1992, 1993), but in order to demonstrate such an effect in the field, first anti-predator behaviours must be detected, and second the behaviours must be shown to affect population dynamics. Snowshoe hares may have anti-predator behaviours that lead to demographic consequences which affect the dynamics of the ten-year cycle (Hik 1994, 1995, Krebs 1996, Boonstra et al. 1998b). Snowshoe hares, their predators, and their food-plants all fluctuate in a ten-year cycle, and some interaction of food and predation is the most likely explanation for the hares' dynamics (Finerty 1980, Bryant 1981a, Fox and Bryant 1984, Bryant et al. 1985, Smith et al. 1988, Keith 1990, Boutin et al. 1995, Krebs 1996). The period of low densities has proven to be difficult to explain by invoking direct effects of these trophic interactions: the low phase can last two to four years, but food availabilities are high and predation risk is low, both of which 54 would suggest that there is no barrier to population increase (Chapter 2, Smith et al. 1988, Sinclair et al. 1988b, Keith 1990, Boonstra et al. 1998a, 1998b). If, however, snowshoe hares used an anti-predator behaviour that affected their ability to use the available food, or if the stress of avoiding predators led to physiological changes, then it is possible that hare densities remain low through behavioural limitation rather than through resource or predation limitation (Hik 1994, 1995, Krebs 1996, Boonstra et al. 1998b). Dietary changes could link behaviour, physiology, and demography. Changes in diet can easily affect physiological condition and fecundity (Windberg and Keith 1976b, Cary and Keith 1979, Vaughan and Keith 1981, Hik 1995, Sinclair and Arcese 1995, Boonstra et al. 1998a) and many other anti-predator behaviours—such as movement, activity, and habitat selection—are thought to lead to changes in diet because of reductions in available foraging time or an altered array of available foods in a given habitat patch (McNamara and Houston 1987, Godin 1990, Lima and Dill 1990, Sih and Moore 1990, Anholt and Werner 1995, Hik 1995). The patchy habitats in the boreal forest lead to a mosaic of areas with varying safety from predation and food availabilities, and during the decline phase of the cycle, hares ate poorer quality foods in habitats that protected them from predators, thus trading-off risk and food (Hik 1994, 1995). This behaviour apparently affects hares physiologically; hares protected from mammalian predators by a fence did not show these behavioural shifts to safer habitats and poorer foods, and they were in better physiological condition than were hares not protected from mammalian predators (Boonstra and Singleton 1993, Hik 1994, 1995, Boonstra etal. 1998a, K.E. Hodges, E.A. Gillis and C.I. Stefan, ms. in prep). 55 My objectives in this chapter are therefore two-fold: first, to describe the diet of snowshoe hares during the low phase, to see if their diets are poor enough to suggest behavioural limitation as a factor in the cycle; and, second, to determine how predation risk and food distribution affect hares' diets during this phase. I will consider three alternative explanations for hares' diet selection: that food quality and distribution explain hares' diets, with no influence of predation risk (food hypothesis); that hares minimize predation risk at all times, restricting their diet to the foods that expose them to the least risk of predation (predation hypothesis); and that both food and risk affect hares' diet selection (risk interaction hypothesis). Each of these possibilities has been shown to apply to the diets of other small mammals (Boutin 1990, Lima and Dill 1990). Diet selection by snowshoe hares Hares are generalist herbivores: in summer they eat leaves, forbs, grasses, and twigs, and in winter they eat woody browse. Hares can change their diets by changing the species composition, choosing particular growth forms (adult vs. juvenile form twigs), or selecting smaller twigs (Bryant and Kuropat 1980, Reichardt et al. 1984, Sinclair and Smith 1984b, Rogowitz 1988). Hares tend not to change the quantity that they eat (Holter et al. 1974, Pease et al. 1979, Sinclair et al. 1982), so if food quality is reduced, hares cannot compensate by increasing their intake. Low quantities of energy in the browse or low digestibility therefore lead to weight loss (Rodgers and Sinclair 1997). Snowshoe hares must eat adequate amounts of food on a short-term basis, since they cannot store fat readily and their body reserves maintain them at resting metabolic rates for about four days at most (Whittaker and Thomas 1983). Protein appears to be especially limiting in the diets of hares in the boreal forest (Sinclair et al. 1982, Sinclair and Smith 1984a), but secondary chemicals can inhibit the digestion of 56 protein (Bryant and Kuropat 1980, Sinclair et al. 1982, Risenhoover et al. 1985, Sinclair et al. 1988b). Hares may also select food based on the fibre content and energy content, but protein and secondary chemicals appear to be the main characteristics by which hares choose their diets (Cheeke 1983, 1987, Sinclair and Smith 1984b, Sinclair et al. 1988a, 1988b, Jogia et al. 1989, Bryant etal. 1992, Schmitz etal. 1992, Rangen etal. 1994, Rodgers and Sinclair 1997). Experimental design, dietary variables, and predictions Dietary analyses can consider two broad types of variables: what is eaten and the nutritional quality of what is eaten. For hares, what is eaten means the species composition of hares' diets, the twig size, the twig growth forms, and how selective hares were in their browsing. Nutritionally, I consider the fibre and protein content of hares' diets, estimated directly from the twigs that hares ate and indirectly via the relationship between faecal fibre and dietary composition. I used experimental manipulations—predator reduction, food addition, the joint manipulation of both, and controls—to create areas with distinct predation risk and food availability, so that I could determine the effects of risk and food availability on hares' diet selection. The predictions are summarized in Table 3.1, and elaborated upon below. If hares chose their diets based on food distribution and quality, but were not influenced by predators, then hares on areas with supplemental food should be able to obtain much of their protein and some of their energy from the rabbit chow; since this food was readily available, hares should have more time for twig selection. Compared to hares on control and fenced sites, hares on food-supplemented areas should be more selective in their diets, eat twigs of smaller diameter, and eat less bark. Because of their high selectivity, hares on food supplemented sites should have low food intake rates, low dietary fibre, high dietary protein, and low faecal fibre. 57 c "XS e T3 O O fc V fc V U V ai o c ai fc ai c ai fc fc" fc V U -d o o fc -o o o fc V fc fc V u V ai o c ai fc 1 3 O O fc V fc fc V U V di O c fc ai O c fc V U V fc fc V -a o o fc U U V ai o c ai fc V -a o o fc v fc fc V <D O c ai fc V •a o o fc fc <u o e o > \< © <u t. PH U •d o o fc V fc fc o c ai fc ai _> o JD 13 (« -4—* o c ai c U T3 O O fc V fc fc ai o c ai fc U -d o o fc V fc fc ai Ci c ai fc fc fc ai o c u fc V U T3" O O fc u u -a o o fc V fc fc di o c di fc T3 O O fc V fc fc ai O c ai fc '•8 u ai o c ai fc V fc fc 73 O O fc _> o JJ 13 V2 di c di fc fc -d O O fc V di O c fc U ai" o c <u fc V fc fc •d o o fc fc fc O o fc V U 0) o c ai fc U U ai O c D fc V fc fc -d o o fc O e ai fc V fc fc •d o o fc IM X ) ai o o a 4M \> * 4M o JM 13 oo ai o s CM Oi ^3 Q £ •4—» ai Q c '53 —^* o IM OH — ai Q ai -4—» Cj IM ai c3 X ) "3 o fc 58 If the predation hypothesis were true, hares on the fence and food+fence treatments have a relaxed constraint, so can be more selective in their diets than hares exposed to predators. Compared to hares exposed to predators, hares protected from predators should eat smaller twigs, eat less willow and spruce bark, and have lower intake rates. Their dietary protein should be higher, dietary fibre should be lower, and faecal fibre should be lower. If the risk interaction hypothesis were true, and hares are willing to sustain greater predation risk when the food rewards are greater, then hares on all treatments should differ from each other. Unlike the food hypothesis, food addition should lead to a d e c r e a s e in the quality of the remaining diet: because larger twigs are more available, eating these reduces time spent foraging and increases time in safe refuges. Food addition relaxes the constraint for protein. Protecting hares from predators should result in increased diet quality, since hares can forage more selectively without a cost of predation; the time constraint is relaxed. Hares on the fenced site should eat the smallest twigs (they can afford time, but need protein), followed by control hares (they should try to limit time, but they need protein), food+fence hares (they can afford time, but do not need protein from twigs), and food hares (they should minimize time, and protein is provided). Protein and fibre intake from browse should decrease and increase, respectively, with twig diameter. Regardless of strategy, hares' diets should be higher in protein during November-December than during January-April. Snow cover is lighter in the fall, and more species are available for browsing. Temperatures also tend to be warmer in the fall, imposing fewer thermal constraints on hares, which reduces the physiological costs of selectivity. 59 M E T H O D S A N D M A T E R I A L S Study area, treatments, and study animals I studied snowshoe hares at treatment sites maintained by the Kluane Boreal Forest Ecosystem Project, located near Kluane Lake, Yukon (60°57'N, 138° 12'W; Krebs et al. 1995). The treatment areas were located in the Shakwak trench, spread along 30 km of the Alaska Highway; each site was at least 1 km away from other experimental sites (Figure 2.1). The forest is predominantly composed of white spruce (Picea glauca) with some aspen and balsam poplar stands (Populus tremuloides, P. balsamifera). Grey-leaf willow (Salix glauca), bog birch (Betula glandulosa), and soapberry (Shepherdia canadensis) are the main understory shrubs. During March 1993-March 1996, encompassing a population low and early increase, I studied snowshoe hares on four treatments which the Project had established by 1988 (Krebs et al. 1995). Lynx (Lynx canadensis) and coyotes (Canis latrans) were excluded from a 1 km2 area protected by a chicken-wire and electric fence (fence). Within the fence, a 10 ha area was strung with monofilament line to deter avian predators (predominantly goshawks, Accipiter gentilis, and great horned owls, Bubo virginianus). Two 36 ha areas (food) were spread with ad lib. pelleted rabbit chow (minimum 16% crude protein) along four 570 m cutlines 180 m apart. The food+fence treatment combined these two manipulations but had no monofilament line. I also used four unmanipulated control sites located throughout the Shakwak trench. In November 1995, a coyote got inside the fence of the food+fence treatment and killed many hares; in January, food addition was stopped there since the treatment was already effectively destroyed. I use data from this treatment only through fall 1995. 60 Snowshoe hares were trapped in March and October of every year; each treatment area had an effective trapping size of 60 ha. Each hare was given a #3 Monel eartag (National Band and Tag Co., Newport KY). Every time we caught a hare, we recorded location, sex, eartag, weight, length of the right hind foot, and reproductive condition. Some hares were radio-collared with 40 g mortality-sensing collars (Lotek, Newmarket ONT). I estimated hares' diets—species composition, twig size, protein and fibre content—based on the twigs consumed along the snow-tracks of radio-collared hares. The calculations used in the description of hares' foraging patterns are summarized in Table 3.2 and described below. Feeding trials To establish the relationship between protein and fibre in hares' diets and hares' faecal fibre content, I ran two sets of feeding trials. In late October 1994,1 caught five wild snowshoe hares (1 <f, 4?) and housed them outside individually in 60x60x120 cm cages. The cages were wood and chicken wire partially wrapped in burlap and plastic to act as wind and snow barriers. I left ample snow for moisture, and provided excess browse, cut fresh daily. Animals were fed birch and then willow for two days each; twigs were cut at 3 mm diameter. This set of feeding trials was curtailed due to bad weather and the hares' weight loss. In January and February 1995,1 used seven hares (2d", 5?) in modified feeding trials. I placed the cages inside an unheated building, and covered them on four sides to provide visual barriers. I weighed the hares daily, and released hares that had lost >50 g in 24 hr. or more than 100 g total. To counter-act the problem of dietary and weight changes, all hares were also fed commercial rabbit chow (minimum 16% crude protein). I weighed food offered and food remaining daily. Each forage was given for two days; on the first day, chow was unrestricted; on 61 Table 3.2. Calculation of dietary indices. The subscripts are: i, j = species, n = number of species, k = twigs, m = number of twigs, a = sites, r = number of sites, DPB = diameter at point of browse. Tracks were usually the sample unit. Estimate Faecal pellet deposition rate Consumption rate Species biomass (except spruce clips) Spruce clip biomass % browsej in diet Species preference (Manly's Alpha) Dietary fibre per track Dietary protein per track Time along track Calculation Data Sources pellets/hr food/day •+ pellets/day £ DPBk • drywtk,dpb jt-i r pelletsa • dry wt/pellet a-l n (biomasSj-r 53 biomassj)- 100% t-i n eatenj • l / J T (eaten/available;) availablej 1 - 1 5>PB rfibre i > d p b n DPBj • protein,,dpb i - i # pellets -=- pellets/hr feeding trials feeding trials tracking twigs in laboratory feeding trials tracking tracking twigs in laboratory tracking, twigs Table 2.2 twigs in laboratory twigs per track twigs in laboratory twigs per track tracking pellet deposition rate Intake rate biomass/track 4- time/track see above 62 the second day I reduced the amount of chow I offered, so that hares ate -70% the amount of chow as on the day previous. The volume of the test forage was always in excess. I used four forage types: willow and birch cut at diameters of 3 mm, 3-5 mm willow segments, and spruce twigs (from squirrels; see below). I counted faecal pellets daily, and collected a subsample of pellets that were not contaminated by urine for chemical analysis. I present data from the second day of each diet, since faecal pellets from the first day on a diet might still contain material from the previous food type. Faecal pellet deposition rate: calibrating a faecal clock Faecal pellet numbers along hare tracks can be converted into estimates of time, once the deposition rate is known. I counted number of pellets dropped per day during the January feeding trials, and during one 24 hour period I counted pellets dropped per hour to determine the daily rhythm and hourly rate of faecal pellet excretion. Additional faecal pellet drop rate trials were conducted using hares in a laboratory colony in Vancouver, BC (January 1994, 2 c?, 5 ? ; October and November 1994, 3 d \ 5 ?). For the statistical analyses, I excluded the first day from the Kluane hares, since the hares were habituating to captivity. Collection of hare faecal pellets for fibre analysis In winter, I collected hares' faecal pellets from snow tracks, often from individuals identified via radio-telemetry. In summer I collected pellets from hares that I observed foraging; again, these were often known hares that I had used radio-telemetry to locate. Whenever snowshoe hares were trapped, we collected faecal pellets; if a hare was caught multiple times within a week, faecal pellets were collected only from its first capture. 63 Collection of forage samples To establish the weight, fibre content, and protein content of hares' food, I collected samples of willow up to 8 mm diameter, birch up to 5 mm, and soapberry to 3 mm. These maxima are close to the extreme sizes that snowshoe hares browse. I collected complete twigs in 1 mm diameter increments, and also as twig segments (e.g. from 3 mm at one end to 4 mm at the other). Willow and birch were collected in April 1993 and in October and December 1994, and soapberry in May 1994. All twigs were leafless and typical of hares' winter browse. In July and September 1995,1 collected samples of bluebell (Mertensia paniculata), fireweed (Epilobium angustifolium), grass (Festuca altaica), and lupine stem (Lupinus arcticus; only stems were collected since that is the only part hares eat). Hares also ate white spruce extensively. From March 1993 until May 1995, hares predominantly ate twigs from newly fallen trees and small branchlets (spruce clips) discarded by red squirrels (Tamiasciurus hudsonicus) foraging at the tops of the trees. In winter 1995-1996, hares often ate directly from standing trees. I collected samples of spruce clipped by squirrels in April 1993 for chemical analyses, and I used twigs collected in 1995 (by G. Sharam) to establish biomass for different twig diameters. Fibre and protein analyses of plants and hares 'faecal pellets All plant and faecal pellet samples were stored frozen, then dried at 65° for >48 hours in a forced air oven. Samples were ground in a Wiley mill to pass through a 40 mesh screen. I analyzed samples for the percent fibre content using the acid-pepsin digestibility technique (Tilley and Terry 1963, Belovsky 1986, Schmitz 1990, Larter 1992). The acid-pepsin innoculant consisted of 2.00 g pepsin in IL of 0.1 M hydrochloric acid. For each sample, 0.200 g 64 was mixed with 20 ml acid-pepsin innoculant in a 20x150 mm test tube (a few samples were 0.100 g and 10 mL innoculant), then heated in a water bath at 37° for 48 hours. Samples were swirled at the beginning, after one hour, after six hours, and after 24 hours of digestion. After 48 hours, they were vacuum filtered through previously weighed filter papers (Whatman #4), then dried at 90-95 C° for 48 hours and reweighed. Fibre was calculated as (g remaining sample/ sample weight) • 100%. I analyzed forbs, twigs, and faecal pellets from the hare feeding trials for crude protein content using the macro-kjeldahl technique (AOAC 1970, Parkinson and Allen 1975). Sulphuric acid digest solution contains 32.4 g lithium sulfate, 0.84 g selenium, 300 mL hydrogen peroxide, and 840 mL concentrated sulfuric acid. Half a gram of sample and 15 mL of solution were put into 250 mL digestion tubes and catalyzed by 1-2 mL hydrogen peroxide. Samples were digested for 45-75 minutes at 410° C. Digested samples were then analyzed using a Tecator auto-analyzer. Crude protein was calculated as 1.401-6.25-acid molarity k-100% / g sample, where 1.401 is the atomic weight of nitrogen, 6.25 is a constant relating nitrogen to crude protein, and k corrects for the amount of nitrogen the auto-analyzer detected in the distilled water used for diluting samples prior to analysis (k was calculated separately for each set of -20 samples run on the auto-analyzer). Snow-tracking snowshoe hares To determine snowshoe hares' diets during months with snow cover (October-April), I used radio-telemetry to locate hares, then followed backwards along their tracks. At each browse site, I recorded the species eaten, twig diameter at the point of browse (DPB), and number of 65 faecal pellets dropped. I also counted all faecal pellets along each track that were not associated with browse sites. For willow twigs, I recorded whether the twigs were juvenile or adult growth form (see Sinclair et al. 1988b). Tracking data were used to calculate the following dietary measures: species composition, twig size, twig growth form, fibre intake, and protein intake. Some tracks (17%) were from unknown hares; these tracks were chosen when I found feeding areas with >50 faecal pellets (-1.5 hours of hare time) or when tracks of radio-collared hares were undecipherable. From March 1993-March 1996,1 tracked 144 hares along 243 tracks (129?:69d":47 unknown). The average track I followed had 22.5 browse points along it (range 1-105) at 4.9 sites (range 1-30), lasted a distance of 36.0 m (range 0-899), and represented 2.7 hare-hours (range 0-20.6). Browse consumption and species preferences of snowshoe hares For each track, I calculated biomass consumed per plant species and grams of fibre and protein eaten, using the predictive curves generated from the laboratory analyses of twigs. These values were then converted into percentage fibre and percentage protein per track. The total biomass eaten of each species was calculated for each season and treatment. To indicate hares' dietary preferences, I used estimates of availability for willow and birch on each site (Table 2.3) and the biomass eaten by hares to generate Manly's alpha values (Krebs 1989). Estimation of spruce clip biomass The biomass of spruce clips eaten could not be calculated directly, because hares ate entire clips, leaving no evidence of how many they had eaten. I therefore had to estimate the biomass eaten from the number of faecal pellets hares left at spruce clip sites. From the feeding trials, I calculated the amount hares ate per day divided by the faecal output per day, 0.204±0.006 66 g/faecal pellet. This value was consistent no matter what hares were eating (spruce clips, >3mm willow, <3mm willow, <3mm birch: one-way ANOVA on drywt/pellet by food-type, F 3 39=0.554, p=0.65). Multiplying this rate by the number of faecal pellets per site therefore estimates the biomass eaten per site. If a hare had eaten both spruce clips and another species at the same site, I calculated the weight of the other species eaten (from twig diameters) and subtracted that from the estimate of total biomass eaten at that site. Intake rate For each track, biomass was estimated from twig diameters and time was estimated from the number of faecal pellets. These values were then combined to estimate the hare's intake rate along each track. I present data for tracks with >50 pellets (-1.5 hr), since shorter tracks are more likely to be biased. Strip-barking of willow and spruce Hares occasionally ate the bark of willow and spruce. In March-May of each year—the exact time depended on snow melt—I ran five parallel transects 120 m apart per grid; every 60 m, I recorded any debarking on each of the nearest four willow and spruce within a 10 m radius. If there were fewer than four individuals within the radius, I did not search for other individuals to examine. I considered only spruce <80 mm DBH, since hares did not eat bark from larger trees. Stems did not need to be fully ringed to be counted. I measured the diameter (at the area of debarking) of each spruce tree debarked, and for willow I counted the number of stems debarked and the total number of stems. 67 Statistical analyses The data were analyzed separately for fall and winter, because through December the snow was still shallow and hares had access to birch, soapberry, and forbs. During January-April, these species were mostly buried in snow. In most cases, data from replicate grids of the same treatment were joined together prior to analysis. Statistical analyses were conducted using Statistica (StatSoft 1995), with a nominal significance level of 0.05. When the assumptions of ANOVA could not be met adequately, analyses were confirmed by general linear models (GLIM 1985), with binomial error distributions where appropriate. The qualitative results were identical using ANOVA or GLM models; for statistical simplicity, I present results from ANOVAs in this chapter. Interaction terms could not always be calculated because of the lack of data for food+fence in 1996, so just main effects were considered. Means were compared using post-hoc Tukey tests, corrected for unequal sample sizes. Means are presented with standard errors. RESULTS Browse weight, protein content, and fibre content The laboratory analyses of twigs generated curves relating twig diameter to twig weight, fibre content, and protein content. These relationships were then used to convert the twig diameters of what hares ate into biomass, protein, and fibre in their diets. Twigs of willow, birch, and soapberry had similar weights for each diameter <5 mm (Figure 3.1, Appendix 2), with 5 mm twigs of birch and willow both weighing 3.7 ± 0.2 g (dry weight). The fibre content of birch and willow increased with twig diameter (Figure 3.2, Appendix 2), but for each twig diameter, birch was more fibrous than willow; 8 mm willow twigs still had a lower percentage of fibre than 3 mm birch twigs. For both species, 1 mm twigs were more fibrous than 2 mm twigs, which might have 68 Figure 3.1. Dry weight of hares' main winter food species. Note different scales. A. Willow, y = 0.20 - 0.16x + 0.004x2 + 0.036x3. 95.7% of the variance was explained. Willow is fit with a cubic equation rather than a quadratic to avoid negative estimates of biomass. B . Spruce, y = 0.82 - 1.43x + 0.94x2 - 0.10x3. Variance explained, 77.2%. The cubic equation fits the data better than does the quadratic. C . Birch, y = 0.42 - 0.55x + 0.24x2. Variance explained, 91.1% D . Soapberry, y = 0.45 - 0.72x + 0.3lx2. Variance explained, 84.5%. 69 O O N C D W ^ C O C M T - O (6) }i|6!9M LO O LO O LO O CN CN t- T -(6) m6j9M o co CN cp "^r o c\i T — o o d (6) m6|9M co LO co CM . o (6) lL|6l9M 70 Figure 3.2. Fibre content of browse. The lines are regression and 95% confidence intervals. The relatively high values for twigs of diameter 1 mm may be due to buds. Note different x-axis scales. A. Willow, y = 59.5 + 1.78x; r2=0.45; F{ I58=130.3, p<0.001. B . Birch, y = 69.8 + 1.46x; r2=0.31; F, 106=48.3, p<0.001. 