Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An experimental analysis of juvenile survival and dispersal in snowshoe hares Boutin, Stanley A. 1983

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

Item Metadata

Download

Media
831-UBC_1983_A1 B69.pdf [ 5.85MB ]
Metadata
JSON: 831-1.0095839.json
JSON-LD: 831-1.0095839-ld.json
RDF/XML (Pretty): 831-1.0095839-rdf.xml
RDF/JSON: 831-1.0095839-rdf.json
Turtle: 831-1.0095839-turtle.txt
N-Triples: 831-1.0095839-rdf-ntriples.txt
Original Record: 831-1.0095839-source.json
Full Text
831-1.0095839-fulltext.txt
Citation
831-1.0095839.ris

Full Text

e. I EXPERIMENTAL ANALYSIS OF JUVENILE SURVIVAL AND DISPERSAL SNOWSHOE HARES by STAN A. BOUTIN B.Sc.(Hons.) Zoology/Edmonton,University Of Alberta,1977 M.Sc. Zoology,Vancouver,the University Of B r i t i s h Columbia,1979 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA @ 1983 THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in Stan A. Boutin, 1983 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by h i s or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of ^-pd (d~ci(£) The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6 (3/81) ABSTRACT If spacing behaviour of snowshoe hares l i m i t s juvenile survival and recruitment during summer, removal of t h i s behaviour should produce an increase in these parameters. During the summers of 1980 and 1981 I removed a l l adults from an 8 ha trapping grid and a l l f i r s t l i t t e r juveniles from another. Experiments were conducted in the ^southwestern Yukon during a period when hare populations were at peak densi t i e s . The experimental removals did not increase survival, but recruitment r e l a t i v e to control areas was higher to the adult removal grid in 1980 and to both the adult removal and juvenile removal grids in 1981. To determine whether juveniles trapped for the f i r s t time were residents or immigrants, I implanted adult females with ,calcium-45. This was passed to nursing young and could be detected by s c i n t i l l a t i o n counting of a sample of bone tissue taken from new r e c r u i t s . Any juvenile without radioactive calcium was classed as an immigrant. The increase in recruitment on the removal areas was due to increased immigration. The number of resident recru i t s was equal on a l l study areas. Results support the hypothesis that spacing behaviour l i m i t s juvenile immigration but not s u r v i v a l . However, immigration to control areas was also high with immigrants making up 70% of the t o t a l number of juveniles present on the areas in October. If food l i m i t s snowshoe hare numbers, addition of food should lead to increased numbers through higher survival and immigration. If food supply influences spacing behaviour of hares, home range size should decrease with food addition. I supplied peak (1980) and declining (1981) hare populations on 8 ha grids (one in 1980 and 2 in 1981) with laboratory rabbit chow for 1 - 4 months during March through June. Population size was determined by live-trapping and movements of animals were monitored by radio telemetry. Food addition decreased weight loss and improved survival of hares in both years. Onset of breeding was advanced in males but not females. In 1980, the number of males on the food addition area was 1.4 times higher than those on the control area while the number of females did not d i f f e r . In 1981, numbers of males and females were up to 3.6 and 3.2 times higher respectively on the food addition area as compared to those on ;the control area. The differences were due mainly to increased immigration. Residents responded to food addition by decreasing home range size in 1980 but not in 1981. Movement of immigrants, as monitored by telemetry, to the food addition area indicated that some established home ranges there while others returned to their old home ranges. Results support the hypothesis that hare densities are limited by winter food supply during the early decline phase of the cycle and possibly during the peak phase as well. A decrease in home range size was not necessary for immigration to occur. To examine the rel a t i o n s h i p of dispersal to changes in snowshoe hare numbers, I monitored dispersal of hares during a population increase, peak, and early decline (1978-1982). Two methods were used: 1) a conventional removal grid in which a l l animals caught each trapping session were removed and 2) telemetry monitoring of radio-collared individuals. The number of animals caught on the removal area was correlated with density on the control area but per capita dispersal rate was not. Both the number of dispersers and the per capita dispersal rate were highest during the period of peak densities on the control area. Dispersal, as measured by the removal grid , was not density dependent. Only 23 of 265 radio-collared animals dispersed during the study. Dispersal accounted for an average of 11% of the losses of radio-collared animals during the population decline. Results from both telemetry and the removal gri d indicated that the decline in hare numbers was not due to d i s p e r s a l . The amount of dispersal as determined by the removal grid was much higher than that determined by telemetry. The difference was more pronounced during the population peak and early decline. This was due to the removal grid over-estimating the average amount of dispersal that was occuring because i t attracted animals to i t . These results point to the need to be more c r i t i c a l of the underlying assumptions of the removal grid method as a way of monitoring d i s p e r s a l . iv TABLE OF CONTENTS ABSTRACT i i LIST OF TABLES v i LIST OF FIGURES , ..vii ACKNOWLEDGEMENTS v i i i GENERAL INTRODUCTION . 2 General Spacing! Behaviour of Snowshoe Hares 6 SECTION 1. THE EFFECT OF SPACING BEHAVIOUR ON JUVENILE SURVIVAL AND RECRUITMENT 7 Introduction 7 Study Area ...... 9 Methods . . 14 Trapping 14 Radioactive Implants . .... 15 Experimental Design 16 Results 19 Numbers Prior to Removals 19 Removals 19 Changes in Numbers 25 Survival 28 Mean Weight in F a l l 30 Effect of Removals on Recruitment 30 Immigration to Control Populations 44 . Discussion 50 Survival . . . 50 Immigration 51 Immigration to Control Populations 54 Social Organization and Immigration 55 SECTION 2. EFFECT of LATE WINTER FOOD ADDITION on SNOWSHOE HARE SPACING BEHAVIOUR and NUMBERS 58 Introduction . .. . ... 58 Study Area . ... 60 Methods 60 Trapping ; . . . . 60 Experimental Design 63 Telemetry 63 Results 66 Changes in Numbers 66 Immigration 74 Survival . ... 80 Weight Loss 83 Onset of Breeding 83 Home Range Size and Dispersal 87 Use of the Food Addition Area 91 Movements of Immigrants 94 Discussion 99 Food Supply and Hare Movements 105 SECTION 3. THE RELATIONSHIP OF DISPERSAL TO THE POPULATION DYNAMICS OF SNOWSHOE HARES 108 Introduction 1 08 Study Area 110 Methods 113 Trapping ! 113 Telemetry 113 Results . . . . . 115 Dispersal and Density .....115 Dispersal as Determined by Telemetry .......122 Population Losses and Dispersal 124 Sex Ratio of Dispersers ......125 Age of Dispersers 127 Weight and Body Size of Dispersers ... 127 Di sc us s i on 1 37 Characte r i s t i c s of Dispersers 139 Habitat Heterogeneity and Dispersal 143 Methods of Studying Dispersal 145 GENERAL DISCUSSION , 150 Hypotheses Explaining Population Cycles .150 Dispersal . . ,. . . 151 LITERATURE CITED . ; 154 ! I LIST OF TABLES Table 1.1 Predictions of hypotheses 10 Table 1.2 Percentage cover of plant species on the study areas 13 Table 1.3 Minimum number a l i v e on study areas prior to removals. , 22 Table 1.4 Number of residents recruited per female on each study area 29 Table 1.5 Minimum ,28 day survival rates of juveniles during the summer. 31 Table 1.6 Minimum 28 day survival rates of juveniles during the f a l l 32 Table 1.7 Total number of juveniles recruited to the study areas . . 36 Table 1.8 Proportion of l a c t a t i n g females receiving calcium-4 5 . 37 Table 1.9 Ratio of immigrants to dispersers. 48 Table 1.10 Minimum .28 day survival rates of resident and immigrant juveniles. . 49 Table 2.1 Mean net weight change of hares 84 Table 2.2 Time periods used to calculate home range si z e . .. 88 Table 2.3 Summary of experimental results , 100 Table 3.1 Population size on S i l v e r Creek and dispersal rates to the Removal. , .119 Table 3.2 Correlation c o e f f i c i e n t s between population size ! on S i l v e r Creek and dispersal rate to the Removal. ...... .120 Table 3.3 Fates of radio co l l a r e d hares .123 Table 3.4 Sex r a t i o of hares on Si l v e r Creek and the Removal , ...126 Table 3.5 Proportion of t o t a l number , caught that were adults on S i l v e r Creek and the Removal .128 ! v i i i i i LIST OF FIGURES Figure 1.1 Location of study areas 11 Figure 1.2 Experimental design .17 Figure 1.3 Changes in numbers on the control areas .20 Figure 1.4 Pattern of removals 23 Figure 1.5 Changes in numbers during the removal period. .. 26 Figure 1.6 Mean weight of juveniles caught in Mid- : September 33 Figure 1.. 7 Number of immigrants caught before September 30. i 39 Figure 1.8 Monthly number of immigrant re c r u i t s 42 Figure 1.9 Proportion of juveniles that were resident on SC 45 Figure 2.1 Location of study areas. 61 Figure 2.2 Experimental design , 64 Figure 2.3 Minimum number a l i v e on study areas from 1978-1981 67 Figure 2.4 Changes in numbers in response to food addition in 1980 70 Figure 2.5 Changes in numbers in response to food addition in 1981 . 72 Figure 2.6 Number of immigrants to the study areas in 1980. 75 Figure 2.7 Number of immigrants to the study areas in 1981. 77 Figure 2.8 Minimum 28 day survival rates . . . 8 1 Figure 2.9 Onset of breeding 85 Figure 2.10 Changes in home range size 89 Figure 2.11 Proportion of radio locations found on the food gr i d . ... 92 Figure 2.12 Distance from gri d of resident and immigrant home ranges 95 Figure 2,. 13 Location of resident and immigrant home ranges. 97 Figure 3.1 Location of study areas 111 Figure 3.2 Minimum number a l i v e on S i l v e r Creek and monthly j number of dispersers caught on the Removal ...116 Figure!3.3 Average weight of male residents and dispersers. . 129 Figure 3.4 Average weight of female residents and dispersers. . ., .131 Figure 3.5 Mean body size of male dispersers and residents. 133 Figure 3.6 Mean body size of female dispersers and residents. 1 35 v i i i ACKNOWLEDGEMENTS I wish to thank my supervisory committee; Dr. J. H. Myers, Dr. C. J. Walters, Dr. J. N. M. Smith, Dr. C. J. Krebs, and Dr. A. R. E. S i n c l a i r for advice throughout the study. These people along with Dr. Mary T a i t t , Dr. Dennis Chitty, and Dr. C. Lee Gass provided helpful comments on e a r l i e r drafts of the thesis. Special thanks to Dr. C. F. Cramer who provided laboratory f a c i l i t i e s and technical advice which allowed me to carry out the radioactive calcium work. The Biology Data Centre Staff also provided invaluable help with aspects of analysis and data organization. I thank the Kluane Lake F i e l d Station s t a f f ; Andy and Carol Williams, P h i l and Claudette Upton, for their time and helpfulness throughout the study. Very special thanks to Scott G i l b e r t , Jean Carey, John Krebs, E l s i e Krebs, Rich Moses, Anne Helbig, Louise and Wally Walker, Maj DePoorter, and Carlos Galindo for f i e l d assistance.. Each was a colleague and friend and contributed immensely to making the research enjoyable. Many thanks to Carlos Galindo, Alison Hart, Mike Maxwell, and Mike Henderson for always having time to l i s t e n to my thoughts and problems. I wish to thank Dr. A.R.E. S i n c l a i r , Dr. C.J. Krebs, and Dr. J.N.M. Smith for monetary support. Special thanks to Tony S i n c l a i r for giving me room to make my own mistakes and to Charlie Krebs for his generosity and consideration. I thank a l l of the above for s a c r i f i c i n g s'ome of their time to help me. I wish to thank the National Science and Engineering Research Council, A r c t i c Institute of North America, Canadian Sportsman's Fund, and the A r c t i c and Alpine Research Committee for monetary support. F i n a l l y , a very special thanks to the Yukon and the snowshoe hare for making this thesis an adventure. 2 GENERAL'INTRODUCTION Snowshoe hares (Lepus americanus) show population cycles with a period of 9-10 years (Keith 1963). Only two long term demographic studies have been conducted on th i s species (Green and Evans 1940; Keith and Windberg 1978). The most recent study, conducted by Lloyd Keith and co-workers, monitored hare populations for f i f t e e n years and spanned two population declines. This work led Keith (1974) to hypothesize that hare declines are i n i t i a t e d by overwinter food shortage. Peak populations of hares over-eat the woody plant species that form their diet in winter and t h i s leads to reduced adult and juvenile survival through starvation. After hare numbers have declined for T-2 years, hare predators become a major cause of losses. This forces hare numbers down to-levels below that dictated by vegetation and allows the vegetation to recover. Once hare numbers are at low l e v e l s , predator numbers also drop. With favorable vegetation and a low predator-hare r a t i o , the number of hares again begins to increase to another peak. The Keith hypothesis attempts to explain changes in hare numbers by e x t r i n s i c factors alone. A l t e r n a t i v e l y , behavioural changes in the animals themselves may cause changes in density (Krebs and Myers 1974). There has been no attempt to determine the influence of behaviour on snowshoe hare population dynamics. Vaughan and Keith (1981) manipulated density of hares in large enclosures and found that i t had no influence on survival and reproduction. These results allow us to reject the hypothesis that density alone can a l t e r survival and 3 reproduction by determining the l e v e l of stress in animals as proposed by Christian (1970). It does not however, test the effects of behaviour. Density per se, may not be important but rather the type of animals present and the behaviour they exhibit. , Evidence that hares encounter food shortage at peak densities i s i n d i r e c t . Pease et a l . (197.9) measured the amount of browse required by hares and compared this to the amount that was available. They concluded that food supply was inadequate during the peak and early decline phase of the cycle. Problems associated with this approach are discussed by S i n c l a i r et a l . (1982), the most notable of which i s the d i f f i c u l t y of measuring how much food i s available to hares whose diet i s variable and highly dependent on changing snow l e v e l s . Wolff (1980) measured the percentage of available twigs that were browsed by hares during a decline in central Alaska. He concluded that 100 % of the available twigs were browsed up to 2 years after the peak. However, Wolff (1980) included only hardwood species in the analysis even though he had previously found that 40 % of the hares' winter diet consisted of :spruce needles (Wolff 1978). As well, spruce was the most abundant species on his study area. If spruce had been included in the analysis browsing intensity would have been much lower. 1 Vaughan and Keith (1981) manipulated food supply of hare populations in large enclosures. They were able to produce demographic changes in populations supplied with low food that were similar to those observed in natural populations at peak 4 and declining densities. However, no one has supplemented food supplies for natural populations nor has anyone investigated the relationship between behaviour and food supply. Windberg and Keith (1978) found that dispersers were l i g h t e r in weight and had a higher incidence of scarring than did animals in control populations. They suggested that t h i s may be due to i n t r a s p e c i f i c competition for food,. Futhermore, Dolbeer and Clark (1975) hypothesized that s o c i a l interactions with adults forced juvenile hares to move into unfavorable habitats. There i s evidence then, to suggest that behaviour may determine which animals have access to food or favorable habitats. I began to investigate the influence of behaviour on snowshoe hare population dynamics by asking whether female spacing behaviour could l i m i t breeding density in spring (Boutin 1980). Results suggested that i t did not. In this study I have examined the influence of spacing behaviour on disper s a l , immigration, and juvenile survival,. Changes in juvenile survival are one of the most important causes of changes in numbers of snowshoe hares (Keith and Windberg 1978; Green and Evans 1940) and c y c l i c microtines (Krebs and Myers 1974). Spacing behaviour in rodents can reduce juvenile recruitment by lowering survival (Sadleir 1965; Healey 1967; Boonstra 1978), increasing dispersal (Beacham 1979), and preventing immigration (Redfield et a l . 1978),. I examined the effe c t of adult and juvenile spacing behaviour on summer recruitment and survival of juvenile snowshoe hares. The results are presented in Section 1. Keith (1974) hypothesized that food a v a i l a b i l i t y dictates changes in snowshoe hare numbers and a number of studies have shown that food a v a i l a b i l i t y can l i m i t numbers of various species (Gilbert and Krebs 1981; Cole and B a t z l i 1978; T a i t t 1981; T a i t t and Krebs 1981; Mares et al:. 1982). However, i t has also been found that changes in food supply may act to change spacing behaviour, and thus lead to changes in numbers (Taitt 1981, T a i t t and Krebs 1981; Mares et a l . 1982).' I examined how food addition during late winter affects the population dynamics and spacing behaviour of a free ranging population of snowshoe hares. The results are presented in Section 2. Krebs et a l . (1969) fenced populations of Microtus  ochrogaster and M. pennsylvanicus . The result was abnormally high population densities and habitat destruction inside the fence. This suggests that dispersal is necessary for the population regulation of these species. Other studies of dispersal have shown that dispersal can account for moderate population declines but i s absent during major c y c l i c a l declines (Krebs and Boonstra 1978, Beacham 1980). Dispersal can also a l t e r population composition by dispersers being a non-random sample of the population (Myers and Krebs 1971; Krebs et a l . 1976). I examined ;how dispersal was related to changes in numbers of snowshoe hares during an increase, peak, and decline in numbers. I asked whether dispersal is responsible for declines in hare ^numbers. The results are presented in Section 6 General Spacing Behaviour of Snowshoe Hares The spacing behaviour of snowshoe hares results in individuals occupying permanent home ranges which overlap between and within sexes (Boutin 1979). Behaviour studies of hares in large outdoor enclosures and around a r t i f i c i a l food sources indicate that they form a dominance hierarchy with males being dominant to females in winter and females dominant to males in spring and summer (Graf 1981). Graf also found that resident animals were dominant to introduced in d i v i d u a l s . On my study area, males become reproductively active in late February or early March but females do not conceive u n t i l early A p r i l (Boutin 1979). The gestation period i s 37 days (Severaid 1942), so the f i r s t l i t t e r s are born in mid to late May. Three l i t t e r s are produced over the summer (Boutin 1979). Males go out of breeding condition in July or August but females may nurse young into September. Young receive l i t t l e parental care, are nursed only once a day, and are weaned at three to four weeks of age (Graf 1981, Severaid 1942). 