71 72 resulted from the ratio of bud weight to total twig weight. Soapberry twigs had similar percentages of fibre at all twig diameters <3 mm; the average fibre content for soapberry twigs was 57.0±0.6% (Appendix 2, but see Appendix 3). Protein content decreased with twig diameter for birch and willow (Figure 3.3, Appendix 2). Birch had a higher protein content than willow for each twig diameter; the maximum for birch was 6.4±0.3%, while the maximum for willow was 4.9±0.3%. Soapberry's protein content ranged from 9.6±0.7% to 11.5+0.2% (Appendix 2). Spruce clips had 54.6±0.5% fibre and 5.7±0.5% protein. In midsummer, lupine had 5.5±0.3% protein; bluebell had 7.4±0.5%; and fireweed had 11.0+0.7%. Fibre values were 55.0±0.7%, 58.8±0.6%, and 46.7±2.7%, respectively. By the fall, these species had significantly reduced their protein content and increased fibre content (Appendix 2). Their fibre values still remained lower than those for most of the twigs hares ate, but protein levels were comparable to browse protein levels. Thus of the woody browse species, spruce had the lowest fibre content, followed by soapberry, willow, and birch. For protein, willow had the lowest levels, spruce and birch had similar levels, and soapberry had the highest levels. Within a species, smaller twigs generally had higher protein and lower fibre than larger twigs did. Snowshoe hare diets: size and species Especially during the fall, hares ate forbs and some uncommon woody species, such as kinnikinnik (Arctostaphylos uva-ursi) and cinquefoil (Potentilla fruticosa) (Table 3.3). Hares on the control and food treatments ate more forbs and more of the other woody species than did hares on the fence or food+fence treatments. These species were not used in the calculation of 73 Figure 3.3. Protein content of browse. The lines are regression and 9 5 % confidence intervals. Note different x-axis scales. A. W i l l o w , y = 5.30 - 0.26x; r 2 =0.45; F 1 7 6 = 6 1 . 9 , p<0.001. B . B i r c h , y = 6.77 - 0.46x; r 2=0.46; F , 4 9 =42.0, p<0.001. 74 75 Table 3.3. Snowshoe hares' consumption of forbs and incidental woody species, n is the total number of browse points from which the percentages are derived. Forbs are primarily bluebells (Mertensia paniculata), fireweed (Epilobium angustifolium), lupine (Lupinus arcticus), grass (Festuca altaica), and northern sweet-vetch (Hedysarum boreale). Other includes kinnikinnik (Arctostaphylos uva-ursi), Labrador tea (Ledum groenlandicum), poplars and aspen (Populus spp.), and shrubby cinquefoil (PotentUlafruticosa). November-December January-Apri l n* % forbs % other n* % forbs % other Control 1992-1993* 233 0 0 1993-1994 298 9.40 0 242 3.72 0 1994-1995 100 32.10 0 795 6.42 1.64 1995-1996 204 23.04 12.25 452 0.22 0.66 Food 1992-19931 86 0 0 1993-1994 184 7.63 4.90 234 0.86 0 1994-1995 136 13.24 0 682 4.40 1.61 1995-1996 50 38.00 34.00 562 0 1.96 Fence 1992-19931 125 0 0 1993-1994 349 2.01 2.01 113 1.77 0 1994-1995 218 8.74 0 346 1.73 0.87 1995-1996 277 9.75 3.25 31 0 0 Food + Fence 1992-1993f 166 0 22.29 1993-1994 93 2.15 0 233 0 0 1994-1995 60 13.33 0 474 0.63 0 1995-1996 143 17.48 2.80 0 — — n was calculated by summing all browse points from each species. Spruce clip numbers were ' included in this summation, from the estimate of their weight • clips/gram. rNo data were collected in fall 1992, and all browse points for winter 1992-1993 are from after March 21. 76 hares' dietary characteristics, because the incidence of each species in the diet was low and it is difficult to determine the amount of forb biomass eaten from observation of the remnant stems. Willow and spruce were the primary foods of hares (Figure 3.4). Hares often ate birch instead of willow in the fall, especially on the fence treatment, but little birch was eaten in the winter. Hares ate trace amounts of soapberry, mostly on control areas in the fall. Through the three winters, hares' consumption of willow declined and spruce consumption increased. Dietary preferences could be calculated only for small (< 5 mm diameter) and large (> 5 mm) willow and birch twigs, since these were the only categories for which availability estimates existed (Table 2.3). Hares preferred small willow and small birch, usually avoided large willow, and avoided large birch entirely (Table 3.4). Hares' preference for small willow declined when birch was present, since small birch was highly preferred. During January-April, hares' preference for small birch disappeared, but birch was buried by snow at that time. Large willow was preferred more in winter than in fall. During both fall and winter 1994-1995, most of the spruce hares ate came in the form of spruce clips dropped by squirrels (Figure 3.5). During 1995-1996, squirrels dropped far fewer clips and hares ate almost all of their spruce directly from the trees. Hares on control and food sites ate a higher proportion of spruce clips in 1993-1994 than did hares on the fence or food+fence sites. Hares ate willow twigs of both juvenile and adult growth forms (Figure 3.6). During winter 1994-1995, most of the willow eaten by hares was of the juvenile growth form, but in the other winters hares ate mostly adult-form twigs. During the fall, hares on the food and 77 Figure 3.4. Species composition of snowshoe hares' diets. The proportions are calculated from biomass estimates. Spruce includes both twigs direct from the tree and squirrel clips. The x's indicate none of that species was eaten, and the O's represent the data missing from food+fence after that treatment was stopped. 78 13 control • food • fence M food+fence birch willow soapberry spruce 1.2 1 "S -° 0.8 o o 0.6 ti o 8" 0.4 1— D_ 0.2 0 fall 1994 X X X birch willow soapberry spruce 1.2 1 4— 0.8 o c o 0.6 o 0.4 CL o 0.2 0 fall 1995 X x x x birch willow soapberry spruce 79 1 T winter 1994 * 0.8 birch willow xx soapberry ffl control • food • fence Bf + f spruce 0.8 CD T J M— o c o ~n o o o. 0.2 0.6 0.4 0 winter 1995 birch willow x x soapberry spruce ~ 0.8 CD T J o 0.6 c o o 0.4 CL O " 0 . 2 0 winter 1996 xO birch willow xO soapberry spruce 80 X i -a .<L> s s A OH v—' uo <U 0) bo fe X ! o S C OH O •J3 -OH T 3 s ° o -uo <u b u n xi ^ ° s X ! S O <u 03 o3 II CN -a i - £ >• o X ! O PH 5 - d 03 "C <D OX) OA <u Cu > <U VH O H O i i x s rv J H X T OH * -3 uo X > o3 <H- O •B ° 03 [ T . X ^ C 0303 X uo T3 '3 > uo £ 03 X i H—» 03 X i O O uo *T3 C 03 =5 O X U c o U 0) >-.SH OH c 3 .5 >n II C N 2, o _ uo I E I E S v OH W >> ^ VH 03 s e (U uo -a c 03 X ^ 03 £3 - C O u Q "g . 03 a> "S3 2 | « 03 OH o c 03 '3 > 03 <U O c <u T J O <U O MM OH C8 ft! £j -2 2 •"•j X »o Mil ca ^ * -^ =2 =3 1 1 X i 941 —H O N r - v o c n v o VO c n VO c n C N o o c n c n c n ^ M O o o d d d d d d d d d O N o o C O o c n T t Tt C N T t v o q q o o u n O N O N C N VO v o v o o o d d d d d d d d d d o o o o o o O o o o o o c n C N o o o o o d o O o o d o o C N ON C N O O O C N d o o i i d d O o d d d o r - o o o o O o u n c n o o p p N O o o O p T t c - - p d i d d d d d o o o i i o o O o o o o o o O N o o o i T t m T t in O N O N ON O N O N O N ON O N >-1 1 >-1 c n T t c n T t O N O N O N O N O N O N O N O N ^ H ^ M c n c n "3 "3 "3 "3 •— •— •— c c c c o o o o U U U U O C N c n o d d o o T t u n T t u n O N O N ON ON O N O N O N ON —" M"H c n T t c n T t O N O N O N O N O N O N O N O N ^ M ^ H ^ M ^ M — - H C N C N -d T 3 -d -o o O o o o O o o PH PH PH PH r - o 00 T t d d T t u n O N O N ON O N I i c n T t O N O N — ~ O N O N <U O C PH PH O O T t u n O N O N O N O N I i c n T t O N O N O N O N u o c c <u u PH PH + + o o o o PH PH 8 1 Figure 3.5. The proportion of spruce eaten as spruce clips. The proportions are based on biomass estimates. Missing values on the figure are due to periods when hares ate no spruce either from trees or from squirrel clips. Open symbols are for November-December and filled symbols are for January-April. 82 Figure 3.6. W i l l o w eaten by hares: the proportion o f w i l l o w twigs o f juven i le growth form. The values are proportions of tw ig numbers. Open symbols are for November -December and f i l l ed symbols are for January-Apr i l . 84 0) = c o o & H—» c o CD CD *-> » C CO C • > CO »»- > >*- > CD CD •a TJ O O o O c c o O CD CD it— 6 * + * o CD O C CD + + CD TJ TJ -~ O = O C l»— »4— 4— > LUJOJ L|yv\oj6 aiiuaAnf jo M O I H M J O uorjjodojd 85 food+fence treatments tended to eat a higher proportion of juvenile form twigs than did hares on the unfed treatments, but this pattern was non-existent in winter. The mean diameter of twigs eaten by hares was greatest for spruce, then willow, then birch and soapberry (Figure 3.7, Appendix 4). In the fall, hares on control areas had the smallest mean diameter for willow twigs, and hares on the food+fence treatment had the largest (two-way ANOVA, treatment F3680=14.3, p<0.001; year F2680=7.04, p<0.001, interaction F6680=3.68, p<0.01). The mean diameter of willow twigs eaten was larger in fall 1993 than in the other falls. In winter, hares on the food+fence treatment continued to have a larger mean diameter of twigs than hares on other treatments, but the mean of willow browsed by hares on control areas was no longer distinguishable from those on the fence or food treatments (two-way ANOVA, treatment F32201=15.82, p<0.001; year F32201=18.33, p<0.001). Willow browse diameter was greatest in winter 1995-1996, followed by 1993-1994; the other two winters were lower and indistinguishable from each other. The mean diameter of browsed birch or spruce did not appear to differ among treatments or years (Figure 3.7b, d). Hares on both the food and food+fence treatments sometimes ate bark from spruce and willow, but debarking on the other treatments was negligible to non-existent (Table 3.5). Willow bushes were more likely to be debarked than were spruce trees. There was no significant difference between the food and food+fence treatments in the size of the spruce trees debarked (t-test, t= -0.328, d.f.=48, p=0.74); mean tree diameter was 43.2±2.4 mm (range 15-75 mm). Willow bushes with debarking damage had 43.0±2.6% of their stems debarked (range 5-100%). 86 Figure 3.7. Mean diameter at point of browse for hares' main browse species. Values for all species are in Appendix 4; here I only display major species for which >5 twigs were browsed. Missing symbols therefore indicate little of that species was browsed. Note different values on axes. A. Willow in November-December, n = 692. B . Birch in November-December, n = 649. C. Willow in January-April. n = 2216. D . Spruce in January-April, n = 824. 87 c o o TD O O CD O c 0 0 o c + TD O O X < < m •x TT CN CN CD LO tD CD CD CD CD 4 & CD CD CO ^ " CD CD CJ) 0 5 CN (LULU) J9J9WBJP 9SM0jq < • 0 O D I— O-co 0 +H c • X • CD Tt CD CO CD CN CD (LULU) j9;9Lue;p 9SM0jq < X D < a • CN 00 CO CN T T CN CN CD CN LO CD CD CD CD 0 0 CD ° > CO ^ " CD CD CD 0 5 (LULU) J919LUBJP 9SM0jq 15 o 0 O X « < • X D < • X D LO CO CD g ) CD CD Tf J O CD CD CO ^ CD CD CD CD CN £ CD CD CD CD I 1 1 1 1— LO LO CO LO CN TT' CO CN (LULU) J919LUBJP 9SM0jq 88 Table 3.5. Hares' consumption of willow and spruce bark, n is number of plants examined, barked is number of plants with debarking damage, and % is the percentage of plants with debarking damage. Values represent the cumulative total for the winter ending in the year given. Spruce Willow n barked % n barked % control 1993 584 0 0 597 1 0.2 1994 400 0 0 392 0 0 1995 400 0 0 400 0 0 1996 400 0 0 400 2 0.5 food 1993 761 31 4.1 663 37 5.6 1994 380 0 0 317 0 0 1995 384 0 0 317 0 0 1996 400 2 0.5 374 18 4.8 fence 1993 168 0 0 196 0 0 1994 168 0 0 196 0 0 1995 164 0 0 196 0 0 1996 200 0 0 200 0 0 food + fence 1993 318 21 6.6 400 58 14.5 1994 116 0 0 200 0 0 1995 129 0 0 200 0 0 1996 180 0 0 200 0 0 89 Snowshoe hare fibre and protein intake In the fall, hares on control areas had the least fibre in their diets, and hares on the fence treatment had the most (Figure 3.8a, Appendix 5; two-way ANOVA treatment F373=22.34, p<0.001; year F273=3.46, p<0.05). Hares on control and food sites ate diets with a similar fibre content in fall 1995, unlike the other falls; hares on all treatments except food had diets with higher fibre content in fall 1995 than in fall 1994. In the winter, hares on control areas had the lowest fibre content in their diets in the last two winters only (Figure 3.8b), and the food+fence hares had the highest fibre content through-out. The only statistically significant treatment difference was between the control and the food+fence treatments (two-way ANOVA treatment F3130=6.28, p<0.005; year F3130=7.37, p<0.001). During the last two winters, hares had lower fibre levels in their diets than they had during the first two winters on all treatments except fence. In the fall, hares on control and fenced sites ate twigs with higher protein content than did hares on the food addition and food+fence treatments (Figure 3.9, Appendix 5; two-way ANOVA treatment F373=9.07, p<0.001; year N.S.). Although hares on the fence and control treatments did not differ from each other statistically in their protein intake, hares on control sites had higher protein levels during two of three falls. In winter, the only statistically significant treatment difference was that hares on control sites had higher protein intake than did hares on the food+fence treatment (two-way ANOVA treatment F3130=5.71, p<0.001; year F3130=4.48, p<0.005). Statistically, hares had more protein in their diets during the last two winters, but most of this difference is due to a large increase in the protein intake of control hares. 90 Figure 3.8. Dietary fibre of snowshoe hares. Values are means of tracks ± SE. A. November-December. Data from 85 tracks. B . January-April. Data from 145 tracks. 91 75 A. November-December 92 Figure 3.9. Dietary protein of snowshoe hares. Values are means of tracks ± SE. A. November-December. Data from 85 tracks. B. January-April. Data from 145 tracks. 93 A. November-December Faecal pellet deposition rate as an index of time; hares' intake rates When hares were active (in winter, -1700-0800 hr), hares dropped 35.2±2.2 faecal pellets/hr (Figure 3.10). During the inactive hours, they dropped 5.0±1.8 pellets/hr. There was no effect of food type on number of pellets dropped in a day (one-way ANOVA, F 3 38=0.06, p=0.98). The wild-caught Kluane snowshoe hares excreted 579.3±15.5 pellets/day. This value was significantly higher than the excretion rate of the Vancouver hares, which was 456.0±19.1 pellets/day (t-test, t=5.053, d.f.=66, p<0.001), but Vancouver hares demonstrated the same daily rhythm of excretion. To convert number of pellets along hare tracks into time, I used the value from Kluane hares while they were active. Hare intake rates were variable (Figure 3.11). In the fall, the only significant treatment effect showed that hares protected from predators had a higher intake rate than did hares on control sites (two-way ANOVA, treatment: F 3 22=3.39, p<0.05; year N.S.). In the winter, no treatment effects were significant, but during the winter of 1994-1995 hares on all treatments except fence had lower intake rates than in the other winters (two-way ANOVA, treatment N.S., year F2.,4=8.28, p<0.001). Faecal fibre Faecal fibre increased with dietary fibre intake (Figure 3.12a; r2=0.63; F! 21=35.2, p<0.001) and decreased with dietary protein intake (Figure 3.12b; r2=0.51; F121=21.5, p<0.001). Faecal fibre can thus be used as an indicator of the nutritional status of wild hares. During January-April, hares on the food+fence treatment had the lowest levels of faecal fibre (Figure 3.13a; two-way ANOVA, treatment F 3 708= 3.65, p<0.001; year F3708=63.94, p<0.001; interaction F9708=3.32, p<0.001). Faecal fibre from hares on control sites did not differ 95 Figure 3.10. Faecal pellet deposition rate of snowshoe hares. Values are mean numbers ± SEM. The Kluane line is based on four wild-caught hares. From 1100-1900 hrs, hares were checked once every three hours, and pellet numbers divided by three. The mean faecal pellet output per day was significantly lower for Vancouver hares; the dotted line of their hourly mean rate is shown for comparison but without error bars (data from 15 hare-days on eight hares). 96 CO > o o o o o o o o o co CD m co CN jnoL) jed si9||9d |B08B; 97 Figure 3.11. Intake rates of snowshoe hares. Values are track means ± S E M , grams dry weight/hour. Only tracks representing more than -1.5 hr (>50 faecal pellets) were included. 98 0 + -+ 1993-1994 1994-1995 1995-1996 9 9 Figure 3.12. The relationship o f faecal fibre to dietary protein and dietary fibre. The lines are regression and 9 5 % confidence intervals. A. Die tary fibre, y = 44 + 0.55x; r 2 =0.63; F[ 2l=35.'. p<0.001. B . Dietary protein, y = 88 - 1.05x; r 2 =0.51; F , i 2 ,=21 .5 , p<0.001. 100 Figure 3.13. Faecal fibre of snowshoe hares. The values are means (% of dry weight) ± SEM. A. January to April, n = 724. B . May-August, n = 672. C. September-October, n = 492. D . November-December, n = 176. 102 85 A. January-April 103 85 „ , 80 -T J CD 75 -it— : aecal 70 65 C. September-October ii TFT- -•— control •a - food • * - -fence A ~ - food+fence 1993 1994 1995 65 -I 1 1 : 1 1993 1994 1995 104 statistically from faecal fibre from hares on the fence or food treatments, but hares on the fence treatment had higher levels of faecal fibre than hares on the food treatment in three of four years. Faecal fibre differed in all years, with fibre lower in the first two winters than in the last two. During the summer, hares on the food+fence treatment again had the lowest level of faecal fibre in two years (Figure 3.13b), but their faecal fibre levels were not significantly different than those from hares on the fence treatment. On all treatments, faecal fibre levels were lower in 1994 than in the other summers (two-way ANOVA, treatment: F3660=4.40, p<0.005; year F2660=22.41, p<0.001; interaction F6660=4.60, p<0.001). Faecal fibre levels were at their highest in September-October (Figure 3.13c), but there were no treatment or annual differences. Faecal fibre levels were also high in November-December (Figure 3.13d), but fibre values were different in all years and the fenced hares had higher faecal fibre levels than hares on any other treatment; additionally, hares on control sites had higher levels of faecal fibre than hares on food addition sites (two-way ANOVA, treatment F3164=23.55, p<0.001; year F 2 1 6 4 = 142.0, p<0.001; interaction F6164=3.66, p<0.005). In general (see also Appendix 6), faecal fibre levels were highest in September-October and lowest in May-August. There was least treatment and annual variation in September-October. Hares on the food+fence and food treatments tended to have lower faecal fibre levels than hares on the control or fence treatments, and during the months with snow cover (November-December and January-April), fenced hares had higher faecal fibre than hares on control areas. 105 DISCUSSION Hares' diets during the low phase Hares' main foods when snow was on the ground were spruce and willow. Hares ate birch in the fall, but birch was only a minor component of the winter diet. Hares also ate forbs and other shrubs in the fall. Hares on the various treatments did not differ much in the amount of spruce they ate in winter. Snowshoe hares ate many spruce clips; this pattern was unexpected since there was no way to predict that red squirrels or hares would forage in this manner. It is difficult to tell whether hares would have consumed as much spruce had squirrels not dropped spruce clips, or whether the spruce clips encourage hares to consume more spruce. During the decline phase, hares on control areas had more spruce browse sites in each successive year (1990-1992), whereas hares on food+fence had fewer spruce sites in each year of decline (Hik 1994). Although the methodologies differ between the two studies, hares on these two treatments were more similar to each other in their spruce consumption during the low phase than in the decline phase. The differences in hares' spruce consumption between these two phases could be linked to spruce clips, anti-predator behaviour, or availability of other foods. The mean diameter of twigs that hares ate was largest on food+fence and smallest on control areas, but mean twig diameters on all treatments were lower than the 5 mm cut-off which has been used as a proxy measure of nutritional stress (Sinclair et al. 1988b, Smith et al. 1988). The values I observed for control hares were similar to those obtained during a previous low phase at Kluane (Smith et al. 1988). The variation among treatments indicates that either the distribution of twigs available differed among treatments, or that hares were selecting different sizes of twigs on the treatments. Because willow and spruce availabilities were broadly similar 106 among treatments (Table 2.3), and plant biomass far exceeded hares' needs (Figure 2.7), it is more likely that the diameter differences represent a behavioural choice rather than a limitation imposed by the landscape. In terms of nutrition, both species composition and twig size mattered. Birch was high in both protein and fibre; when hares ate a lot of birch, as they did on the fence treatment in fall, their dietary fibre was high, their dietary protein was low, and their faecal fibre high. Spruce was high in protein (although not so high as birch) and low in fibre, and thus control hares that ate spruce had low dietary fibre and high dietary protein. This pattern makes it difficult to assess dietary quality on the basis of species composition, since none of the food species at Kluane was either superior (lowest fibre, highest protein) or inferior (highest fibre, lowest protein). Secondary compounds would also complicate nutritional assessment, since these compounds can interfere with digestion. Hik (1994) used the amount of spruce in the diet as a proxy for nutritional quality of the diet, but the data presented here do not support that determination. Dietary protein estimates ranged from 4.0-6.8% of the dry matter intake of hares. These values are far below the 9.9% that Sinclair et al. (1982) estimated hares need to maintain their weight. The weight and body condition of hares at this phase of the cycle (Figure 2.9, K.E. Hodges, E.A. Gillis and C.I. Stefan ms. in prep) suggest that hares are not suffering from malnutrition, so one or both estimates must be in error. The 9.9% estimate is based on 150 g of dry matter intake daily, but this intake amount has not been verified from field data. The protein values I have calculated may be underestimates, given the forbs and other shrubs which hares ate. Additionally, hares forage preferentially on nitrogen-fertilized areas (Nams et al. 1996), and the twigs hares ate may have had more than the average protein content for their size. 107 Hares on the food and food+fence treatments ate more juvenile-form willow than hares on the fence and control treatments. These treatments had higher hare densities at the peak of the cycle (Boutin et al. 1995), resulting in a higher proportion of the adult-form willow twigs being eaten; it takes several years of regrowth before adult-form twig numbers rebound, but juvenile-form twigs regrow quickly (Smith et al. 1988, Sinclair et al. 1988b). It is possible, therefore, that hares on the food and food+fence treatments responded to twig availability rather than preferring juvenile-form twigs. It has been postulated that juvenile-form twigs are higher in secondary compounds, and that these regrowth patterns could force hares to eat poor diets during the low phase (Bryant and Kuropat 1980, Bryant 1981a, Fox and Bryant 1984), but that seems unlikely. During the low phase hares ate many species of plant, which are differentially chemically defended, and on control sites there was only one winter in which over half the willow twigs in hares' diet were of juvenile growth form. How did hares choose their diets? I premised this chapter on the argument that hares might select their diets on the basis of food quality, predation risk, or both. If hares selected their diets on the basis of food quality, then hares given supplemental food should have been more selective in their diets and thus should have obtained nutritionally better diets. Similarly, if hares sought to avoid predators, hares protected from predators should have had higher selectivity and better diets. In the trade-off situation, hares with supplemental food should choose poorer foods to gain extra time in cover, hares protected from predators should be more selective of their diets, and hares given food and protected from predators should be able to obtain the best diets. The data do not support any of these hypotheses unequivocally (Table 3.6). 108 Table 3.6. Summary of snowshoe hare diets during the cyclic population low. FF = food+fence treatment; C = control. Commas indicate no difference between treatments; < indicates a treatment is statistically lower than another one for that parameter. November-December January-April Browse diameter Debarking* Dietary fibre Dietary protein Intake rate C < Food, Fence < FF C < Food, FF < Fence Food, FF < Fence, C C < Fence C, Food, Fence < FF C, Fence < Food, FF C < F F F F < C no difference Faecal fibre Food < C < Fence FF < Fence FF < Food < Fence F F < C *Debarking technically refers to the entire six month period, but most debarking occurred in March and April. 109 Twig diameter, dietary fibre, and dietary protein showed differences between the diets of hares on control sites and hares on the food+fence site, with hares on the fence and the food sites usually falling somewhere in between. This pattern suggests that food and predators interact in their effects on hares' diets. The dietary patterns do not, however, support an interpretation of a food-predation risk trade-off, since the predicted patterns did not occur on the treatments where the single factors were manipulated. Additionally, the distribution of food species may account for many of the observed dietary differences; birch had an especially strong effect on hares' nutritional intake, and birch was unevenly distributed among the treatments (cf. Table 4.1). Food availability was critical to hares' diets. When snow was deep, hares ate spruce and willow; when it was shallow, they also ate birch, soapberry, and forbs (see also Keith 1983, Smith et al. 1988). Hares ate more birch and soapberry in areas where these species were abundant (see also Smith et al. 1988). The spruce clips again demonstrate the role of opportunity in hares' diets, as does the fact that hares are generalist herbivores that range across a continent, eating a variety of species (de Vos 1964, Bryant and Kuropat 1980, Rogowitz 1988). Density might also be related to foraging behaviour; the densities during the last two years were higher than during the first two (Figure 2.2), and faecal fibre and to a lesser extent dietary fibre and dietary protein showed more of a difference between the first two and the last two winters than within each pair of winters. Additionally, although food per hare was always higher than the amount of food needed per hare, the distribution of food and other hares across the habitats may have affected how hares foraged. This argument is similar to the availability argument, but at the much finer scales of hares' home range sizes and nightly foraging paths. 110 Methodological ramifications Estimates of biomass of each species eaten, dietary protein, and dietary fibre are limited by the accuracy of tracking data, and the strength of the relationships between twig size, protein and fibre. Despite these biases, these indices are likely to be better estimators of hares' diets than the more common methods of counting number of twigs eaten per species or number of sites at which a species was eaten (Bookhout 1965, Telfer 1972, Pease etal. 1979, Fox and Bryant 1984, Rogowitz 1988, Smith etal. 1988, Sinclair etal. 1988b, Hik 1994). Twig weight and nutritional composition are heavily dependent on twig diameter, hares eat twigs of different sizes, and the various species do not have the same relationships between biomass, nutritional composition, and twig diameter (Figure 3.7, Appendices 2, 4; see also Smith et al. 1988a). Determination of dietary preferences depends on the accuracy of the availability estimates. That is difficult methodologically, and for hares almost anything can become food when they are food-stressed (Sinclair et al. 1982). Previous field and cafeteria trials support the results that hares prefer birch over willow, and small twigs over large (Sinclair and Smith 1984b, Smith et al. 1988). Also, since this methodology was applied uniformly to all treatments, relative differences in preferences should be detectable even if it is impossible to derive absolute preference values. Faecal fibre obviates the problems of tracking and relying on regressions between twig diameter and twig nutritional content, but faecal fibre could be a biased index of hare nutrition if hares ate large amounts of species not considered in the calibration of faecal fibre and dietary protein and fibre. At Kluane, this potential bias is minimal, since hares ate almost exclusively the species I used in the calibration. Previous research has used faecal protein as an index of hares' dietary quality to avoid the problems of tracking data (Sinclair et al. 1982, Sinclair and Smith 111 1984a), but this application is controversial because faecal protein may reflect dietary protein poorly since faecal protein can contain endogenous protein and may also be affected by the presence of secondary compounds in the diet (Pehrson 1984, Risenhoover et al. 1985, Sinclair et al. 1988b). Faecal fibre is not affected by secondary compounds, but is reflective of both dietary protein and dietary fibre, thus making it a more reliable dietary index than faecal protein (Hobbs 1987, Wofford etal. 1985). Can diets during the low phase create population consequences? If hares had poor diets during the low phase, that could result in prolongation of low densities through condition-induced changes in fecundity or through changes in survival or predation rates (Boonstra and Singleton 1993, Hik 1994, 1995, Krebs 1996, Boonstra etal. 1998a, 1998b). Poor diets have been posited to result from the delayed regrowth of food plants, increases in their secondary compound content, or behavioural switches as hares avoid predators (Bryant 1981a, Keith 1990, Hik 1995). The data presented here support neither the supposition that hares had poor diets during this phase nor any of the hypotheses relating food and predation to diets. Although availability of particular foods influences hares' diets, general food availability was high and there was little indication that hares were nutritionally stressed (Figure 2.9, K.E. Hodges, E.A. Gillis, and C.I. Stefan, ms. in prep). Plant secondary compounds do not seem able to create poor diets at this phase. Predation risk similarly did not create substantive differences in hares' diets or nutritional intake. Although these results do not directly test for any linkages between diet and condition, fecundity, or survival, they do show clearly that it is difficult to justify looking for poor diets as an explanation for a prolonged low phase. 112 C H A P T E R 4 H A B I T A T USE B Y SNOWSHOE H A R E S DURING T H E L O W PHASE: A T E S T O F P R E D A T O R SENSITIVITY A N D T H E R E F U G I U M HYPOTHESIS INTRODUCTION During the ten-year snowshoe hare (Lepus americanus) cycle, hares appear to shift their use of habitats. During the decline and low phases, they are thought to use 'refugia': areas of dense cover which offer more protection from predators than do more open areas (Keith 1966, Wolff 1980, 1981, Hik 1994). As densities increase, hares reputedly disperse back into more open areas, until at peak densities hares use all habitat types (Wolff 1980, 1981). This pattern of habitat use may contribute to the dynamics of the cycle, since survival and perhaps fecundity are linked to habitat type (Buehler and Keith 1982, Hik 1994, Rohner and Krebs 1996, Villafuerte et al. 1997). The habitats that offer better protection from predators may be poorer in food (Wolff 1980, 1981, Hik 1994, 1995), which could lead to malnutrition, reduced fecundity, and reduced juvenile survival (Hik 1994, 1995, Boonstra etal. 1998a, 1998b). Habitat shifts have been used to explain the two to four year duration of the cyclic low phase (Wolff 1980, 1981, Keith 1990). The low phase persists longer than one might expect, given that food availability is high, survival good, and predators scarce (chapter 2, Krebs et al. 1995, Boonstra et al. 1998b). In addition to the direct effect of differential mortality in the different habitats, hares' habitat choices can lead to deleterious stress effects if they are chronically exposed to predators (Boonstra and Singleton 1993, Boonstra et al. 1998a). Hares in areas protected from predators but poor in food might suffer nutritional stress. Maternal stress from 113 either source might influence the survival of their offspring via maternal effects or a reduced ability of mothers to care for their offspring (Boonstra et al. 1998b). If there is behavioural structure in the population such that juveniles are relegated to poorer habitats (Dolbeer and Clark 1975, Boutin 1984a, Graf 1985, Graf and Sinclair 1987), that could additionally decrease juvenile survival. Since small changes in juvenile survival can have large impacts on hare population dynamics (E.A. Gillis, D. Haydon, C.I. Stefan, and C.J. Krebs, ms. in prep), the mortality consequences of hares' habitat choices might suffice to keep the population densities low for several years. These cyclic habitat shifts could occur if hares moved into particular habitats to avoid predators (Wolff 1980, 1981, Hik 1994, 1995), but differential survival in the varying habitats could also contribute to the habitat contraction observed during the decline phase (Keith 1966). Although numerous studies have documented that hares prefer dense habitats, with dense understory more important than dense overstory (Buehler and Keith 1982, Orr and Dodds 1982, Wolfe et al. 1982, Pietz and Tester 1983, Keith et al. 1984, Litvaitis et al. 1985a, 1985b, Monthey 1986, Rogowitz 1988, Scott and Yahner 1989, Litvaitis 1990), relatively few studies have addressed predators' habitat use at the same time (Litvaitis et al. 1986, Koehler 1990, Murray et al. 1994, O'Donoghue 1997). Proportionally more hare deaths occur in open habitats than in closed habitats (Murray et al. 1994, Hik 1994, Rohner and Krebs 1996), and survival is apparently higher in dense habitats (Sievert and Keith 1985, Litvaitis et al. 1985b); many authors therefore assume that hares select dense habitats for protection from predators, but this idea has seldom been tested experimentally (but see Hik 1994). 114 Predator-induced habitat selection has, however, been shown for other small herbivores (Lima and Dill 1990). Small rodents usually increase their use of bushy habitats when predators are present or when increased illumination (i.e moonlight) increases predation risk (Kotler 1984a, 1984b,Brown 1988, Brown etal. 1988, Dickman et al. 1991, Kotler et al. 1991, 1992, Longland and Price 1991, Kotler 1992, Lagos et al. 1995). Differential attack rates or survival can occur in the different microhabitats (Kotler et al. 1991, Longland and Price 1991, Dickman et al. 1991, Dickman 1992). My aims in this chapter are two-fold: first, to determine whether predators and food supply influence hares' habitat selection; and second, to determine whether hares use 'refugia' during the low phase of the cycle. To answer these questions, I address both hares' use of habitats and their selection of habitats. I use three habitat indices: the amount of open ground at a location ('open' versus 'closed' habitat), the degree of association between individual plants ('light' versus 'dense' clusters), and the amount of cover immediately around hares ('exposed' versus 'thick' cover). I studied snowshoe hares' habitat choices during the low and early increase phases of a cycle; food and predation risk were manipulated in a factorial design, allowing me to discriminate their effects on hares' habitat use patterns. If the refugium hypothesis is correct, the following predictions should be upheld: 1) hares exposed to predators should select closed habitats with dense species clusters and stay closer to thick cover than do hares protected from predators; 2) relative to habitat availability, more predation should occur in open than in closed habitats; 3) closed habitats should have less food than open habitats; and 4) as hare densities increase, hares should use more open habitats and areas with lighter species clusters. 115 M E T H O D S AND M A T E R I A L S Study sites, treatments, and habitat availability I assessed snowshoe hare habitat selection from May 1993 to April 1996, encompassing the low and early increase phases of a ten-year cycle (Figure 2.2). I used study sites established and maintained by the Kluane Boreal Forest Ecosystem Project, in the Shakwak Trench, southwestern Yukon (Krebs et al. 1995). The study sites were located in white spruce (Picea glauca) forest; aspen (Populus tremuloides) and balsam poplar (P. balsamifera) occurred sporadically, sometimes achieving locally dense populations. Feltleaf willow (Salix glauca), greyleaf willow (S. alaxensis), a short (<5 m high) tree willow (S. bebbiana), bog birch (Betula glandulosa), and soapberry (Shepherdia canadensis) were the predominant understory shrubs. Creeping juniper (Juniperus horizontalis), shrubby cinquefoil (Potentilla fruticosa), and Labrador tea (Ledum groenlandicum) occurred rarely, and seldom achieved sizes sufficient to act as cover for snowshoe hares. Most of the study area was forested, but there were occasional marshes and swamps which had dense willow and birch clusters with few or no trees. Four experimental manipulations were established by the Project during the previous cyclic increase (Krebs et al. 1995). Two 34 ha sites (food) were provisioned year-round with ad lib. commercial rabbit chow spread along four 570 m cutlines that were 180 m apart. One 1 km2 area (fence) was enclosed in a chicken wire and electric fence, which excluded terrestrial predators (lynx, Lynx canadensis and coyote, Canis latrans). Within the fence, 10 ha were covered with monofilament strung through the trees to deter great horned owls (Bubo virginianus), goshawks (Accipiter gracilis), and Harlan's hawks (Buteo jamaicensis). One site (food+fence) combined the fence and food manipulations, but had no monofilament. I used two 116 34 ha control sites. All sites were at least 1 km away from each other (Figure 2.1). In keeping with the nomenclature of the Project, food 1 refers to Gravel Pit; food 2, Agnes; control 1, Sulphur; and control 3, Chitty. Each treatment grid was divided into a 30 m by 30 m checkerboard and marked by 400 permanent grid stakes. I used these stakes as locations around which to estimate habitat availability. To assess the relative amount of food available within open and closed habitats, I determined the number of locations of each habitat type which had dense clusters of food species within them. Snowshoe hares and their locations Hares were trapped for population estimates in March and October, and sporadically at other times to maintain the number of radio-collared hares. At each experimental grid, 86 Tomahawk live traps (Tomahawk Live Trap Co., Tomahawk WI) were located in four lines, with 30-60 m between each trap and 180 m between trap lines. Traps were baited with alfalfa; on food addition grids, rabbit chow was also used. Hares were eartagged with number 3 monel eartags (National Band and Tag Co., Newport, KY). For each captured hare, we recorded eartag, location, weight, length of right hind foot, sex, and reproductive status. Some hares were fitted with 40 g radio-collars (Lotek, Newmarket ONT) which were equipped with mortality censors that doubled the pulse rate when the collar had not moved for four hours. During the snow-free period of each year (-May-September), I used radio-telemetry to locate snowshoe hares. I and my assistants used handheld receivers and antennae (Telonics, Mesa AZ) and we located snowshoe hares visually whenever possible, by approaching slowly and quietly. Snowshoe hares became habituated to our presence, and moved away from us rarely (for 117 a full discussion of the effect of observers on snowshoe hare behaviour, see chapter 5). Whenever we located a snowshoe hare, we recorded the habitat characteristics for it. When snow was on the ground, I recorded habitat information for each browse site along snowshoe hare tracks (chapter 3). Habitat classification: open ground, species clusters, and immediate cover In a circle of 15 m radius around each gridstake and hare location, I visually assessed the amount of ground free of trees, shrubs, and deadfall. I described the amount in 10% increments (a grass meadow is 90-100% open while a willow swamp is 0-10% open). If the amount of open ground bordered two measurements, I designated it as the less open one. Also in a circle of 15 m radius, I separately considered the degree of clustering for each of the following: birch, willow, other shrubs, Populus species, mature spruce, spruce shrub-seedlings, and spruce deadfall. 'Other species' includes soapberry, cinquefoil, and tree willow; the value given for 'others' at a location was the highest that occurred for any of those three species at that location. 'Shrub-seedling' is a size category rather than an age category, and refers to spruce trees under ~2 m tall and/or with stem diameter (dbh) of <7 cm. For each species, the degree of clustering was assigned as: 0, absence; 1, a few isolated individuals; 2, some individuals, most within 0-2 meters of another of the same kind; 3, densely associated individuals; 4, impassable thickets of associated individuals. Since 3 and 4 were similar categories, and were often difficult to tell apart in the field, they are lumped for the analyses in this chapter; I hereafter refer to them simply as 'dense' clusters. Immediate cover was described within a 5 m radius of each gridstake or hare location. I recorded the identity and distance to the nearest cover item in each of four directions. Cover was 118 counted only if it was large enough to obscure a hare, such as a tree, large bush, or deadfall; small items like single willow stems and grass were not counted. If no cover item was encountered within five meters, I described it as 'no cover'. I define 'thick cover' to mean that in all four directions hares were < 1 m away from a cover item, and 'exposed cover' to mean that in >2 directions, the nearest cover item was >3 m away. All other combinations were considered to be 'intermediate cover'. Statistical analysis I present two main sorts of data: I describe habitat availability and hares' use of habitat, and I address hares' preferences for certain types of habitat. I use Kolmogorov-Smirnov statistics to compare the annual distribution of hares' locations in habitats with different amounts of open ground. Statistical analyses were conducted in Statistica (StatSoft 1995). I used Manly's alpha (Krebs 1989) to describe hares' preferences for open ground, dense spruce, willow, birch, and deadfall clusters. Manly's alpha statistic is calculated as: = 1 n. 2(r. In) where a{ is the preference for habitat i, r, and r- are the proportion of habitat i and j used by hares (/ and j = 1,2, 3...m), n, and n- are the proportion of habitat i and j in the environment, and m is the number of habitat types. An a{ greater than 1/m indicates preference, and an a{ less than 1/m indicates avoidance. Manly's alphas were calculated for each experimental trapping grid separately, with availability based on the 400 gridstake sample per grid and hares' habitat use based on telemetry locations in the summer and browse sites along tracks in winter. 119 RESULTS Habitat availability The treatment grids differed from each other substantially in their habitat characteristics (Table 4.1). The understory characteristics broadly separated the six grids into two groups. Food 1, fence, and food+fence had more locations with dense birch and willow than did the control sites or food 2. Both control sites and food 2 had scattered or no birch around >93% of their gridstakes, whereas over half of the gridstakes on food 1 and food+fence had dense birch. A similar pattern occurred for willow: both controls and food 2 had no or scattered willow at 68-94% of the gridstakes, whereas food 1, food+fence, and fence had dense willow at 70-83% of gridstakes. Soapberry was the most common 'other shrub', and on control 3 it occurred at 85% of the gridstakes, but here and elsewhere 'other shrubs' seldom formed dense clusters. In general, food 1, food+fence, and fence had more shrub cover than did control 1, control 3, and food 2. Spruce provided most of the tree cover; aspen and poplar were never prevalent on any treatment grid, with 75-99% of gridstakes having no Populus whatsoever. Control 3 and food 2 had dense clusters of spruce at 55 and 67% of gridstakes, respectively; these values are much higher than occurred on the other grids. The same pattern was true for spruce shrub/seedlings. The fence grid had the highest occurrence of dense deadfall (18.5%), and control 1 had the most gridstakes without any deadfall (47%). Control 3, food 1, and the fence grid had -30% of their gridstakes in closed habitat (<40% open ground), while the other treatment areas each had only -13% of their gridstakes in closed habitats. The greatest amount of open habitat (>70% open ground) occurred on control 1 and food+fence, with 21.25% and 11.75%, respectively. 120 Table 4.1. Habitat characteristics of the treatment areas. The values are the percentage of locations on each trapping grid that had the given species ranking; circles of 15 m radii were surveyed around 400 gridstakes per grid. Species were ranked on a 0-4 scale for how they were distributed: 0, no individuals present; 1, few, scattered individuals; 2, some individuals present, most within 2 m of each other; 3, fairly dense clumps; 4, tight, almost impassable clumps. Other shrubs includes cinquefoil (Potentillafruticosa), soapberry (Shepherdia canadensis), and a short tree willow (Salix bebbiana). Spruce shrub/seedlings includes spruce seedlings and young trees up to ~2 m tall and smaller than ~7 cm DBH. Willow includes S. alaxensis and S. glauca. Control 1 Control 3 Food 1 Food 2 Fence Food+Fence B I R C H 0 - absent 80.00 78.50 12.75 74.00 30.00 8.50 1 - scattered 15.00 18.75 31.00 19.25 30.50 32.25 2 - clumped 5.00 2.75 44.75 6.75 30.00 50.50 3 - dense 0 0 10.25 0 9.50 8.25 4 - very dense 0 0 1.25 0 0 0.25 W I L L O W 0 - absent 0 2.50 0 36.75 1.75 0 1 - scattered 68.75 91.00 30.00 51.75 15.75 17.50 2 - clumped 29.75 6.25 67.75 10.50 71.75 72.75 3 - dense 1.50 0.25 2.00 1.00 8.50 9.50 4 - very dense 0 0 0.25 0 2.25 0 O T H E R SHRUBS 0 - absent 51.75 4.00 25.50 41.00 71.00 81.00 1 - scattered 38.25 50.00 56.25 47.25 16.50 14.00 2 - clumped 9.25 38.75 14.00 9.50 10.00 4.25 3 - dense 0.75 7.00 4.25 1.50 2.25 0.50 4 - very dense 0 0.25 0 0.75 0.25 0 ASPEN/POPLAR 0 - absent 99.25 75.00 83.75 85.25 90.00 98.75 1 - scattered 0.25 12.25 6.50 6.00 5.75 0.50 2 - clumped 0.50 6.75 5.75 3.75 1.75 0.50 3 - dense 0 4.00 1.75 2.50 1.75 0 4 - very dense 0 2.00 2.25 2.50 0.75 0 121 Table 4.1 continued. Habitat characteristics of the treatment areas. Control 1 Control 3 Food 1 Food 2 Fence Food+Fence SPRUCE 0 - absent 2.00 0.25 1.50 0.50 5.00 0 1 - scattered 30.75 2.50 19.50 8.75 32.25 56.00 2 - clumped 47.25 42.50 61.25 24.25 49.50 42.50 3 - dense 18.75 38.00 16.25 37.25 13.00 1.25 4 - very dense 1.25 16.75 1.50 29.25 0.25 0 SPRUCE SHRUB/SEEDLINGS 0 - absent 31.50 0.25 20.25 17.00 30.00 71.75 1 - scattered 40.75 32.50 55.75 26.25 55.25 27.25 2 - clumped 17.75 25.25 18.00 20.75 9.50 0.50 3 - dense 8.50 21.75 4.75 15.75 3.50' 0 4 - very dense 1.50 20.25 1.25 20.25 1.75 0.25 D E A D F A L L 0 - absent 47.00 19.50 24.50 28.00 20.25 17.50 1 - scattered 33.00 49.00 58.75 44.50 28.50 67.25 2 - clumped 13.75 19.00 15.25 16.25 32.75 11.50 3 - dense 4.50 8.50 1.50 8.50 14.50 3.25 4 - very dense 1.50 4.00 0 2.75 4.00 0.25 O P E N GROUND 0-10% 0 0.75 0 0 0.75 0 11-20% 0 3.50 1.50 0.25 3.25 0.50 21-30% 2.75 9.00 9.25 1.25 8.25 3.25 31-40% 10.50 17.25 19.50 11.00 17.50 7.75 41-50% 16.00 27.50 25.50 21.75 22.75 15.75 51-60% 24.75 22.75 23.25 33.00 24.25 28.75 61-70% 24.75 14.00 14.50 23.25 16.75 32.00 71-80% 14.75 5.00 5.50 7.25 4.50 9.50 81-90% 5.50 0.25 1.00 1.50 0.25 2.25 91-100% 1.00 0 0 0.75 1.75 0 mean ± SE 63.0 ± 0 . 8 51.8 ± 0 . 7 53.0 ± 0 . 7 51.4 ± 1.3 53.4 ± 0 . 8 61.5 ± 122 Species clusters and amount of open ground Closed habitats were more likely to have dense clusters of spruce, birch, and willow than were open habitats (Table 4.2). Control 3 and food 2 tended to have dense spruce clusters in closed habitats (75-78% of closed habitats had dense spruce clusters), whereas fence and food+fence, and to a lesser extent food 1, had dense birch and willow clusters in closed habitats. These patterns suggest that closed habitats have higher food availability than open habitats. Snowshoe hare habitat use and preferences in summer Hares on the different treatments did not differ substantially in their use of open habitats either between years or between treatments (Figure 4.1). On all treatments more than half of hares' locations were in habitats with 40-60% open ground. Hares on control sites did not use increasing amounts of open habitat as their densities increased (Kolmogorov-Smirnov statistics comparing year separately for control 1 and control 3: all pair-wise comparisons gave p>0.10, except for control 1 1994 to 1995, which had a p<0.01, with hares in 1994 using more open habitats; the Bonferroni-corrected alpha level is 0.0083). Hares on all grids tended to use closed and intermediate habitats; fewer than 5% of hares' locations were in open habitats. On all treatments, hares preferred closed habitats (Figure 4.2). In 1994, hares on three sites slightly avoided closed habitats, but in all other years and all other sites, hares used closed habitats in greater proportions than these occurred. For all sites except control 1, hares showed the least preference for closed habitat in 1994, the final summer of the low phase. Habitat at sites where hares were killed by predators Hares were killed by avian and mammalian predators in habitats with roughly similar distributions (Figure 4.3; Kolmogorov-Smirnov statistic, p>0.10). Manly's alphas show a 123 Table 4.2. Relationship between the clustering of hares' main food species and total ground cover. The amount of open ground is divided into closed (0-40%), moderate (41-70%), and open (71-100%); within each grouping, the values indicate the percentage of those locations in which each species is densely clustered (ranks of 3-4). The number of cases indicates the number of gridstakes (out of 400/grid) which occurred in each category of habitat. Control 1 Control 3 Food 1 Food 2 Fence Food+Fence Closed habitat n cases 53 122 121 50 119 46 dense birch 0 0 22.31 0 20.17 19.57 dense willow 7.55 0.82 4.13 2.00 24.37 52.17 dense spruce 35.85 74.59 19.83 78.00 13.45 0 Intermediate habitat n cases 262 257 253 312 255 306 dense birch 0 0 6.32 0 5.49 8.17 dense willow 0.76 0 1.58 0.96 4.71 4.58 dense spruce 21.37 48.64 17.79 71.15 13.73 1.63 Open habitat n cases 85 21 26 38 26 47 dense birch 0 0 11.54 0 0 0 dense willow 0 0 0 0 7.69 0 dense spruce 5.88 14.29 7.69 13.16 7.69 0 124 Figure 4.1. Hares ' use of open habitat. Va lues are percentages o f hare locations that occurred in each category of open ground. Open ground was assessed in 10% increments. Da ta are f rom hares located from May-September o f each year; habitat availabil i t ies were based on a sample of 400 gridstakes per gr id . A n average o f 216.4 hare locations were col lec ted per gr id per year (range 13-486). 125 CD ilabl ilabl CO LO cu CD CD CD > CJ) CD CD CO T _ T~ • o i i •? i i o o o o o o o o o o o o LO TT CO CN v - LO Tf CO CN s u o i i B o o i p l u e o j a d 126 Figure 4.2. Preference of hares for closed habitat. Data are from radio-telemetry locations of snowshoe hares from May-September of each year. Three categories of open ground—0-40% (closed), 41-70% (intermediate) and 71-100% (open)—were used to calculate Manly's alpha index of preference. The solid bars indicate the means of the three summers for each grid. The dashed line indicates no preference: values above the line indicate selection, and values below the line indicate avoidance; the magnitude indicates the amount of preference or avoidance. Sample sizes are the same as in Figure 4.1. 