7 SECTION 1. THE EFFECT OF SPACING BEHAVIOUR ON JUVENILE SURVIVAL AND RECRUITMENT Introduct ion Snowshoe hares (Lepus americanus) show 9-10 year cycles in numbers (Keith 1963). Declines in hare numbers are correlated with reduced juvenile survival and i t is hypothesized that t h i s i i s due to e x t r i n s i c rather than behavioural factors (Keith 1974). Keith postulated that winter food supply determines juvenile survival in two ways. F i r s t l y , juveniles born to females that survive winters of food shortage grow more slowly and enter the following winter at a l i g h t e r body weight than those born in years when winter food supply i s adequate (Keith and Windberg 1978). Secondly, juveniles lose weight at a greater rate during winters when food supply i s inadequate. Both of the above lead to more juveniles dying of starvation in peak and early decline phases of the cycle. Keith (1974) implies that spacing behaviour i s unimportant in determining juvenile s u r v i v a l . Evidence for this comes from experiments where hare densities were manipulated in large enclosures (Vaughan and Keith 1981). Juvenile survival was not affected by d i f f e r e n t levels of density. However, behaviour may influence survival by determining how many and which juveniles are forced to disperse from a population. If so, enclosing the population would prevent dispersal and no response to density would be expected. Wolff(1980) has argued that movement of juveniles between 8 habitats i s determined by food and cover rather than density. However, the number of juveniles he monitored was inadequate to test whether behaviour influences juvenile movements. There i s some evidence from hares and from rodents that spacing behaviour may influence juvenile survival and movements. Boutin (1979) found that juvenile hares s e t t l e d in areas free of adults, suggesting that adults may exclude juveniles from some areas. Dolbeer and Clark(l975) postulated that juvenile hares in populations in Colorado were forced into unfavorable habitat by resident adults. Graf (1981) found that both resident adult and juvenile hares in large enclosures were more aggressive toward introduced animals than to fellow residents. As well, f i r s t l i t t e r juveniles survive better than second and t h i r d l i t t e r animals (Boutin 1979; Keith and Windberg 1978). This suggests that older juveniles may influence the survival of younger ones. In rodents, spacing behaviour of adults during the breeding season can reduce juvenile recruitment by lowering surviv a l ( S a d l e i r 1965; Healey 1967), increasing dispersal (Beacham 1979), and preventing immigration (Redfield et a l . 1978) . In t h i s section I asked two questions. The f i r s t was: Does spacing behaviour l i m i t juvenile survival and o v e r a l l numbers? The second question was: ;Does spacing behaviour l i m i t juvenile recruitment by a l t e r i n g rates of emigration and immigration. If so, experimental manipulation of spacing behaviour should lead to changes in survival and d i s p e r s a l . As well, I wanted to determine i f adults and f i r s t l i t t e r juveniles affected these 9 factors s i m i l a r l y . To answer these questions, I experimentally removed adults or f i r s t l i t t e r juveniles from two d i f f e r e n t areas. I had 4 working hypotheses. They were that juvenile survival and recruitment were limited by: A. Adult spacing behaviour B. Juvenile spacing behaviour C. Spacing Behaviour (both adults and juveniles) D. Other Factors (spacing behaviour unimportant) The pre'dictions from each hypothesis are l i s t e d in Table 1 . 1.. Study Area The study was conducted at Kluane Lake Yukon T e r r i t o r y , Canada (61 ° North, 138 ° West). Seven areas were used during the two years of study (Fig. 1.1). The vegetation cover was open spruce (Picea glauca) boreal forest (Douglas 1974). I matched control and experimental study areas by vegetation type. Table 1.2 provides a comparison of vegetative cover on the 7 s i t e s . Vegetation was sampled by estimating cover of each species within a c i r c l e 6 m in diameter and centred at stations uniformly situated over the study areas. The number of spruce trees greater than 5 cm in diameter were also counted, and the diameter of the largest tree was measured to determine tree density and s i z e . I 10 Table 1.1. Predictions from the hypotheses tested in thi s study. A l l factors are expressed r e l a t i v e to levels on the control area. PREDICTIONS HYPOTHESIS POP. SIZE JUVENILE RECRUITMENT SURVIVAL AR i JR AR JR AR J.R A. Adult + - = + + Spacing Behaviour B. Juvenile - = + = + = + Spacing Behaviour C. Both -= - = + + + + D. Other Factors = = = = = = Figure 1.1. Location of study areas at Kluane Lake, Yukon. Sil v e r Creek-SC (1980 contr o l ) , Beaver Pond-BP (1:981 control), AR-Adult Removal, JR-Juvenile Removal, TR-Total Removal. 12 1 3 Table 1.2. Precentage cover of the major plant species on the study areas. SC - 1980•control, BP - 1981 control, AR -adult removal, JR - juvenile removal, TR - t o t a l removal. The number of quadrats sampled are in brackets. 1980 1981 Species SC AR JR TR BP AR JR (100)(50)(49)(100) (100)(50)(50) Picea glauca  Salix glauca  Betula glandulosa  Shepherdia canadensis Populus tremuloides  Grami nae Lupinus arcticus  Pleurozium sp. Arctostaphylus uva-ursi  Arctostaphylus rubra Picea glauca density (per ha) 788 418 1104 294 263 411 705 Av. diam. (cm) 20 23 16 7 18 14 14 18 14 20 1 4 8 1 6 1 4 8 23 19 1 5 15 1 6 19 0 8 0 4 12 1 3 4 8 4 3 3 0 0 0 0 0 0 4 0 0 0 0 20 0 21 16 1 7 7 5 8 1 3 2 6 6 25 1 0 27 1 1 4 4 10 5 5 7 20 8 18 1 5 10 1 7 3 1 0 5 14 Methods Trapping Each study area consisted of a 300 m X 300 m trapping g r i d with 100 stations spaced 30 m apart in a 1 0 X 1 0 pattern. Between 50 and 60 double door l i v e - t r a p s were placed at alternate stations over the gri d , set for two consecutive nights, and checked each morning. This was done each week in 1980 and biweekly in 1981. Traps were baited with a l f a l f a and apples and locked open when not in use. The location, tag number, sex, reproductive condition, weight, and length of the right hind foot were recorded for each capture. Reproductive condition of females was determined by the size and appearance of nipples. Medium or large nipples with matted fur indicated that the individual was nursing a l i t t e r . Juveniles appearing in traps for the f i r s t time were c a l l e d r e c r u i t s . I lumped juveniles into 3 l i t t e r groups by their body weight and length of right hind foot. This method was r e l i a b l e u n t i l late September, after which recrui t s were c l a s s i f i e d as juvenile or adult only. Population size was determined by the complete enumeration (Krebs 1966) and J o l l y (1965) techniques. Only complete enumeration values are provided because both methods showed similar changes in numbers and values determined by the J o l l y method averaged only 10% higher than those from the complete enumeration technique. 15 Radioactive Implants To determine whether juvenile r e c r u i t s were residents or immigrants, I implanted adult females with radioactive calcium-45. The calcium-45 was transmitted to the young during l a c t a t i o n and deposited in their bone tissue. In 1980 I injected 60-70 microcuries of calcium-45 subcutaneously as calcium chloride in water. Any females captured during the period beginning 1 week prior to b i r t h and ending 2 weeks after b i r t h of the f i r s t l i t t e r received an i n j e c t i o n . This was repeated for l i t t e r 2 and 3 on the control and juvenile removal grids. In 1981 calcium-45 was implanted as calcium oxylate in the manner described by Rongstad(1965). Females required only one implant to mark a l l l i t t e r s . This more e f f i c i e n t method was not used in 1980 because I encountered problems in converting the calcium-45 to calcium oxylate. Implanting began two weeks prior to b i r t h of the f i r s t l i t t e r . Samples for s c i n t i l l a t i o n counting were taken from each rec r u i t by c l i p p i n g a toe from their l e f t hind foot. Juveniles with detectable r a d i o a c t i v i t y were termed residents. The remainder were termed immigrants. To determine i f juveniles were receiving calcium-45 from implanted females I implanted or injected calcium-45 into pregnant females ! held in large pens. Young from these were tested for r a d i o a c t i v i t y . Six d i f f e r e n t females were injected and a l l of their 11 young were radioactive. Nine females were implanted and a l l resulting 22 young were radioactive. 16 Experimental Design Experiments began in May 1980 and were repeated in 1981. The study areas, and the experimental treatments they received, are shown in F i g . 1.2. In each year, one area had a l l adults removed (AR), a second had a l l f i r s t l i t t e r juveniles removed (JR), and a t h i r d had no manipulation (SC in 1980, BP in 1981). These grids were changed between years because another set of experiments was conducted on the areas in February 1981. I f e l t that these experiments might confound results of t h i s study. Removals began in the f i r s t week of July and continued u n t i l October 15 in 1980. In 1981, removals began in late June and continued to September 30. A fourth gr i d on which a l l animals were removed each trapping session (Total Removal) was maintained throughout the study. Animals caught on t h i s grid were considered to be dispersers from the surrounding habitat and represented a sample of the potential immigrants available to s e t t l e on the other study areas. The Total Removal (TR) had been trapped since May 1979 while S i l v e r Creek (SC), Beaver Pond (BP), and the 1980 Juvenile Removal (JR) had been trapped since May 1978. The other grids were established in May of the year they were used. 1 7 i Figure 1.2.. Experimental design used to test the hypotheses that adult or juvenile spacing behaviour l i m i t s juvenile survival and recruitment. i 18 GRID 1980 M Silver Creek (SC) <-Juvenile Removal ( J R ) Adult Removal (AR) Tota l Removal (TR) <-Beaver Pond (BP) Juvenile Removal ( JR ) Adult Removal (AR) Total Removal (TR ) J j A | S Control -N I Remove Litter 1-| I Remove Adults-H Remove all Hares. 1981 Control Remove Litter "M h-Remove AdultsH . Remove all Hares 19 Results Numbers Prior to Removals Figure 1.3 shows long term changes in Minimum Number Alive (MNA) for the control areas (SC and BP). Numbers on SC increased to a peak in September 1980 and then declined u n t i l production of young began in summer 1981. There was some recovery but numbers in f a l l were only 57% of what they were one year before. i Numbers on BP also peaked in 1980 but did not decline l a s much as those on SC and by f a l l of 1981, were equal to those on the grid one year e a r l i e r . The manipulations then, were done at peak population densities in both years. Population size prior to i n i t i a t i o n of removals d i f f e r e d s l i g h t l y between study areas but were similar between years(Table 1.3). Relative to control areas, numbers on the AR were 35% lower in 1980 and 15% lower in 1981. Numbers on the JR were 10% lower than those on SC in 1980. It appears then, that there were no major differences in population size between ^grids prior to i n i t i a t i o n of the experiments. i • i Removals For t h i s experiment to be e f f e c t i v e , i t was important that a l l adults and f i r s t l i t t e r juveniles be removed ' from the respective experimental areas. Adult removals i in 1980 jbegan la t e r and were less complete than in 1981 (Fig. 1.4).' In 1980 j a l l adults were present at the time of b i r t h of l i t t e r 2 and 30% 20 Figure 1.3. Changes in MNA (minimum number ali v e ) of snowshoe hares on the control grids between 1978-1981. 21 22 Table 1.3. MNA(minimum number al i v e ) on the study areas prior to the removals.. SC (Silver Creek Control), BP (Beaver Pond Control), JR (Juvenile Removal), AR (Adult Removal). 1980 GRID MALES FEMALES TOTAL SC 24 26 50 JR 22 23 45 AR 16 1 7 33 1981 BP 20 25 45 JR 20 26 46 AR 1 5 23 38 23 Figure 1.4. Pattern of removals on the experimental grids 2U 1 980 M J J A S O 25 remained at the time of weaning. In contrast, 75% of the 1981 adult population had been removed before b i r t h of l i t t e r 2 and only 1 adult remained at the time of weaning. F i r s t l i t t e r removals began at the same time as adult removals. The pattern of large numbers of individuals being removed in early sessions followed by few thereafter, was similar on a l l experimental grids except the 1981 JR, where large numbers of l i t t e r 1 juveniles were removed each trapping session throughout the summer. Changes in Numbers If either adult or juvenile spacing behaviour alone l i m i t s survival and recruitment, population size should be higher on one of the experimental grids r e l a t i v e to that on the control areas. The removals produced no consistent change in numbers on the grids (Fig. 1.5). After the removals, numbers on the experimental grids during summer, with the exception of the 1981 JR, never recovered to those on SC. The pattern of changes in numbers on the AR areas p a r a l l e l l e d those on the control areas through to November but actual values were 15% lower in 1980 and 40% lower in 1981. In November of 1980, numbers on SC declined rapidly while those on the AR remained high. This produced November densities 80% higher on the AR than on SC. In contrast,' 1981 November densities on the AR and BP were equal. The pattern of changes in numbers on the JR grids d i f f e r e d markedly between years. In 1980, numbers peaked in mid-July at levels 40% below those on SC. In contrast, despite continual 26 Figure 1.5. Changes in t o t a l minimum number a l i v e (MNA) on the control (SC and BP), adult removal (AR), and juvenile removal (JR). V e r t i c a l drops in values represent sessions when animals were removed. 27 — i 1 1 1 1 1 1 i i 1 i r i i i r May Jun Jul Aug Sept Oct Nov 28 removal of large numbers of animals, numbers on the 1981 JR were able to recover and remain near those on BP. November densities on the JR grids were 30% lower than SC in 1980 and equal to those on BP in 1981. The removal of adults or juveniles did not lead to the predicted increase ; i n population size on the removal grids. Survival If spacing behaviour l i m i t s juvenile s u r v i v a l , juveniles should survive better on the removal grids r e l a t i v e to the control areas. The number of r e c r u i t s per la c t a t i n g female i s an index of juvenile survival from b i r t h to f i r s t trapping. However, th i s index i s subject to bias by d i f f e r e n t i a l immigration and emigration. To avoid this bias I used the number of resident r e c r u i t s ( j u v e n i l e s having calcium-45 in their bones) per l a c t a t i n g female as an index of early s u r v i v a l . This measure is only an index of survival because losses to emigration were included. Table 1.4 shows that with the exception of the 1980 JR, the number of residents recruited per la c t a t i n g female was higher on the experimental grids than on the control areas but not s i g n i f i c a n t l y so. •' Once juveniles are f i r s t captured, their minimum survival rate can be determined by recapture rates in subsequent trapping sessions. I analyzed survival during two periods; summer (July through September) and f a l l (October through December). Removal of adults or f i r s t l i t t e r juveniles did not improve survival of juveniles remaining on the g r i d . Table 1.5 shows that summer 29 i Table 1.4. Number of residents recruited per lactating female on each of the study areas. SC - 1980 control, BP - 1981 control, JR - juvenile removal, AR - adult removal. GRID 1980 LITTER 1 LITTER 2 SC 2.0 1.28 JR 2.75 .75 AR 3.8 1 981 BP .5 .71 JR 1.16 1.36 AR . 5 I 1 30 survival rates did not d i f f e r between grids with the exception of l i t t e r 2 females surviving better on the 1981 AR than on BP. Both sexes survived equally well on a l l grids. Monthly survival rates from October to December were similar between control and removal grids(Table 1.6). The only exception was that male survival was s i g n i f i c a n t l y higher on the 1980 AR(chi-square=8.37, P<.. 001). It appears then, that the experimental removal of adults or f i r s t l i t t e r juveniles did not a l t e r juvenile survival rates. Mean Weight in F a l l Keith and Windberg (1978) found that juveniles grew more slowly and thus had lower f a l l weights during peak and early decline years. To determine i f juvenile weight was affected by spacing behaviour, I compared the weights of juveniles caught in mid-September on each of the study areas (Fig. 1.6 ). In both 1980 and 1981 l i t t e r 1 and 2 juveniles had lower weights than those on the control areas. Weights of l i t t e r 2 juveniles on the JR were lower than those on the control grids in 1980 but higher in 1981. These results were opposite to the prediction that spacing behaviour could reduce juvenile growth rates. Effect of Removals on Recruitment If adults or l i t t e r 1 juveniles l i m i t recruitment, more recruits should appear on the removal areas than on the control areas. Table 1.7 shows the t o t a l number of juveniles recruited 31 Table 1.5. Minimum 28 day survival rates of juveniles during summer (July-September). Sample sizes are in brackets. GRID 1980 LITTER 1 LITTER 2 Males Females Males Females Si l v e r .44(66) .37(68) .43(69) .35(61) Creek Adult .43(109) .37(69) .33(101) .47(60) Removal Juvenile .44(65) .25(41) . Removal 1981 Beaver .84(13) .46(22) ,69(23) .42(20) Pond Adult .61(65) .67(60) .51(25) .86(15)* Removal Juvenile , .61(38) .43(35) Removal * P < .05 i 32 Table 1.6. Minimum 28 day juvenile survival rates during f a l l (October-December). S i l v e r Creek (1980 control), Beaver Pond (1981 co n t r o l ) . Sample sizes are in brackets. GRID 1980 MALES FEMALES Si l v e r .43 (65) .45 (57) Creek Adult .69 (110)** .59 (75) Removal Juvenile .38 (42) .56 (24) Removal 1981 Beaver .37 (40) .38 (37) Pond Adult .46 (48) .45 (33) Removal Juvenile .59 (32) .28 (14) Removal ** P < .01 3 3 F i g u r e 1.6. Average (± 2 S.E.) weights of j u v e n i l e s caught i n mid-September on each of the study areas. J u v e n i l e s were lumped a c c o r d i n g :to the l i t t e r they were born i n (1 or 2). SC - 1980 c o n t r o l , BP - 1981 c o n t r o l , AR -a d u l t removal, JR - j u v e n i l e removal. 1980 14 0 0 1200 1000: 800 I o i { i i 1981 • S C , B P o AR A J R c S 14 00. 1200 10 0 01 800 35 i to the study areas before September 30. In 1980, increased recruitment of l i t t e r 2 caused t o t a l recruitment to the AR to be 1.2 times that to SC. This was not s i g n i f i c a n t but the number of l i t t e r 2 r e c r u i t s to the AR was s i g n i f i c a n t l y higher than that to SC (chi-square=4.29, df=1, P<.05). In 1981 both the JR and AR had higher recruitment than BP(2.4 and 1.5 times respectively). In the case of the AR, only l i t t e r 1 responded whereas both l i t t e r s did so on the JR. More males(87) than females(50; c h i -square=9.99, df=1, P<.01) recruited to the 1980 AR. In a l l other cases the sex r a t i o of r e c r u i t s was even. Recruits were of two types: residents and immigrants. My a b i l i t y to co r r e c t l y i d e n t i f y resident and immigrant juveniles depended on the proportion of resident females that received implants of calcium-45. Table 1.8 shows the number of la c t a t i n g females for each l i t t e r that were trapped on the grids and the proportion of these that received implants. In 1980 this proportion was between .67-.71, with the exception of l i t t e r 2 on the AR. No females were implanted on the AR, because they were to be removed before th e i r juveniles would be old enough to survive on t h e i r own. A l l l i t t e r 2 juveniles appearing on the AR could then be considered immigrants. However, 9 lac t a t i n g females were present long enough to raise juveniles to independence.. The net result of this i s that some l i t t e r 2 resident r e c r u i t s were probably classed as immigrants. In 1981, implanting was done more e f f i c i e n t l y and 88-100% of the lact a t i n g females were implanted except for BP which had a lower value of 78%. If we assume that no la c t a t i n g females were on the 36 Table 1.7. Total number of juveniles recruited to the study areas before September 30. Grand t o t a l s include t h i r d l i t t e r juveniles. CO - control ( S i l v e r Creek in 1980 and Beaver Pond in 1981), AR - adult removal, JR - juvenile removal. GRID 1980 1981 LITT 1 LITT 2 GT LITT 1 LITT 2 GT CO AR JR 57 58 56 55 79* 48 1 23 1 48 1 15 23 48** g 1 *** 35 31 53 61 80 1 ^ g*** * P < .05-** P < .01 *** P < .001 37 Table 1.8. Proportion of females captured while lac t a t i n g that received calcium-45. S i l v e r Creek (1980 control) Beaver pond (1981 c o n t r o l ) . GRID 1980 LITTER 1 LITTER 2 No. lactat ing (A) No. B/A imp. (B) No. No. B/A lact a t i n g imp. S i l v e r 14 Creek 10 71 21 15 .70 Juv Removal 12 67 20 14 .71 Adult Removal 10 .70 1 4 Beaver 10 Pond Juv Removal Adult Removal 25 24 1 0 1 . 0 24 21 .96 .88 1981 1 4 19 1 1 18 7 8 95 S i l v e r C r e e k 8 1 . 0 1 6 16 1 . 0 38 g r i d but avoiding capture, 30% more of the t o t a l number of rec r u i t s in 1980 should have been classed as residents. In 1981, this error was reduced to 12% or less on a l l grids except BP where i t was 22%. With the exception of l i t t e r 2 on the 1980 AR the proportion of la c t a t i n g females that were implanted was equal on a l l grids. The increased recruitment observed on the experimental grids r e l a t i v e to the controls was due to increased immigration. Total immigrant recruitment to September 30 was s i g n i f i c a n t l y higher on a l l experimental areas except the 1980 JR(Fig. 1.7). In the case of the 1980 AR t h i s was due to increased recruitment of l i t t e r 2 immigrants. This result is biased because no females on t h i s grid were implanted for l i t t e r 2. However, i f the number of immigrants i s decreased by the number of young that the remaining resident females could have produced, values are s t i l l s i g n i f i c a n t l y higher than on SC. In 1981, l i t t e r 1 was responsible for the increased recruitment seen on both the AR and JR. Sex r a t i o of immigrants was even on a l l grids with the following exceptions. In 1980 more female than male l i t t e r 1 juveniles immigrated to SC (chi-square=7.7, df = 1, P<.01) and the AR(chi-square=7.2, df=1,P<.0l). In the same year more male than female l i t t e r 1 (chi-square=7.2,P<.01) and l i t t e r 2 (chi-square=7.78, df=1, P<.01) also immigrated to the AR. The timing of immigration d i f f e r e d between study areas. F i g . 1.8 shows that the increased recruitment to the AR areas, r e l a t i v e to the control areas, occurred in one month only; i 39 Figure 1.7. Number of immigrants (juveniles caught for the f i r s t time and having no calcium-45 present) caught before September 30 on the control (SC in 1980 and BP in 1981), adult removal (AR), and juvenile removal (JR) grids. One and 2 refer to l i t t e r groups. Notice that the number of immigrants was s i g n i f i c a n t l y higher (chi-square tests) on 3 of 4 experimental areas r e l a t i v e to that on the control areas. 1 9 80 1981 10Q _ 80. — o 6 0 cc 4 0. d _ 20. S C AR JR B P AR J R a - P < .01 b - P < . 0 5 41 August in 1980 and July in 1981. This short pulse of immigration occurred immediately after the majority of the adult removals had taken place (Fig. 1 . 4 ) . In contrast, recruitment of immigrants to the 1981 JR was higher than that to BP in July, August, and September. As well, unlike the AR areas, removals continued throughout the summer with 10 or more animals being removed each trapping session (Fig. 1.4). On the AR areas, because l i t t l e adult immigration occurred, removals took the form of large numbers of animals in early sessions followed by few thereafter. The pattern on the AR areas then, was one of rapid removals followed by a short pulse of immigration whereas on the 1981 JR, removals were continuous throughout the summer as was increased immigration. Could immigration to the removal grids have been higher? F i g . 1.8 shows the number of animals ; caught on the TR each month. This number serves as an index of the maximum number of immigrants that could r e c r u i t to the other study areas. In 1980, the number of animals caught each month on the TR was always higher than the number of immigrants that recruited to the other grids, suggesting that immigration to these areas was r e s t r i c t e d . The difference between the number caught on the TR and the number of immigrants to the AR in August was s l i g h t however (68 vs 78). In 1981, the number of animals caught on the TR during July and August was equal to or less than the number of immigrants on the other grids during the same time. In both 1980 and 1981, the number of animals caught on the TR was higher during September to November r e l a t i v e to the preceding months. 42 Figure 1.8. Number of immigrants (animals caught for the f i r s t time) caught each month on the study areas. SC -1980 control, BP - 1981 control, AR - adult removal, JR - juvenile removal, TR - t o t a l removal. cc UJ CO 1 9 81 __l S C . B P • A R • J R S T R J A S O N 44 At the same time, immigration to the other study areas decreased. This resulted in the number caught on the TR being at least 5 times higher than the number of immigrant r e c r u i t s on the other grids in 1980 and 2-3.4 times higher in 1981. Overall,, the number of animals caught each month on the TR in 1.981 was half that in 1980. To summarize, removal of adults or juveniles did lead to an increase in immigration. This increase occurred after a substantial proportion of the population had been removed, but ceased i f immigrants were allowed to remain in the population. The number of animals caught on the TR increased from summer to f a l l in both 1980 and 1981. There was no similar increase in the number of immigrants to the other study areas. Immigration to Control Populations I next asked whether immigration to control populations occurred. Because I could i d e n t i f y immigrants, i t was possible to determine the amount of immigration to control populations. I used SC for t h i s analysis because i t was the control g r i d with the highest proportion of implanted females. F i g . 1,. 9 shows the proportion of the t o t a l number of juveniles present on the grid that were resident. The proportion of residents was never greater than .70 and values decreased throughout the summer, so that by October less than 30% of the juveniles on the grid were actually born there. Juveniles then, were moving between populations throughout the summer, and most animals present on the control area in the f a l l had come from elsewhere. 45 Figure 1.9. Proportion of juveniles present on S i l v e r Creek that were classed as resident r e c r u i t s (calcium-45 present). Notice that less than.30% of the juveniles present in October were resident. R e s i d e n t s / T o t a l (Juv only) p — i - Jo .co > ,cn pi s i ,co co ft 47 The number of animals caught on the TR gives a measure of the number of dispersers moving through the control areas. The number of immigrant recruits(non-radioactive) caught on the control area provides a measure of how many dispersers appear in normal populations. The r a t i o of: the number of immigrants to SC number caught on TR gives an index of the p r o b a b i l i t y that a disperser w i l l successfully s e t t l e in a new area. A high value would indicate that most dispersers can recruit to a control population. Table 1.9 shows that in both 1980 and 1981, the r a t i o of immigrants/dispersers decreased from summer (July-August) to f a l l (September-November). Summer values in 1981 were higher than those in 1980. Most dispersers were able to immigrate to new populations (73-100%) in July and August of 1981. During the remainder of 1981 and a l l of, 1980, the maximum number able to do so was only 39% of the t o t a l number of dispersers. Did immigrants survive as well as residents once they were trapped on the study areas? Table 1.10 shows that survival rates d i f f e r e d between residents and immigrants in 2 cases. L i t t e r 2 residents survived better than immigrants on SC (chi-.square=6.25, df=1, P<.05) L i t t e r 2 immigrants on the 1981 JR had higher survival(chi-square=4.97, df=1, P<.05) than those, on BP. Survival rates of residents did not d i f f e r between grids or sexes, nor was t h i s the case for immigrants. Removal of adults 48 Table 1.9. Ratio of immigrants (as determined by absence of calcium-45) caught on S i l v e r Creek (control grid) to dispersers ( animals caught on the t o t a l removal (TR). MONTH 1980 SC TR SC/TR No. Imm. No. Caught JULY 7 23 .30 AUG 31 79 .39 SEPT 17 103 .16 OCT 6 109 .04 NOV 6 118 .05 1981 JULY 7 4 1.75 AUG 16 23 .73 SEPT 3 48 .06 OCT 6 73 .08 NOV 17 59 .29 49 Table 1.10. Minimum 28 day survival rates of resident and immigrant juveniles on the study areas during summer (July-September). Sample sizes are in brackets. SC -1980 control, BP - 1981 control, AR - adult removal, JR -juvenile removal. GRID 1980 LITTER 1 RESIDENT IMMIGRANT LITTER 2 RESIDENT IMMIGRANT SC .45 (52) AR .40 (80) JR . .39 (81 ) .44 (98) .51 (78) .45 (22) ..27 (51)* .38 (61) .39 (63) 1981 BP AR JR .39 (8) .47 (29) .62 .69 (19)* (103) 49 (20) 41 (34) .62 (24) .52 (73) .70 (37)* * P<.05 50 or juveniles then, did not improve survival of immigrants. To determine i f immigrants d i f f e r e d from either residents or dispersers, I compared weights and body condition (weight/length of right hind foot) of residents, immigrants, and dispersers. I analyzed data for 1981 and lumped animals from SC, AR, and JR. There was no difference in in mean weights or body condition between these groups at any time during the study. Di scussion Survival Keith and Windberg (1978) found that changes in juvenile survival were highly correlated with changes in numbers of snowshoe hares. I attempted to increase juvenile survival and al t e r t o t a l numbers by manipulating spacing behaviour. Removal of adults or f i r s t l i t t e r juveniles did not a l t e r numbers in any consistent fashion, nor did i t improve juvenile survival during summer or f a l l . The experiments were conducted when hare numbers were at peak l e v e l s , the time when juvenile survival rates drop to i n i t i a t e the decline. If spacing behaviour was important in causing t h i s reduction, the experimental removals should have been most e f f e c t i v e at thi s time. As a re s u l t , I reject the hypothesis that spacing : behaviour l i m i t s summer and f a l l survival rates of juvenile snowshoe hares. My findings d i f f e r from those for rodents. Sadleir(1965) and Healey(l967) found that adult male Peromyscus reduced recruitment during the breeding season by lowering juvenile 51 s u r v i v a l . Chitty and Phipps(l966) thought that adult males influenced survival of juvenile Microtus agrestis and Clethrionomys glareolus. Adult females appear to lower survival of juvenile M. townsendi i (Redfield et a l . 1978; iBoonstra 1978). These species d i f f e r from snowshoe hares in that young are able to breed in the season of their b i r t h . Juvenile snowshoe hares may not e l i c i t the same l e v e l of aggressive response that maturing juvenile rodents do. Young hares are not competing for females or nest s i t e s and, as a consequence, their survival rate may not be influenced by adults. Immigration The experimental removal of adults and f i r s t l i t t e r juveniles did increase recruitment through immigration in 3 of 4 cases. Immigration to the removal grids coincided with the removal of large numbers of animals, regardless of whether the animals removed were adults or f i r s t l i t t e r juveniles (Fig. 1.8). When few or no animals were removed, immigration returned to normal lev e l s even though there were no adults or f i r s t l i t t e r juveniles present on the respective grids. It appears that increased immigration occurred when free space was created by the removals. Once immigrants f i l l e d the vacated space, the population quickly became less permeable to immigrants. These results support the hypothesis that spacing behaviour can l i m i t immigration but reject the hypothesis that adults or juveniles alone, are responsible for t h i s l i m i t a t i o n . It could be argued that the increase in immigration after removals was not a 52 response to removal of spacing behaviour per se, but rather to a favorable per capita l e v e l of resources created by the reduced density. If immigration rates were solely determined by the a v a i l a b i l i t y of resources, one would predict that population size on the experimental grids should have returned to control l e v e l s after each removal. This did not occur ( F i g . 1.5 ). The a v a i l a b i l i t y of resources may influence immigration but results of t h i s study indicate that spacing behaviour i s also involved. Few studies have examined the effect of spacing behaviour on immigration in lagomorphs. In european rabbits (Oryctolagus  cuniculus), immigrants are unlikely to recruit into new warrens unless a l l or most of the resident population has been removed (Daly 1981). Others have found that adult females are aggressive toward unfamiliar young (Mykytowycyz and Dudzinski 1972; Myers and Poole 1962). Mykytowycz and Gambale (1965) showed that resident juveniles are also aggressive toward immigrants. Graf(1981) introduced juvenile snowshoe hares into an enclosed colony of adults and juveniles. He found that both adults and juvenile residents were dominant to introduced individuals and interacted at a greater rate with these than with fellow residents. Keith and Surrendi(1971) found a low proportion of juvenile snowshoe hares (29% vs 79% for control areas) in an area where f i r e had forced adults to concentrate at three times their normal density. They attributed t h i s to juvenile exclusion by adult spacing behaviour. Dolbeer and Clark(l975) hypothesized that adults forced juvenile hares to use open less favorable :53 habitat. Windberg and Keith(1976a) manipulated sex ratios of snowshoe hares by removing adults in A p r i l . There was no increase in immigration on these areas during summer but by December, immigration to the female removal area was higher than expected. Wolff(1980) proposed that movements of juveniles between habitats are determined by a v a i l a b i l i t y of food and cover rather than s o c i a l interactions. However, his evidence for t h i s i s circumstantial. This study is the f i r s t to provide experimental evidence that movements of juvenile snowshoe hares are affected by spacing behaviour. Wolff(1980) may be correct in that food a v a i l a b i l i t y and cover are important factors but i t i s l i k e l y that spacing behaviour determines how many and which animals gain access to favorable areas. As yet, we do not know what q u a l i t i e s allow some animals to do so and others not. This study found no differences in body weight or condition between residents, immigrants, and dispersers. Other species also show increased immigration in response to removals. Davis et al.(l964) found higher immigration to exploited woodchuck populations. Krebs(l966) and Smyth(1968) removed adult voles to keep the population in a state :of increase but were unsuccessful because of high immigration rates. Dunford(1977) provided suggestive evidence that adult aggression, rather than decreased food a v a i l a b i l i t y , inhibited immigration of juvenile round-tailed ground s q u i r r e l s in late summer. 54 Immigration to Control Populations The study of immigration in small mammal populations has suffered from the l o g i s t i c a l ! d i f f i c u l t y of ide n t i f y i n g immigrants. This i s especially true during the breeding season when conventional live-trapping cannot .distinguish between resident and immigrant r e c r u i t s . This study was the f i r s t to use radioactive markers to monitor immigration s p e c i f i c a l l y . The effectiveness of thi s technique depends on being able to implant a l l resident females and assumes . that a l l such animals are trappable. Trapping e f f o r t in thi s study was much higher than that used in any other study of snowshoe hares. Also, in an e a r l i e r study (Boutin 1980), I placed radio transmitters on a l l animals caught in traps. The transmitters were v i s i b l e on free ranging hares. I saw no animals without transmitters during the study. This suggests that most animals present in the area were I captured. A common assumption of live-trapping studies is that a l l juveniles caught for the f i r s t time during the breeding season were born on the study area, and are thus residents. Results 'ot t h i s study indicate that many snowshoe hare juveniles had dispersed prior to f i r s t capture. This early movement would have gone undetected i f calcium-45 had not been used to mark residents. The transfer of calcium-45 from mother to offspring represents a useful method by which r e l a t i v e population losses and gains of pre-trappable animals can be monitored. It i s important that t h i s method be used in other studies to determine i f the early juvenile movement detected in t h i s study i s i 55 common f e a t u r e of many small mammals or i s unique to snowshoe hares. S o c i a l O r g a n i z a t i o n and Immigration T h i s study i n d i c a t e s that immigration to snowshoe hare p o p u l a t i o n s was r e s t r i c t e d by spacing behaviour, but u n l i k e m i c r o t i n e s , a d u l t spacing behaviour alone was not r e s p o n s i b l e . Instead, the presence of any i n d i v i d u a l a l t e r e d the p r o b a b i l i t y that animals would immigrate to an area. At the same time however, many j u v e n i l e s moved from where they were born to new areas before they were f i r s t c a ptured. By l a t e summer, most of the j u v e n i l e s on the c o n t r o l area had come from elsewhere ( F i g . 1.9). T h i s e a r l y d i s p e r s a l was not a f f e c t e d by the presence or absence of a d u l t s . ( T a b l e 1.4) i n d i c a t i n g that some other f a c t o r i was i n v o l v e d : p o s s i b l y that d i s p e r s a l has a strong g e n e t i c component (Howard 1960). T h i s study d i d not examine the mechanism by which spacing behaviour i n f l u e n c e s immigration. It i s p o s s i b l e that animals present i n an area a c t u a l l y r e p e l immigrants or l i t may be that immigrants merely a v o i d areas occupied by hares. In f a l l (September - November), the number of d i s p e r s e r s i n c r e a s e d s u b s t a n t i a l l y while the number of n o n - r a d i o a c t i v e r e c r u i t s ( i m m i g r a n t s ) to the c o n t r o l area remained the same or d e c l i n e d . T h i s i n d i c a t e s that a smaller p r o p o r t i o n of d i s p e r s e r s are able to Immigrate to another p o p u l a t i o n at t h i s time. T h i s c o u l d be due to a b e h a v i o u r a l change in members of the r e s i d e n t p o p u l a t i o n which in t u r n , i s r e l a t e d to a change i n food supply. 