127 CO CD c CO O C O Tf L O CJ) ( J ) C D C D C D C D O c CD CD E 0) CD CO o c • o < I • < I I • I < o < CD O c CD C O ci CO ci c i CN O iejjqeij pssop JOJ souejsjejd .SSJBII 128 Figure 4.3. Habitat at snowshoe hare mortal i ty sites. The data are f rom 34 raptor and 50 mammal ian k i l l s which occurred on the six study grids. Habitat avai labi l i ty is the average o f gr id availabil i ty weighted by the number o f k i l l s per grid. 129 130 slightly different pattern, however, with mammalian predators killing disproportionately more hares in closed and intermediate habitats (a=0.45, 0.36, and 0.19 for closed, intermediate and open habitats respectively; 0.33 indicates no selection). Raptors, in contrast, killed disproportionately many hares in intermediate habitats (a=0.27, 0.51, and 0.22 for closed, intermediate and open habitats respectively; 0.33 indicates no selection). Only 11.8% of hares killed by raptors died in closed habitats, compared to 24% of hares killed by mammalian predators. Both mammals and raptors had -6% of their kills in open habitats. These results suggest that raptors may have hunted or had higher hunting success in more open habitats than did mammalian predators. Species clusters around snowshoe hares Under 40% of hares' summer locations had dense clusters of birch and willow (Figure 4.4). Dense birch clusters only occurred on the fenced grids and food 1, but on both fenced grids hares were more likely to be found in locations with dense clusters of willow than of birch. Hares' use of dense spruce clusters was a rough inverse to their use of dense shrub clusters—hares on control 3 and food 2 were located far less often in locations with dense shrub clusters, and much more often in dense spruce clusters, than were hares elsewhere, but these two sites also had at least triple as many gridstakes situated in dense spruce clusters. Hares on the food+fence treatment used almost no dense clusters of spruce, but dense spruce clusters occurred at only 1.3% of the gridstakes. Hares used dense deadfall clusters similarly on all sites, except for the fence treatment where hares were found in many more locations with dense deadfall. On most grids in most years, snowshoe hares showed strong preferences for dense willow and deadfall clusters (Figure 4.5). For the three treatment grids that had dense birch clusters, 131 Figure 4.4. Snowshoe hares' summer use of dense birch, willow, spruce, and deadfall. Species clusters were assessed for circles of 15 m radius. Data are from radio-telemetry locations of snowshoe hares in May-September. The open bars indicate availability for each treatment grid, assessed from 400 gridstakes on each grid. The x's indicate none of the hare locations were in sites with dense clusters of that species. Sample sizes are the same as in Figure 4.1. Confidence limits were estimated by binomial approximation (Krebs 1989, p. 473). A . Birch. There was no dense birch on either control grid or on Food 2. B. Willow. Includes Salix glauca and S. alaxensis. C . Spruce. Includes trees of D B H > 7 cm. D. Deadfall. Includes fallen spruce trees of a size sufficient to hide a hare. 132 100 90 CO 1— CO co o CD CO c CD TD CO c o CO o o 80 4 • i 70 60 50 40 30 20 10 0 A. Birch x x x x x x x x il X X X X • avail. E1993 • 1994 • 1995 C 1 C 3 food 1 food 2 fence FF 100 90 B. Willow co 80 CD 1 70 o CD £ 60 CD TD £ 50 g 40 g S 30 20 10 0 J C 1 C 3 food 1 133 food 2 fence FF 134 Figure 4.5. Snowshoe hares' summer preferences for dense birch, willow, spruce, and deadfall. Species clusters were assessed for circles of 15 m radius. Data are from radio-telemetry locations of snowshoe hares in May-September. Manly's alpha values of preference were calculated using four categories of clustering: absent, lightly clustered, moderately clustered, and densely clustered. The solid bars display the means for each grid for the three summers. The dashed lines indicate no preference: values above the line indicate preference, and values below the line indicate avoidance. Sample sizes as in Figure 4.1. A. Birch. No dense birch occurred on either control grid or on Food 2. B. Willow. Includes Salix glauca and S. alaxensis. C. Spruce. Includes trees of DBH > 7 cm. D. Deadfall. Includes fallen spruce trees of a size sufficient to hide a hare. 135 A. Birch CD -5> 0.8 _ D O CD CO I 0.6 CD O p 0.4 • 1993 o 1994 o 1995 - mean no selection g) 0.2 CO o o 0 + + + C 1 C 3 food 1 food 2 fence FF 1 0.9 CO I 0.8 f 0.7 to I 0.6 o O 5 o | 0.4 CD £ 0.3 to 2 0.2 co 0.1 B . Willow o • o B O • 1993 o 1994 o 1995 - mean no selection 0 + C 1 C 3 food 1 food 2 fence FF 136 C. Spruce o • • 1993 o 1994 o 1995 - mean — no selection 8 • "6" a + o + C 1 C 3 food 1 food 2 fence FF 0.9 D. Deadfall CO CO CO _D P CO CO c CO TJ 0.8 0.7 0.6 = 0.5 0.4 1» 0.3 0 o £ 0 a 0 k_ Q . "8 0.2 L_ co • o a 1993 o 1994 o 1995 - mean — no selection 0.1 0 -I H 1 \— 1 1 1 C 1 C 3 food 1 food 2 fence FF 137 hares only occasionally showed a preference for locations with dense birch. Hares favoured dense spruce in some years and avoided it in others, except on food 2, where hares consistently preferred dense spruce. Dense spruce was favoured in 1995 and slightly avoided in 1994. Immediate cover of snowshoe hares When snowshoe hares were resting, they were almost always (72-100% of cases) in locations where their nearest cover items in four directions were < 1 m away ('thick' cover; Figure 4.6, Appendix 7). In the summer, active hares were in thick cover 20-75% of the time. Hares' browse sites were much less likely to be in thick cover, especially in January-April. In summer, hares tended to use more thick cover than habitat availabilities would predict, whereas at their fall and winter browse locations, hares used less thick cover than the habitats would have afforded them. In the summer, spruce, willow, and deadfall were all common as the species providing hares with immediate cover (Figure 4.7). Birch was neither common in the habitat nor in hares' locations. Hares used willow and spruce less often than these species occurred in the habitat, but hares used deadfall for cover much more often than it occurred in the habitat at large, especially when they were resting. In the months with snow cover, snowshoe hares shifted away from the heavy reliance on deadfall to a high association with spruce (Figure 4.8). Willow, although still accounting for up to 47% of the species near to hares, was utilized the same as or lower than its occurrence in the habitat at large. Deadfall dropped in utilization, to levels similar to those in the habitat. Hares used birch for cover less than predicted from its occurrence on the grids. 138 Figure 4.6. Hares' microhabitat choices: use of thick cover . T h i c k cover means that in al l four directions hares were < 1 m away from cover. E a c h point represents the annual percentage o f hare locations in thick cover. A v a i l a b i l i t y o f thick cover was derived f rom 400 gridstakes per gr id . Sample sizes for resting hares range f rom 1-264, x = 83; for active hares, 3-157, x = 54; for fa l l browse sites, 9-53, x = 33; for winter browse sites, 2-68, x = 35 (see also A p p e n d i x 7). 139 1— CO CO E CO E es E CO ' E es CO CO CO "co "co O) c "co CO 9> _> res CD res CO o . c res CD CO CO 0 CO 2 CO CO CO CO CO CO 5 2 _ .a ~ 0) 5 TO CD > CD • D • •4 •* COi o a a CO o c CO D • CN T3 O O J 9 A 0 0 >|0!m in suorjeooi jo 96ejuaoj9d T3 o o CO o » c o o c o o o o o oo o CO o ""ST o CN 140 Figure 4.7. Snowshoe hares' summer use o f cover species. E a c h point represents the percentage occurrence for each species as cover at hare locations f rom one summer (May-September) . The lines indicate habitat availabil i ty derived from a sample of 400 gridstakes per gr id . The sample size o f species (four per hare location) for resting hares ranged f rom 12-1056, x = 504; for active hares, 48-652, x = 326. 141 CO CO $ 2 CO 5" J - CD o CO c CO • om 4 -4 * . « • -4 -41 •cv.' -4 illow • <*>«M • om 4 •4 CD O c CD TJ O O omm-m irch CQ i o o CD O c CD FF 4 41 ••*> •4 4 • • • <> CM TJ O 4 4» 4*9 4 • • • O CO control Deadfall 44 M « 4 « • • *6 CN TJ O O TJ O O CO 2 •#-» C o o c o o CD O c. CD CN TJ O O • .<m -4«4 4 0.' 4 « * • -4 Spruce •4 o\ m 4» b • — W4M TJ O O CO 2 C o o cz o o CD O c CD CN TJ O O TJ O O CO 2 ' c o o c o o o o o 00 o CO o o CN o o o 00 o CO o o CN sapeds jeaup % 142 Figure 4.8. Snowshoe hares' winter use of cover species. Each point represents the percentage occurrence for each species as cover at hares' browse sites. The lines indicate habitat availability derived from a sample of 400 gridstakes per grid. The sample size of species (four per location) for fall ranged from 72-304, x = 156; for winter, 8-464, x = 219. 143 CO l _ CD i • CO CJ) H—< » CO 'to "co >*— CD > 0 "to CO "co CO CD CD i _ CO 2 t_ CO g x : x a x : X2 • 4 * 41 Willow V 4 * • O C 0 CN T J O O T J O O CO 2 H—< C o o c o o ? • ' 6 • • irch CD < u. 0 o c 0 CN T J O O O O CO 2 c o o c o o o o o CO o CD o T f o CN • 441 4 4> 4 m CO 04 tt— T3 CO 0 Q 4Ji 8 c .0 CN T J O O 4 4«M * 0 4« o o. T J O 2 C Z 8 2 -*—< c o o 0 o c 0 CN T J O O T J O O CO 2 c o o c 8 o o o CO o CD o T f o CN S 9 l 0 9 d S JB9U JO % 144 DISCUSSION Did snowshoe hares use refugia? The refugium hypothesis predicted that fenced hares would use more open habitats, with lighter species clusters, than would hares exposed to mammalian predators. Hares were also predicted to use more open habitats and lighter species clusters as hare densities increased (Wolff 1980, 1981, Hik 1994). Neither of these patterns occurred (Table 4.3). Hares on the fenced treatment used marginally more dense habitats than did control hares, but hares on food+fence and controls resembled each other in their habitat use patterns. On control areas, hares used similar amounts of open and closed habitat in each year. There was a very slight annual difference on one control area, but it was opposite to the prediction. Hik (1994) suggested that dense spruce cover might define refugia; during the low phase, hares on both fenced grids used fewer dense spruce clusters, but hares' preferences for dense spruce did not differ between treatments. Many of these differences in habitat use patterns were due to habitat availability differences rather than to differences in hare behaviour. Hares on all treatments preferred closed habitats to open habitats and preferred locations with dense willow or dense deadfall clusters. Hares did not consistently avoid or prefer dense spruce. Although hares on control 3 and food 2 were located in dense spruce more than were hares on any other treatment, their preferences were similar; these two grids simply have dense spruce clusters occurring in a greater proportion of their total area. What constitutes a refugium? The mortality patterns suggest that during the low phase hares at Kluane had no habitat that was safer than others. Although most raptor kills occurred in intermediate habitats, 145 CO CD O G <u ,P 4-H D '•8 X rt rt rt X "5b j2 3 <+H <U CD co O <D O -f i ^ 3 0 s -<4-> <<-> O O c X J X J X J "3 C O P H CO CO D U B 1H S-H S~ ca cd l " a a .2 c o -t—» c CD X J C CD P H CD X J CO cd 5 CD CO 3 <+- X O cd 8 g P cd a CD S-. CD CO 3 +-» O c X J x3 — . (D cd X CD - 3 CD X J cd C • f i ^ s—i O CD i <+H I, cd ^ CO CO > C CO rt rt rt 4) X C •2 -fi 2 H CD X ) rt lH •4—' X J O & M CO O C CO cd CD '— CD cd -t—» x cd X o .1) > "•H G rt -J3 CD O *H CD fc co CD N C O co rt CD CO C O -t—» c CD X J C t & g P CJ - f i rt CD l-fi CO •— O , rt X J CD '— P H I V, o > rt CO •4—» rt CO C rt P H CD CO 3 X "co CD Id l - f i CD CO 3 C O T J CD V -P H C o CD 'co rt cj 1 'd c rt O > CJ — . rt X J *-* CD O SH _ P H O S i 2 — rt rt X C CD i . s B » u +3 CD co n e 3 rt K * -CD Q rt 13 J_> ~ O fe X rt G rt HD "5 S2 co co £ » CD H X rt cd X X X J X J C U CD § " " P H c CO rt JS C .CD CD C 6 x rt - 3 C CD P H o c 3 O •a CD o CD SH r-.CD X J CD x3 £ o g G P H CD „CD « ' CD rt X 3 X J co CD CD O •— C - 3 < B 3 CD 3 CJ C C rt ,<D JS co CD rt X CD CO 3 T J CD O c CD CD b * rt c -f i 3 TJ C CD rt ^ - f i CD CD O S CO co CD CD co t2 fi 3 CD 5 X J CJ CD CJ c e .CD rt 3 C c rt O c CD o 3 CD t-rt 3 C c rt O c co CD CO rt e CD O co 3 •= co >-> CD . G H CO rt C CO • H SH rt CD •k-> •!-> 1 3 J 3 O T3 cD CD O co 3 ^ P H O co C rt X i >> ^ , 2 c — C U 3 ^ & B C rt co • — e 1-H !_, G pl 3 | 2 8 6 CD O G CD ^tS cd X 3 rs: rt CD O G G SS JS CD o -4—» CD _> *4—» rt (rel 4-H CO r3 rt lab X ) '3 rt > - C rt CO CD rt O 6 "2 55 rt o c rt X ) c 3 X J fc rt O X J c o o .o O X J co rt 2 CD — G rt CD X ! bO CO C CD P H O C ^ .9 O G 5 O rt ^ X i - g CD _ g J— i CO H G g td X J X _ r t ' rt > rt X J O o CD > o o M o CD 3 O XI ' H « rt y > CD £ «i E S c CD E —^* rt CD rt S CD £ O C X J in CD > O CD CD * P H .2 cT ?D rt O E CD co 3 o -X J CD IH P H G CD X CD td X J CD C LH CD fi > g O o * 9 X J rt CD X J CD P \JS o o '•S co CD LH rt I X CD o P H X o > rt — , 3 CD CJ 3 P H CO CO o C G rt CD * >> >• O rt CD 3 X ? 3 i-CO CD > £>• O o CO CD 'u CD P H CO •— CD > O CJ CD X X X J B o 1x3 CD LH P H lH . o o G rt CD CD s rt >-. CO rt G P H O CD '-3 co o • f i H cD LH * P H 146 mammalian predators killed more hares in closed areas than in open areas. Concurrent research at Kluane showed that lynx and coyotes were located more often in areas with less overstory cover than were hares (O'Donoghue 1997), but several studies have found that predators preferentially use habitats that have the highest hare densities (Koehler 1990, Murray et al. 1994). Coyotes had high rates of successful chases in closed spruce forests during the late increase and peak phases of a cycle, while lynx had equal success rates in various habitat types (Murray et al. 1994). These patterns suggest that large-scale habitat characteristics may not have much to do with safety from predation, especially if predators select their habitats on the basis of the presence of hares (Koehler 1990, Murray et al. 1994). Since the efficacy of a habitat type as a refuge depends upon both the abundance of predators and their hunting ability within that habitat, the low predator densities (<0.07/km2) during the low phase and the slight differential hunting ability of raptors and mammalian predators may mean that many habitat types offered the same degree of safety to hares during this phase. The refugium hypothesis has also posited that hares must trade-off food acquisition and safety, since no habitat offers both excellent food and safety. This position does not appear to be tenable for Kluane hares during the low phase: no habitat offered superior safety, but nor did any habitat result in clearly reduced food. In fact, the putatively risky open habitats may also have less food in them, since open ground occurs partially because fewer food plants grow there. The food density during the low phase was so high that it is likely that even these habitats offered sufficient food for hares (Table 2.3). 147 What scale is appropriate for assessing refugia? The refugium hypothesis for hares was constructed on the basis of habitat patches of multiple hectares (Wolff 1980). Similarly, many of the studies which have found dense understory to be favoured by hares have considered habitats at forest stand levels, and hence have compared hares' use of plots of multiple hectares (Conroy et al. 1979, Orr and Dodds 1982, Wolfe et al. 1982, Monthey 1986, Fuller and Heisey 1986, Scott and Yahner 1989). In contrast, the median diameter of overstory habitat patches at Kluane ranges from 79-188 m (O'Donoghue 1997); hares' home range diameters were -300-500 m (see chapter 5). Kluane hares' home ranges therefore can seldom be contained within a patch of any given overstory type. During a population decline at Kluane, Hik (1994) used a 15 m scale to find that hares on control sites and the food+fence sites used denser habitats as the decline continued. Similarly, most of the evidence for hares' differential survival in various habitats occurs at the scale of 15-20 m (Hik 1994, Murray et al. 1994, Rohner and Krebs 1996, Cox et al. 1997). The 15 m data I have presented do not indicate that hares used refugia during the low phase. At microscales, large trees and deadfall may provide complete safety for hares, since hares are both difficult to detect and difficult to catch in these microhabitats (M. O'Donoghue, pers. comm). When hares rest, they almost exclusively chose places in the midst of dense foliage or under deadfall, again suggesting that these are places of safety. In an experimental study, increasing the abundance of such mini-refuges did not increase hare survival rates relative to hares on control sites, but most mortalities occurred in areas away from the manipulated brush piles (Cox etal. 1997). 148 What habitat characteristics did hares select? During the low phase, neither food addition nor exclusion of mammalian predators influenced hares' habitat choices. Instead, hares routinely preferred closed habitats with dense clusters of willow and deadfall; within these habitat types, hares used thick cover. These conclusions are similar to those of many other studies—hares like dense, brushy habitats (Conroy et al. 1979, Orr and Dodds 1982, Wolfe et al. 1982, O'Donoghue 1983, Fuller and Heisey 1986, Rogowitz 1988, Koehler 1990, Ferron and Ouellet 1992). Hares appeared to select habitats based on willow, deadfall, and ground closure. Since species clusters and open ground were inversely correlated, it is difficult to distinguish which specific attribute (if any) hares were selecting. Dense brush may have been the primary characteristic hares were selecting; Rogowitz (1988) and Ferron and Ouellet (1992) have suggested that total cover and not species composition is the critical determinant of habitat selection for hares. The data presented here, however, suggest that species composition mattered to the hares: hares had strong preferences for willow and deadfall, but not so strong for birch or spruce. Hares' preference for deadfall indicates that hares are at least in part selecting habitats for their protective cover. Since dense brush could be beneficial to hares as protection against predators, good foraging grounds, and even thermal protection (Belovsky 1984, Hik 1994, Villafuerte et al. 1997), it is possible that hares simply try to use dense habitats, without being particularly sensitive to either predation risk or food supply. Cyclic habitat shifts and demographic implications The refugium hypothesis suggests that during the low phase, hares shift their use of habitats in a way that affects their survival, nutrition, body condition, and perhaps fecundity 149 (Wolff 1980, 1981, Hik 1994, Boonstra et al. 1998b). Although the evidence presented here does not support the existence of predator-sensitive habitat shifts by hares during the low phase, hare habitat use could still differ between phases of the cycle. The patterns during the low phase indicate that hares are not constrained by predators or food, but that does not imply that they do not shift habitats in response to these factors at other times. Examining hare habitat use into the increase phase would be useful, since the range of hare densities examined here is much smaller than the range that occurs through a cycle, and expansion into other habitats may occur at much higher densities than I observed. The lack of differential habitat use by hares between treatments, however, indicates either that hares did not use refugia during the low phase, or that refugia are created and maintained by some mechanism other than a direct response to the numbers of terrestrial predators in the surrounding area. These data do not support the argument that constricted habitat use during the low phase leads to a cascade of negative dietary, physiological, and population consequences (Hik 1995). I did not find evidence of habitat choice being constricted, nor did I find evidence that use of any particular habitat limited hares nutritionally. In order to verify this suspected lack of refugia during the low phase, hare habitat use in other phases of the cycle, and mechanisms behind their habitat choices, need to be compared with the data presented here. 150 C H A P T E R 5 T H E E F F E C T S O F F O O D A N D P R E D A T I O N R I S K O N S N O W S H O E H A R E S ' M O V E M E N T P A T T E R N S I N T R O D U C T I O N Movement is a critical component of behaviour, since an animal's movements determine the habitats, foods, and potential mates it will encounter. Home range size similarly demarcates an area of opportunity for the animal, and dispersal creates a new array of options. Movements may also be crucial to population dynamics: dispersal is the clearest case of movement affecting demographics (Lidicker 1985a, 1985b, Wolff 1997), but changes in home range size and movement rates can affect predation and starvation rates, density, and even fecundity if there are nutritional changes associated with the movement changes (Carpenter 1987, Ives and Dobson 1987, McNamara and Houston 1987, Lima and Dill 1990, Hik 1994, 1995, Boonstra et al. 1998b). Cyclic population dynamics in small mammals may be partly determined by their spacing behaviour and dispersal (Lidicker 1985c, Taitt and Krebs 1985, Krebs 1996, Wolff 1997). Home range size, movement rates, and dispersal may be affected by food distribution, predators, and conspecifics (Hixon 1980, Boutin 1990, Lima and Dill 1990, Larsen and Boutin 1994). Much recent work has focussed on the indirect effects of predation on movements: small mammals often reduce their activity levels, movement rates, and home range sizes when risk of predation is high (Brown et al. 1988, Desy et al. 1990, Jedrzejewski and Jedrzejewska 1990, Kotler et al. 1991, Saarikko 1992, Hughes and Ward 1993, Jedrzejewski et al. 1993, Fenn and 151 M a c d o n a l d 1995). In at least one case, however, animals increase their activity and home range size i n the presence o f predators (degus, Octodon degus, L a g o s et al. 1995). The dynamics o f the 10-year snowshoe hare (Lepus americanus) cycle may result partially f rom hares' behavioural shifts in response to predators ( W o l f f 1980, 1981, Sievert and K e i t h 1985, K e i t h et al. 1993, H i k 1994, 1995, Krebs 1996). The ten-year cyc le has two to four years during w h i c h hare densities are low, despite h igh food availabil i t ies and l o w predation rates (chapter 2, K r e b s 1996, Boons t ra et al. 1998b). The low phase cou ld be explained behavioural ly, since changes o f movement patterns in response to predators cou ld lead to dietary shifts and reduced body condi t ion, wh ich could in turn reduce fecundity or juveni le survival (Boonstra and Singleton 1993, H i k 1994, 1995, Boons t ra et al. 1998a, 1998b). Changes in juveni le survival are cr i t ical to the maintenance of the snowshoe hare cycle , wi th smal l changes having large populat ion consequences ( E . A . G i l l i s , D . Haydon , C . I . Stefan, and C . J . Krebs , ms. in prep). In this chapter, I describe the movement patterns o f snowshoe hares during a c y c l i c l o w phase, and I test whether hares' movements are pr imar i ly determined by food supply, predation risk, hare density, or a combinat ion o f these factors. I use a factorial manipulat ion o f food addition and predator reduction to assess the relevance o f each factor to hares' movements. If hares move more when food is scarce or wide ly distributed, then food addition should result in a reduction in movement (see Table 5.1 for detailed predictions). If hares reduce movements when predation r isk is h igh, then hares protected f rom predators should move more than do unprotected hares. If hares trade-off food and predation risk, then hares on food addit ion areas should move the least (small movements do not compromise food acquisi t ion, but do enhance safety from predators) and hares protected f rom predators should move the most (there is a reduced mortal i ty 152 Table 5.1. Predictions for the effects of food supply, predation risk, and density on snowshoe hare movements. Predictions are relative to the values for hares on control areas. If both food supply and predation risk affect movements as predicted, then hares on the fence and food treatments should show the predicted changes in these movement indices, but the response of hares on the food+fence treatment is impossible to predict given the opposing predictions for the two factors. factor predictions rationale food supply on food and food+fence treatments: smaller home ranges shorter distances between browse sites lower movement rates reduced activity reduced dispersal rate longer time per browse site movement by sexes similar higher local density of quality food-less effort is needed to obtain dietary requirements. predation risk on fence and food+fence treatments larger home ranges longer distances between browse sites higher movement rates higher activity higher dispersal rate shorter time per browse site movement by sexes similar lower predation risk-cost of movement is lower since movement does not affect predation rates when there are no predators. density on higher density areas*: smaller home ranges shorter distances between browse sites lower movement rates reduced activity reduced dispersal rate similar time per browse site males move more than females more hares in neighborhood-less effort is needed to find mates and aggressive interactions can be avoided by reducing encounter rates with other hares. * Average densities were largest on the food+fence, then food, fence, and control treatments. Densities generally increased from 1993-1996. See Figure 2.2, Table 5.2. 