56 In summer, food supplies are not l i m i t i n g (Windberg and Keith 1976b), and although adults may compete with each other for mates, juveniles do not represent competitors because they do not breed u n t i l the following season. As a r e s u l t , juveniles are allowed to move between populations and do so extensively. In f a l l though, when hares begin to feed on less nutritious woody vegetation, animals with an established home range may become more resistant to dispersers. This may result in a smaller proportion of animals moving through an area being able to s e t t l e . Juveniles dispersing at a younger age then, are more l i k e l y to r e c r u i t to a new population. Unlike most mammals, hares do not show male biased dispersal (Greenwood 1980). Graf(1981) found that hares exhibit a dominance heirarchy in summer with females usually dominant to males. This system did not result in more males disappearing from control areas or appearing on the removal area. As well, there was no indication that adult females allowed their female offspring to remain in their home range as is the case in voles (Madison 1980). This may be due to the fact that female ihome ranges overlap extensively (Boutin 1980) and juveniles s t i l l encounter unrelated adults even i f they remain on their mother's home range. I found no evidence that juveniles gain a survival advantage by remaining in their natal area as opposed to dispersing. Very few resident juveniles remained on , the control area by October (Fig. 1.9), suggesting that most had either died or emigrated. These results may be due to the fact that the study was 57 conducted during peak population densities. A l l available habitat i s occupied at t h i s time and juvenile survival is low everywhere. However, as populations decline, more and more areas become vacant u n t i l hares exist only in isolated pockets of habitat (Wolff 1980; Keith 1966). The r e l a t i v e favorableness of habitats during increases and declines could change rapidly, depending on food supply, predation rates, and spacing behaviour of hares. S i m i l a r l y , the r e l a t i v e advantages of remaining resident versus dispersing w i l l change with phase of the cycle and with habitat also. For example, changes in hare numbers in some habitats show a r e l a t i v e l y small amplitude. These may be areas where residents are resistant to immigration and adults expel most of the juveniles produced in the area. Other habitats may become temporarily more favorable and immigration to these would be high and emigration low. These p o s s i b i l i t i e s point to the need to monitor immigration and emigration of marked animals in a variety of d i f f e r e n t habitats at the same time. To conclude, results indicate that spacing behaviour of hares does not affect juvenile survival but is one factor that l i m i t s immigration to established populations. However, at least during peak population densities, young juveniles move extensively between populations. Future work should concentrate on determining whether immigration to populations d i f f e r with phases of the cycle and with various habitats. The use of radioactive markers to monitor early juvenile immigration i s an e f f e c t i v e way to proceed. 58 SECTION 2 . EFFECT OF LATE WINTER FOOD ADDITION ON SNOWSHOE HARE SPACING BEHAVIOUR AND NUMBERS Introduction Populations do not increase without l i m i t and much e f f o r t has been devoted to determining what factors are responsible for this l i m i t a t i o n . In the case of small mammals there is increasing experimental evidence that populations are limited by food (Taitt 1 9 8 1 ; T a i t t and Krebs 1 9 8 1 ; G i l b e r t and Krebs 1 9 8 1 ; Cole and B a t z l i 1 9 7 8 ; Mares et a l . 1 9 8 2 ; Flowerdew 1 9 7 3 ) . Gilbert and Krebs ( 1 9 8 1 ) went so far as to propose that 2 - 3 f o l d increases in density after food addition are the norm for a l l small mammal populations. However, food addition to unenclosed populations have been done on only a limited number of small mammal species ( c r i c e t i d and s c i u r i d rodents). This hypothesis needs testing on other groups before i t s generality can be accepted. Watson and M O S S ( 1 9 7 0 ) present evidence that some populations of mammals and birds are limited , by spacing behaviour. More importantly, recent studies have suggested that food supply can af f e c t behaviour which in turn, can l i m i t density (Taitt 1 9 8 1 ; T a i t t and Krebs 1 9 8 1 ; Mares et a l . 1 9 7 6 , 1 9 8 2 ) . T a i t t ( 1 9 8 1 ) found that food addition to populations of deermice (Peromyscus maniculatus) and voles (Microtus  townsendi; (Taitt and Krebs 1 9 8 1 ) led to an increase in numbers, primarily through immigration. Mares et a l . ( 1 9 8 2 ) obtained similar results in chipmunks (Eutamias s t r i a t u s • ) In each case, 59 residents decreased their home range size after food addition, and the authors argue that i t is this change in spacing behaviour that allows immigration to occur. Mares et a l . (1976) found that during food addition, immmigrants se t t l e d in parts of residents' home ranges that the residents had stopped using after food was added. After supplemental feeding was stopped, residents were able to force most of the immigrants to emigrate. Snowshoe hares (Lepus americanus) cycle in number with a period of 9-10 years. Keith (1974) hypothesized that hares decline from peak numbers because they suffer winter food shortage which increases weight loss and lowers s u r v i v a l . Food addition to populations in large enclosures support this hypothesis (Vaughan and Keith 1981) but no one has supplied supplemental food to natural populations. Wolff (1980) suggests that hare movements are determined by food a v a i l a b i l i t y . This section examines the effect of short term (1-4 months) food addition in late winter through spring on snowshoe hare spacing behaviour and demography. I f i r s t tested the hypothesis that snowshoe hares are limited by food. If so, hares should respond to food addition by 1) 2-3 fo l d increases in numbers, 2) increased immigration, and 3) e a r l i e r i n i t i a t i o n of breeding. Secondly, I tested the hypothesis that home range size i s related to food supply. If so, average home range size of hares should decrease and immigration should increase after food addition. I also monitored movements of immigrants to determine where they se t t l e d in the population. F i n a l l y , I asked i f food addition to unenclosed populations can produce similar e f f e c t s i 60 to those found in enclosed populations; namely decreased weight loss and increased s u r v i v a l . Study Area Figure 2.1 shows the r e l a t i v e locations of the study areas. Vegetation is similar to that described in section 1. Habitat surrounding the study areas is heterogeneous but a l l is used by hares except that immediately south of S i l v e r Creek where a 250 m s t r i p of bare ground e x i s t s . Radiotelemetry indicated that hares rarely crossed this area. In 1980, snow covered the study s i t e s u n t i l May 1-7 and hares did not have access to new leaves or herbs u n t i l the last week of May. These events occurred roughly one week later in 1981. Methods Trapping Each study s i t e had a 300 X 300 m trapping grid with 100 stations arranged 30 m apart in a 10 X 10 pattern. Between 50 and 60 double door liv e - t r a p s were di s t r i b u t e d at stations spaced evenly over the gri d . Traps were baited with a l f a l f a , set for two consecutive nights, and checked each morning. This was done every week in 1980 but was reduced to bi-weekly in 1981 to avoid trap deaths. Data col l e c t e d from each capture were as described in section 1. 61 Figure 2.1. Location of study areas at Kluane Lake, Yukon. S i l v e r Creek (SC) - control, Telemetry (TE) - food addition, 1980 and 1981, Removal (RE) - a l l animals removed, G r i z z l y (GR) - food addition, 1981. ' i 62 63 Experimental Design Figure 2.2 shows the experimental design used in the study. In 1980, the Telemetry trapping area received laboratory rabbit chow from Feb. 29 to March 29. The chow was placed in ten 20 X 250 X 10 cm feeders which were d i s t r i b u t e d evenly over the g r i d . S i m i l a r l y , on March 7 1981, chow was again placed on Telemetry as well as on a second grid, G r i z z l y . On March 30 the chow was s h i f t e d from the feeders to the traps themselves. Each trap was locked open and had 500 g of feed placed in i t . During regular trapping sessions, the traps were set but not moved. Food was supplied to Telemetry in this manner u n t i l July 1 and to G r i z z l y u n t i l May 15. Throughout the study I removed a l l animals caught from a fourth gr i d (Removal). The Removal was established in May 1979. S i l v e r Creek and Telemetry had been trapped since May 1978, and G r i z z l y since May 1980. Telemetry Hares on S i l v e r Creek and Telemetry were f i t t e d with radio transmitters weighing 30 g. The number of animals receiving radios depended on the a v a i l a b i l i t y of transmitters. Radios on each gri d were located by triangulation from 2 permanent towers (Fig.. 2.1). The true bearing of the radio from each tower was within ± 1.5 0 of the recorded bearing 95% of the time (Boutin 1979). Each radio was located synchronously by two people, one at each tower. A l l locations were taken between 0600 and 2400. 64 Figure 2.2. Experimental design. 65 1 9 8 0 1 9 8 1 F M M A M J J 2 8 2 9 7 3 0 1 5 1 S C ~ C o n t r o l T E A T A - A T -G R I • A 1 R E A l l A n i m a l s — R e m o v e d — A F o o d A d d i t i o n T Food A d d i t i o n S t o p p e d A Food Moved to T r a p s 66 Twenty-30 locations were obtained every 2 weeks. Home range size was determined by the minimum polygon method (Hayne 1949), I modified t h i s technique to include only 90% of the t o t a l number of locations for each animal. Locations farthest from a l l others were excluded. Locations f a l l i n g in a region where error polygon length (Heezen and Tester 1967) was greater than 75 m were disregarded. Any animal with greater than 20% of i t s locations in t h i s zone was not included in the analysis because locations were not considered accurate enough to determine home range si z e . I determined home range size of only those ianimals for which I had 15 or more locations during the period of analysis. Results Changes in Numbers The number of hares on the grids were determined by the methods described in section 1. The response of hare numbers to food addition may depend on the phase of the cyc l e . Figure 2.3 shows that changes in numbers on S i l v e r Creek and Telemetry were similar p r i o r to the start of the study. Late winter(March) numbers increased each year u n t i l 1980. Numbers of hares on Si l v e r Creek and Telemetry in March 1981 were half those present in 1980. Trapping on G r i z z l y did not cover a long enough period to determine when hare numbers on this area reached a peak. In March 1981, numbers on G r i z z l y were twice that on Telemetry and Si l v e r Creek. Telemetry received supplemental food then, in the 67 Figure 2.3. Changes in minimum number a l i v e of hares on the control (SC) and food addition (TE and GR) areas. . SC • TE • GR , i — i — i — i — ' i i " i — i — i — i — i r~i ' r r r F Jul - Sept F Jul - Sept 1 9 8 0 1 9 8 1 69 winters of peak (1980) and early decline phases (1981) of the cycle. i Did numbers increase on the experimental grids after food was added? Changes in numbers after food addition varied between sexes and between years (Figs. 2.4, 2.5). During food addition in 1980, the number of males on Telemetry was 1.4 times higher than that on Sil v e r Creek. This difference decreased to 1.2 times after food addition stopped and was zero by early May. The number of females remained equal on the two grids throughout the spring of 1980. In February 1981, numbers on S i l v e r Creek and Telemetry ; were again equal, but half that of one year before. After food addition to Telemetry, male and female numbers there, increased 1 while those on S i l v e r Creek remained constant. Numbers were up to 3.6 and 3.2 times higher on Telemetry than on S i l v e r Creek for males and females respectively. The number of hares on Telemetry in May 1981 was s l i g h t l y higher than that present one year e a r l i e r . Prior to food addition, the number of males on G r i z z l y was twice that on S i l v e r Creek. Following food addition, male numbers increased b r i e f l y in early A p r i l but declined o v e r a l l and remained double those on S i l v e r Creek. The number of females on G r i z z l y jprior to food addition, was equal to that on S i l v e r Creek but was twice as high during food addition. To summarize, food addition increased numbers of males but not females on Telemetry in 1980, and in both sexes in 1981. Food addition to G r i z z l y had no effect on male numbers but 70 i Figure 2.4. Changes in minimum number a l i v e of males and females on the control (SC) and food addition (TE) grids in 1980. Shaded area is the food addition period. Minimum Number A l i v e 72 i Figure 2.5. Changes in numbers of males and females on the control (SC) and food addition (TE and GR) grids in 1981. V e r t i c a l l i n e s indicate the period when food was added in large feeders. Shaded area is when food was placed in traps. CO . j j ! 74 increased female numbers. Immigrat ion ! The higher number of animalsi on Telemetry and G r i z z l y r e l a t i v e to that on Sil v e r Creek could be due to increased immigration or s u r v i v a l . I f i r s t examine immigration. New animals tagged each trapping session were classed as immigrants. This assumes that these individuals were not l i v i n g on the grid and avoiding capture. Immigration to the food grids varied between years (Figs. 2.6, 2.7). In 1980, except for an i n i t i a l pulse of males immediately after food addition, the number of immigrants to Telemetry per trapping session did not d i f f e r from that to S i l v e r Creek. However, the t o t a l number of male immigrants to Telemetry during the food addition period was s i g n i f i c a n t l y higher than that to S i l v e r Creek (1.2 vs. 3, c h i -square = 4.26, df=1, P <.05).. During the food addition period in 1981 the number of male immigrants to S i l v e r Creek was never more than 2 animals per trapping session and averaged 0.75 animals per trapping session. i Average values were 4.6 and 3.6 times higher on Telemetry and G r i z z l y respectively. A t o t a l of 28 males immigrated to Telemetry during the food addition period as compared to 16 on I G r i z z l y and 6 on Silver Creek. j The number of female immigrants to S i l v e r Creek averaged 1.8 animals per trapping session during the food addition period. Average values were 5.0 and 6.0 times higher to Telemetry and Gr i z z l y respectively. Seventy-five females i i 75 Figure 2.6. Number of immigrants (animals caught for the f i r s t time) caught on the control (Silver Ck) and food addition (Telemetry) grids in 1980. The number of animals caught on a t o t a l removal (Removal) grid i s also shown. Shaded area i s the food addition period. 77 | Figure 2.7. Number of immigrants (animals caught for the f i r s t time) caught on the control (Silver Ck) and food addition (Telemety and Gri z z l y ) areas in 1981. The number of animals caught on a t o t a l removal area (Removal) i s also shown. V e r t i c a l l i n e represents the time when food was added. Shaded area is when food was placed in traps. 27. c 2 1 J O i_ ._? | 15J 1981 • Silver Ck • Telemetry • Grizzly • Removal T 79 immigrated to Telemetry during the food addition period as compared to 37 on G r i z z l y and 13 on S i l v e r Creek. Figures 2.6 and 2.7 also shows the number of animals caught on the. Removal (dispersers). In 1980, an average of 8.5 males and 7 females were caught per trapping session during the food addition period. This was well above the number of immigrants caught on Si l v e r Creek or Telemetry. In 1981, the number of animals caught on the Removal during the food addition period was similar to the number of immigrants caught on Telemetry. In 1981 there was a marked increase in the amount of immigration to a l l grids in late A p r i l and early May (Figs. 2.6, 2.7). The increase was more pronounced in females ;than in males. This increase occurred two weeks after food had been shifted from the feeders to the traps in 1981. During the 3 weeks of food addition prior to the s h i f t , there was no change in immigration rates on Telemetry or Gr i z z l y r e l a t i v e to those on Si l v e r Creek. After the changeover, there was an immediate increase, a f u l l two weeks before the seasonal increase in immigration on S i l v e r Creek and the iRem'oval. A similar but less pronounced increase in immigration occurred in A p r i l 1980. It was most apparent on Si l v e r Creek where the number of immigrants went from less than 3 per trapping session to as high as 10 per session in late A p r i l . Males and females responded equally. To summarize, food addition increased immigration of males to the food grids in 1980 and 1981. The number of female immigrants to the food grids increased; in 1981 only. In both years there was an increase in immigration to a l l grids in mid-80 A p r i l to early May. This increase was more pronounced in 1981 than 1980 and involved more females than males. Survival I next asked i f food addition improved survival of resident hares. I defined residents as any animal caught at least once prior to the food addition period. The recapture rate of marked animals can be used as an index of s u r v i v a l . It i s a minimum estimate because emigrants are 'considered as losses. I performed a 3-way ANOVA on 28-day minimum survival rates of hares with grid , time, and sex as the main e f f e c t s . In 1980, survival was lower on a l l grids in the post food addition versus the prefood addition and food addition periods (Newman-Keuls test, P<.05). There were no differences in survival between grids or between sexes in 1980 or 1981. However, a closer examination showed that survival of hares on Telemetry improved after food addition in each year (Fig. 2.8).. The difference was s i g n i f i c a n t in 1981 ( t -test, t=3.54,d.f.=14„ P = .003). Over the same time period in 1980, survival of hares on S i l v e r Creek decreased and in 1981, i t improved only s l i g h t l y (.56 to .61). Survival of hares on Gr i z z l y did not improve after food addition. This may have been due to unusually high lynx predation on the g r i d . At least two animals were known to hunt there regularly and hares caught in traps were disturbed by lynx a number of times. This did not occur on the other study areas. To summarize, food addition improved survival of resident hares on Telemetry but not on G r i z z l y . ! 81 ! Figure 2.8. Average minimum 28-day survival rates of hares caught on the control (SC) and :food addition (TE and GR) grids. Narrow bars are 95% C 82 1.0. 1980 • 8 J 4) • 6J .4 1 0> O or .2. D > > i_ ZD Ul >^  D "O 1.0 CO CNJ .8. Pre Food Add. Food Add. 1981 Food Add. Stopped • SC . TE i G R .6. .4. • 2J 0 Pre Food Food Add. Add. 83 Weight Loss • i Keith and Windberg (1978) hypothesized that overwinter weight loss in hares lis related to food a v a i l a b i l i t y . To test t h i s , I measured the change in weight of hares trapped both one ; i i week prior to food addition and one week after food addition had stopped. This was done in 1980. In 1981, because food addition continued for a longer period than in 1980, I used weights of animals caught in mid to late A p r i l rather than those of animals caught after food addition stopped. Table 2.1 shows that both males and females on the food grids lost s i g n i f i c a n t l y less weight than those on Si l v e r Creek. Females on the food grids actually gained weight. The lower weight loss of animals on the food grids occurred during the food addition period only. Prior to t h i s , there was no s i g n i f i c a n t difference in weight change between animals from the food addition and control g r i d s ( t -tests, P>.05). , Onset of Breeding Food addition advances the breeding season in deermice (Taitt 1981). I next asked i f supplemental food could cause this to happen in snowshoe hares. The onset of breeding in males was defined as the time when greater than 50% of the males had scro t a l testes. Males on the food grids came into breeding condition e a r l i e r than did those on Si l v e r Creek (Fig. 2.9). In 1980, greater than 50% of the males were breeding on Telemetry 3 weeks before the same situ a t i o n occurred on Si l v e r Creek. In 84 Table 2.1. Mean net weight (g) change of hares caught both one week prior to and one week after food addition. Sample sizes are in brackets. Grid Before Food Addit ion During Food Addition 1980 Males Si l v e r Creek Telemetry •55.0 •63.5 (10) (10) -152.73 -111.25 (11) (12) Females Si l v e r Creek . Telemetry +10.83 (6) -32.14 ( 7) -131.89 (9) + 12.0 (10)** 1981 Ma 1 e s S i l v e r Creek Telemetry Gr i z z l y 142.14 ( 8) 98.00 ( 5) -103.89 ( 9) -67.5 ( 8 ) ~ 69.44 ( 9) Females Si l v e r Creek Telemetry G r i z z l y -23.50 (3) + 1 1 1.. 00 ( 5) *** + 43.75 ( 4)** ** P<.01 *** P<.001 85 Figure 2.9. Timing of onset of breeding of hares on the control (SC) and food addition (SC and GR) areas. Males were classed as be in breeding condition i f they had scrotal testes; females by i f they were l a c t a t i n g . Narrow bars - no animals breeding, medium bars - less than 50% , wide bars - greater than 50%. 1981 87 1981 onset of breeding was one week e a r l i e r on Telemetry and 3 weeks e a r l i e r on G r i z z l y . Supplemental food did not change the time at which females gave b i r t h (Fig. 2.9). Home Range Size and Dispersal In order to test i f home range size changed with food supply, I compared home range size of radio-collared resident hares after food addition. The actual time periods in which the data were lumped are shown in Table 2.2. To increase sample size during the 1980 food addition period I included animals which may not have had 15 or more locations in the designated period but did so at some time during which food was present on Telemetry (February 29 - March 29). Only those animals caught at l e a s t once pr i o r to the food addition period (residents) were included in the analysis. Males and females were lumped because size of their home ranges did not d i f f e r s i g n i f i c a n t l y . Residents on Telemetry decreased their home range size in 1980 after food addition (t-test,t=3.18,df=26, P=.003, F i g . 2.10). In 1981, they did so in May only (t-test,t=2.04,df=24, P=.05) but thi s seemed due to season rather than the presence of food as hares on S i l v e r Creek also decreased home range size in May ( t -test,t=2.29,df=15, P = .03). It seems then, that resident hares responded to increased food by decreasing home range size in 1980 but not in 1981. I next asked i f radio-collared resident hares with supplemental food were less l i k e l y to disperse than those with normal food. An animal was considered to have dispersed i f i t 88 Table 2.2 Time p e r i o d s f o r which home range s i z e of r a d i o -c o l l a r e d hares was c a l c u l a t e d . Sample s i z e i s the number of r e s i d e n t s (animals caught at l e a s t once before food was added) f o r which home range s i z e c o u l d be determined. Sample S i z e 1980 C o n t r o l (SC) Food (TE) February 20-29 - pre-food a d d i t i o n 12 1 1 March 12-21 . - food added 1 6 17 Mar 30-Apr 9 - po s t - f o o d a d d i t i o n 1 2 1 5 I 1 981 Feb 20 - Mar 6 - p r e - f o o d a d d i t i o n 1 2 1 1 March 14-27 - March food a d d i t i o n 16 1 5 A p r i l 10-26 - A p r i l food a d d i t i o n 1 1 1 2 May 20-June 14 - May food a d d i t i o n 5 1 1 89 Figure .2.10. Mean home range size (as determined by radio telemetry) of residents (animals caught at least once prior to food addition) on the control (Silver Creek) and food addition (Telemetry) study areas.. Narrow bars represent 95% confidence l i m i t s . 90 1 9 8 0 S i l ve r C reek 1 9 8 1 h a 6 4. h a 6 4_ V ^ 1 r l -v 6 4. 2J T e l e m e t r y 6_ 4 2J A A A M A M A Food Added V Addi t ion Stopped A A A M A M 1 91 moved to occupy a home range that did not overlap with the one i t occupied when f i r s t radio-collared. The proportion of residents which dispersed was the same each year so I lumped data from both years. Three of 36 radio c o l l a r e d residents dispersed from S i l v e r Creek. This compares with 5 of 41 dispersing on Telemetry. The addition of food did not a l t e r dispersal rate of residents. Use of the Food Addition Area To determine i f hares provided with supplemental food increased their use of the food addition area, I compared the proportion of radio locations found on Telemetry before and during food addition. In 1980 the average proportion increased from .20 to .48 (t - t e s t , t = 3.18, df = 36, P = .003) when food was added (Fig. 2.11). After food addition stopped the proportion dropped to .27. In 1981 hares did not increase their use of the food addition area (Fig. 2.11). I also compared use of the food gri d by residents with that of animals caught after food was added. The two groups did not d i f f e r in their use of the food addition area ( t - t e s t s , P > .05), indicating that residents and immigrants had equal access to the food addition area. 92 Figure 2 . 1 1 . Mean proportion of telemetry locations of radio-collared hares that were on the food addition (Telemetry) g r i d . Narrow bars represent 95% confidence : l i m i t s . 1 9 8 0 1981 Pre Food Post Food Food Pre Food Mar Apr May F O O D 94 Movements of Immigrants I I next asked i f the increase in immigration to Telemetry in 1981 was due to animals just off the grid expanding their home range when food was added. The observed increase in numbers then, would r e f l e c t an increase in r e l a t i v e trapping area rather than actual density. If the above i s true, immigrants should have had larger home ranges than residents.. This was not the case at any time during the food addition period ( t - t e s t s , P > ..05) . As a further test, I compared the location of residents and immigrants r e l a t i v e to the g r i d . To do t h i s , I calculated the distance from the arithmetic centre of each home range to the centre of .the trapping g r i d . Home ranges located in areas where the error polygon length of locations was greater than 75 m were assigned values of 500 m. Mean distances of residents and immigrants from the grid were not s i g n i f i c a n t l y d i f f e r e n t ( t -test, t=1.55, df=2l, P > .05) However, the frequency d i s t r i b u t i o n of distance from the grid for each group d i f f e r e d (Fig. 2.12). Home ranges of residents had a unimodal d i s t r i b u t i o n . Only one animal was located greater than 500 m away. In contrast, the frequency d i s t r i b u t i o n of distance of immigrants from the grid was bimodal. (Figs. 2.12 and 2.13). Of the 20 immigrants for which s u f f i c i e n t information was obtained to determine home range location, 12 settled within 500 m of the gr i d . Of the 8 that s e t t l e d farther away, 4 were trapped only once and none of these were trapped in May. Further, of the 52 female immigrants caught in late A p r i l to early May, 23 were 95 Figure 2.12. Frequency d i s t r i b u t i o n of distance from the centre of the food g r i d (Telemetry) to the arithmetic centres of home ranges of residents and immigrants during A p r i l and May, 1981, APRIL R e s i d e n t s MAY I m m i g r a n t s 75 225 375 150 3QO 450 0 7 5 225 375 150 3 0 0 .450 DISTANCE (m) gure 2.13. L o c a t i o n s of a r i t h m e t i c c e n t r e s of r e s i d e n t and immigrant home ranges on Telemetry. Square i n d i c a t e s the t r a p p i n g g r i d . A P R I L o 150 m 1 N o o • R e s i d e n t s o I m m i g r a n t s M A Y o 99 i captured only once. It appears then, that immigrants were not animals which expanded their ranges to include the gr i d after food was added. However, only a portion of a l l immigrants caught on the gr i d actually remained there. Others spent only a brief time there before s e t t l i n g elsewhere. Immigrants that did remain near the gri d had movements similar to residents. Did immigrants remain on the food gr i d after food addition stopped? In 1980, only one of 16 immigrants was caught after food addition was stopped. In 1981, only 35 of 105 immigrants did so. Discussion Table 2.3 summarizes the findings of t h i s chapter. Increased immigration led to higher numbers on the food grids in 1981 only. Why did the response to food addition d i f f e r between years? One p o s s i b i l i t y is that food was not l i m i t i n g in 1980. Natural food supplies may have been adequate in 1980 but heavy browsing could have reduced food to inadequate levels in the following year. The l a t t e r may be true but evidence suggests that food was in short supply during the 1980 food addition period. During the food addition period, bait l e f t in the traps after the regular trapping session disappeared rapidly on S i l v e r Creek but often remained on Telemetry u n t i l the next session. As well, a second check of traps in the afternoon of the f i r s t day of trapping produced as many as 15 animals on S i l v e r Creek but none on Telemetry. This high number of captures i s unusual as hares are normally inactive during the day and i t suggests that 100 Table 2.3. Summary of effects of late winter food addition on snowshoe hare demography and s p a t i a l arrangement. Response Demography 1980 1981 Numbers and Immigrat ion s l i g h t increase in males Survival Weight Loss Onset of Breeding condi t ion males females increased decreased advanced unchanged Telemetry: males increased females increased G r i z z l y : females increased males decreased increased decreased advanced unchanged Spatial Arrangement Home Range Size decreased unchanged 101 the bait was important enough for the hares to s h i f t their a c t i v i t y patterns. The above observations suggest that foodlwas in short supply during the food addition period in 1980. A second p o s s i b i l i t y is that the animals themselves d i f f e r e d in aggressiveness between years (Chitty 1967). Hares may have been more aggressive in the year of peak numbers and consequently were able to exclude immigrants from s e t t l i n g ] In the f i r s t year of decline however, surviving individuals may have been less aggressive and so unable to prevent immigration. However, t h i s i s contrary to the Chitty Hypothesis which predicts that animals should be more aggressive in the decline than in the increase and peak phases of the cycle. A t h i r d p o s s i b i l i t y is that the response to food addition is density related. Numbers on S i l v e r Creek and Telemetry, prior to food addition in 1980 were twice that in 1981 (Fig. 2.4, 2.5). The 1980 levels may r e f l e c t a l i m i t set by spacing behaviour, with the result that additional food produced no increase. S i m i l a r l y , male numbers on G r i z z l y began at levels equal to those on S i l v e r Creek and Telemetry in 1980 and again showed no increase in numbers after food addition. However, two factors may have influenced the re s u l t s from G r i z z l y . F i r s t l y , trapping success was reduced by: 1) a moose in the ; area frequently setting off traps while attempting to get at the feed and 2) hares may have avoided traps because of frequent v i s i t s by lynx hunting in the area. Secondly, the unusually high predation pressure created by the presence of at least two lynx may have reduced survival of hares on the g r i d . Immigration of ; 102 males to Gr i z z l y was higher than to Si l v e r Creek and i f not for the above factors, t h i s would probably have led to an increase in numbers. A fourth explanation for the differences in response to food addition in 1980 versus 1981 i s that immigrants may have been untrappable in 1980. They may have been present on the food gri d but avoided traps because they had access to food at feeders. This i s supported by results in 1981, when immigration did not increase on Telemetry or Gri z z l y u n t i l food was placed in traps (Fig. 2.7), even though supplemental food had been available a f u l l 2 weeks e a r l i e r . ' Further experimentation i s needed to d i f f e r e n t i a t e between the above p o s s i b i l i t i e s . At present one can only say that supplemental food can increase numbers of hares 2-3 fo l d during the early decline phase of the cycle and l i k e l y may also do so during the peak phase. Why did hares on Telemetry reduce their home range size in 1980 but not in 1981? Home range size may be a function of a variety of factors, only one of which i s food supply. Hares on supplemental food may s t i l l require natural browse to obtain an optimal d i e t . In 1980, the natural browse available on Telemetry may have been such that hares could contract their range and s t i l l obtain their requirements. Browsing in the year of peak densities may have altered the vegetation enough to make this impossible for hares in 1981. It i s also possible that predation pressure may have influenced the home range size of hares on Telemetry in 1981. 103 Predation pressure appeared higher in 1981 than in 1980. Predators were responsible for k i l l i n g 6 of 23 resident radio col l a r e d hares on S i l v e r Creek in 1981 versus 2 of 14 in 1980. Keith and Windberg (1978) also found that predation pressure was higher in post peak versus peak years. Wolff (1980) argued that i hares choose habitats which afford protection from predation in post jpeak years. In 1981, hares on Telemetry spent considerable time off the trapping grid in areas of thick immature spruce and willow. This may have been to avoid predation and hares were unwilling to s h i f t to the more exposed food addition area as evidenced by the lack of increase in use of the food addition area in 1981. These explanations are speculative, but they can be tested. Supplemental food could be added to a grid and natural food and cover could be increased by cutting large spruce trees. Home ranges of hares in t h i s area should decrease i n size after the manipulation. This study showed that a decrease in home range size i s not necessary for hare numbers to increase. This d i f f e r s from findings in deermice (Taitt 1981), voles (Taitt and Krebs 1981), and chipmunks (Mares et a l . 1982). In each case supplemental food led to a decrease in home range size and an increase in immigration. T a i t t (1981) postulated that the change in spacing behaviour caused by the addition of food was responsible for the increased immigration. Mares et a l . (1976) provide strong evidence that this i s true in chipmunks. Findings of this study indicate that immigration can occur in the absence.of a decrease 1 04 in home range si z e . Hares form dominance heirarchies and have home ranges that overlap extensively (Graf 1981; Boutin 197.9). It seems that immigration i s less dependent on changes in home range size in a so c i a l system such as thi s as compared to the more s i t e exclusive systems of chipmunks (Mares et a l . 1976), voles (Madison 1980), and deermice. Home range size i s only one measure of spacing behaviour and may not be the best one. The behavioural changes brought about by food addition, i f they occur in hares, may not necessarily be refl e c t e d by changes in home range s i z e . Supplemental food decreased weight loss and improved survival of both sexes, and advanced onset of breeding condition in males in ;both 1980 and 1981. Vaughan and Keith (1981) provided supplemental food to hare populations in large enclosures. Theirs was a long term food addition but they obtained results similar to thi s study. Empirical evidence c o l l e c t e d over 15 years has led Keith (Keith and Windberg 1978) to hypothesize that overwinter food shortage i s the c r i t i c a l factor i n i t i a t i n g snowshoe hare declines. Results from th i s study indicate that food supply can affe c t survival and weight loss of hares in free ranging populations. Further, food addition on Telemetry in 1981 restored breeding numbers to above the peak leve l s present in 1980. This suggests that the change leading to the decrease in numbers between 1980 and 1981 was food related. This may have been a direct habitat change or a change in the animals' a b i l i t y to survive on a given amount of food. 105 Food Supply and Hare Movements This study confirms Wolff's (1980) contention that food supply is an important factor in determining movement of hares between habitats. Food addition increased rates of immigration but did not a l t e r emigration rates. Dobson (1979) obtained similar results in C a l i f o r n i a n ground s q u i r r e l s . However, emigration in t h i s study was low on both S i l v e r Creek and Telemetry. A better test of the effect of food supply on emigration rates would be to reduce rather ! than increase food supply. The increase in the number of new animals appearing on the study areas in A p r i l of 1980 and 1981 suggests that dispersal increased at t h i s time. This coincides with the period when the f i r s t matings of the year take place and probably r e f l e c t s increased movement by animals searching for favorable areas to breed. The increase in movement in A p r i l 1981 was much more pronounced than in 1980. As well, in 1981, females were moving more than males whereas males and females were moving equally in 1980. I suggest that the difference between years was due to the r e l a t i v e a v a i l a b i l i t y of food in each year. In A p r i l 1980, food supplies were s u f f i c i e n t to support most of the hares present on the study areas. As :a re s u l t , the increase in dispersal at t h i s time was r e l a t i v e l y small. In contrast, food shortage was more pronounced in 1981. Consequently, Telemetry, because food was added, and the Removal, because the r e l a t i v e browsing pressure on the area was reduced by removal of animals, represented areas of favorable food supply. This served to a t t r a c t hares to both 106 areas. The higher response of females was probably due to increased energy demands re l a t i v e to males because they were start i n g to produce embryos. I think the removal area and food addition grids actually attracted animals in 1981 for the following reasons. F i r s t l y , the increase in movement indicated by the removal grid was not reflected in an increase in dispersal of radio c o l l a r e d animals on S i l v e r Creek. Secondly, a large number of immigrants to Telemetry during May were captured only once. These along with immigrants that had home ranges located well off the grid (Fig. 2.13) suggest that many animals were making forays onto the food area. Most did not remain there, possibly because residents did not allow them to do so. It seems then, that hares normally increase t h e i r movements in A p r i l . However, animals do not necessarily disperse at t h i s time but may make forays off their .range for 1-2 days. They then return unless they find a more i favorable area. In 1981, when food supply was limited, the apparent amount of movement was exaggerated by both Telemetry and the Removal because they represented areas of favorable food supply. Animals reaching these areas remained long enough to be captured at least once and some actually established home ranges on or near the g r i d . Food a v a i l a b i l i t y then, i s one factor that determines where hares w i l l s e t t l e in spring and in so doing, can l i m i t density at this time. Not a l l immigrants remain in the area though. One possible reason i s that animals already present a c t i v e l y prevent settlement. In previous experiments I t r i e d to a l t e r spring breeding densities of females by a l t e r i n g spacing ; 107 behaviour (Boutin '1980). They were unsuccessful. However, from findings in thi s study, the experiments may have been done after the period of spring re-organization and so were i n e f f e c t i v e . As well, the experiments were done during the increase phase of the hare cycle. The r e l a t i v e importance of spacing behaviour in determining density may change with phases of the cycle as might the importance of food a v a i l a b i l i t y . Experiments involving both factors during various phases of the cycle are required i f we are to f u l l y understand how they influence snowshoe hare population dynamics. To conclude, supplemental food in late winter decreased weight loss and improved survival in hares as predicted by the Keith (1974) hypothesis. Food a v a i l a b i l i t y can l i m i t numbers in spring but the effect of i n i t i a l densities on the response of hare numbers to food remains unclear. Not a l l animals that enter an area with supplemental food remain there. Some appear to return to their i n i t i a l home range. This may be due to residents on the food addition area preventing them from remaining. 