153 cost to moving, and moving enhances food acquisition). Finally, if high density results in reduced movements, then the smallest movements should occur on the highest density areas, and movements should decrease during the three years of the study, as hare densities increased. A priori, I do not anticipate males and females to exhibit different movement patterns, because many previous studies have not found differences (reviewed in Keith 1990). M E T H O D S A N D M A T E R I A L S Study area, treatments, and study animals I studied snowshoe hares near Kluane Lake, Yukon, from March 1993-April 1996, using study sites maintained by the Kluane Boreal Forest Ecosystem Project (Krebs et al. 1995). I used two 34 ha. food addition areas (food); one 1 km2 area (fence) surrounded by a chicken wire and electric fence which kept out the main terrestrial predators, lynx (Lynx canadensis) and coyotes (Canis latrans), and had monofilament spread over ten ha to deter avian predation; one combination food+fence treatment, but without monofilament; and two control grids (Figure 2.1). I also used hares from five other control sites, but these five sites were not used in all years, nor were their population densities known. All sites were in white spruce (Picea glauca) forest, with scattered aspen (Populus tremuloides) and poplar (P. balsamifera) stands; bog birch (Betula glandulosa), willow (Salix glauca and S. alaxensis), and soapberry (Shepherdia canadensis) provided the predominant understory cover. Snowshoe hares were trapped in Tomahawk live traps (Tomahawk Live Trap Co., Tomahawk WI) baited with alfalfa (and rabbit chow on food and food+fence treatments) on each study grid every March and October for population estimates, and in other months as needed to maintain the numbers of radio-collared hares. Eighty-six traps were spaced in four rows, with 154 traps 30-60 m apart, and rows of traps 150 m apart. The effective trapping area was -60 ha per trapping grid. Each hare was given a monel eartag (No. 3, National Band and Tag Co., Newport, KY). Most adult hares and some juveniles heavier than 1200 g were fitted with 40 g radio-collars (Lotek, Newmarket ONT); each collar was equipped with a mortality sensor that doubled the pulse rate when the collar was inactive for four hours. For every capture, we recorded the animal's eartag, location, weight, right hind foot length, sex, and reproductive status. Locations and activity of snowshoe hares My assistants and I used handheld receivers (TR2, Telonics, Mesa AZ ) and antennae (Yagi or H) to find radio-collared hares. We had good success using telemetry to find animals; many hares (58%) were located visually, and many of the remainder were hiding in thick deadfall and thus were difficult to see. Walking in to locate the hares did not disturb them. In winter, I could see from snow tracks that most hares were resting in cover and did not startle as I approached them. In summer, when we observed hares foraging, we often found fresh faecal pellets, indicating that the hares had been foraging undisturbed and had not altered their behaviour on our approach. We startled hares away from their original positions in only 7.7% of cases (noticeable by a loud telemetry signal becoming suddenly soft). Even when that occurred, we were within 30 m of the hare, given the nature of the equipment. We were thus able to determine both location and activity of hares without undue influence upon either. During April-September, we located each hare every one to three days. During October-March, hares were located sporadically for the purposes of snow-tracking (see Chapter 3). Hares' locations on the trapping grids were estimated to 3 m (l/10th grid stake). When hares were within 150 m of the grid, their locations were estimated to the nearest 30 m. Locations further 155 than 150 m off-grid were determined via GPS (Trimble Basic Navigation System) which was usually corrected using simultaneous readings on a fixed waystation at the Kluane Lake Research Base of the Arctic Institute of North America (corrected locations were accurate to 30 m; uncorrected locations to 150 m). When a hare's location could not be defined within 30 m, I did not use that location in analyses, except for a few dispersing animals, when the dispersal distance was several kilometers and the location error was trivial in comparison to the distance travelled. Whenever hares were located, we recorded the time and their activity. 'Active' hares were eating or moving. 'Resting' hares were sitting under deadfall, bushes, or trees. We classified activity as 'unknown' when we could not tell what the hare was doing. I categorized times into hourly intervals and calculated the proportion of animals active during each hour. Summer movement rates In the summers of 1994 and 1995, we followed 6-21 hares per treatment for 6-9 hours a day usually starting at 0500-0600 h, sometimes starting at 1900 h. We located each hare an average of 5.4 times (range 2-14) within that time span, at intervals of 30-75 minutes. I calculated movement rates by dividing total distance by total time. If a hare settled into its resting period, I did not use the subsequent relocations since the animal clearly had stopped moving. Home range estimation Home ranges were estimated by minimum convex polygons (Mohr 1947, Harris et al. 1990) and kernel methods (Worton 1987, 1989, 1995), using CALHOME (Kie et al. 1994). Polygons were calculated using 100% and 95% of the locations, and kernels were calculated using 95% and 50% of the locations. Home ranges were calculated only for hares with 12 or 156 more locations, since previous researchers at Kluane have found that estimates of hares' home ranges do not increase much after 10-20 locations (Boulanger 1993, Allcock 1994, Hik 1994). I calculated summer home ranges for each year from telemetry data collected from April through September. At a longer temporal scale, I calculated a home range for each hare based on all the trapping and telemetry data collected during its life. Because the snowshoe hare literature contains numerous home range estimates derived from trapping data, I also calculated home ranges derived from trapping data to compare these with home ranges derived from telemetry. Kernel and convex polygon home range estimators depend on sample size and assume lack of autocorrelation (Boulanger and White 1990, Harris et al. 1990, Worton 1995). I found no relationship between estimates of home range size (95% minimum convex polygons) and number of hare locations for ranges with > 12 locations (r2 = 0.02, F u 2 9 = 2.48, p = 0.12), so the number of locations does not appear to be a problem for these home range estimates. The only locations which were potentially autocorrelated were those collected to generate movement estimates, from hares which were active. Hares can easily traverse their home ranges in the 30-75 minute intervals at which locations were taken, thus minimizing potential autocorrelative problems. Winter movement rates, distances, and time spent at browse sites I calculated winter movement rates from snow-tracks of hares in winter. I used radio-telemetry to locate known individuals, then followed their tracks backwards. Browse sites were identified by the 45"angle of browsed twigs and the green colour of the twigs' ends. Along hares' tracks, I recorded the number of faecal pellets at each browse site and the distance between browse sites. I calculated winter movement rates for tracks longer than -1.5 hr by dividing total distance by total time per track (time was estimated by multiplying pellet number x 35.2 pellets/hr; 157 see Chapter 3, Figure 3.10). Mean time at browse site and mean distance between sites were calculated for each track, and hare tracks were then compared across treatments, sexes, and years. Dispersal and forays Ideally, animals would be counted as dispersers if they moved from one clearly defined home range to another home range or died after leaving the first home range. Because the requisite information is difficult to obtain, proxy distances are often substituted (e.g. Beacham 1981, Jones 1986, Krohne and Burgin 1987, Gillis 1997), and animals which move more than the given distance are classified as dispersers. In this case, I define the minimum dispersal distance to be greater than two home range diameters. This distance is arbitrary, but is larger than would occur if hares moved into home ranges adjacent to their original ranges. Home range diameters were defined individually where possible, from empirically derived home range areas (using 95% minimum convex polygons). For hares for which I could not calculate individual home ranges, I derived the minimum dispersal distance from the average home range size for their experimental treatment and sex. Some animals travelled at least the minimum dispersal distance but returned to areas they had used before, so I classified their movements as forays. Statistical analyses I used analyses of variance to test for treatment, sex, and annual differences, and regression to assess the effects of density on movement parameters. Analyses were performed using Statistica (StatSoft 1995). For winter movement rates, time at browse sites, and distance between browse sites, the ANOVAs test only main effects since I do not have data for all of the categories. Activity data were analyzed for times between 701-1700, since only these intervals had adequate sample sizes for all treatments. 158 RESULTS Snowshoe hare home ranges: methodological choices Kernel estimates (95%) of hare home ranges were approximately 1.9 times as large as minimum convex polygon estimates (Appendix 8). This pattern occurs as a function of the way the estimators calculate range size, since the kernels often included large areas which had no hare locations within them. Since the two estimators were strongly correlated with each other (Table 5.2A), I subsequently use data from minimum convex polygons using 95% of the data points to avoid the inclusion of non-utilized areas; the conclusions are robust to the estimator chosen. Snowshoe hare home ranges estimated from trapping data were smaller than ranges estimated from telemetry data (Figure 5.1) on all treatments except food+fence. Hares on the food addition treatments had particularly low trapping ranges relative to their telemetry range sizes, and the correlation between telemetry and trapping ranges is understandably poor (Table 5.2B). The correlation between home ranges estimated from telemetry data and home ranges estimated from telemetry+trapping data is good, so I use the latter since that includes more information for each hare. Snowshoe hares' home ranges Hares did not utilize all areas within their home ranges evenly; half of hares' locations occurred within 1.4-5.3 ha, or -20-30% of the total area contained within their total home ranges (Figure 5.2, Appendix 8). Males had home ranges about double the size of females' ranges on the control and food addition treatments, but males and females had similar range sizes on the fence and food+fence treatments. Male hares on the food and control sites had larger home ranges than 159 Table 5.2. Spearman's coefficients of rank correlation for snowshoe hare home ranges calculated using different methods. All home range estimates had at least 12 locations. A. Home range estimators. Correlations were calculated for home ranges estimated from both trapping and telemetry locations. Total n = 140. B. Method of collecting hare locations. Correlations are for 95% convex polygons. Total n = 33. A . ESTIMATORS Convex Polygons Kernel Analysis 100 % 95% 95% 50% 100% polygon 1.00 95% polygon • 0.88* 1.00 95% kernel 0.88* 0.89* 1.00 50% kernel 0.73* 0.75* 0.73* 1.00 *p<0.001. B . DATA SOURCE Trap Telem. Trap+Telem. Trapping 1.00 Telemetry 0.30 1.00 Trap+Telem. 0.63* 0.83* 1.00 *p<0.001. 160 Figure 5.1. Snowshoe hare home ranges: comparison of data sources. Points are means for each treatment and sex. Total home range size (x-axis) refers to estimates made using all telemetry and trapping data for each individual. The y-axis is the home range size calculated from only one of the two data sources. Trapping data provided poor estimates of home range, as the dashed line indicates; telemetry home ranges were much larger, and were close to total home range sizes. 161 i O LO O LO o CN T - T -ejep Ajiaw9|8i. J O 6indde.u. (BLJ) az;s aBuBJ 162 Figure 5.2. Snowshoe hare home ranges and core utilization. The bars represent mean home range area for hares on each treatment (ha ± SE), estimated by 95% minimum convex polygons from trapping+telemetry data; each hare's home range was based on all of the locations obtained during its life. Home range cores were calculated using a 50% kernel estimator. The number at the bottom of each bar is the average percentage of the home range area contained within the core range, and the number at the top of each bar is the sample size. F = female, M = male. 163 0- 2 IT) O CD IT) • 3 <3* I A I ^ — i i ^ 1 a 1 ^ L L - C O U _ C O u_ o CM CM + CD O c + TJ O O CU O c 0 TJ O O c o o LO CN O CN m o X — T — (Bl|) 9ZJS 9 6 U B J 9UU0L) LO 164 hares on the fence and food+fence treatments (two-way ANOVA, treatment F3129=6.71, p<0.01; sex F,,129=14.61, p<0.01; sex x treatment F3129=4.43, p<0.01). Females had statistically similar home range sizes on the four treatments (Figure 5.2), but both sexes on the food+fence treatment had smaller home ranges than did hares on other treatments (6.6 ha for both sexes, compared to female means of 8.4-9.4 ha and male means of 10.5-18.9 ha). Annual variation in home range size was slight (Figure 5.3). One-way ANOVAs for each sex on each treatment showed that the only significant difference was for females on the food+fence site, with larger home ranges in 1995 than in 1993 and 1994 (F223=10.12, p<0.01). Summer activity patterns Hares were less active during the afternoon than in morning or evening (Figure 5.4), with more hares resting than active between 1100 h and 2000 h. Hares on the fence treatment were generally less active than other hares (three-way ANOVA on data from 701 h to 1700 h: hours F9136=54.15, p<0.001; treatment F3136=6.81, p<0.001, sex and all interaction terms N.S). There was more variation in activity level among treatments during evening than in morning. Movement rates In summer, hares on food addition sites had higher movement rates than hares on the fence or food+fence sites, and males had higher movement rates than females (Table 5.3; three-way ANOVA, treatment F3384=3.50, p<0.05; sex F[ 3g4=6.18, p<0.05; year and all interaction terms N.S). Hares on control areas also had high movement rates, comparable to the rates of hares on food treatments, but this difference from the fence and food+fence treatments was not statistically significant. Males also had higher movement rates than females in winter (about 165 Figure 5.3. Snowshoe hare home ranges in summer. These range sizes were calculated us ing telemetry data collected from Apr i l -September in each year, wi th the 95% m i n i m u m convex po lygon estimator. E a c h point represents a single hare's home range size. 166 : A. Females • • • • * t * A 1 control food fence food+fence 1993 1994 1995 Mean B. Males • V • A • A • A control food fence food+fence 1993 1994 1995 Mean 167 Figure 5.4. Snowshoe hare activity patterns in summer. The y-axis is the percentage o f hares observed foraging or active (hares with unknown activity were excluded). T i m e s were categorized into hour- long periods (e.g. 1001-1100 was counted as 1100). The total sample size is 4639 observations o f hares, f rom telemetry data col lected dur ing Apr i l -September , 1993-1995. 168 Table 5.3. Movement rates of snowshoe hares. Summer is May-September; winter is November-April. Values are meters/hour ± SEM. In summer, values are means for each hare per day. Days, fixes, and hares are all sample sizes—number of days on which hares were followed, total number of fixes gathered, and number of hares from which fixes were collected. In winter, values are means of track rates. Tracks were included only if they were ~1.5 hr long (> 50 faecal pellets along the track). Events refer to places where hares ate something along a track, and are analagous to fixes. Days and tracks are the sampling units for summer and winter, respectively. SUMMER FEMALES MALES days fixes hares movement rate days fixes hares movement rate (m/hr) (m/hr) control 1994 1995* 17 114 92 768 11 12 113.6 ± 15.7 56.4 ± 2.9 8 167 44 719 5 9 124.6 ± 27.7 113.9 ±6.3 food 1994 1995 6 19 38 118 2 9 98.0 ± 19.2 69.7 ± 6.8 10 2 46 6 5 2 106.7 ± 18.3 169.5 ±91.4 fence 1994 1995 11 5 62 25 5 4 68.3 ± 15.1 47.7 ± 15.2 1 5 6 25 55.1 73.9 ± 18.6 food + fence 1994 1995 21 7 110 31 13 4 42.4 ± 4.0 45.4 ± 8.7 3 4 9 19 3 4 127.3 ± 14.5 46.4 ± 12.1 Data are used by permission from T. Chu and J. Jekielek, who used this technique for males and females, respectively, for their 1996 undergraduate Honours Theses at the University of British Columbia. 170 Table 5.3 cont. Movement rates of snowshoe hares. WINTER FEMALES .. tracks events movement rate (m/hr) MALES tracks events movement rate (m/hr) control 1993- 1994 1994- 1995 1995- 1996 4 4 4 39 54 23 12.1 ±4.1 29.5 ±8.0 16.4 ±6.4 1 7 6 16 65 37 35.5 20.2 ± 8.0 22.7 ± 7.0 food 1993- 1994 1994- 1995 1995- 1996 2 14 12 9 96 84 5.9 ±4.5 9.4 ±3.2 10.6 ± 1.8 1 5 2 5 26 18 10.5 13.8 ±4.7 29.7 ± 7.9 fence 1993- 1994 1994- 1995 1995- 1996 3 4 0 18 27 0 3.3 ±2 .0 4.6 ±2.1 2 1 4 12 7 33 43.4 ±9.1 9.4 16.8 ±4.3 food + fence 1993- 1994 1994- 1995 1995- 1996 6 2 2 32 9 14 7.7 ± 1.9 13.5 ±0.1 10.2 ±5.1 0 3 2 0 18 14 29.8 29.9 14.9 10.5 171 double), but winter movement rates did not vary annually or with treatment (Table 5.3; three-way ANOVA, sex Fli69=14.19, p<0.001; treatment and year N.S). These data were also analyzed via multiple regression (Table 5.4, Figure 5.5). Movement rates and home ranges were larger for males than for females, but hare density, food addition, fencing, and year did not show relationships with these movement parameters. Summer movement rates (Table 5.4, Figure 5.6) were lower on the fenced grids, but winter movement rates were similar for all treatments and both sexes. Foraging patterns Male and female snowshoe hares spent about the same time per browse site (Figure 5.7; three-way ANOVA, year F2149=9.06, p<0.001, sex and treatment N.S). Hares spent about twice as long at browse sites during the winter of 1994-1995 on all treatments except the food+fence. The mean distance between browse sites showed no significant annual, treatment, or sex variation (Figure 5.8; three-way ANOVA, all effects N.S). Dispersal and forays Females snowshoe hares dispersed at 1.3 to 2.3 times the rate at which males dispersed. (Table 5.5). Hares on the fence treatment had the highest dispersal rates (20.8% for males and 33.3% for females), and hares on control sites had the lowest (8.4% and 4.7% for males and females). We recorded only a few forays (0-9.5% of the population), and all except one occurred during the summer. The one exception was a control female in October 1994, who then dispersed and died within a week of the foray. Females made more forays than males. Female hares made the longest dispersal movements (Figure 5.9), with hares travelling 9.5 km, 8.6 km, and 6.1 km. The longest movements by males were 4.4 km, 4.2 km, and 4.1 km. 172 Table 5.4. Regression analyses of snowshoe hares' movement rates. Graphical representations of these data are in Figures 5.5 and 5.6. The regression tables indicate the intercept, the slopes for each variable, and t-test results for the hypothesis that each parameter differs from 0. Significant parameters are indicated in boldface. A. Summer movement rate as a function of spring hare density. ANOVA summary: F514=1.97, p<0.15,r2=0.41. (See Figure 5.5a). variable df estimate t p-value intercept 1 25592 0.80 0.43 food 1 2.99 0.12 0.91 fence 1 -24.24 -1.16 0.27 sex 1 29.29 2.20 0.04 year 1 -12.79 -0.80 0.44 spring density 1 -6.35 -0.15 0.88 B. Winter movement rate as a function of fall hare density. ANOVA summary: F 5 24=2.04, p<0.11,r2=0.30. (See Figure 5.5b). variable df estimate t p-value intercept 1 4327 0.55 0.58 food 1 -6.03 -1.17 0.25 fence 1 -3.76 -0.73 0.47 sex 1 13.03 2.95 0.007 year 1 -2.16 -0.55 0.59 fall density 1 2.09 0.83 0.41 C. Home range size as a function of spring hare density. ANOVA summary: F520=4.24, p<0.01, r2=0.51. (See Figure 5.5c). variable df estimate t p-value intercept 1 191 0.09 0.93 food 1 2.54 0.78 0.44 fence 1 -3.57 -1.47 0.16 sex 1 4.63 2.68 0.01 year 1 -0.09 -0.08 0.94 spring density 1 -5.79 -1.34 0.20 173 D. Summer movement rate as a function of home range size. A N O V A summary: F412=5.44, p<0.01,r2=0.64. (See Figure 5.6a). variable df estimate t p-value intercept 1 70.90 5.08 0.0002 food 1 -8.04 -0.69 0.50 fence 1 -30.34 -2.29 0.04 sex 1 18.89 1.46 0.17 home range size 1 1.75 1.39 0.19 E . Winter movement rate as a function of home range size. A N O V A summary: F 4 1 7 =l .33, p<0.30, r2=0.24. (See Figure 5.6b). variable df estimate t p-value intercept 1 15.44 2.51 0.02 food 1 -1.63 -0.33 0.74 fence 1 -3.24 -0.56 0.58 sex 1 10.93 1.86 0.08 home range size 1 -0.18 -0.33 0.75 174 Figure 5.5. Hares' movement rates as functions of their density. Each point represents an individual hare's movement rate or summer home range size. F F denotes food+fence. Spring densities are for March, and fall estimates for October. A. Summer movement rates. B. Winter movement rates. C. Summer home range sizes. 175 200 3 CD E > O E 160 120 80 40 o • • • A. S u m m e r m o v e m e n t • a 0.0 0.2 0.4 0.6 0.8 1.0 spring hare dens ity (hares\ha) 1.2 1.4 control f ema le control ma le food f ema le food ma le fence f ema le fence ma le F F f ema le F F ma le 60 50 t 40 cu to ~ 30 cu E cu 5 20 E 10 • o B. W in te r movemen t 2 3 4 5 fal l hare density (hares\ha) o • A A control f ema le control ma le food f e m a l e food ma l e fence f ema le fence ma le F F f ema le F F ma le 176 40 cu N to cu cn c CO 35 30 25 20 15 10 5 • o Oo °* • °% 0 * C. Home range size a • • • • • A A 0.0 0.2 0.4 0.6 0.8 1.0 spring hare density (hares\ha) 1.2 1.4 1.6 control female control male food female food male fence female fence male FF female FF male 177 Figure 5.6. The relationship between home range size and movement rates. E a c h point represents data for an indiv idual hare. H o m e range sizes are calculated f rom al l o f each animal's locations. F F denotes food+fence. A. Summer movement rates. B. W i n t e r movement rates. 178 CD 250 225 200 175 150 125 100 75 50 25 o A. Summer • o o • A A A A 0 10 20 30 home range size (ha) 40 control female control male food female food male fence female fence male FF female FF male 0) c cu E CD > o E 250 225 200 175 150 125 100 75 50 25 • • o B. Winter o A • 0 ° A • 10 20 30 home range size (ha) 40 control female control male food female food male fence female fence male FF female FF male 179 Figure 5.7. Time snowshoe hares spent at winter browse sites. Values are mean ± SEM; time estimates were derived from counts of faecal pellets at browse sites, transformed to time by a conversion factor of 35.2 pellets/hr (Figure 3.10). Means were calculated for each track, and the number of tracks is in each bar. 180 120 -r A. Females 100 80 60 40 20 0 111993-1994 ^ 1994-1995 • 1995-1996 I 14 I control food fence food+fence 140 - r 120 i 100 80 60 F 40 20 0 B. Males 13 control 5 food fence food+fence 181 Figure 5.8. Distance snowshoe hares travelled between browse sites in winter. Va lues are metres ± S E M . The years on the x-axis represent winters, f rom N o v e m b e r - A p r i l . Means were calculated for each track, f rom a total o f 163 tracks. 182 CO cn i m CD <U O CO « 3 -g .o OD i CO O) CD CD i i n CO CO o ' c CO cn co cn in cn m cn cn ro cn co cn m cn m cn cn cn i ro cn •a o o c o o o CM CM O O CM CM S9|BLUaj S9|BLU ( L U ) s\u\od asMOjq uaaMiaq aoueisjp 183 Table 5.5. Long distance movements of snowshoe hares. All animals were radio-collared for at least two weeks sometime between 31 March 1993 and 31 March 1996. Dispersal distance is the distance hares had to travel to exceed 2x the diameter of the average home range size for that treatment. Dispersers exceeded that distance* and did not return; hares which went on forays exceeded that distance but returned within a month1. dispersal total DISPERSAL FORAY distance (m) hares n % (95% CL) n % (95%CL) FEMALES control 650 95 8 8.4 (3.7-15.9) 3 3.2 (0.7-9.0) food 660 55 11 20.0(10.4-33.0) 4 7.3 (2.0-17.6) fence 690 21 7 33.3 (14.6-57.0) 2 9.5 (1.2-30.4) food + fence 580 76 17 22.4(13.6-33.4) 2 2.6 (0.3-9.2) MALES control 980 64 3 4.7 (0.9-13.1) 3 4.7 (0.9-13.1) food 930 49 7 14.3 (5.9-27.2) 1 2.0 (0.1-10.9) fence 730 24 5 20.8 (7.1-42.2) 0 0 food + fence 580 39 5 12.8 (4.3-27.4) 0 0 *Nine animals were included as dispersers if they exceeded two times their own home range diameter even though that was smaller than the average value listed here. t A n additional five animals exceeded the dispersal distance, but returned to their former home ranges after a month had elapsed. These hares are not counted above in the table; they were: two food 9-, one food ci", one food+fence ?, and one food+fence <?. 184 Figure 5.9. Dispersa l distances o f snowshoe hares. Da ta are for radio-col lared hares that dispersed between M a r c h 1993 and A p r i l 1996. 185 C/> £ " co ra E ra CD JB E E • 0 o o o o O OOCjOCDCD CD • • • !•• o o <f oo o • 0 o / o o <4>co • • o o CD | oo o o o oo CD CN C (tu>i) aoueisip lesjadsjp 186 58% of dispersing females and 80% of dispersing males moved less than 2 km. Of the 40 animals for which I could determine a season of dispersal, 37.5 % (females: 11/27; males: 4/13) dispersed between April and September, and the rest dispersed between October and March. DISCUSSION Factors influencing movement rates of snowshoe hares Sex and the presence of mammalian predators were the main factors affecting snowshoe hares' movements (Table 5.6). Males and females differed from each other in home range size, summer and winter movement rates, and proportion dispersing. The food hypothesis anticipated that increasing the food supply would reduce hares' movements. Females on food addition sites had lower winter movement rates than females on control sites, but that was the only support for the food hypothesis. For predation risk, hares should have had higher movement rates when risk was reduced; the opposite pattern occurred. Male hares protected from predators had smaller home ranges than males exposed to predators, and summer movement rates were smaller for protected hares of both sexes. Predation risk therefore had a clear effect, but one that was opposite to the patterns commonly found in other small mammals (Lima and Dill 1990). The density hypothesis predicted that movements would be inversely related to density. The increasing densities throughout the three years of the study did not result in systematic decreases in movements within each treatment type. Since the treatments differed in hare density, but movements did not differ with density, movement differences between hares on the various grids are more likely due to treatment than they are to density. Density alone does not appear to be a good explanation for the movement patterns of snowshoe hares. 