108 SECTION 3. THE RELATIONSHIP OF DISPERSAL TO THE POPULATION DYNAMICS OF SNOWSHOE HARES Introduction Population size and composition are dependent on inputs from births and immigration and output from deaths and emigration. In the past immigration and emigration were considered equal and thus unimportant. This view i s no longer i acceptable (Lidicker 1962, 1975, Krebs and Myers 1974) and to understand a species' population dynamics we must analyze dispersal (Watson and Moss 1979). Evidence that dispersal i s necessary for population regulation comes from studies in which i t has been a r t i f i c i a l l y eliminated. Numbers in increasing populations of M. ochrogaster and M. pennsylvanicus reached abnormally high lev e l s and caused habitat destruction when enclosed (Krebs et a l . 1969). This effect has been observed in M. pinetorum (Gentry 1968) and in peak populations of M. townsendi i (Boonstra and Krebs 1977). Various workers have found no correlation between density and number of dispersers (Myers and Krebs 1971, Windberg and Keith 1976a) while others have found a positive relationship (Van Vleck 1968, Krebs et a l . 1976; see Gaines and McClenaghan 1980 for a review). In c y c l i c species, dispersal i s highest in increasing populations and low during declines (Myers and Krebs 1971., Krebs et a l . 1976, Hilborn and Krebs 1976, Beacham 1980). Garten and Smith (1974) thought dispersal could be important in l i m i t i n g peak densities of old f i e l d mice, but that i t was not the major cause of 109 declines. However, spring declines in some species can be accounted for by dispersal (Beacham 1980). Dispersal can also influence population composition. Dispersers can d i f f e r from residents in age, weight, reproductive condition, behaviour, and genetics (Gaines and McClenaghan 1980),. Further, the c h a r a c t e r i s t i c s of dispersers can vary with density and population phase in c y c l i c species (Myers and Krebs 1971). The study of dispersal i s d i f f i c u l t because of the problem of i d e n t i f y i n g and following dispersers. The approach to th i s problem in small mammal studies has been to define dispersers operationally as those individuals moving into an area which i s kept vacant by removal trapping (Myers and Krebs 1971). This removal grid method has been used to study dispersal in voles (Myers and Krebs 1971; Krebs et a l . 1976; Tamarin 1978), lemmings (Gaines et a l . 1979 a and b), deermice(Fairbairn 1978), ground s q u i r r e l s (Dobson 1979) and snowshoe hares (Windberg and Keith 1976a). This technique assumes that animals caught on the removal area provide a measure of the r e l a t i v e number and type of animals dispersing from surrounding populations. It also assumes that the sample of dispersers obtained in this manner is unbiased, or at least, that the bias remains the same. There have been few attempts to monitor dispersal by alternate methods (Gaines and McClenaghan 1980) and only one study has tested the assumptions of the removal gri d method. Dobson (1981) presents evidence to suggest that removal grids may not provide an accurate measure of dispersal in populations of ground s q u i r r e l s . 1 10 This section examines the role of dispersal in the population dynamics of snowshoe hares. Windberg and Keith (1976a) used a removal grid to monitor dispersal in hares during the peak and decline phase of a population cycle. They concluded that dispersal was not density dependent and could not account for the population decline. However, they used a removal area to monitor dispersal and this may not have sampled a l l dispersers during the decline. I examined dispersal in snowshoe hares during a population increase, peak, and early decline by two methods. These were: 1) the conventional removal gri d method and 2) by telemetry monitoring of radio collared animals. I asked the following questions: 1. Is dispersal related to density and or rate of population increase? 2. Was the decline in hare numbers due to dispersal? 3. Are dispersers similar to residents in age, sex, weight, and body size? 4. Do the removal grid and telemetry techniques show the same amount and type of dispersal? Study Area Two areas were used during the study (Fig 3.1). S i l v e r Creek served as a control while the Removal served as an area in which a l l animals captured each session were removed. The location of the Removal was changed (SCR to R, F i g . 3.1) in 1979 because radio telemetry monitoring of hares on S i l v e r Creek suggested that animals may have been drawn toward the trapped 111 Figure 3.1. Relative locations of the study areas at Kluane Lake, Yukon. S i l v e r Creek (SC), Removal (RE), Pre-1979 Removal (SCR). ! 112 113 i out area. Both removal areas were trapped intensively to remove the resident population before monitoring of dispersal began. Vegetative cover on the grids is described in section 1. Methods Trapping Each study s i t e had a 300 X 300 m trapping gri d with 100 stations arranged 30 m apart in a 10 X 10 pattern. Between 50 and 60 double door l i v e - t r a p s were dis t r i b u t e d at stations spaced evenly over the gri d . Traps were baited with a l f a l f a in winter and apples in summer, set for two consecutive nights, and checked each ' morning. Traps were sometimes set for longer periods on the Removal, i f animals s t i l l remained on the gri d . Trapping began in 1977 and is continuing at present. Frequency of trapping varied over the study but was generally more frequent between February to September and from 1979-1981 than at other times. I t r i e d to keep the Removal free of animals which meant that the grid was trapped more often when more animals were caught. Data col l e c t e d from each capture i s as described in section 1. Telemetry Hares caught on the control and weighing over 700 g were equipped with radio transmitters.; I t r i e d to tag a l l e l i g i b l e animals but was unable to do so because of the a v a i l a b i l i t y of 1 1 4 transmitters. Transmitters were located by triangulation from permanent towers near the study area (Fig. 3.1,see Boutin 1980 for a f u l l d e s c r i p t i o n ) . The frequency of locations varied but was generally 1 per day between February 1 and September 30 and between mid to late November. At other times animals were located at least once a week. I also recorded whether or not an animal was active each time i t was located. This was done by holding the receiving antenna stationary and monitoring changes i n signal strength. E r r a t i c changes in signal strength indicated that the animal was a l i v e . Animals thought to be dead were tracked with a handheld antenna u n t i l they were sighted. As well, locations of radios which were considered unusual because of a major change from previous locations were v e r i f i e d by actually locating the radio. The range of the telemetry system was variable but radios were regularly detected as far as 5 km away from the permanent towers. Range was greatly increased when the receiving antenna was placed well above vegetation and topographic b a r r i e r s . In spring;of each year I t r i e d to locate lost transmitters by monitoring from a high point of Hand (1700 m) overlooking the study areas. Radios were detected as far as 10 km away. In May 1981, the study s i t e s and surrounding area were searched by attaching the receiving antenna to the wing of an a i r c r a f t and f l y i n g over the area. To summarize, most transmitters functioning within a 3-5 km radius of S i l v e r Creek were detectable from the permanent towers and some radios within a 10 km radius of S i l v e r Creek were detected by special searches. 1 15 Thus hares had to disperse at least 5 km and probably further before the transmitter signal was undetectable. Results Dispersal and Density Numbers on Silv e r Creek were estimated by complete enumeration (Krebs 1966) (see section 1). The population was increasing when the study began in 1978, and continued to do so u n t i l f a l l of 1980 (Fig. 3.2). In each year, numbers increased between July and October due to recruitment of juveniles. In 1978 and 1979 numbers remained constant or declined s l i g h t l y (1% per week) over winter (November - March). In 1980-81 however, the overwinter rate of decline was higher (3% per week) u n t i l mid-March. Numbers remained stable u n t i l mid-May when the number of males declined by 10% per week for 4 weeks while females i increased at a rate of 13% per week. Total numbers again increased in July through juvenile recruitment. After a short decrease the population remained constant u n t i l mid-December after which i t declined at a rate of 11% per week to early February. Values remained constant but low for the duration of the study. To summarize, numbers increased from 1978 to a peak in f a l l of 1980. This was characterized by high juvenile recruitment i n July to October and r e l a t i v e l y small declines over winter. Following the peak, the overwinter declines became increasingly severe in 1980-81 and 1981-82 and caused the population to 1 16 Figure 3.2. Changes in minimum number a l i v e (MNA) on S i l v e r Creek (top)land monthly number of dispersers caught on the Removal (bottom). 117 118 decline to below 1978 l e v e l s . Figure 3.2 also shows the number of animals caught per ' month on the Removal. Dispersal occurred throughout the study , but was generally low from 1978 to 1979. It then increased to a peak in the f a l l and winter of peak numbers on S i l v e r Creek. The number of dispersers was low i n May and June but increased during summer and f a l l 1981, before becoming low in January 1 982. I next asked whether dispersal was related to numbers on Si l v e r Creek. I considered three measures of dispersal rate: number of immigrants to the Removal per month, recovery r a t i o , and r e l a t i v e recruitment index. Recovery r a t i o was defined after Krebs et al.(1976) as: number of hares removed from the removal gri d at time t population size on S i l v e r Creek at t The r e l a t i v e recruitment index was defined as: number of untagged hares on S i l v e r Creek at t number of hares removed from the Removal at t I divided the year into fi v e periods (Table 3.1) and lumped data in each. Values for the d i f f e r e n t measures of dispersal rate are shown in Table 3.1. Correlation c o e f f i c i e n t s between population size and measures of dispersal rate are shown in Table 3.,2. Only the average number of immigrants per month was s i g n i f i c a n t l y correlated with population size on S i l v e r Creek. None of the indices of dispersal rate were s i g n i f i c a n t l y correlated with 119 Table 3.1. Population size and rate of increase for S i l v e r Creek (SC) and dispersal rates for the removal area. Recovery r a t i o - number of hares removed from removal area / population size on S i l v e r Creek (SC), Relative recruitment index - number of untagged hares on S i l v e r Creek / number of hares removed from the removal g r i d . Average Inst. Av. no. Recovery Relative no. on rate immigrants r a t i o recr. SC inc. per mo. index 1978 May-June 9 .0270 6. 5 .72 .77 July-Aug-Sep 23 .0561 15. 3 .66 .54 Oct-Nov 23 -.0591 1 1 . 5 .49 .22 1979 Dec-Jan-Feb 1 7 -.0385 3 .18 .33 Mar-April 1 7 .0571 4 .24 1 .25 May-June 18. 6 -.0138 1 1 . 5 .61 .52 Jul-Aug-Sept 34 .0608 39. 6 1.16 .36 Oct-Nov 49 .0372 34 .69 .41 1 980 Dec-Jan-Feb 49 -.0716 1 1 .22 1 .42 Mar-April 50. 5 -.0505 66 1 .30 .14 May-June 48 .0255 22. 5 .47 .42 Jul-Aug-Sept 85 .0684 89 1 .05 .52 Oct-Nov 85 -.0683 101 .5 1.19 . 1 1981 Dec-Jan-Feb 41 -.0484 97. 3 2.39 .09 M a r - A p r i 1 26 -.0492 55 2.15 .07 May-June 25 -.0017 32. 5 1.. 31 ..27 J u l - A u g - S e p t 43. 8 . 0348 40. 6 .92 .75 Oct-Nov 31 . 6 -.0100 35. 5 1.12 .17 1982 Dec-Jan-Feb 9. 2 -.1273 10 1 .09 .16 Mar-April 3. 3 .0817 5 1 .51 .40 May-June 3 1 .33 1 20 Table 3.2. Correlation c o e f f i c i e n t s between population number on the Control and three measures of dispersal rate. Number of sampling periods are in brackets. Av. no. immigrants Recovery r a t i o Relative recruitment index Density .78* (21) .08(21 ) .003(20) Rate of increase 20(20) .12(20) .32(20) * P<.0001 121 rate of population growth. Recovery r a t i o i s a measure of the per capita rate of dis p e r s a l . Values were highly variable but were generally lower (:mean=.65, n=12) in pre-peak than peak and post-peak (mean=1.33, n=9) phases of the cycle. The highest recovery ratios occurred in December to A p r i l 1980-81: the period when Si l v e r Creek numbers were highest. The r e l a t i v e recruitment index provides two useful pieces of information. F i r s t l y , during the period of juvenile recruitment (July - September), i t provides a r e l a t i v e index of how many juveniles disperse from S i l v e r Creek. The r e l a t i v e recruitment index was always less than 1 during t h i s period. Thus, in spite of having no resident juvenile production, recruitment to the Removal was greater than to Si l v e r Creek. This implies that juvenile dispersal from normal populations was high. During the rest of the year, the r e l a t i v e recruitment index provides a measure of the pr o b a b i l i t y that a disperser (animals caught on the Removal) w i l l successfully immigrate to a population (untagged animals caught on S i l v e r ; Creek). The r e l a t i v e recruitment index was greater than one in March - A p r i l 1979 and in December - February 1979-80. In a l l other cases i t was less than 0.55, suggesting that most dispersers did not successfully recruit to established populations. This was p a r t i c u l a r l y true during October-April 1980-81 when the r e l a t i v e recruitment index was less than 0.10. To summarize, only the average number of dispersers per 122 month was correlated with the number of animals on S i l v e r Creek. The recovery ratio was highest during the peak f a l l and ; winter, indicating that the greatest per capita dispersal occurred at this time. Low re l a t i v e recruitment index values during t h i s period indicated that most of the dispersers were unable to immigrate to established populations. Dispersal as Determined by Telemetry Hares with radio c o l l a r s that were known to have died, dispersed, or for which radio contact was lost were lumped into four categories. They were: 1. Animals which had dispersed. A disperser was defined as any individual which l e f t i t s home range to occupy an area d i s t i n c t from i t s i n i t i a l range. 2. Animals found dead on their home range. 3. Animals found dead off their home range. 4. Animals whose fate was unknown (radio transmitter l o s t ) . The number of individuals in each category i s shown in Table 3.3. The most s t r i k i n g finding i s that of 265 animals receiving transmitters during the study, only 23 (9%) were classed as dispersers. The per capita rate of dispersal (columns D/A, Table 3.3) was the same during the population increase and peak (mean=0.05, n=12) as during the decline (mean= 0.03, n=5). Radio-collared dispersers could be missed in two ways. F i r s t l y , contact with radio transmitters could be lost by animals dispersing beyond the range of the telemetry receiver. Most transmitter losses (colunm E, Table 3.3) (72%) occurred in 1 23 Table 3.3. Fates of radio collared hares and the proportion of losses explained by dis p e r s a l . a No. radio Deaths Disp. Trans. % losses co l l a r e d On Off losses explained by Grid Grid dispersal A B C D E F 1978 July-Aug-Sep 16 3 0 2 0 .4 Oct-Nov 23 2 1 1 1 .25 1979 Dec-Jan-Feb 23 3 1 4 0 .50 Mar-April 1 1 5 0 0 2 .0 May-June 29 1 0 0 0 .5 Jul-Aug-Sept 24 •1 1 0 0 .0 Oct-Nov 29 2 0 2 1 .5 1980 Dec-Jan-Feb 47 5 2 2 1 2 .22 Mar-April 33 9 2 . 2 0 .15 May-June 33 2 0 1 3 .33 Jul-Aug-Sept 54 9 1 2 1 . 17 Oct-Nov 59 1 0 7 0 0 .0 1981 Dec-Jan-Feb 56 24 3 2 8 .07 Mar-Apri1 30 8 0 0 0 .0 May-June 33 7 2 1 0 . 1 Jul-Aug-Sept 41 5 1 0 0 .0 Oct-Nov 37 2 0 3 1 .6 1 9 8 2 Dec-Jan-Feb 2 6 1 7 2 0 3 . 0 Grand Total 265 1 1 5 22 23 32 .14 a. calculated as: the number of dispersers ( t o t a l losses - transmitter losses) I 124 the December to February period of 1979-80 and 1980-81. This coincides with the time when the animals were monitored less intensively, thus increasing the pr o b a b i l i t y of los s . It i s doubtful that a l l transmitter losses could be attributed to dispersal but the exact proportion i s unknown. A second group of dispersers which may have been missed is animals that were found dead off their home range. These individuals may have been k i l l e d while dispersing or may have been k i l l e d on their home range and c a r r i e d off by a predator. If hares found dead off the grid are classed as dispersers (column C + D, Table 3.3), the maximum percentage of radio c o l l a r e d animals dispersing during any one time period was s t i l l only 22%. To summarize, few radio-collared animals dispersed in any time period during the study. Some dispersers may have been missed in the December-February period because of transmitter losses. The proportion of radio c o l l a r e d animals dispersing did not change over the cycle. Population Losses and Dispersal In each time period during the study, radio c o l l a r e d animals were lost from the population by death or di s p e r s a l . I next asked what percentage of these losses (columns B + C + D, Table 3.3) could be attributed to dispersal ( column F, Table 3.3). Hares with unknown fates because the transmitter was lost (column E) were excluded from the analysis. Generally, the per cent losses explained by dispersal was greater during the pre-125 peak (mean=0.27, n-11) vs. the post-peak period (mean=0.11, n=6). Dispersal accounted for as much as 60% of the losses in October-November 1981 but accounted for only 14% o v e r a l l . Dispersal explained a large proportion of losses only when the t o t a l number of losses was low. The percentage of losses explained by dispersal was not correlated with density (r=0.08, n=l8,p>0.l) or with rate :of population growth (r=0.3, n=l8, p>0 . 1 ) . To summarize, losses due to dispersal were higher, in general, when the population was increasing. Dispersal accounted for an average of only 11% of the losses during the decline. Sex Ratio of Pispersers Sex r a t i o of animals on Silv e r Creek was determined by t a l l y i n g each animal every time i t was trapped and summing within time periods. Table 3.4 shows the sex r a t i o of animals on Si l v e r Creek and on the Removal. Sex ra t i o of the two areas was s i g n i f i c a n t l y d i f f e r e n t in December-February 1980-81 (chi-square=7.35, df=1, P<.01) when there were more males on S i l v e r Creek but s l i g h t l y more females on the Removal. The proportion of dispersing females was s i g n i f i c a n t l y greater than 0.5 in May-June 1981 (chi-square=6.78, df=1, P<.01). Overall, dispersal was not sex-related. 