187 CO rt «3 E C E E U rt cu . LH u £ U 3 " E fi -B 0) rt + r. X ) cu O X MH C II 0 3 fc 2 fc CU l l — CU .2 c 4 E ? E — X J o cu ° -5 T J rt cu ta t" 3 X J •c J 3 - f i X J b i2 cu c >• CU L -E © £ fi o g E E cd CU cd cu 2 cd X CU O co cd £ .a O X J c c co • — e g V 8 rt 3 CU CU X H t>0 c 'S rt CU E rt C O CU '5b E o cu .2 o X £ .52 cu rt s -= rt c X ) o 2 »-> D H 2 *-> o rt ,rt X ~ cu § ^ CU W-l E ° •4-4 C cj • — CU cu It: CJ OJ fi r-CU C j-H rt fe o X J P H rt §< x 2^ cu " 3 CJ CO C P cu 55 w *H ,UJ _fH X J <u • 3 X rt -jri CJ cu rt £ o E E => fe 00 X J • o m cu -4—4 cu rt es X J H .S CU O c CU rt o CO C 3 O cj co -CU X J CU -t—> X rt fi X 2 U rt P P H .5 CU es u I U o u es fc 2 « CU TJ C c o o rt CU c _o "-4—4 rt X J CU G *- O P H • — 1 -*—> JH rt X J X J o cu c»Z P H CU o CU LH ,«J H-H X J o 2 b V O f cu o a fi 3 CU 2 fc cu V fc fc X J o 2 V 4J o C CU •• fc CU O V g fc fc fc CU s CU w> c es © 33 3 _o rt X J CU LH P H X J CU o 2 X J CU •i—> CO ,2 o 2H U T J " o o fc V CU o 3 CU fc C CU « 0 . 3 O X J CU CU o c CU ^ cu CO W CU ^ T J o U X J o o fc V fc fc CU o 3 CU fc c cu E CU >• O ha CU E E 3 3 rt X J cu LH P H X J " o o cu o 3 CU LH ,u "+-4 X J o U V fc fc x J cu O cj O 3 fc 2 cu i fe O X J fi o fc 2; of " b S CU E CU >• o IM CU CU 3 O 3 ON V >/-> O N ON O N O N S CU O 3 e .cu o CU . ™ cu cn O u » es cu cu 3 o 3 CU o 3 LH , C J X J o CU o c LH .CU O 2 CU CJ C es cu CB © CQ 3 _o '-4—» rt X J CU X J o .0 X J CU 4—4 co CU -4-4 o 2 " b •b •b " b O f V V II V V O f O f O f O f b CU o 3 CU fc V fc fc xf o o fc V U Q -a 188 High movement rates as an anti-predator strategy Snowshoe hares had higher movement rates when they were exposed to lynx and coyote predation than when they were protected. That could occur if high movement rates of hares do not affect encounter rates, or encounter rates and mortality rates are not closely linked. Both of these scenarios are possible for snowshoe hares. Encounter rates are a function of both predators' movements and hares' movements; raptors had equal access to the treatments (the monofilament was largely ineffective), so I shall focus on terrestrial predators. During the low phase of the cycle, predators were uncommon: there were between 14 and 24 lynx and coyotes in the 350 km2 study area (Table 2.4, O'Donoghue 1997, O'Donoghue et al. 1997). During this phase, lynx and coyotes had large home ranges, they were active less than half of each day, they did not travel far daily, and they used areas with less spruce cover than did hares (O'Donoghue 1997). All of that means that any individual hare seldom encountered a predator, simply because predators were scarce on the landscape. Hares therefore would have periods of time with non-existent to low predation risk, then periods of high risk when a particular lynx or coyote was hunting in that area (M. O'Donoghue, pers. comm). For most of the time, then, hares' movement rates simply would not affect their encounter rates with predators, because predators were not there to be encountered. When predators were present, hares may have temporarily decreased their movements rates; anecdotally, when predators were on our trapping grids, we caught fewer hares, suggesting that they had reduced their movement or activity on a nightly basis in response to risk, since trappability is related to activity (Boulanger 1993). If hares during the decline phase also reduced movements when they encountered predators, the high number of predators could lead to permanently 189 reduced movements on areas where hares were exposed to predators, since encounters with predators and signs of predators would be frequent (Boutin et al. 1995, O'Donoghue et al. 1997). Encounter rates and mortality rates might also be decoupled. From snow-tracking data, O'Donoghue (1997) found that lynx successfully killed hares in only -30% of attempts; the rates for coyotes varied from -30-65% of attempts. These data are for hares actually chased by the predators, and it is also possible that hares and predators might come into close proximity yet without a chase. Therefore even if hares encounter predators, death is far from guaranteed. Hares might benefit from higher movement rates for a number of reasons. Rates of encounter with potential mates would increase, and ability to select from an array of foods and habitats would also presumably increase (Wolff 1980). Snowshoe hares may also be able to evade predators by moving more rather than by moving less. Among the leporids, snowshoe hares are unique in that they are the only forest-dwelling hare. Their congeners are open country animals that often escape predation by running rather than by hiding (e.g. brown hares, Lepus europaeus, Hutchings and Harris 1995). Snowshoe hares might move away whenever they encountered predators or predator sign, thus leading to higher movement rates for hares exposed to predators. Food availability's effect on hares' movements Hares did not change their movement rates or home range sizes in response to food addition, unlike many other small herbivores which have strong responses to food (Mares et al. 1976, 1982, Taitt and Krebs 1981, Sullivan etal. 1983, Boutin 1990). In another study of hares, food addition affected home range size in one of two years (Boutin 1984b), suggesting that environmental conditions or density may affect hares' response to food supplementation. Fine-scale food distribution did affect time at browse sites; in the winter of 1994-1995, hares spent 190 longer at browse sites, mainly due to their high consumption of bits of spruce branch that were clipped by red squirrels (Tamiasciurus hudsonicus); the squirrels ate the buds and dropped the remaining branchlets to the ground, and the hares ate many of these (Figures 3.4, 3.5). These spruce bits were highly clumped spatially, leading to longer browse times. Hares also did not change the distance between their browse sites when food was added. This pattern could occur because of constraints on foraging behaviour: if hares travelled too far between browse sites, they would be unable to sustain themselves; a lower boundary on hares' travel distances is set by the distribution of food plants. During the decline phase, hares on food+fence moved shorter distances between browse points than control hares did (Hik 1994), and hares on both sites moved two to four times longer distances than hares on all treatments during the low phase. This difference is probably due to the depletion of small twigs which occurs during the peak and decline phases, making food less available then (Smith et al. 1988). Sex differences in movement patterns Males had larger winter and summer movement rates than females, and on the two fenced treatments, males had larger home ranges. In contrast, females dispersed at a higher rate and for longer distances. This variation is probably due to the mating system of hares. Snowshoe hares can have as many as four litters in a summer, and breed immediately post-partum. Females breed synchronously (O'Donoghue 1994), and restrict their movements when they have young litters (Graf and Sinclair 1987, O'Donoghue and Bergman 1992, Allcock 1994, Jekielek 1996), perhaps so they can defend leverets from predation by arctic ground squirrels (Spermophilus parryii plesius) and red squirrels (O'Donoghue 1994, Stefan 1998). Several males attempt to breed with each female, and for males to locate and breed with many females, they need to travel farther than 191 the females; males increase their movements when females are oestrous (Chu 1996). Winter movement rates were lower than summer movement rates, and this pattern is again probably due to mating; in winter, hares are not breeding and there would be no reason for hares to seek each other out. The low temperatures in winter probably encourage reduced movements when possible, to reduce physiological stress. Various other studies have found no differences between male and female hares in terms of their movement patterns and home range sizes (O'Farrell 1965, Dolbeer and Clark 1975, Tompkins and Woehr 1979, Wolff 1980, Boutin 1984b, 1984c, Barta et al. 1989, Boulanger 1993), but several studies have found that males have larger movements and larger home ranges than females (Bider 1961, Ferron and Ouellet 1992, Hik 1994). Since these studies were done across the continent, at different times of the cycle, in different seasons, and with differing methods, it is difficult to know which factors cause the differences in behaviour between males and females. When differences occur, it is consistently the case that males move more than females, which supports the interpretation that mating behaviour is key to the differences. Dispersal The high, female-biased dispersal rates were not expected because mammalian dispersal is usually thought of as a juvenile and a male phenomenon (Gaines and McClenaghan 1980, Lidicker 1985a, Johnson and Gaines 1990, Wolff 1997). More juveniles than adults disperse in snowshoe hares; -50% of young hares dispersed prior to breeding during the early increase phase (Gillis 1997). It is not clear why adults would disperse at all, especially if dispersal decreases survival as has been commonly suggested (Tamarin 1978, Krebs 1978a, Taitt and Krebs 1985). 192 In other small mammal species, dispersal is rare from areas with food supplementation or protection from predators (Mares et al. 1976, Taitt and Krebs 1981, 1983). It could be that the high rates of dispersal from the manipulated areas are related to the higher densities there, but the adult dispersal I observed was higher than the dispersal observed in the increase and peak phases two cycles previously (Boutin et al. 1985), even though the densities I had were lower. Scale of movements and potential demographic effects I have presented data that represent within day, within season, and within lifetime movement patterns of snowshoe hares, but transient responses to risk—even as long as a week-were not detectable. Other small mammals have clear behavioural responses to risk that decay quickly after specific encounters with predators (Desy et al. 1990, Jedrzejewski and Jedzrejewska 1990, Kotler 1992). Hares show transitory responses to moonlight, and their nightly trappability is affected by the presence of predators (Gilbert and Boutin 1991, Boulanger 1993), so at least some of hares' responses to risk may be transitory. Although predation risk can affect behaviour for days (Kotler 1992) or even months (Hik 1994), many studies have not considered long-term demographic consequences of behavioural shifts (McNamara and Houston 1982, Houston et al. 1988). Predation rates can be affected by movements that increase encounter rates, and dispersal is thought to be costly for this reason (Cockburn and Lidicker 1983, Shields 1983, Moore and Ali 1984, Johnson and Gaines 1990). High daily movement rates or activity can also affect predation mortality (Schmutz et al. 1979, Dickman et al. 1991, Stuart-Smith and Boutin 1995, Koivunen et al. 1996). Changes in condition or fecundity probably require longer time periods (Boonstra et al. 1998a), which implies that transitory responses to immediate risk are unlikely to affect these parameters. 193 Synthesis and directions for further research Sex and predation risk account for most of the differences in movement patterns between hares during the low phase. Food supply did not affect hares' movements and density had slight effects; in this hares are unlike other small mammals, such as chipmunks, voles, and mice, which are strongly affected by food and density (Mares etal. 1976, 1982, Taitt and Krebs 1981, 1983, Sullivan et al. 1983, Boutin 1990, Wolff 1997). Predation risk usually causes small mammals to reduce their activity and movements (Brown et al. 1988, Desy et al. 1990, Jedrzejewski and Jedrzejewska 1990, Kotler et al. 1991, Saarikko 1992, Hughes and Ward 1993, Jedrzejewski et al. 1993, Fenn and Macdonald 1995), but hares either increased or did not change their movements. Hares' responses to predation risk were at the scales of daily movement rates and home range sizes. These scales are large both spatially and temporally, which suggests they may be able to influence demography through survival and fecundity changes. Short-term responses may occur which would affect survival, but these would be unlikely to influence fecundity. The observed patterns of change, however, do not support the supposition that predator-sensitive movement patterns can prolong the cyclic population low phase, especially since females showed no responses at several behavioural levels. 194 CHAPTER 6 SNOWSHOE HARE BEHAVIOUR AND DEMOGRAPHY DURING THE LOW PHASE: IMPLICATIONS FOR CYCLIC DYNAMICS Were there behavioural shifts with demographic consequences? In this thesis, I have examined the idea that predators might cause hares to adopt particular behaviours with negative physiological and demographic repercussions. Potentially, these behavioural patterns might affect demography enough to explain why snowshoe hare densities remain low for two to four years despite high food availabilities and low predator densities (Hik 1995, Boonstra et al. 1998). The data I have presented do not indicate that predators cause hares to adopt different behaviour, nor do they support the argument that hare behaviour during this phase leads to negative demographic consequences. The essence of the 'refugium' hypothesis links the observed habitat shifts through the cycle to changes in reproduction and survival (Wolff 1980, Hik 1994, 1995, Boonstra et al. 1998). Habitats are presumed to influence the stress, nutrition, fecundity, and survival of hares. Since food availability and predation risk influence habitat selection in other small mammals (Lima and Dill 1990) and were observed to do so for hares during the decline phase (Hik 1994), it seemed a reasonable supposition that predator-induced habitat selection during the low phase could produce the dietary and physiological responses that would ultimately lead to a lag in the population cycle. The data presented here, however, show that neither food supply nor risk of predation clearly influenced hare behaviour during the cyclic low phase. 195 Snowshoe hare behaviour during the low phase During the low phase, snowshoe hares used most if not all habitat types. They preferred areas with less open ground and areas with dense clusters of spruce and willow, but these patterns of use and preferences did not change with hare density, the addition of food, or the elimination of mammalian predators (Table 4.3). Patterns of mortality, habitat use by predators, and predator hunting success indicate that during the low phase habitats may not differ from each other in predation risk for hares (Figure 4.3, Murray et al. 1994, O'Donoghue 1997). Predation risk may have affected the movement patterns of hares, in that hares exposed to mammalian predators moved more than did hares protected from mammalian predators (Table 5.6). These results were inconsistent among years and sexes, however, so it is difficult to know whether hares genuinely increased their movement rates in response to the risk they experienced. For several indices of movement, hares did not differ between treatments or years. Females, in particular, showed few changes in their movement patterns in response to predation risk. That indicates that the supposed behaviour to demography link could not work via a mechanism of movements, since there were no behavioural differences to affect female fecundity. The refugium hypothesis suggests that hares obtain poorer food as a result of some behavioural change. Quite aside from the lack of habitat or movement changes that could be expected to lead to dietary shifts, two lines of evidence suggest that hares were not experiencing dietary limitations during the low phase. First, food was abundant, not only overall but also within most habitat types (Tables 2.3, 4.2). There did not appear to be a habitat type which would constrict hares' diets. Second, the dietary results showed that hares on control sites actually had lower fibre intake and higher protein intake than hares on areas with supplemental 196 food or without predators (Table 3.6). Although these results may be due to availability differences between the treatment sites, it is difficult to reconcile the observed data with the dietary patterns predicted by the predator-induced refugium hypothesis. Furthermore, hares did not change their foraging behaviour as their densities increased. These patterns are opposite to the idea that predation risk constrains hares to poorer diets. The strength of the interaction between behaviour and demography The low phase therefore does not appear to be prolonged by predator-induced behavioural limitation that causes demographic changes. This conclusion is troubling at first glance, since hare behaviour during the decline phase does appear to affect their survival and fecundity, and hares display increased stress levels in the presence of predators (Hik 1994, 1995, Rohner and Krebs 1996, Boonstra et al. 1998a). The reproductive and survival patterns also seemed to indicate that hare behaviour might be important to hare dynamics at this phase (Boonstra et al. 1998b). The alternative is that during the low phase, life is as good as it gets for hares: predators are scarce, food is abundant, and most behavioural patterns are functionally equivalent in their implications for survival and fecundity. The corollary to this interpretation is that the prolongation of the low phase is caused by some other factor than behavioural limitation. This alternative viewpoint has the following supporting evidence. First, the food and predator data clearly show that hares experience high food densities and low predator densities during the low phase (Chapter 2, Smith et al. 1988b, Keith 1990, O'Donoghue et al. 1997). Second, food densities appear to be high in many habitat types, and, more critically, food densities do not appear to be inadequate in areas that offer more protective cover (Table 4.2, C.J. Krebs, A.R.E. Sinclair and W. Hochachka, unpublished data). Third, predators utilized many of the same 197 habitats that hares did, and the success of their hunting was not clearly linked to the openness or refuge characteristics of the habitat type (Figure 4.3, O'Donoghue 1997). These three points taken together suggest that 'refuges' did not occur during the low phase, since safety was not a function of the use of particular habitats. Furthermore, it does not appear that hares had to trade-off risk and food, since risk was so low and food so abundant. As two further confirmations, hares—especially females-moved in similar ways on all experimental manipulations. In terms of hares' food choices, availability and season are better able to explain the observed patterns than are patterns of risk or food distribution. Taken altogether, the lack of hares' behavioural responses to food addition, reduction of predation risk, and changes in hare density, suggest that their behaviour simply was not strongly affected by these factors during the low phase. The questions then become whether and when behaviour affects demography during the hare cycle, how strong these effects are, and what suite of characteristics allow behaviour to influence demography. I suggest the following model (Figure 6.1). When hare densities are low, hares do not exhibit predator-sensitive behaviour, because most behaviours are functionally equivalent. As predator densities increase during the increase phase, to some predator density, pd, at which hares start employing predator-sensitive behaviours. Hares will continue using predator-sensitive behaviour until predator densities have again dropped below pd in the late decline phase or early low phase. The demographic consequences of the predator-sensitive behaviour will partially depend upon the food availability. During the increase phase, food availability is still high (fa), since the years of the low phase have been characterized by plant regrowth. Thus even though hares change their behaviour in response to predators, their nutrition is likely to remain good, and 198 Figure 6.1. A conceptual model of the role of behaviour in hare demography. The solid line indicates a hare cycle. During the low phase, predator densities are below the dangerous density, pd, at which hares shift their behaviour. Sometime during the increase phase, the predator density becomes high enough for hares to change their behaviour, but food density is still high (fa), so mortality is probably much more affected than fecundity. During the decline phase, however, food densities are low (fl), so mortality and fecundity are both likely to be affected by predator-induced behavioural shifts. 199 200 fecundities should similarly remain high. During the peak and decline phases, however, after the food plants have suffered heavy browsing ifl), predator-sensitive behaviour should exacerbate the problems of low food availability and cause even greater condition and fecundity changes. This model partially decouples demographic changes and behavioural changes, in that behaviour should shift quickly in response to predator densities, but demographic changes would probably be delayed at least until the next breeding season. Additionally, changes in behaviour probably only amplify rather than explain demographic trends. Anti-predator behaviour by the hares may be able to explain the asymmetry between the increase and decline phases (i.e. the rapidity of the decline) because of the conjoined negative effects of food limitation, heavy predation, and behavioural limitation during the decline, compared to the behavioural limitation and moderate predation of the increase phase. This model appears to fit the observed patterns of reproduction and survival (Keith 1990, Krebs et al. 1995, Stefan 1998). Hares had four litters during the late low phase, which corresponds to the model's prediction of 'ideal' conditions (Stefan 1998). During the increase, hares had three litters, which is consistent with predator-induced behavioural changes. During some decline years, hares have only two litters, consistent with behavioural limitation and food constraints. This model would explain why reproduction follows the pattern it does, whereas models looking only at food availability and predation have difficulty predicting the early reduction of litters. Survival patterns also support this model: starvation deaths—an extreme indicator of hares' nutritional status—very seldom occurred during the low phase (Table 2.5), but during the previous population cycle at Kluane, some hares starved in the late increase phase, and the highest proportion of hares starved in the peak and decline phases (Boutin et al. 1986). 201 The continuing problem of the extended low phase This potential outline for the interaction of behaviour and demography still leaves the problem of explaining why the low phase persists for several years. Adult survival was high and similar during the low phase (Figure 2.5). During the summers of 1994 and 1995, females had four litters, with similar annual natalities (Stefan 1998). Yet the rate of increase was higher from 1995 to 1996 than it was for 1994 to 1995 (Figure 2.3). If adult survival and natality did not differ between the two periods, that implicates juvenile survival (see also Keith and Windberg 1978, Krebs etal 1986, Gillis 1997, E.A. Gillis, D. Haydon, C. Stefan, and C.J. Krebs, ms. in prep). The juveniles in 1994 were part of a smaller total hare population; the predators may have been able to take a high enough percentage of them to keep the total population from increasing as much. In 1995, the starting hare population was higher, and hares may have been able to swamp the predator populations. Another explanation for changes in juvenile survival comes from the maternal stress hypothesis (Boonstra et al. 1998a, 1998b), in which it is argued that the stress experienced during the decline phase affects animals for several generations, thus producing the necessary lag in the cycle. It is difficult to tell whether stress experienced in the winter of 1991-1992 could still be affecting juvenile survival after summer 1994, since we lack data on the relationships between hares' stress levels, their reproduction and the survival of their leverets. The next generation: research questions for another cycle Juvenile survival is the critical missing link in studies of the snowshoe hare cycle. Juvenile survival is generally lower than adult survival (Keith 1990, Gillis 1997), but small changes in juvenile survival can lead to disproportionately large demographic changes (E.A. Gillis, D. 202 Haydon, C.I. Stefan and C.J. Krebs, ms. in prep). Since the prolonged low phase does not appear to be explicable via food availabilities, amount of predation, or behavioural limitation, describing the cyclic changes in juvenile survival may do much to clarify what can produce a cyclic lag. Second, the refugium and maternal stress hypotheses both invoke a linkage between food consumption, body condition, fecundity, and perhaps survival of the young (Hik 1995, Boonstra et al. 1998a, 1988b). These linkages are so far mainly correlational, whereas Stefan (unpublished data) found no direct link between age or condition and the litter sizes of females in an early increase phase. Explicit descriptions of this pathway would help to clarify whether such a mechanism could create the reproductive changes observed during the cycle. Behaviourally, tracing the linkage between the diet of hares and their habitat choices would strengthen the predator sensitive foraging argument (Hik 1995). Third, if the model described in this chapter is accurate, then hares should shift their behaviour in response to predators during the increase and peak phases. These behavioural shifts are not predicted to lead to large demographic changes, because of the still abundant food supply, whereas similar behaviour in a decline phase should have greater negative effects on fecundity. Information on hare behaviour during these phases may lead to greater insight about the relative roles of food plants, predators, and hares themselves in the generation of the cycle. 