1 26 Table 3.4. Sex ra t i o on S i l v e r Creek and the Removal. Sample sizes are in brackets. S i l v e r Creek Removal 1978 May-June Jul-Sept Oct-Nov .44(34) .36(115) .48(48) .57(26) .44(45) ,50(24) 1979 Dec-Feb Mar-Apr i1 May-June July-September October-November .54 (1 1 ) .57(37) .41(87) .42(95) .57(51 ) ,67(9) ,62(8) .25(4) .45(110) .45(81 ) 1 980 December-February March-April May-June July-September October-November .43(124) .49(239) .56(145) .51(410) .53(207) 56(34) 55(131 ) 42(45) 45(254) 58(215) 1981 December-February March-April May-June July-September October-November .61* (105) ,.65(75) .39(51 ) •45(196) .56(25) 45(227) 54(94) 34(63) 57(90) 45(130) 1982 December-February ,40(67) March-April ,37(8) * chi-square = 7.35, P<,01 .59(44) .3(10) 1 27 Age of Dispersers Animals caught on the Removal were classed as adults or juveniles during the July-September period. I could not distinguish between adults and large juveniles in other periods, so a l l animals were lumped together. Table 3.5 shows that dispersal was biased toward juveniles in the July-September period in a l l years. The proportion of adults was 2.5-8 times higher on S i l v e r Creek than on the Removal. This was probably true during October-November also. Of the radio c o l l a r e d animals that dispersed, 13 of 23 did so before their f i r s t breeding season. Weight and Body Size of Dispersers I next asked whether dispersers d i f f e r e d in body weight from that of residents (Figs. 3.3, 3.4). Adults were excluded from July-September samples because few adults were dispersing at ; this time. Both male and female dispersers were l i g h t e r than residents during October-November in a l l years except 1978. Male dispersers were also l i g h t e r than male residents in March-April of 1981. In females, dispersers were l i g h t e r than residents in December-February 1980-81. In March-April 1982, female dispersers were heavier than female residents. I used the length of hares' right hind foot as a measure of body size. Differences in body size between dispersers and residents followed the same pattern as differences in : body weight (Figs. 3.5, 3.6). Dispersers were smaller on average than 1 28 Table 3.5. .Adults as a proportion of the t o t a l number of animals caught on S i l v e r Creek and the Removal during July-September. Sample sizes are in brackets. S i l v e r Removal Creek 1978 .23(118) .09(44) 1979 .35(95) .07(119) 1980 .48(410) .06(278) 1981 .59(196) .14(125) 129 Figure 3.3 Average(± 2SE) weights of male dispersers caught on the Removal (RE) and male residents caught on S i l v e r Creek (SC),. O CD CD ro o 00 - o > _ I- _ CD i_ Z O •n D > _ «- _ CO t_ Z O -n D > _• <- _ _ o -n o > _ 00 ro Mean Weight (g) o o r • o Jo. I — t — I -O 1 • — — l - •—I r - C — i K H t-CH i - O t t-O-i o o • 3J CO m o CO O 131 Figure 3.4. Average (± 2SE) weights of female dispersers caught on the Removal (RE) and female residents caught on S i l v e r Creek (SC). CD ^1 oo CD oo o oo oo ro <- _ CO (_ z O -n O > _ <- _ CO <_ Z O — a > _ <- _ C0i_ 2 O -n a > _ <- _ co <_ Z O •n a > _ Mean Weight (g) CD o o cn O O P I • 1 K H I—O—I t -OH I • K X K H I • 1 I • 1 t - O - l I i - O - t o • 3J cn m o CO 1 33 Figure 3.5. Average (± 2SE) length of right hind foot as a measure of body size for male dispersers caught on the Removal (RE) and male residents caught on S i l v e r Creek (SC). Length of RHF (mm) o JL co C O Z O -n o > 2 CO co o co CO <_ Z O -n o > ^ co t_ z O -n o - _ CO <_ Z O -n o > Z 00 ro t - O -KM K H ! — • — I i - O - H K M o • 73 cn m o 4>* Figure 3.6. Average (± 2SE) length of right hind foot as a measure of body size for female dispersers caught on the Removal (RE) and female residents caught on S i l v e r Creek (SC) . o o _ ^ - 2 CD •S CO c_ 00 Z O •n O > 2 CO CO i _ Z O — o > 2 oo o oo 00 r o cn <— z O n o > 2 <- 2 CO i _ z o •n O > 2 Length of RHF (mm) -o 1 -o-i — • — i -o 1 I • 1 K5H t — • — i K> -CM K M K > o • 30 CO m o CO c n 137 residents in October to November. This was true for females in December-February also. Male dispersers were of larger body size in July-September of 1980 and 1981. This was also true of i females in 1979. To summarize, male and female dispersers were smaller and li g h t e r than residents in October-November and in the winter of 1980-81. i Discussion I began th i s chapter asking the questions "Is dispersal related to snowshoe hare numbers and i s dispersal responsible for the declines observed in snowshoe hare numbers?" The number of animals caught on a t o t a l removal area was correlated with the density of the S i l v e r Creek population but not with i t s rate of growth. The per capita rate of dispersal (recovery r a t i o ) showed no r e l a t i o n to density or rate of population growth. It was density independent. Recovery rat i o s were highest in the peak winter as was the number of dispersers. Dispersal accounted for only 11% of the t o t a l number of losses of radio-collared animals during the decline, Thus, I conclude that the observed decline in snowshoe hare numbers was not due to d i s p e r s a l . Findings of this study agree with those of Windberg ; and Keith (1976a) who monitored immigration to a removal grid during peak and declining hare dens i t i e s . They found a higher number of dispersers ,during the peak year and recorded a recovery r a t i o i ! 2.4 times higher in the peak winter than during the decline. The relationship of dispersal to changes i n snowshoe hare 138 numbers follows the pattern observed in c y c l i c microtines (Gaines and McClenaghan 1980). The number of dispersers i s correlated with density and dispersal explains a higher percentage of the losses from control populations during phases of population increase than during phases of decline (Myers and Krebs 1971; Krebs et a l . 1976; Tamarin 1977). Microtine populations c h a r a c t e r i s t i c a l l y exhibit spring declines in numbers which vary in duration and degree (Krebs and Boonstra 1978).The rate of decline usually increases at the beginning of the breeding season, and i t has been suggested that t h i s i s due to increased aggression (Krebs and Boonstra 1978). Krebs and Boonstra (1978) hypothesized that losses during moderate declines in Microtus townsendii were due to dispersal but that severe declines were due to death. Beacham (1980) la t e r found experimental evidence to support t h i s . Results of t h i s study suggests that snowshoe hares do not exhibit yearly spring declines, although increased movement may occur at this time. This i s suggested by events on S i l v e r Creek in spring 1981, when males showed a rapid decrease while females increased (Fig. 3.2). The decrease in males was due to death as indicated by telemetry and the increase in females was due to immigration since juvenile recruitment had not yet begun. Further evidence for an increase in movement in spring was discussed in section 2. 139 Characteristics of Dispersers Dispersers were not a random sample of the control population. Most were individuals less than one year old. As i well, dispersers were li g h t e r in weight and had smaller body size than residents. The difference was most pronounced in October-November of 1979, 1980 and 1981. This was probably due to a disproportionate number of juveniles dispersing at t h i s time. Windberg and Keith (1976a) found that dispersers were lig h t e r than residents during the winters of the peak and f i r s t year of population decline. However, they found that dispersers had larger body size than residents at this time. They argued that the l i g h t e r body weight of dispersers and a high dispersal rate (recovery ratio) during the winter of peak density reflected overwinter food shortage and that hares which were unsuccessful at competing for food supplies (lighter weight) were forced to disperse. This study found that dispersers in the peak winter (males in March-April and females in December-February) were l i g h t e r than residents but not in the winter of early decline. This suggests that i f food shortage i s r e f l e c t e d by l i g h t e r weight dispersers, food was in short supply during the peak winter only.. Snowshoe hares did not show sex-biased dispersal in t h i s study. This i s unusual for mammals in which most species show male biased dispersal (Greenwood 1980). Graf(1981) found that hares form dominance heirarchies with females being dominant to males in summer and vice versa in winter. These behavioural heirachies did not lead to one sex dispersing more than the 1 40 other however. \ i I n i t i a l l y , workers characterized dispersers as individuals which had been forced to leave their home area by intraspec1fic s t r i f e (Errington 1946). These animals were normally subordinate individuals in poor physical condition and with l i t t l e chance of surviving, Lidicker (1975) was the f i r s t to point out that dispersal at certain times may be adaptive and should not always be equated with mortality. He coined the term presaturation dispersal to refer to dispersers which are in good physical condition and have a r e l a t i v e l y high p r o b a b i l i t y of • surviving even though they disperse. Dispersers of t h i s type tend to leave their i n i t i a l home area before conditions there force them to do so (the carrying capacity i s reached). Lidicker (1975) termed the more conventional type of dispersal in which subordinate individuals are forced to disperse, saturation d i s p e r s a l . These concepts have created much confusion in recent dispersal l i t e r a t u r e for two reasons. F i r s t l y , carrying capacity of a habitat is a nebulous concept, especially when the environment i s very heterogeneous. Workers have (employed a c i r c u l a r argument to define carrying capacity of an area as being reached when saturation dispersal occurs, and ; in turn saying that the carrying capacity of the environment has been reached so I saturation dispersal is occurring (Wolff 1980). In c y c l i c species, the amount of habitat that i s occupied can change d r a s t i c a l l y over a cycle, and certain areas w i l l f i l l with animals sooner than others. Dispersal before a l l habitats have been f i l l e d , is by d e f i n i t i o n , presaturation dispersal (Lidicker 141 1975; Wolff 1980). This interpretation misses the c r u c i a l point that dispersers at this time could s t i l l be s o c i a l subordinates which have been forced from occupied areas. The second source of confusion i s that workers have attempted to argue that the type of dispersal that i s occurring can be inferred from certain c h a r a c t e r i s t i c s of the dispersers. For example, Lidicker (1975) argued that the dispersal of young ' i pregnant females probably indicates that presaturation dispersal is occurring. Myers and Krebs (1971) found a disproportionate number of young females dispersing from vole populations. However, Bujalska (1973) and Saitoh (1981) have shown that older female voles suppress maturation of young females. These individuals then, are subordinate and may s t i l l be forced to disperse from the population by s o c i a l i n t e r a c t i o n . S i m i l a r l y , Krebs (1978) argued that a high proportion of juvenile dispersers suggests that saturation dispersal i s occurring. However, juvenile dispersal i s c h a r a c t e r i s t i c of most species and does not imply that the young are forced to leave their home area. The c r u c i a l issue that the presaturation dispersal concept brings attention to is that the cause of dispersal can f a l l into two categories. In the case of presaturation dispersal, the animal i s not forced to leave by s o c i a l aggression. Stenseth (1983) points out that t h i s type of dispersal should have a strong genetic component. In contrast, saturation dispersers are forced to emigrate by i n t r a s p e c i f i c competition. These individuals would have a low genetic propensity for 142 d i s p e r s a l . The c r u c i a l issue then, i s to determine what causes animals to disperse. This cannot be implied from c h a r a c t e r i s t i c s of dispersers or from the density of animals in an area. ;A more useful approach would be to manipulate factors which could cause dispersal and observe corresponding changes in the number and type of dispersers that r e s u l t . The concept of saturation and pre-saturation dispersal also predicts that survival of the two types of dispersers should d i f f e r , with pre-saturation types having r e l a t i v e l y higher survival than their saturation counterparts. Again, this does not necessarily follow. During periods of population increase, some unoccupied habitats may become temporarily favorable. Both presaturation and saturation dispersers w i l l have a r e l a t i v e l y high p r o b a b i l i t y of surviving in these. S i m i l a r l y , at peak densities, both types of dispersers w i l l survive poorly because there are no unoccupied habitats a v a i l a b l e . It i s not the type of disperser then, that determines survival but rather the a v a i l a b i l i t y of favorable habitats, which in turn, i s determined by phase of the population cycle. The r e l a t i v e recruitment index provides a measure of the success of dispersers at immigrating to established populations (excluding the period for juvenile recruitment) and i s thus a crude index of survival of dispersers. Dispersing hares were more successful at r e c r u i t i n g to established populations during population increase (mean=0.61, n=9) than during the peak and decline (mean=0.l8, n=7, Table 3.1, July-September periods excluded). 143 Habitat Heterogeneity and Pispersal Wolff (1979) used the concepts of presaturation and saturation dispersal to produce a model to explain changes in hare d i s t r i b u t i o n during a population cycle. The model defines three types of habitat; refugia, which are occupied by hares throughout the cycle, suboptimal habitat which is used by hares during periods of increase, and marginal habitat which i s used at peak densities. Wolff (1980) concluded that presaturation dispersal i s responsible for habitat expansion during the increase phase of the cycle. At peak densities, hares encounter "frustrated d i s p e r s a l " because they are unable to find vacant suitable habitat. This results in carrying capacity of a l l areas being exceeded and severe overbrowsing r e s u l t s . This leads to a decline in hare numbers which i s extended by predation. Wolff (1980) argued that d i f f e r e n t i a l predation pressure in more open habitats lead to the contraction of habitat use during the decline. Evidence for the model is inconclusive. The c r u c i a l information on loss rates due to predation in various habitats is lacking, and conclusions drawn by Wolff from browsing intensity data are suspect because spruce, the species used most by hares, was excluded from the analysis (Wolff 1978, 1980). Also, data presented by Wolff (1980) suggests that emigration of hares from suboptimal areas to refugia could explain the contraction in habitat use during the decline rather than d i f f e r e n t i a l predation pressure. The model i s useful though because i t points out that the amount, timing, and type of 144 dispersal could d i f f e r between habitats. Future studies should monitor dispersal in a variety of habitats at the same time. Wolff (1979) postulated that cycles in hare numbers were more prominent in the northern part of the species' range because of the lack of "dispersal sinks". Dispersal sinks are areas of habitat into which animals disperse and suffer high rates of mortality. Wolff (1980) argues that hares in the north are able to disperse between habitats with l i t t l e loss from predation because predation numbers are low r e l a t i v e to those in more southerly portions of the hare's range. Dispersal sinks in the north then, are i n e f f e c t i v e at removing ;dispersers. This leads to a buildup in hare numbers in a l l habitats and eventual habitat destruction through overbrowsing. In the south, habitats are more discontinuous, and predation pressure i s r e l a t i v e l y high. Most dispersers are unable to survive in suboptimal habitat and high hare densities do not occur. These ideas are d i f f i c u l t to test. They require measures of r e l a t i v e survival rates of ;dispersers over the population cycle and between northern and southern populations. Tamarin (1978) argued that the presence of "dispersal sinks" are important in determining whether a population w i l l show cycles. Dispersal sinks are areas in which animals disperse and suffer higher rates of mortality. High numbers of animals disperse into these sinks and t h i s leads to population declines and c y c l i c a l population behaviour. In the absence of dispersal sinks, as on islands, dispersers are not removed by predators and are accepted back into the population (Tamarin 1978). This 145 leads to stable population numbers. This study does not support Tamarin's arguments. Dispersal was less prominent during the population decline and most losses were due to death in s i t u (Table 3.3) as opposed to dispersal and predation thereafter. Therefore, the population should not have shown c y c l i c a l behaviour. Methods of Studying Dispersal The amount of dispersal measured by the : removal grid technique d i f f e r e d from that determined by telemetry in two ways. F i r s t l y , the large number of immigrants appearing on the removal g r i d suggested that dispersal was prominent while telemetry indicated that few animals dispersed , at any time during the study. Secondly, the per capita rate of dispersal (recovery r a t i o ) , as measured by the removal grid, indicated that dispersal was highest in peak and early decline phases of the cycle. However, the proportion of radio c o l l a r e d hares that dispersed did not change with phase of the cycle. I w i l l discuss the reasons for these differences in the following section. The use of a removal gri d as an experimental method to monitor dispersal is based on a number of assumptions. These are that: 1) the removal area does not a f f e c t animals on the control area, 2) the majority of dispersers are caught each trapping session, and 3) the animals being removed are actual dispersers as opposed to individuals making brief forays from their home range (Krebs et a l . 1976). The location of the removal g r i d was moved in t h i s study because radio telemetry suggested that 146 movements of animals on S i l v e r Creek were affected by the the presence of the removal. After i t s relocation in 1979, only 9 tagged individuals from S i l v e r Creek were captured there. This made i t impossible to use the number of animals tagged on S i l v e r Creek and l a t e r caught on the Removal as a measure of the percent losses from S i l v e r Creek due to d i s p e r s a l . Given random di r e c t i o n of dispersal, the number of tagged dispersers encountering a removal area is highly dependent on i t s distance from the control area. The removal area samples an unknown percentage of the t o t a l number of animals dispersing from the control area. This percentage though, is probably well below 100%. My use of telemetry to monitor d i s p e r s a l , recorded v i r t u a l l y a l l losses due to d i s p e r s a l . Yet the percentage loss explained by dispersal in this study was lower than most studies using recovery of tagged individuals on a removal area (Gaines and McClenaghan 1980). This means that dispersal was either much less common in snowshoe hares than in other species of small : mammals or that removal gri d studies overestimate losses due to dis p e r s a l ; possibly because they are too close to the control area. Animals may be attracted to the removal area because of the reduced density there. Tamarin (1978) provides evidence that more tagged voles were caught on a removal gri d used in his study than was expected by random movement. Dobson (1981) argued that an abnormal ,sex r a t i o of ground squirrels on a control area < was due to dispersal of animals to a nearby removal area. As well, the removal of animals from an area a l t e r s i t s f a v o r a b i l i t y r e l a t i v e to that of surrounding areas (section 2). 147 The r e l a t i v e difference in density between that on the removal area and that i n surrounding areas w i l l change over a population cycle. During the increase phase of the cycle, the removal grid may merely represent a patch among many which i s free of hares. As numbers build however, and more habitats become occupied, the removal g r i d w i l l become more favorable and thus overestimate the average amount of dispersal that i s occurring. Also, i f hares encounter winter food shortage during peak populations, the removal gri d may have r e l a t i v e l y more food because density and browsing pressure are lower. The removal area then, attracts hares, and would do so more strongly during peak de n s i t i e s . This explains why there was an increase in the recovery r a t i o at this time but no increase in dispersal from S i l v e r Creek as measured by telemetry. Further, I argue that the large short term increases in dispersal to the removal area in late A p r i l to early May, 1981 (section 2) and the coresponding increase in immigration on the food addition grids at thi s time, was due to the two areas representing r e l a t i v e l y favorable habitats. Both were areas of high food a v a i l a b i l i t y and thus were favorable areas for hares to s e t t l e . How can the removal area "att r a c t " animals? It i s possible that hares make forays off of their home ranges to assess the f a v o r a b i l i t y of surrounding areas. I recorded 5 instances in which radio c o l l a r e d animals made long range movements off their home ranges for one to two days and then returned. Whether an animal actually becomes a disperser (leaves i t s home range ; 148 permanently) then, is dependent on conditions in surrounding areas as well as in i t s home area. A removal g r i d may overestimate dispersal by preventing animals that would normally return to their home range from doing so (caught in traps) or by having them remain on the grid because conditions are unusually favorable. This is more l i k e l y to occur during periods of high density when animals would not normally f i n d vacant areas to se t t l e and so would return to their o r i g i n a l home range. If thi s situation occurs, the removal gri d violates the assumption that i t i s sampling true dispersers rather than individuals making brief forays off their home range. These animals are potential dispersers but the a v a i l a b i l i t y of vacant habitat dictates whether they actually become dispersers. These ideas can be tested. Movements of animals could be monitored by telemetry in areas near those that have been manipulated by lowering density or adding food. The prediction i s that more dispersal should occur near the manipulated areas as compared to areas located farther away. The use of telemetry as a method of monitoring dispersal has much appeal. Individuals can be monitored on a da i l y basis and their exact whereabouts can be known at a l l times. As well, the range over which an animal can be monitored is much greater than that using only live-trapping grids. Theoretically a l l dispersers in a sample of a radio-collared population can be detected rather than an unknown sample as is the case when removal grids are employed. As well, dispersers need not be defined operationally as those individuals caught on a removal 1 49 area (Myers and Krebs 1971) or as those individuals crossing a s t r i p of unfavorable habitat (Beacham 1980). The major disadvantage of telemetry i s that, at least for snowshoe hares, only a small proportion of animals disperse from the population. As a result a large number of animals must be tagged to obtain many instances of d i s p e r s a l . The most s t r i k i n g result of t h i s study is the small number of radio-collared animals that did disperse. It would require a p r o h i b i t i v e number of c o l l a r e d animals to obtain enough cases of dispersal to examine differences in body size and condition between residents and immigrants. It seems then, that both the removal area and telemetry methods of monitoring dispersal have advantages and disadvantages. Studies of dispersal should use both methods. To conclude, dispersal was not responsible for the observed decline in snowshoe hares. Comparison of dispersal as monitored by the conventional removal g r i d method with that determined by telemetry monitoring of animals indicated that the removal area over-estimated the amount of dispersal that occurred. This over-estimation became more pronounced in the peak and early decline of the cycle. This was because r e l a t i v e l y high food a v a i l a b i l i t y on the removal area served to att r a c t hares. The study indicated that r e l a t i v e f a v o r a b i l i t y of various habitats must be considered when studying d i s p e r s a l . ! 150 j GENERAL DISCUSSION i • Hypotheses Explaining Population Cycles ! A number of hypotheses have been advanced to explain c y c l i c a l population fluctuations (Krebs and Myers 1974). I w i l l now examine how findings of t h i s study relate to these general hypotheses. Keith (1974) advanced the hypothesis that snowshoe hare declines are i n i t i a t e d by food shortage. Findings in this study are consistent with t h i s hypothesis. Food addition to free ranging populations caused decreased weight loss and improved survival as predicted. As well, food addition to Telemetry in 1981, the f i r s t year of decline, restored numbers to peak l e v e l s . This suggests that the reason for the differences in numbers between peak and decline years was food related. However, the actual mechanism could have been lower quantity of food ;(Keith and Windberg 1978), quality of food (Bryant and Kuropat 1980), or that the animals themselves changed, such that they required more food to survive in 1981 than in 1980. An al t e r n a t i v e hypothesis to explain hare declines i s that spacing behaviour acts through d i f f e r e n t i a l dispersal of certain genotypes to change the composition of the population from non-aggressive to aggressive animals (Krebs et a l . 1976). Aggressive , genotypes are less able to cope with normal environmental 'pressures and t h i s leads to a decline (Chitty 1967). I found that manipulation o:f spacing behaviour did not a l t e r juvenile survival or emigration rates. Further, this hypothesis predicts that residents should d i f f e r from dispersers during the increase i • '• • . 151 i ; phase of the cycle-. Except for a higher proportion of juvenile ! i dispersers t h i s was not the case. However, genetic and i behavioural c h a r a c t e r i s t i c s were not examined. Charnov and Finerty (1980) argue that dispersal acts to i ! a l t e r the average relatedness among individuals in a population. During phases of .lowj numbers, animals exist in isolated groups in which relatedness i s high and immigration i s low. As numbers increase, more animals disperse to other populations. These strangers increase the amount of aggression in the population and t h i s leads to even more di s p e r s a l . F i n a l l y , at high densities, aggression acts to i n h i b i t breeding and the population declines. This hypothesis predicts that movement of animals between populations should be high at peak population densities. Results from section 2 support t h i s prediction. To summarize, findings in t h i s study are consistent with the hypothesis that the decline in hare numbers was food related. I found no evidence that spacing behaviour played a s i g n i f i c a n t role in the decline. Di spersal i This study used two new techniques to monitor hare movements between populations. F i r s t l y , calcium-45 was used to detect immigration by juveniles before they were trapped for the f i r s t time. Results indicated that considerable immigration did occur. This early movement would have gone undetected i f calcium-45 had not been used and suggests that we cannot assume that most juveniles caught on a trapping gr i d for the f i r s t time 1 52 are residents. Secondly, the use of radio telemetry to monitor dispersal of animals greater than 700 g indicated that very few of these animals dispersed. It also showed that the removal g r i d over-estimated dispersal and this was more pronounced during the peak and early decline phase of the cycle. This was because the f a v o r a b i l i t y of the removal gri d as a place for hares to s e t t l e , increased as other areas became occupied and food became l i m i t i n g . Further, some hares appeared to leave their home range to check surrounding areas but then returned, possibly because there were no vacant areas. This suggests that dispersal then, is not only a function of factors in the animals' home environment, but of surrounding areas also. Potential dispersers may return to their i n i t i a l home range i f a l l habitats are occupied. In snowshoe hares i t appears that animals dispersing in .summer are more l i k e l y to immigrate to a new population than those moving in f a l l . The questions s t i l l remains as to what causes some animals to disperse, and what allows some to enter a new population while others are prevented from doing so. This study found evidence that food and spacing behaviour can aff e c t immigration rates but neither appeared to influence emigration rates. It i s possible that dispersal in hares has a strong genetic component, possibly as an adaptation to coping with rapidly changing environments. The dynamic nature of these changes along with the heterogeneity of the environment must be taken into account in future studies of snowshoe hares and c y c l i c species in general. Models of habitat heterogeneity such as the one developed by Wolff (1980) are useful beginnings but 153 i t i s only through the development of hypotheses with s p e c i f i c a p r i o r i predictions and subsequent c r i t i c a l experiments that the interaction between behaviour and external factors w i l l be understood. 1 54 LITERATURE CITED Beacham, T.D. 1979. Size and growth c h a r a c t e r i s t i c s of dispersing voles, Microtus townsendii . Oecologia 42: 1-10. Beacham, T.D. 1980. Dispersal during population fluctuations of the vole, Microtus townsendii J. Anim. Ecol. 49:862-877. Boonstra, R. and C.J. Krebs. 1977. A fencing experiment on a high density population of Microtus townsendi i . Can. J. Zool.. 55: 1166-1175. Boonstra, R. 1978. Effect of adult townsend voles on survival of young. Ecol. 59: 242-248. Boutin, S. 1979. Spacing behaviour of snowshoe hares in r e l a t i o n to their population dynamics. M.Sc. Thesis. U.B.C. Boutin, S. 1980. Effect of spring removal experiments on the spacing behaviour of female snowshoe hares. Can. J. Zool, 58:2167-2174. Bryant, J.P. and P.J. Kuropat. 1980. Selection of winter forage by subarctic browsing vertebrates: the role of plant chemistry. Annu. Rev. Ecol. Syst. 11:261-285. Bujalska,G. 1973. The role of spacing behaviour among females in the regulation of reproduction of the bank vole. J. Repr. and Fert. 19:465-474. Charnov, E.L. and J.P. Finerty. 1980. Vole population cycles: a case for kin-selection? Oecologia 45: 1-2. Chitty, D.H. 1967. The natural selection of self-regulatory behaviour in animal populations. Proceedings Ecol. Soc. of Australia 2: 51-78. Chitty, D.H. and E. Phipps. 1966. Seasonal changes in survival in mixed populations of two species of voles. J. Anim. Ecol. 35:313-331. 1 55 Chr i s t i a n , J.J. 1970. Social subordination, population density, and mammalian evolution. Science 168; 84-90. Cole, F.R. and G.O. B a t z l i . 1978. Influence of supplemental feeding on vole populations. J. Mammal. 5 9 : 8 0 9 - 8 1 9 . Daly, J . 1981. Eff e c t s of so c i a l organization and environmental d i v e r s i t y on determining the genetic structure of the wild rabbit, Oryctolagus cuniculus . Evol. 35:689-706. Davis, D.E., J.J. Chris t i a n , and F. Bronson. 1964. Effect of exploitation on b i r t h , mortality, and movement rates in a woodchuck population. J. Wi l d l . Manage. 28:1-9. Dobson, F.S. 1979. An experimental study of dispersal in the C a l i f o r n i a ground s q u i r r e l . Ecol, 60: 1103-1109. Dobson, F.S. 1981. An experimental examination of an a r t i f i c i a l dispersal sink. J. Mammal. 62:74-81. Dolbeer, R.A. and W.R. Clark. 1975. Population ecology of snowshoe hares in the central rocky mountains. J . Wildl. Manage. 39: 535-549. Douglas, G. 1974. Montane zone vegetation of the Alsek River region, southwestern Yukon. Can. J. Bot. 52:2505-2532. Dunford, C. 1977. Behavioural l i m i t a t i o n of round-tailed ground s q u i r r e l density. Ecol. 58:1254-1268. Errington, P. 1946. Predation and vertebrate populations. Q. Rev. B i o l . 21:144-177, 221-245. Fai r b a i r n , D.J. 1978. Dispersal of deer mice, Peromyscus  maniculatus . Oecologia 32: 171-193. Flowerdew, J.R. 1973. The effect of natural and a r t i f i c i a l changes in food supply on breeding woodland mice and voles. J. Repr. Fert. Supp. 1.9:259-269. 156 Gaines, M.S., C.L. Baker, and A.M. Vivas. 197.9a. Demographic attributes of dispersing southern bog lemmings Synaptomys  cooperi in Eastern Kansas. Oecologia 40: 91-101. Gaines, M.S., C.L. Baker, and A.M. Vivas. 1979b. An experimental analysis of dispersal in fluctuating vole populations: demographic parameters. Ecol. 60:814-828. Gaines, M.S. and L.R. McClenaghan J r . 1980. Dispersal in small mammal's. Ann. Rev. Ecol. Syst. 11:163-69. Garten, C T . and M.H. Smith. 1974. Movement by o l d f i e l d mice and population regulation. Acta. T h e r i o l . .19: 513-514, Gentry, J.B. 1968. Dynamics of an enclosed population of pine mice, Microtus pinetorum . Res. Pop. Ecol. 10: 21-30. Gi l b e r t , B.S. and C.J. Krebs. 1981. Effects of food on Peromyscus and Clethrionomys populations in the southern Yukon. Oecologia 51:326-331. Graf. R.P. 1981. Some aspects of snowshoe hare behavioural ecology, unpubl M. Sc. Thesis, University of B r i t i s h Columbia. Green, R.G. and CA. Evans. 1940. Studies on a population cycle of snowshoe hares on the Lake Alexander area. J . W i l d l . Manage. 4: 220-238, 267-278, 247-258. Greenwood. P.J. 1980. Mating systems, philopatry, and dispersal in birds and mammals. Anim. Behav. 28:1140-1162. Hayne, D.W. 1949. Calculation of size of home range. J. Mammal. 30:1-18. Healey, M.C 1967. Aggression and self - r e g u l a t i o n of population size in deermice. Ecol. 48: 377-392. Heezen, K.L. and J.R. Tester. 1967. Evaluation of radio-tracking by triangulation with special reference to deer movements. J. W i l d l . Manage. 31:124-141, 157 Hilborn, R., and C.J. Krebs. 1976. Fates of disappearing individuals in fluctuating populations of Microtus townsendii . Can. J. Zool. 54:1507-1518. Howard, W.E. 1960. Innate and environmental disperal of individual vertebrates. Am. Midi. Nat. 63: 152-161. J o l l y , G.M. 1965. E x p l i c i t estimates from capture-recapture data with both death and immigration stochastic model. Biometrika 52:225-247. Keith, L.B. 1963. W i l d l i f e ' s ten-year cycle. Univ. Wisconsin Press, Madison, Wis. 201 pp. Keith, L.B. 1966. Habitat vacancy during a snowshoe hare decline. J. Wildl. Manage. 35:16-20. Keith, L.B. and L.A. Windberg. 1978. A demographic analysis of the snowshoe hare cycle. Wild. Monogr. 58. Keith, L.B. 1974. Some features of population dynamics in mammals. Proc. Internatl. Cong. Game B i o l . 11:17-58. Keith, L.B. and D.C. Surrendi. 1971. Effects of f i r e on a snowshoe hare population. J. Wildl. Manage. 35:16-26. Krebs, C.J. 1978. Aggression, dispersal, and c y c l i c changes in populations of small rodents, i n : Aggression, dominance, and individual spacing Krames, L., P. Pl i n e r , and T. Alloway eds. Plenum Publishing Corporation, 1978. Krebs, C.J. and J.H. Myers. 1974. Population cycles in small mammals. Advances in Ecological Research. 8: 267-399. Krebs, C.J. I. Wingate, J. Leduc, J. A. Redfield, M. T a i t t , R. Hilborn. 1976. Microtus population biology: dispersal in fluctuating populations of M. townsendi. Can. J. Zool. 54:79-95. 158 Krebs, C.J. 1966. Demographic changes in fluctuating populations of Microtus c a l i f o r n i c u s . Ecol, Monogr. 36: 239-273. Krebs, C.J., B.L. Kel l e r , and R.H. Tamarin. 1969. Microtus population biology: Demographic changes in fluctuating populations of M. ochrogaster and M. pennsylvanicus in southern Indiana. Ecol. 50: 587-607. Krebs, C.J. and R. Boonstra.' 1978. Demography of the spring decline in populations of the vole, Microtus townsendi i J . Anim. Ecol. 47:: 1 007-1 01 5. Lidicker, W.Z. 1975. The role of dispersal in the demography of small mammals. In: Small mammals : productivity and dynamics of populations (K. Petrusewicz, F.B. Golley, L, Ryszkowski, eds.). Chapt. 5. New York. Cambridge Univ. Press. Madison. D. M. 1980. Space use and s o c i a l structure in meadow voles, Microtus pennsylvanicus. Behav. Ecol. and Sociobiol. 7:65-71. Mares, M.A., M.D. Watson, and T.E. Lacher, J r . 1976. Home range perturbations in chipmunks (Tamius s t r i a t u s ) . Food supply as a determinant of home range size and density. Oecologia 25:1-12. Mares, M.A. , T.E. Lacher, J r . , M.R. W i l l i g , N.A. Bitor. 1982. An experimental analysis of s o c i a l spacing in Tamias s t r i a t u s . Ecol. 63:267-273. Myers, J.H. and C.J. Krebs. 1971. Genetic, behavioural, and reproductive attributes of dispersing f i e l d voles Microtus  pennsylvanicus and M. ochrogaster. Ecol. Monogr. 41: 53-78. Myers, K. and W.E. Poole. 1961. A study of the biology of the wild rabbit, Oryctolagus cuniculus (L.) in confined populations. II. The effects of season and population increase on behaviour. CSIRO Wildl. Res. 6:1-41. 159 Mykytowycz, R. and S. Gambale. 1965. A study of the inter-warren a c t i v i t i e s and dispersal of wild rabbits, Oryctolagus  cuniculus, l i v i n g in a 45-ac. paddock. CSIRO Wi l d l . Res. 10:111-123. Mykytowycz, R. and M.L. Dudzinski. 1972. Aggressive and protective behaviour of adult rabbits, Oryctolagus  cuniculus (L.) towards juveniles. Behav. 43:96-120. Pease, J.L., R.H. Vowles, and L.B. Keith. 1979. Interaction of snowshoe hares and woody vegetation. J. Wi l d l . Manage. 43: 43-60. Redfield, J.A., M.J, T a i t t , and C.J. Krebs. 1978. Experimental a l t e r a t i o n of sex ratios in populations of Microtus  townsendi i , a f i e l d vole. Can. J. Zool. 56: 17-27. Rongstad, O.J. 1965. Calcium-45 l a b e l l i n g of mammals for use in population studies. Health Physics 11: 1543-1556. Sadleir, R.M.F.S. 1965. The rel a t i o n s h i p between organistic behaviour and population changes in the deermouse Peromyscus maniculatus (Wagner). J. Anim. Ecol. 39: 331-352. Saitoh, T. 1981. Control of female maturation in high density populations of the red-backed vole, Clethrionomys rufocanus bedfordiae. J. Anim. Ecol. 50:79-87. Severaid, J.H. 1942. The snowshoe hare: i t s l i f e history and a r t i f i c i a l propagation. Maine Dept. Inland Fisheries and Game. S i n c l a i r , A.R.E., C.J. Krebs,and J.N.M. Smith. 1982. Diet quality and food l i m i t a t i o n in herbivores: the case of the snowshoe hare. Can. J. Zool. 60:889-897. Smythe, M. 1968. The effects of the removal of individuals from a population of bank voles, Clethrionomys glareolus . J. Anim. Ecol. 37: 167-183. 160 Stenseth, N.C. 1982. Causes and consequences of dispersal in small mammals. In: Swingland, I. and P. Greenwood (eds.) The ecology of animal movements. Oxford Univ. Press. T a i t t , M.J. 1981. The effect of extra food on small rodent populations: Deermice (Peromyscus maniculatus). J . Anim. Ecol. 50:111-124. T a i t t , M.J. and C.J. Krebs. 1981. The effect of extra food on small rodent populations: I I . voles (Microtus towmsendii). J. Anim. Ecol. 50:125-137. Tamarin, R. H, 1977. dispersal in island and mainland voles. Ecol, 58:1044-1054. Tamarin, R.H. 1978. Dispersal, population regulation, and k-selection in f i e l d mice. Am. Nat. 112: Van Vleck, D.B. 1968. Movements of Microtus pennsylvanicus in rela t i o n to depopulation areas. J. Mammal. 49: 92-103. Vaughan, M.R. and L.B. Keith. 1981. Demographic response of experimental snowshoe hare populations to overwinter food shortage. ;J. W i l d l . Manage. 45:354-380. Watson A. and R. Moss. 1970. Dominance, spacing behaviour, and aggression in rel a t i o n to population l i m i t a t i o n in vertebrates. Animal Populations in re l a t i o n to Their Food  Resources (Ed. A. Watson), pp. 167-218. Blackwell S c i e n t i f i c Publications, Oxford. Watson, A. and R. Moss. 1979. Population cycles in the Tetraonidae. Ornis Fennica 56: 87-109. Windberg, L.A. and L.B. Keith. 1976a. Experimental analyses of dispersal in snowshoe hare population. Can. J. Zool. 54: 2061-2081. Windberg, L.A. and L.B. Keith. 1976b. Snowshoe hare population response to a r t i f i c i a l high den s i t i e s . J. Mammal. 57:523-553. 161 Wolff, J.O. 1978. Food habits of snowshoe hares in i n t e r i o r Alaska. J. Wildl. Manage. 42:148-153. Wolff, J.O. 1979. Refugia, d i s p e r s a l , predation, and geographic va r i a t i o n in snowshoe hare cycles, pp. 441-449. i n : Proc. of World Lagomorph Conference Aug. 12-16, 1978, University of Guelph. Wolff, J.O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecol. Monogr. 50: 111-130. 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0095839/manifest

Comment

Related Items