203 L I T E R A T U R E CITED Abrahams, M.V. and L.M. Dill. 1989. A determination of the energetic equivalence of the risk of predation. Ecology 70:999-1007. Abrams, PA. 1992. Why don't predators have positive effects on prey populations? Evolutionary Ecology 6:449-457. Abrams, PA. 1993. Why predation rate should not be proportional to predator density. Ecology 74:726-733. Adamcik, R.S. and L.B. Keith. 1978. Regional movements and mortality of great horned owls in relation to snowshoe hare fluctuations. Canadian Field-Naturalist 92:228-234. Adamcik, R.S., A.W. Todd and L.B. Keith. 1978. Demographic and dietary responses of great horned owls during a snowshoe hare cycle. Canadian Field-Naturalist 92:156-166. Allcock, K., 1994. Do predation risk and food affect the home ranges of snowshoe hares? BSc Honours Thesis, University of British Columbia. Anholt, B.R. and E.E. Werner. 1995. Interaction between food availability and predation mortality mediated by adaptive behavior. Ecology 76:2230-2234. Association of the Official Analytical Chemists. 1970. Official methods of analysis of the Association of Official Analytical Chemists. 11th ed., Washington, D.C . Barta, R.M., L.B. Keith and S.M. Fitzgerald. 1989. Demography of sympatric arctic and snowshoe hare populations: an experimental assessment of interspecific competition. Canadian Journal of Zoology 67:2762-2775. Beacham, T.D. 1981. Some demographic aspects of dispersers in fluctuating populations of the vole Microtus townsendii. Oikos 36:273-280. Belovsky, G.E. 1984. Snowshoe hare optimal foraging and its implications for population dynamics. Theoretical Population Biology 25:235-264. Belovsky, G.E. 1986. Optimal foraging and community structure: implications for a guild of generalist grassland herbivores. Oecologia 70:35-52. Belovsky, G.E. and O.J. Schmitz. 1994. Plant defenses and optimal foraging by mammalian herbivores. Journal of Mammalogy 75:816-832. 204 Bider, J.R. 1961. An ecological study of the hare Lepus americanus. Canadian Journal of Zoology 39:81-103. Bookhout, T A . 1965. Feeding coactions between snowshoe hares and white-tailed deer in northern Michigan. Transactions of the North American Wildlife and Natural Resources Conference 30:321-335. Boonstra, R., D. Hik, G.R. Singleton and A. Tinnikov. 1998a. The impact of changing predator risk on the ability to handle stress: population declines and lows in the snowshoe hare cycle. Ecology in press. Boonstra, R., C.J. Krebs and N.C. Stenseth. 1998b. Population cycles in small mammals: the problem of explaining the low phase. Ecology in press. Boonstra, R. and G.R. Singleton. 1993. Population declines in the snowshoe hare and the role of stress. General and Comparative Endocrinology 91:126-143. Boulanger, J., 1993. Evaluation of capture-recapture estimators using a cyclic snowshoe hare population. MSc Thesis, University of British Columbia. Boulanger, J. and C.J. Krebs. 1994. Comparison of capture-recapture estimators of snowshoe hare populations. Canadian Journal of Zoology 72:1800-1807. Boulanger, J.G. and G.C. White. 1990. A comparison of home-range estimators using Monte Carlo simulation. Journal of Wildlife Management 54:310-315. Boutin, S. 1980. Effect of spring removal experiments on the spacing behavior of female snowshoe hares. Canadian Journal of Zoology 58:2167-2174. Boutin, S. 1984a. The effect of conspecifics on juvenile survival and recruitment of snowshoe hares. Journal of Animal Ecology 53:623-637. Boutin, S. 1984b. Effect of late winter food addition on numbers and movements of snowshoe hares. Oecologia 62:393-400. Boutin, S. 1984c. Home range size and methods of estimating snowshoe hare densities. Acta Zoologica Fennica 171:275-278. Boutin, S. 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems, and the future. Canadian Journal of Zoology 68:203-220. Boutin, S., B.S. Gilbert, C.J. Krebs, A.R.E. Sinclair and J.N.M. Smith. 1985. The role of dispersal in the population dynamics of snowshoe hares. Canadian Journal of Zoology 63:106-115. 205 Boutin, S., CJ. Krebs, R. Boonstra, M.R.T. Dale, S.J. Hannon, K. Martin, A.R.E. Sinclair, J.N.M. Smith, R. Turkington, M. Blower, A. Byrom, F.I. Doyle, C. Doyle, D. Hik, L. Hofer, A. Hubbs, T. Karels, D.L. Murray, M. O'Donoghue, C. Rohner and S. Schweiger. 1995. Population changes of the vertebrate community during a snowshoe hare cycle in Canada's boreal forest. Oikos 74:69-80. Boutin, S., C.J. Krebs, A.R.E. Sinclair and J.N.M. Smith. 1986. Proximate causes of losses in a snowshoe hare population. Canadian Journal of Zoology 64:606-610. Brand, C.J. and L.B. Keith. 1979. Lynx demography during a snowshoe hare decline in Alberta. Journal of Wildlife Management 43:827-849. Brand, C.J., L.B. Keith and C A . Fischer. 1976. Lynx responses to changing snowshoe hare densities in central Alberta. Journal of Wildlife Management 40:416-428. Brown, J.S. 1988. Patch use as an indicator of habitat preference, predation risk, and competition. Behavioural Ecology and Sociobiology 22:37-47. Brown, J.S., B.P. Kotler, R.J. Smith and W.O. Wirtz II. 1988. The effects of owl predation on the foraging predation of heteromyid rodents. Oecologia 76:408-415. Bryant, J.P. 1981a. The regulation of snowshoe hare feeding behaviour during winter by plant antiherbivore chemistry, pp. 720-731 in Proceedings of the World Lagomorph Conference, eds. K. Myers and CD. Maclnnes. University of Guelph, Guelph. Bryant, J.P. 1981b. Phytochemical deterrence of snowshoe hare browsing by adventitious shoots of four Alaskan trees. Science 213:889-890. Bryant, J.P. and P.J. Kuropat. 1980. Selection of winter forage by subarctic browsing vertebrates: the role of plant chemistry. Annual Review of Ecology and Systematics 11:261-285. Bryant, J.P., P.B. Reichardt, T.P. Clausen, F.D. Provenza and P.J. Kuropat. 1992. Woody plant-mammal interactions, pp. 343-370 in Herbivores: Their Interactions with Secondary Metabolites, eds. G.E. Rosenthal and M.R. Berenbaum. Academic Press, Inc., London. Bryant, J.P., G.D. Wieland, T. Clausen and P. Kuropat. 1985. Interactions of snowshoe hare and feltleaf willow in Alaska. Ecology 66:1564-1593. Buehler, D.A. and L.B. Keith. 1982. Snowshoe hare distribution and habitat use in Wisconsin. Canadian Field-Naturalist 96:19-29. Carpenter, F.L. 1987. The study of territoriality: complexities and future directions. American Zoologist 27:401-409. 206 Cary, J.R. and L.B. Keith. 1979. Reproductive changes in the 10-year cycle of snowshoe hares. • Canadian Journal of Zoology 57:375-390. Cheeke, P.R. 1983. The significance of fiber in rabbit nutrition. Journal of Applied Rabbit Research 6:103-106. Cheeke, P.R. 1987. Rabbit Feeding and Nutrition. Academic Press, Inc., Orlando. Chitty, D. 1967. The natural selection of self-regulatory behaviour in animal populations. Proceedings of the Ecological Society of Australia 2:51-78. Chu, T.K.Y., 1996. Fluctuations in home range size of male snowshoe hares during summer breeding season. BSc Honours Thesis, University of British Columbia. Cockburn, A. and W.Z. Lidicker Jr. 1983. Microhabitat and population ecology of an herbivorous rodent, Microtus californicus. Oecologia 59:167-177. Conroy, M.J., L.W. Gysel and G.R. Dudderar. 1979. Habitat components of clear-cut areas for snowshoe hares in Michigan. Journal of Wildlife Management 43:680-690. Cox, E.W., R.A. Garrott and J.R. Cary. 1997. Effect of supplemental cover on survival of snowshoe hares and cottontail rabbits in patchy habitat. Canadian Journal of Zoology 75:1357-1363. Desy, E.A., G.O. Batzli and J. Liu. 1990. Effects of food and predation on behaviour of prairie voles: afield experiment. Oikos 58:159-168. de Vos, A. 1964. Food utilization of snowshoe hares on Manitoulin Island, Ontario. Journal of Forestry 62:238-244. Dickman, CR. 1992. Predation and habitat shift in the house mouse, Mus domesticus. Ecology 73:313-322. Dickman, CR., M. Predavec and A.J. Lynam. 1991. Differential predation of size and sex classes of mice by the barn owl, Tyto alba. Oikos 62:67-76. Dolbeer, R.A. and W. R. Clark. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. Doyle, F.I. and J.N.M. Smith. 1994. Population responses of northern goshawks to the 10-year cycle in numbers of snowshoe hares. Studies in Avian Biology 16:122-129. 207 Elton, C. and M. Nicholson. 1942. The ten-year cycle in numbers of the lynx in Canada. Journal of Animal Ecology 11:96-126. Fenn, M.G.P. and D.W. Macdonald. 1995. Use of middens by red foxes: risk reverses rhythms of rats. Journal of Mammalogy 76:130-136. Ferron, J. and J.-P. Ouellet. 1992. Daily partitioning of summer habitat and use of space by the snowshoe hare in southern boreal forest. Canadian Journal of Zoology 70:2178-2183. Finerty, J.P. 1980. The Population Ecology of Cycles in Small Mammals. Yale University Press, New Haven. Fox, J.F. and J.P. Bryant. 1984. Instability of the snowshoe hare and woody plant interaction. Oecologia 63:128-135. Fuller, T.K. and D.M. Heisey. 1986. Density-related changes in winter distribution of snowshoe hares in northcentral Minnesota. Journal of Wildlife Management 50:261-264. Gaines, M.S. and L.R. McClenaghan Jr. 1980. Dispersal in small mammals. Annual Review of Ecology and Systematics 11:163-196. Gilbert, B.S. and S. Boutin. 1991. Effect of moonlight on winter activity of snowshoe hares. Arctic and Alpine Research 23:61 -65. Gillis, E.A., 1997. Natal dispersal and post-weaning survival of juvenile snowshoe hares during a cyclic population increase. MSc Thesis, University of British Columbia. GLIM. 1985. Version 3.77. Royal Statistical Society, London. Godin, J.-G.J. 1990. Diet selection under the risk of predation. pp. 739-769 in Behavioural Mechanisms of Food Selection, ed. R.N. Hughes. Springer Verlag, Berlin. Graf, R.P. 1985. Social organization of snowshoe hares. Canadian Journal of Zoology 63:468-474. Graf, R.P. and A.R.E. Sinclair. 1987. Parental care and adult aggression toward juvenile snowshoe hares. Arctic 40:175-178. Green, R.G. and C A . Evans. 1940. Studies on a population cycle of snowshoe hares on the Lake Alexander area. I. Gross annual censuses, 1932-1939. Journal of Wildlife Management 4:220-238. 208 Harris, S., W.J. Cresswell, P.G. Forde, W.J. Trewhella, T. Woollard and S. Wray. 1990. Home-range analysis using radio-tracking data—a review of problems and techniques particularly as applied to the study of mammals. Mammal Review 20:97-123. Hik, D.S., 1994. Predation risk and the 10-year snowshoe hare cycle. Ph.D. Dissertation, University of British Columbia. Hik, D.S. 1995. Does risk of predation influence population dynamics? Evidence from the cyclic decline of snowshoe hares. Wildlife Research 22:115-129. Hixon, M.A. 1980. Food production and competitor density as the determinants of feeding territory size. American Naturalist 115:510-530. Hobbs, N.T. 1987. Fecal indices to dietary quality: a critique. Journal of Wildlife Management 51:317-320. Holter, J.B., G. Tyler and T. Walski. 1974. Nutrition of the snowshoe hare Lepus americanus. Canadian Journal of Zoology 52:1553-1558. Houston, A., C. Clark. J. McNamara and M. Mangel. 1988. Dynamic models in behavioural and evolutionary ecology. Nature 332:29-34. Hughes, J.J. and D. Ward. 1993. Predation risk and distance to cover affect foraging behaviour in Namib desert gerbils. Animal Behaviour 46:1243-1245. Hutchings, M.R. and S. Harris. 1995. Does hunting pressue affect the flushing behaviour of brown hares (Lepus europaeus)? Journal of Zoology 237:663-667. Ives, A.R. and A.P. Dobson. 1987. Antipredator behavior and the population dynamics of simple predator-prey systems. American Naturalist 130:431-447. Jedrzejewski, W. and B. Jedrzejewska. 1990. Effect of a predator's visit on the spatial distribution of bank voles: experiments with weasels. Canadian Journal of Zoology 68:660-666. Jedrzejewski, W., L. Rychlik and B. Jedrzejewska. 1993. Responses of bank voles to odours of seven species of predators: experimental data and their relevance to natural predator-vole relationships. Oikos 68:251-257. Jekielek, J., 1996. Changes in spatial distributions and movement rates of female snowshoe hares during the breeding season. BSc Honours Thesis, University of British Columbia. Jogia, M.K., A.R.E. Sinclair and R.J. Andersen. 1989. An antifeedant in balsam poplar inhibits browsing by snowshoe hares. Oecologia 79:189-192. 209 Johnson, M.L. and M.S. Gaines. 1990. Evolution of dispersal: Theoretical models and empirical tests using birds and mammals. Annual Review of Ecology and Systematics 21:449-480. Jones, W.T. 1986. Survivorship in philopatric and dispersing kangaroo rats (Dipodomys spectabilis). Ecology 67:202-207. Kaplan, E.L. and P. Meier. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 53:457-481. Keith, L.B. 1963. WUdlife's Ten-year Cycle. University of Wisconsin Press, Madison. Keith, L.B. 1966. Habitat vacancy during a snowshoe hare decline. Journal of Wildlife Management 30:828-832. Keith, L.B. 1974. Population dynamics of mammals. International Congress of Game Biologists . Keith, L.B. 1983. Role of food in hare population cycles. Oikos 40:385-395. Keith, L.B. 1990. Dynamics of snowshoe hare populations, pp. 119-195 in Current Mammalogy, ed. H.H. Genoways. Plenum Press, New York. Keith, L.B., S.E.M. Bloomer and T. Willebrand. 1993. Dynamics of a snowshoe hare population in fragmented habitat. Canadian Journal of Zoology 71:1385-1392. Keith, L.B., J.R. Cary, O.J. Rongstad and M.C. Brittingham. 1984. Demography and ecology of a declining snowshoe hare population. Wildlife Monographs 90:1-43. Keith, L.B., A.W. Todd, C.J. Brand, R.S. Adamcik and D.H. Rusch. 1977. An analysis of predation during a cyclic fluctuation of snowshoe hares. Proceedings of the International Congress of Game Biologists 13:151-175. Keith, L.B. and L.A. Windberg. 1978. A demographic analysis of the snowshoe hare cycle. Wildlife Monographs 58:1-70. Kie, J.G., J.A. Baldwin and C.J. Evans. 1994. CALHOME home range analysis program. California Department of Fish and Game. Koehler, G.M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Koivunen, V., E. Korpimaki and H. Hakkarainen. 1996. Differential avian predation on sex and size classes of small mammals: doomed surplus or dominant individuals? Annales Zoologica Fennici 33:293-301. 210 Kotler, B.P. 1984a. Harvesting rates and predatory risk in desert rodents: a comparison of two communities on different continents. Journal of Mammalogy 65:91-96. Kotler, B.P. 1984b. Risk of predation and the structure of desert rodent communities. Ecology 65:689-701. Kotler, B.P. 1992. Behavioral resource depression and decaying perceived risk of predation in two species of coexisting gerbils. Behavioural Ecology and Sociobiology 30:239-244. Kotler, B.P., L. Blaustein and J.S. Brown. 1992. Predator facilitation: the combined effect of snakes and owls on the foraging behavior of gerbils. Annales Zoologica Fennici 29:199-206. Kotler, B.P., J.S. Brown and O. Hasson. 1991. Factors affecting gerbil foraging behavior and rates of owl predation. Ecology 72:2249-2260. Krebs, C.J. 1978a. Aggression, dispersal, and cyclic changes in populations of small rodents, pp. 49-60 in Aggression, Dominance, and Individual Spacing, eds. L. Krames, P. Pliner and T. Alloway. Plenum Publishing, New York. Krebs, C.J. 1978b. A review of the Chitty hypothesis of population regulation. Canadian Journal of Zoology 56:2463-2480. Krebs, C.J. 1989. Ecological Methodology. HarperCollins, New York. Krebs, C.J. 1996. Population cycles revisited. Journal of Mammalogy 77:8-24. Krebs, 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 snowshoe hare cycle in Canada's Yukon? pp. 886-896 in Wildlife 2001: Populations, eds. D.R. McCullough and R.H. Barrett. Elsevier Applied Science, London. Krebs, C.J., S. Boutin, R. Boonstra, A.R.E. Sinclair, J.N.M. Smith, M.R.T. Dale, K. Martin and R. Turkington. 1995. Impact of food and predation on the snowshoe hare cycle. Science 269:1112-1115. 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-supplemented populations in the southern Yukon, 1976-84. Journal of Animal Ecology 55:963-982. Krebs, C.J., A.R.E. Sinclair and S. Boutin. 1996. Vertebrate community dynamics in the boreal forest of north-western Canada, pp. 155-161 in Frontiers of Population Ecology, eds. R.B. Floyd, A.W. Sheppard and P.J. DeBarro. CSIRO Publishing, Melbourne. 211 Krebs, CJ. and G.R. Singleton. 1993. Indices of condition for small mammals. Australian Journal of Zoology 41:317-323. Krohne, D.T. and A.B. Burgin. 1987. Relative success of residents and immigrants in Peromyscus leucopus. Holarctic Ecology 10:196-200. Lagos, V.O., L.C. Contreras, P.L. Meserve, J.R. Gutierrez and F.M. Jaksic. 1995. Effects of predation risk on space use by small mammals: a field experiment with a Neotropical rodent. Oikos 74:259-264. Larsen, K.W. and S.B. Boutin. 1994. Movements, survival, and settlement of red squirrel (Tamiasciurus hudsonicus) offspring. Ecology 75:214-223. Larter, N.C. 1992. Forage fibre analyses: a comparison of two techniques. Wildlife Research 19:289-293. Lidicker, W.Z., Jr. 1985a. An overview of dispersal in non-volant small mammals, pp. 369-385 in Migration: Mechanisms and Adaptive Significance, ed. M.A. Rankin. Contibutions in Marine Science, Suppl 27., Port Aransas, TX. Lidicker, W.Z., Jr. 1985b. Dispersal, pp. 420-454 in Biology of New World Microtus. ed. RH Tamarin, American Society of Mammalogists, Special Publication #8, Shippensburg PA . Lidicker, W.Z., Jr. 1985c. Population structuring as a factor in understanding microtine cycles. Acta Zoologica Fennica 173:23-27. Lima, S.L. and L.M. Dill. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68:619-640. Litvaitis, J.A. 1990. Differential habitat use by sexes of snowshoe hares (Lepus americanus). Journal of Mammalogy 71:520-523. Litvaitis, J.A., J.A. Sherburne and J.A. Bissonette. 1985a. A comparison of methods used to examine snowshoe hare habitat use. Journal of Wildlife Management 49:693-695. Litvaitis, J.A., J.A. Sherburne and J.A. Bissonette. 1985b. Influence of understory characteristics on snowshoe hare habitat use and density. Journal of Wildlife Management 49:866-873. Litvaitis, J.A., J.A. Sherburne and J.A. Bissonette. 1986. Bobcat habitat use and home range size in relation to prey density. Journal of Wildlife Management 50:110-117. Longland, W.S. and M.V. Price. 1991. Direct observations of owls and heteromyid rodents: can predation risk explain microhabitat use? Ecology 72:2261-2273. 212 Mares, M.A., T.E. Lacher Jr., M.R. Willig and N.A. Bitar. 1982. An experimental analysis of social spacing in Tamias striatus. Ecology 63:267-273. Mares, M A . , M.D. Watson and T.E. Lacher Jr. 1976. Home range perturbations in Tamias striatus. Oecologia 25:1-12. McNamara, J.M. 1990. The starvation-predation trade-off and some behavioural and ecological consequences, pp. 39-58 in Behavioural Mechanisms of Food Selection, ed. R.N. Hughes. Springer Verlag, Berlin. McNamara, J.M. and A.I. Houston. 1982. Short-term behaviour and lifetime fitness, pp. 60-87 in Functional Ontogeny, ed. D.J. McFarland. Pitman, London. McNamara, J.M. and A.I. Houston. 1987. Starvation and predation as factors limiting population size. Ecology 68:1515-1519. Mohr, CO. 1947. Table of equivalent populations of North American small mammals. American Midland Naturalist 37:223-249. Monthey, R.W. 1986. Responses of snowshoe hares, Lepus americanus, to timber harvesting in northern Maine. Canadian Field-Naturalist 100:568-570. Moore, J. and R. Ali. 1984. Are dispersal and inbreeding avoidance related? Animal Behaviour 32:94-112. Murray, D.L., S. Boutin and M. O'Donoghue. 1994. Winter habitat selection by lynx and coyotes in relation to snowshoe hare abundance. Canadian Journal of Zoology 72:1444-1451. Nams, V.O., N.F.G. Folkard and J.N.M. Smith. 1996. Nitrogen fertilization stimulates herbivory by snowshoe hares in the boreal forest. Canadian Journal of Zoology 74:196-199. Nellis, C H . and L.B. Keith. 1976. Population dynamics of coyotes in central Alberta, 1964-68. Journal of Wildlife Management 40:389-399. Norrdahl, K. and E. Korpimaki. 1995. Does predation risk constrain maturation in cyclic vole populations? Oikos 72:263-272. O'Donoghue, M. 1983. Seasonal habitat selection by snowshoe hare in eastern Maine. Transactions of the Northeast Section, the Wildlife Society 40:100-107. O'Donoghue, M. 1994. Early survival of juvenile snowshoe hares. Ecology 75:1582-1592. 213 O'Donoghue, M., 1997. Responses of coyotes and lynx to the snowshoe hare cycle. Ph.D. Dissertation, University of British Columbia. O'Donoghue, M. and C M . Bergman. 1992. Early movements and dispersal of juvenile snowshoe hares. Canadian Journal of Zoology 70:1787-1791. O'Donoghue, M., S. Boutin, CJ. Krebs and E.J. Hofer. 1997. Numerical responses of coyotes and lynx to the snowshoe hare cycle. Oikos 80:150-162. O'Donoghue, M. and CJ. Krebs. 1992. Effects of supplemental food on snowshoe hare reproduction and juvenile growth at a cyclic population peak. Journal of Animal Ecology 61:631-641. O'Farrell, T.P. 1965. Home range and ecology of snowshoe hares in interior Alaska. Journal of Mammalogy 46:406-418. Orr, CD. and D.G. Dodds. 1982. Snowshoe hare habitat preferences in Nova Scotia spruce-fir forests. Wildlife Society Bulletin 10:147-150. Otis, D., K. Burnham, G. White and D. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildlife Monographs 62:1-135. Parkinson, J.A. and S.E. Allen. 1975. A wet oxidation procedure suitable for the determination of nitrogen and mineral nutrients in biological material. Communications in Soil Science and Plant Analysis 6:1-11. Pease, J.L., R.H. Vowles and L.B. Keith. 1979. Interaction of snowshoe hares and woody vegetation. Journal of Wildlife Management 43:43-60. Pehrson, A. 1984. Faecal nitrogen as an index of hare browse quality. Canadian Journal of Zoology 62:510-513. Pietz, P.J. and J.R. Tester. 1983. Habitat selection by snowshoe hares in north central Minnesota. Journal of Wildlife Management 47:686-696. Pollock, K.H., S.R. Winterstein, C M . Bunck and P.D. Curtis. 1989a. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Pollock, K.H., S.R. Winterstein and M.J. Conroy. 1989b. Estimation and analysis of survival distributions for radio-tagged animals. Biometrics 45:99-109. 214 Rangen, S.A., A.W.L. Hawley and R.J. Hudson. 1994. Relationship of snowshoe hare feeding preferences to nutrient and tannin content of four conifers. Canadian Journal of Forest Research 24:240-245. Reichardt, P.B., J.P. Bryant, T.P. Clausen and G.D. Wieland. 1984. Defense of winter-dormant Alaska paper birch against snowshoe hares. Oecologia 65:58-69. Risenhoover, K.L., L A . Renecker and L.E. Morgantini. 1985. Effects of secondary metabolites from balsam poplar and paper birch on cellulose digestion. Journal of Wildlife Management 38:370-371. Rodgers, A.R. and A.R.E. Sinclair. 1997. Diet choice and nutrition of captive snowshoe hares (Lepus americanus): Interactions of energy, protein, and plant secondary compounds. Ecoscience 4:163-169. Rogowitz, G.L. 1988. Forage quality and use of reforested habitats by snowshoe hares. Canadian Journal of Zoology 66:2080-2083. Rohner, C. 1996. The numerical response of great horned owls to the snowshoe hare cycle: consequences of non-territorial 'floaters' on demography. Journal of Animal Ecology 65:359-370. Rohner, C. and F.I. Doyle. 1992. Methods of locating Great Horned Owl nests in the boreal forest. The Journal of Raptor Research 26:33-35. Rohner, C. and C.J. Krebs. 1996. Owl predation on snowshoe hares: consequences of antipredator behaviour. Oecologia 108:303-310. Saarikko, J. 1992. Risk of predation and foraging activity in shrews. Annales Zoologica Fennici 29:291-299. Schmitz, O.J. 1990. Management implications of foraging theory: evaluating deer supplemental feeding. Journal of Wildlife Management 54:522-532. Schmitz, O.J., D.S. Hik and A.R.E. Sinclair. 1992. Plant chemical defense and twig selection by snowshoe hare: an optimal foraging perspective. Oikos 65:295-300. Schmutz, S.M., D.A. Boag and J.K. Schmutz. 1979. Causes of the unequal sex ratio in populations of adult Richardson's ground squirrels. Canadian Journal of Zoology 57:1849-1855. Scott, D.P. and R.H. Yahner. 1989. Winter habitat and browse use by snowshoe hares, Lepus americanus, in a marginal habitat in Pennsylvania. Canadian Field-Naturalist 103:560-563. 215 Seber, G.A.F. 1982. The Estimation of Animal Abundance and Related Parameters. 2nd ed. Charles Griffin & Co., London. Shields, W.M. 1983. Optimal inbreeding and the evolution of philopatry. pp. 132-159 in The Ecology of Animal Movement, eds. I.R. Swingland and P.J. Greenwood. Clarendon Press, Oxford. Sievert, P.R. and L.B. Keith. 1985. Survival of snowshoe hares at a geographic range boundary. Journal of Wildlife Management 49:854-866. Sih, A. and R.D. Moore. 1990. Interacting effects of predator and prey behavior in determining diets, pp. 771-796 in Behavioural Mechanisms of Food Selection, ed. R.N. Hughes. Springer Verlag, Berlin. Sinclair, A.R.E. 1986. Testing multi-factor causes of population limitation: an illustration using snowshoe hares. Oikos 47:360-364. Sinclair, A.R.E. and P. Arcese. 1995. Population consequences of predation sensitive-foraging: the Serengeti wildebeest. Ecology 76:882-891. Sinclair, A.R.E. and J.M. Gosline. 1997. Solar activity and mammal cycles in the Northern hemisphere. American Naturalist 149:776-784. Sinclair, A.R.E., J.M. Gosline, G. Holdsworth, C.J. Krebs, S. Boutin, J.N.M. Smith, R. Boonstra and M. Dale. 1993. Can the solar cycle and climate synchronize the snowshoe hare cycle in Canada? Evidence from tree rings and ice cores. American Naturalist 141:173-198. Sinclair, A.R.E., M.K. Jogia and R.J. Andersen. 1988a. Camphor from juvenile white spruce as an antifeedant for snowshoe hares. Journal of Chemical Ecology 14:1505-1514. Sinclair, A.R.E., C.J. Krebs and J.N.M. Smith. 1982. Diet quality and food limitation in herbivores: the case of the snowshoe hare. Canadian Journal of Zoology 60:889-897. Sinclair, A.R.E., C.J. Krebs, J.N.M. Smith and S. Boutin. 1988b. Population biology of snowshoe hares. III. Nutrition, plant secondary compounds and food limitation. Journal of Animal Ecology 57:787-806. Sinclair, A.R.E. and J.N.M. Smith. 1984a. Protein digestion in snowshoe hares. Canadian Journal of Zoology 62:520-521. Sinclair, A.R.E. and J.N.M. Smith. 1984b. Do plant secondary compounds determine feeding preferences of snowshoe hares? Oecologia 61:403-410. 216 Smith, C H . 1983. Spatial trends in Canadian snowshoe hare, Lepus americanus, population cycles. Canadian Field-Naturalist 97:151-160. Smith, J.N.M., CJ. Krebs, A.R.E. Sinclair and R. Boonstra. 1988. Population biology of snowshoe hares. II. Interactions with winter food plants. Journal of Animal Ecology 57:269-286. Statsoft, I. 1995. Statistica for Windows, Version 5. Tulsa, OK . Stefan, CI. 1998. Reproduction and pre-weaning survival in a cyclic population of snowshoe hares. MSc Thesis, University of British Columbia. Stephens, D.W. and J.R. Krebs. 1986. Foraging Theory. Princeton University Press, Princeton. Stuart-Smith, A.K. and S. Boutin. 1995. Behavioural differences between surviving and depredated juvenile red squirrels. Ecoscience 2:34-40. Sullivan, T.P., D.S. Sullivan and CJ. Krebs. 1983. Demographic responses of a chipmunk (Eutamias townsendii) population with supplemental food. Journal of Animal Ecology 52:743-755. Taitt, M.J. and CJ. Krebs. 1981. The effect of extra food on small rodent populations: II. Voles (Microtus townsendii). Journal of Animal Ecology 50:125-137. Taitt, M.J. and CJ. Krebs. 1983. Predation, cover, and food manipulations during a spring decline of Microtus townsendii. Journal of Animal Ecology 52:837-848. Taitt, M.J. and CJ. Krebs. 1985. Population dynamics and cycles, pp. 567-620 in Biology of New World Microtus. Ed. RH Tamarin, American Society of Mammalogists, Special Publication #8, Shippensburg, PA . Tamarin, R.H. 1978. Dispersal, population regulation, and K-selection in field mice. American Naturalist 112:545-555. Telfer, E.S. 1972. Browse selection by deer and hares. Journal of Wildlife Management 36:1344-1349. Tilley, J.M.A. and R.A. Terry. 1963. A two-stage technique for the in vitro digestion of forage crops. British Grassland Society Journal 18:104-111. Todd, A.W. and L.B. Keith. 1983. Coyote demography during a snowshoe hare decline in Alberta. Journal of Wildlife Management 47:394-404. 217 Todd, A.W., L.B. Keith and C A . Fischer. 1981. Population ecology of coyotes during a fluctuation of snowshoe hares. Journal of Wildlife Management 45:629-640. Tompkins, D.B. and J.R. Woehr. 1979. Influence of habitat on movements and densities of snowshoe hares. Transactions of the Northeast Section, Wildlife Society, Fish & Wildlife Conference 36:169-175. Trostel, K., A.R.E. Sinclair, CJ . Walters and CJ. Krebs. 1987. Can predation cause the 10-year hare cycle? Oecologia 74:185-192. Vaughan, M.R. and L.B. Keith. 1981. Demographic response of experimental snowshoe hare populations to overwinter food shortage. Journal of Wildlife Management 45:354-380. Villafuerte, R., J.A. Litvaitis and D.F. Smith. 1997. Physiological responses by lagomorphs to resource limitations imposed by habitat fragmentation: implications for condition-sensitive predation. Canadian Journal of Zoology 75:148-151. Whittaker, M.E. and V.G. Thomas. 1983. Seasonal levels of fat and protein reserves of snowshoe hares in Ontario. Canadian Journal of Zoology 61:1339-1345. Windberg, L.A. and L.B. Keith. 1976a. Experimental analyses of dispersal in snowshoe hare populations. Canadian Journal of Zoology 54:2061 -2081. Windberg, L.A. and L.B. Keith. 1976b. Snowshoe hare population response to artificial high densities. Journal of Mammalogy 57:523-553. Wofford, H., J.L. Holechek, M.L. Galyean, J.D. Wallace and M. Cardenas. 1985. Evaluation of fecal indices to predict cattle diet quality. Journal of Range Management 38:450-454. Wolfe, M.L., N.V. Debyle, CS. Winchell and T.R. McCabe. 1982. Snowshoe hare cover relationships in northern Utah. Journal of Wildlife Management 46:662-670. Wolff, J.O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. Wolff, J.O. 1981. Refugia, dispersal, predation, and geographic variation in snowshoe hare cycles, pp. 441-449 in Proceedings of the World Lagomorph Conference, eds. K. Myers and CD. Maclnnes. University of Guelph, Guelph. Wolff, J.O. 1997. Population regulation in mammals: an evolutionary perspective. Journal of Animal Ecology 66:1-13. 218 Worton, BJ. 1987. A review of models of home range for animal movement. Ecological Modelling 38:277-298. Worton, B.J. 1989. Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70:164-168. Worton, B.J. 1995. Using Monte Carlo simulation to evaluate kernel-based home range estimates. Journal of Wildlife Management 59:794-800. Ylonen, H. 1989. Weasels Mustela nivalis suppress reproduction in cyclic bank voles Clethrionomys glareolus. Oikos 55:138-140. Ylonen, H. 1994. Vole cycles and antipredatory behaviour. Trends in Ecology and Evolution 9:426-430. Ylonen, H. and H. Ronkainen. 1994. Breeding suppression in the bank vole as antipredatory adaptation in a predictable environment. Evolutionary Ecology 8:658-666. 219 Appendix 1. Calculation of shrub biomass on each grid from belt transects and clip plots. 1. These data and formulae are courtesy of C.J. Krebs, A.R.E. Sinclair, and W. Hochachka, unpublished. 2. Five to ten 90-180 m long transects were conducted per grid. Within belts of 2 m, all willow and birch bushes were measured for number of stems >5 cm, circumference around the base, and height of the tallest stem. 3. 76 birch bushes and 104 willow bushes were measured, cut down, and weighed. Some bushes were taken from each treatment grid on which belt transects were conducted. 4. Regressions were calculated separately for each species, using circumference (C), height (H), and number of stems (N) to predict biomass of large twigs (>5mm). The grids were considered separately, and were given separate predictive equations if the data warranted separating them from an over-all equation. Geometric formulae were originally considered, but they were not considered acceptable because they predicted negative biomass in some cases, and their improvement in the fit to the data was marginal. Regressions were considered acceptable if they gave r2 values of » 0 . 9 0 and did not predict negative biomass values. 5. The prediction errors are usually <10 g/bush, but can occasionally be as large as 70 g/bush. Total biomass of large twigs per grid is therefore estimated with an error of -5-10%. 6. The actual regression equations are: birch, fence treatment: biomass = -37.73-N2 + 0.022-C-H + 334.9-N + 1.72-H-N birch, Gravel Pit food addition treatment: biomass = 7.10-N2 + 0.022-C-H - 64.64-N + 1.72-H-N birch, all other grids: biomass = -2.03-N2 + 0.022-C-H - 44.01-N + 1.72-H-N willow, all grids: biomass = 9.40-N2 + 0.043-C-H 7. The regressions yield biomass (wet weight) of large twigs per bush. These values are summed for all bushes contained within the belt transects, and kg/ha (dry weight) for each grid is calculated as: total g 1 kg 10000 m2 0.55 g dry width-length 1000 g 1 ha 1 g wet 220 where the the first term gives g/m2 (width and length are for the transect), the second and third terms convert that into kg/ha, and the final term converts wet weight into dry weight. 8. On four treatment grids, all bushes in -50 1 m 2 plots were cut down in May of each year to assess the ratio of <5mm twigs to total biomass. On the remaining three treatment grids, individual bushes were cut down in 1994 and 1995, to calculate the same ratio. 9. The annual biomass of <5mm twigs per grid was calculated as: •L ( S N s = S+L 1-where S is the biomass of <5mm twigs and L is the biomass of >5 mm twigs, both in kg\ha. 10. Confidence intervals were derived from the clip plot data, since these were more variable than the standing crop estimates. 221 Appendix 2. Nutritional quality of hares' major food species: whole twigs and herbs. Values are mean ± SEM; sample size in parentheses (wet weight and dry weight have same sample sizes as in dry wt. %). twig wet dry dry wt. as fibre protein size (mm) weight (g) weight (g) % of wet wt. (% of dry wt.) (% of dry wt.) birch 1 0.154 ±0.006 0.101 ±0.004 64 4 ± 1.03 (113) 72 73 + 0 67 (26) 5.94 ± 0.21 (10) 2 0.528 ± 0.034 0.334 ± 0.023 62 1 ± 1.22 (57) 69 90 ± 0 67 (21) 6.37 ± 0.28 (10) 3 1.441 ±0.093 0.876 ± 0.058 60 7 ± 0.55 (23) 74 50 + 0 50 (21) 5.37 ±0.14 (10) 4 3.131 ±0.166 1.930 ±0.097 61 8 ± 0.46 (21) 75 83 ± 0 76 (21) 5.01 ±0.17 (10) 5 5.986 ±0.349 3.670 ±0.212 61 4 ± 0.49 (20) 77 18 ± 0 54 (20) 4.33 ±0.23 (10) willow l 0.066 ± 0.004 0.036 ± 0.002 58 4 ± 2.00 (65) 65 96 + 0 88 (18) 4.81 ±0.13 (10) 2 0.454 ± 0.032 0.267 ±0.019 58 9 ± 0.73 (52) 59 75 ± 0 92 (20) 4.20 ±0.41 (10) 3 1.280 ±0.081 0.728 ± 0.045 57 0 ± 0.46 (32) 62 14 ± 1 19 (21) 4.91 ± 0.33 (10) 4 3.199 ±0.157 1.960 ±0.097 61 3 ± 0.57 (20) 66 43 ± 0 82 (21) 4.36 ±0.33 (10) 5 6.054 ± 0.343 3.659 ±0.202 60 6 ± 0.42 (20) 69 38 ± 0 83 (20) 3.80 ±0.13 (10) 6 11.96 ±0.555 7.092 ± 0.322 59 4 ±0.51 (20) 70 85 + 0 76 (20) 3.67 ±0.09 (10) 7 20.39 ±0.804 12.12 ±0.492 59 4 ±0.39 (20) 72 00 + 0 77 (20) . 3.50 ±0.20 (10) 8 29.26 ± 1.185 17.45 ±0.655 59 8 ± 0.33 (20) 73 90 ± 0 78 (20) 3.11 ±0.18 (10) soapberry l 0.062 ± 0.005 0.033 ± 0.003 59 3 ± 3.35 (44) 57 00 ± 0 98 (15) 9.62 ±0.74 (10) 2 0.347 ± 0.024 0.226 ±0.014 67 4 ± 0.74 (80) 59 23 ± 1 40 (20) 11.41 ±0.23 (10) 3 1.739 ±0.101 1.029 ±0.060 59 2 ±0.45 (21) 54 88 ± 0 45 (20) 11.53 ±0.21 (10) spruce 0.942 ± 0.034 0.483 ±0.017 51 4 ± 0.36 (32) 54 59 + 0 48 (32) 5.68 ±0.45 (10) fireweed summer 9.291 ±0.816 2.600 ±0.239 27 91 ± 0.69 (25) 46 72 ± 2 73 (25) 10.95 ± 0.73 (6) fall 1.690 ±0.147 1.155 ±0.088 69 35 ± 1.08 (25) 59 65 ± 1 12 (24) 3.43 ± 0.48 (7) lupine summer 3.893 ±0.414 0.942 ± 0.099 24 57 ± 0.37 (26) 55 04 ± 0 74 (25) 5.48 ± 0.26 (5) fall 0.273 ± 0.022 0.200 ± 0.032 61 85 ± 1.63 (25) 56 00 ± 3 05 (6) 3.97 ±0.13 (6) bluebell summer 6.083 ± 0.502 1.206 ±0.108 19 90 ± 0.46 (26) 58 79 ± 0 64 (26) 7.40 ± 0.45 (5) fall 1.984 ±0.152 1.153 ±0.077 59 89 ± 1.80 (25) 67 82 ± 2 93 (25) 6.18 ±0.69 (5) grass fall 3.128 ±0.208 2.006 ±0.132 65 13 ± 1.69 (25) 74 36 ± 0 62 (25) 4.56 ± 0.22 (9) chow 91.51 ± .050(3) 54.11 ± 1.50(18) 16* Manufacturer's minimum estimate. 222 +1 ^ o D CO -4—» c -a e I * •Si 2 « = a WD jg CU . -o o O O O o o © © © o o © o o O CN o CN VO CN O CO o CN Tt in Tt CO CN CN VO Tt r-m m m Tt d d d d d d d d d d d d d d d d +l +1 +1 +l +1 +1 +l +i +i +i +i +l +1 +1 +1 +1 Tt ON ON in in ON r-ON m oo o 00 VO m Tt CN vq CN ON r-Tt CN VO q CN CO in co CN CN Tt Tt Tt Tt co co co CN ON >n oo (26) (25) (29) (25) (49) (20) o o o O o © O in (20) oo t-VO Tt IT) vo in Tt m Tt 00 q oo r-- CN ON CN r- r--co ON 00 in 00 ON CN o VO d d d d d d d CN — d d d +i +i +i +i +i +1 +1 +l +1 +1 +1 +1 +1 +l +l +1 CO r-Tt co r-o CN 00 00 q o q in vo in Tt m in in O 00 o ON o q o q CO ON CO ON <N CN r-vd r- r-- oo r- od vo CN VO in VO in VO vd VO 00 vo r- CO r-in VO in VO (113) (117) (78) (50) m (32) (56) (50) m (50) (50) (50) (50) (oe) (30) (oe) CN CN o r- oo IT) m CN m 00 Tt oo CO in CN CN Tt CO CO oo in oo CN o Tt r--o CO oo in Tt d d d d d CN d d d d d d d d d d +l +l +l +l +l +1 +1 +l +1 +1 +l +1 +1 +1 +1 +1 o r- ON co Tt q o CO oq in VO o co vo o VO co vo VO VO Tt m CO m in m VO VO VO d VO 00 m od in od in VO ON m Tt o o oo o O vo CN o ON CN o r~ m o o o oo o O CO o m Tt o r-Tt o m m o VO vo CN ON o CN o m CO O CO in o d d d d d d d d d d d d d d d d +i +i +i +i +i +l +l +1 +l +l +1 +1 +1 +1 +i +1 o Tt m oo co oo CN IT) CN VO CN t— CN O t-- m CN o oo vo ON q 00 00 co <N CN CN CN ON co VO CN m CN <N co 00 d d d d " d d d d CO co d d d VD o o co o oo CO o oo Tt o VO ON o CN O o vo o in CN o r-o vo r-o CO ON O CO o co CN Tt CO m CO O VO O oo 00 O d d d d d d d d d d d d d d d d +l +1 +1 +l +1 +1 +1 +l +1 +1 +1 +1 +1 +l +1 +l Tt m 00 m CN CN ON in vo >n oo r-VO o o in o CO CO c--in Tt >n Tt CN ON CO ON CN in co m 00 vo Tt Tt CN Tt m o Tt d d d d CN d d d , — 1 i—i CN in in d d CN co Tt m illow CN co Tt m VO r-- oo ha S-CN CO o CN CO Tf o CN co Tt m vo t« o CN 223 O cd CO g .~° rt c '% T J c uT CU X ) E CU o CU Q OH tu CU X 3 3 CU X CU X E cu > o 3 ^ .3 O N <u ^ 3 "~ O >-TJ £ co 3 cd -& £ .£ cu ^ o O 3 cd ,cu L H < + H ~ + O T3 . C | _ TJ c - 3 cd 5 E ^ cu rt cu CO 3 CU J2 CO E £ 5 2 c*x « o O 3 S 1 rt O H = g rt cd T 3 3 °-O CU co CU •*—» CU o cd cd O <u CU i> M c cd cd TJ CU ^ e l m cu CO S cd CU 5. ^ T J CU OH cd D t-O U 3 * H 5 rt OH _ _ co T3 O O 3 CU •33 = co CU u cu CU 2i is cu E * . L H O N ^ O N 3 cu . . M cd 3 PQ -c CU 4—* 3 cu OH Q CU CO I * L H X • ° > , cu X S 3 K CU H-* cd E Q . s & a cd •> o <X. CO — H - 4 3 3 TJ CU H-t co i i CU * O 33 T3 ^ s cd O H £ < 2 CQ ft. cu Q B .cu 1 a <2 £ E £ PQ PH Q =2 £ S B CQ CM O B "8 a eg £ S s CQ ON Q o s s-fi- fi s ^ 2 o m •3-CN d d d d d +l +l +l +l +l o vo o o vo O N Tt- oo O N VO O N 1 1 vd CN 1 1 1 1 CN CN CN CN l l 1 1 o ,_ r~- o o CN CN 00 m o o- o OO m O N CN o in o rt —^  o o O N 00 d oo o o o CN 00 00 cn in m r-- in r-- in in O0 cn o o o O o d d d d d d d d +1 +l +l +l +l +1 +i +i +1 V) O N r- cn .—i vo r- O N '—1 O N vq O N in in vq i i —• —- —; —' CN CN —• —' 1 1 1 1 CN o , o oo 'Ct oo O vo in o o VO 00 Tt- r- ^ « ^ H CN CN o m in cn oq CN oq cn o o vo d cn CN in o o r- r- cn o m O N o O o O 00 t-- CN rt rt CN o O d d d d d d d +1 +i +l +l +1 +1 +1 +l o in o i—i o rt in CN oo oo •—i o o vq m m CN oo CN CN O N CN —* i CN CN CN 1 1 1 cn O cn cn o o CN CN O N CN o CN '—' rt in o m in o O N o rt —\ rt CN d rt o d vo o o VO oo CN m oo m m CN cn vo m OO in CN CN rt CN o o o d d d d d d d d d +1 +l +1 +1 +l +l +l +l +l +l in in vo o oo rt •"cl- O N o t> •—1 cn CN m cn in in CN CN "Ct CN ^H i i CN CN ^ —' CS oo CN o O N O N in cn O N in — CN O N vo O O N cn r~; q vq CN q m d oo d o O N O N in in cn in d CN in vo in VO m vo m vo O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N m 1 Tt" l in X C J l cn in ? — e 1 — cn i in C cn l Tl-1 m O N O N O N La O N O N O N S O N O N O N OS O N O N O N O N O N O N O N O N O N > O N O N O N O O N O N O N rt rt CQ rt >• rt rt t» — 224 £ B » PM cu Q C + cl I'* r-l £ B PH PH Q - si s, pp PH Q '•8 -«2 ^ PQ PH Q o e s-_ a S £ ON oo cn rt ON rt rt rt rt o d d d d d +1 +l +i +i +l 00 rt rt o m m m CN vq cn ! CN ! i i 1 1 l l i i cn cn cn • i i i i i j l l o CN o i o O o i ON rt CN i o o o l IT) cn 1 i CN vo O 1 CN cn cn rt VO o CN vd i o O o i i ON CN ! o o o i CN r- VO "fr m 00 r-- CN CN o CN cn CN >—c io rti cn rt CN cn d d d d d d d d d +l +1 +l +l +1 +i +i +l +1 ON 00 cn o ,—i m CN ON "fr VO o iri ON oo ON cn en ! CN i i CN CN CN cn 1 1 1 1 1 1 l l rt ON rt cn o o o • f r CN "fr 00 o o o o ^ H CN rt rti r- in rtl 00 CN ON o CN ON IT) rt 00 iri en rt o o CN o rt cn "fr 00 o o o o CN ON vo "fr CN vo r- vo CN o ON 00 o ON ON o cn CN o cn >—1 o o o o cn d d d d d d d d d d d d d +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 cn vo cn CN O ON o "fr vo CN r- oo cn — • CN VO CN ir> r- ON Ti- oq vq cn CN CN CN — • CN CN CN en — 1 —< cn r- r- O o cn rt O cn r- "fr o t-- o VO rti CN CN rti rti 00 ON o CN cn rti CN CN ir> cn o ON vq rt 00 IT) oq oo cn q ON rt cn vd o d rt VO fc rt vd o cn CN rt m ON 00 cn cn O 00 CN vo CN o ON IO oo cn r-ON CN o CN CN O rt O >—• o d d d d d d d d d d d +1 +1 +1 +1 +l +l +1 +1 +1 +l +l o m ON ,—i r» IT) rt IO cn ON ON r~- vq vo o rti oo rtl O "fr cn IT) cn CN cn CN ! i CN CN CN en CN cn 1 1 ! — vo ON in o O cn CN r- rt cn O o o VO rt rt rt r- CN CN cn vo ir> "fr CN CN vo ON oo 00 ON "fr CN ir> CN 00 CN CN rt d o o d d cn o o iri d VO ON IT) CN cn cn m vo cn IT) vo cn "fr IT) vo >> cn "fr IT) VO o\ ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON IM en ON ON ON ON 1 CN l cn l 1 m CJ i CN 1 cn l l ir> JO l CN l cn l "fr I IT) •2 ~ CH CN 1 cn I "fr 1 IT) ON ON ON ON u ON ON ON ON rB ON ON ON ON ca ON ON ON ON ON ON ON ON ON ON ON ON > ON ON ON ON o ON ON ON ON rti rt( ' — 1 rti PQ rtl ' — 1 ' — 1 rti 1> rti rti ' — 1 C/3 rti rti ' — 1 rti 225 c o CU 1 / 1. a a < i-9 Z e CU O u a| CU u -O !H cu o5-i a CU (~| C/3 fi B W cu os O o CN (N o Tt CN o Tt cn vo oo cn O N 00 CN CN Tt CN CN m — 00 •— d d d d d d d d d d d d d d d +l +1 +1 +1 +l +1 +1 +1 +1 +l +l +l +l +1 +1 cn Ti- oo vo vo r- r- cn Tt o cn r- CN r- t--cn en Tt in in 00 CN 00 r- oo oo 00 O Tt r-Tt in in in Tt Tt in Tt Tt Tt Tt Tt Tt Tt Tt 1 VO Tt vo CN VO m O N VO CN 00 O N cn cn Tt cn vq oo 00 r-- O vq O cn '—1 00 Tt cn .-I d d d Tt vd d +1 +1 +l +1 +1 +i +1 +1 +1 +1 +l +l +1 +1 r- O N Tt 00 O N Tt cn cn O N — O N oo VO CN — VO t - O N oo '—' cn Tt Tt CN r- CN O r--vd od cn CN 00 — CN —' —1 d t-^  Tt — vo m m m VD VO m vo vo VO VO VO VO VO VO ! cn r-- O N CN VO r- cn cn m oo CN VO cn cn o cn oo CN m O O in vo m CN VO O N cn Tt cn m CN cn Tt CN cn in CN CN in oo vo Tt 00 O cn cn oo CN VO O Tt o CN CN Tt Tt o CN O in Tt Tt CN m o CN (N o <N cn •— <N CN d d d d d d d d d d d d +l +1 +1 +l +1 +1 +i +1 +i +1 +l +1 m Tt vo cn cn CN vo o Tt r- oo t-- cn vo O N r~; cn O N in r~ 00 vd in in in Tt Tt in in in Tt Tt Tt Tt CN O N r- Tt o VO O N cn oo oo o o CN O N VO in r- cn CN O N CN — - d CN O d d CN CN * - H d CN +1 +1 +1 +1 +1 +1 "Tl +1 +1 +1 +1 +1 VO 00 00 r- 00 Tt VO ,—i CN in oo — CN vo cn in vq oq in cn oo Tt oq Tt — cn cn — d Tt O N Tt CN in m in vo vo VO VO r-- VD VO VO VO vo oo cn Tt CN O r- O N CN VO in O cn VD 00 o oo cn cn Tt r- 00 VD Tt CN cn CN r- O O cn m CN in VO r- Tt VO cn Tt m VO O N O N O N O N O N O N O N O N ^ '—' CN cn 4 m O N O N O N O N O N O N O N O N o c2 cn Tt m VO O N O N O N O N O N O N O N O N 1—1 ^ CN cn Tt in O N O N O N O N O N O N O N O N CU w B cn Tt m VO O N O N O N O N O N O N O N O N ^ '—1 ^ CN cn Tt lA O N O N O N O N O N O N O N O N CU C cn Tt m vo cu O N O N O N O N O N O N O N O N -0 O C N cn Tt in O N O N O N O N O N O N O N O N 226 IT) r- o CN vo cn cn ,—i r- CN o VO CN Tt cn Tt '—1 Tt p Tt Tt in m r- in d d d d d d d —- d d d +l +l +l +1 +l +1 +l +l +l +1 +l +1 +l m ITi oo O ON Tt oo CN ON Tt o 00 in ON '—' oo CN Tt ON 00 CN ON in vo in Tt vd vd CN en 00 d od r~ od ON od r- r- r- VD oo r- r- r- r-VO VO m o m ON CN Tt Tt vo oo CN CN r-- vo Tt Tt r- Tt Tt ON CN Tt ON ON Tt CN r- o cn o in VO in VO Tt vo Tt vo in in in ON d d d d CN d d d d d d +l +1 +1 +l +1 +1 +1 +l +l +1 +i +i +1 r- cn cn CN Tt o 00 oo o CN CN oo 00 m vo vo O Tt m o in cn Tt o vd vd vd cn OO Tt CN od CN ON r- r-- r- c-- r- r-- r-- 00 oo r- r- OO ON r- vo m CN VO vo cn m ON vo cn CN cn CN CN o Tt in cn oo CN CN VO oo Tt VO cn Tt oo in Tt Tt cn Tt vo Tt d —- d d —^  d d d d d d d +l +1 +l +l +1 +l +i +l +1 +1 +1 +1 in o O oo cn o CN cn Tt cn ON oq cn p o ON cn ON »—1 VO in cn cn vq d CN cn <N CN d vd d Tt oo oo t— r- r~- r-- 00 oo r- 00 r-oo VO o vo VD m c- ON cn m vo t--<N Tt cn in r- CN VO r- oo cn m CN cn vo cn cn cn O r- VO Tt vo Tt Tt CN Tt CN cn Tt cn d d d d d d d d d d d d +l +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +l +l m .—1 i> ON 00 ON VO cn vo VO o oo oo in in <—> r- r- in ON 00 Tt o CN vd c-^  cn vd Tt i> en d d d od r- r- r- oo 00 r- 00 oo CN oo *—* cn o VO (N *—* oo cn cn cn CN CN — in ON ON vo <—i ON r-C/3 c i t s a<3 '5 3 " U < 9 9 3 ~ tuo +i. O <- 3 i£ U ^ ^ < 9 9 ON § N S <U © H ^ 2 W Z CJ ON ON „ 3 o Q VO i C ON c3 227 rt CO 3 O CU O O 5 2 co b * o x P r-T 3 CU "CJ > M - L - •-j CU O T J . £3 J3 £ E £ £ « 2 ^ o eg cu cd cu cu 1 o X 0 rt ^ • — - — cd -*-» cu w § | cu £ X cd x cu I rt OJO c o cd 5 O c M LH 3 ° 5 c > w t d 4> X (30 — 3 . CO xi 3 co .2 s ts • - © -8 o 3 CU *r? bfi 2 rt 3 H-H 3 l_ cu cu o > fe cu a £ «-O cu 2 £ CU 3 O co rt E « 2 T3 CU co O u •B .2 r -*—» co cd 3 y • i J u 2 "2 •3 bb CN LH Al CU t 3 CU 3 cu T3 CU LH CU =S CO CU CU CO O CU „• X CO 4-* CU H_J II co C >- C H cu cu > -O co O CU o a rt ^> E >, * t O CU fe - f i O H O H rt co o rt -o X c cu rt ° „-co ,rt -> C^H S O rt 3 _cU CO "CJ co - CU LH M cu cd > O T 3 CU .S II a 3 : O , X o CU o w "fr 3 CU 3 a, ^ X T3 cu CU S co • - CU LH CO E 2 cu cu -> cu > O H LH CO L> CU > o o o co — CU CU o o 3 3 cd cd H-» H_» CO CO T 3 T 3 bfl . 3 cd X bb 3 CU > o o 2 2 b0 rt LH c2 CU _> "H—» CU rt ^ "g ! -S O H bo r •S tS rt CU 3 - 3 . . cd CO H-> LH 3 <-O cu • ~ H 4—1 £ | • s i ^ 3 2 § X ^ CU O X S £ — CU O cu •B CU .£ Q ; b 1 * CU CU CU C fa * + ~ o o fa fi CU CU CU fi cu fa fi - I CU X CN -T3 O © fa C cu o © fa fi - I CU © -w e © U fi CU s © U fi q o o q in o o o o "fr in • r i CN o C N c n o o d c n "fr CN CN m c n C N ! VD in q CN v q CN c n rt rt CN "fr CN o o O N c n O N v d o o in v o "fr c n c n c n "fr "fr v d "fr C N c n • O N O N CO "fr N O CN o o v o O N m c n CN in o "fr c n o o O N c n r-c n in o c n "fr in o in o in 100.0 00 O o "fr CN c n CN "fr "fr d CN "fr "fr o o " fr 100.0 q in 00 C N o v o o q CN q 00 o CN o o CN o o O N "fr O N o o O O d v o O N o o m "fr c n o o c n "fr O O o o o "fr v o 00 in v o c n o "fr VO "cl-00 VO o o c n m CN "fr "fr m CN c n c n CN c n m m ON o o "fr m CN o cn o o "fr c n c n oo O N O o "fr r--c n c n o o "fr rt X _s 'es > ca o o d o ^ o d "fr o o d v q oo oo CN in oo «n C N c n r- CN r~~ oo rt r - c n CN ON q O O e n P ^ ° ) in C N CN O N O N t O c n c n " f r r t . c n CO c O N to O o c n 'S ei CJ t-- O N r~ _ ) d 00 v d o o o o r ~ >-. CU rt c n c n fi fi c n v o O s <r> rr 1/5 CO ©> O N O N O N O N O N CO CU i—i •<-< H-< CO CO CO CU CU CU JS U u u CN O N o O CN r - O N "fr v d CN O N v o "fr "fr c n r - O N 00 in O rt CN o o rt o o c n c n c n O N m c n r -O N O o o c n v d O N o q O N d 00 c n m CN v o v o m rt CN CN v o CN o O CN CN CN c n rt O N o O N in d m CN "fr C N r - o C N CN m >r, O N O N O N O N O N O N M M ^ M ^ M — — CU CJ CU ca ca ca rt rt O C N ON ~ 1 S rt. i r -vo vo CU CO • © X "co CU ca "fr O N o r-~ CN e n q "fr in rt CN in d ^ - f r "fr CN "fr O "fr CN O "fr "fr CN in o o "fr O CN CN " fr "fr r-- r-~ in v d "fr VO ^ CN in oo O N rt N O "fr ON t— 0O c-^  v d in m rt m oo o O N in in rt rt CN ^ O N oo r--rt. N O in i> q t-^ v d O N e n v o o O d in oq oo 00 C— ON m rt "fr v q "fr CN v q VO o o "fr in CN e n "fr i> CN d "fr O N rt CN VO 00 O N CN c n CN e n d CN e n CN e n O N e n c n in "fr e n O N CN CN e n 1993 1994 1995 it 1994 it 1995 it 1996 — B B B is is is C M '% wi wi 228 Appendix 8. Snowshoe hare home range sizes. Summer ranges include only telemetry data taken during April-September. Telemetry ranges include all telemetry fixes for each hare. The total ranges include all telemetry and trapping locations for each hare. All ranges have a minimum of 12 locations. Values are mean ± SE. MCP100: minimum convex polygon (Harris et al. 1990), using all locations; MCP95, only 95% of the locations used; adaptive kernel (Worton 1987, 1989) using 95% and 50% of locations. FEMALES hares mcp 100 (ha) mcp 95 (ha) kernel 95 (ha) kernel 50 (ha) CONTROL summer 1993 1 11.76 9.21 11.76 3.68 summer 1994 7 11.90 + 3.09 8.80 + 2.50 16.94 ± 3.31 2.35 + 0.74 summer 1995 11 7.33 + 0.98 5.61 ± 0.79 10.99 + 1.42 1.80 + 0.34 telemetry 23 11.10 + 1.46 7.18 ± 1.02 11.96 + 1.59 1.81 + 0.26 trapping 4 6.06 + 0.65 6.06 ± 0.65 13.32 + 1.53 2.17 + 0.66 total 24 14.26 + 1.95 8.38 ± 1.12 15.82 ± 1.98 2.30 ± 0.43 FOOD summer 1994 2 4.70 + 1.16 4.70 + 1.16 9.70 + 3.35 2.57 ± 0.77 summer 1995 15 8.24 + 1.40 6.35 + 1.29 12.29 + 2.24 1.59 ± 0.32 telemetry 16 9.64 + 1.22 6.85 + 1.04 11.50 + 1.80 1.73 ± 0.30 trapping 3 1.83 + 0.76 1.83 ± 0.76 3.15 + 1.23 0.68 ± 0.26 total 20 12.21 ± 2.07 8.49 ± 2.03 17.86 + 4.82 2.07 ± 0.39 FENCE summer 1993 3 7.30 + 2.26 5.57 ± 1.55 11.16 + 4.58 1.34 ± 0.20 summer 1994 4 4.69 ± 1.59 4.69 ± 1.59 9.95 + 3.32 1.24 ± 0.39 summer 1995 4 18.03 + 7.81 7.52 ± 3.23 . 14.74 + 6.18 1.70 ± 0.41 telemetry 8 15.28 + 3.94 7.74 ± 1.31 15.07 ± 2.91 1.38 ± 0.13 trapping 4 7.67 ± 1.72 7.08 ± 1.79 15.36 ± 3.86 1.48 ± 0.19 total 8 19.50 ± 3.99 9.41 ± 1.50 16.56 ± 3.02 1.66 ± 0.19 FOOD + FENCE summer 1993 16 3.49 ± 1.16 1.76 ± 0.39 5.19 ± 1.68 0.57 ± .18 summer 1994 5 2.23 ± 0.34 2.23 ± 0.34 4.23 ± 0.67 1.13 ± 0.28 summer 1995 5 7.03 ± 0.99 4.99 ± 0.58 9.93 ± 1.13 1.09 ± 0.24 telemetry 22 6.08 ± 1.22 3.52 ± 0.58 8.17 ± 2.20 5.19 ± 1.68 trapping 11 7.18 ± 1.36 6.27 ± 1.13 12.95 ± 2.63 0.75 ± 0.20 total 34 9.13 ± 1.46 6.57 ± 1.03 12.45 ± 2.19 1.49 ± 0.39 229 Appendix 8 cont. Snowshoe hare home range sizes. MALES hares mcp 100 (ha) mcp 95 (ha) kernel 95 (ha) kernel 50 (ha) CONTROL summer 1993 4 21.09 + 4.61 15.17 ± 4.18 24.71 ± 4.65 3.67 + 1.13 summer 1994 3 13.14 + 2.06 13.14 + 2.06 20.29 ± 10.26 3.63 + 2.21 summer 1995 9 20.44 + 4.42 13.50 + 1.55 25.61 ± 3.70 4.95 + 1.15 telemetry 17 30.44 7.09 17.49 ± 1.55 26.54 ± 1.99 4.26 + 0.34 trapping 8 12.37 ± 2.16 11.61 ± 1.87 19.50 ± 2.56 3.11 + 0.56 total 19 32.45 ± 6.28 18.88 ± 1.81 29.55 ± 3.00 5.29 0.89 FOOD summer 1994 2 24.07 ± 12.35 24.07 ± 12.35 44.00 ± 21.50 10.45 + 1.96 summer 1995 3 14.49 ± 5.67 11.36 ± 4.09 23.19 ± 8.06 3.59 ± 1.11 telemetry 7 23.87 ± 6.23 17.99 ± 2.91 35.48 ± 8.55 10.69 + 5.67 trapping 2 2.47 ± 1.22 2.47 ± 1.22 4.41 ± 0.11 0.52 + 0.04 total 11 25.57 ± 6.24 17.01 ± 2.70 32.96 ± 5.90 4.73 ± 0.92 FENCE summer 1993 2 11.54 ± 4.05 8.83 ± 3.32 16.35 ± 6.53 1.99 ± 0.79 summer 1994 1 3.07 3.07 5.53 2.02 summer 1995 5 10.27 ± 2.80 8.29 ± 2.19 15.97 ± 3.66 3.05 ± 1.34 telemetry 9 12.09 ± 2.40 9.28 ± 1.70 17.02 ± 2.81 2.28 ± 0.59 trapping 3 6.65 ± 1.65 5.73 ± 1.03 9.76 ± 3.27 2.32 ± 0.89 total 12 13.02 ± 2.41 10.50 ± 1.88 18.95 ± 3.25 3.02 ± 0.47 FOOD + FENCE summer 1993 1 2.14 2.14 3.52 0.76 summer 1995 3 4.77 ± 2.59 5.01 ± 1.16 8.96 ± 2.12 0.89 ± 0.33 telemetry 7 6.10 ± 1.47 4.85 ± 1.03 9.32 ± 2.13 0.88 ± 0.27 trapping 5 6.04 ± 2.22 6.02 ± 2.21 13.28 ± 4.65 1.27 ± 0.75 total 9 9.10 ± 1.99 6.59 ± 1.43 15.65 ± 4.81 1.41 ± 0.26 230 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088796/manifest

Comment

Related Items