UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparison of a cyclic and non-cyclic population of snowshoe hares in Kluane, Yukon Jardine, Claire 1995

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

Item Metadata

Download

Media
831-ubc_1995-0363.pdf [ 3.05MB ]
Metadata
JSON: 831-1.0099008.json
JSON-LD: 831-1.0099008-ld.json
RDF/XML (Pretty): 831-1.0099008-rdf.xml
RDF/JSON: 831-1.0099008-rdf.json
Turtle: 831-1.0099008-turtle.txt
N-Triples: 831-1.0099008-rdf-ntriples.txt
Original Record: 831-1.0099008-source.json
Full Text
831-1.0099008-fulltext.txt
Citation
831-1.0099008.ris

Full Text

A COMPARISON OF A CYCLIC AND NON-CYCLIC POPULATION OF SNOWSHOE HARES IN KLUANE, YUKON by CLAIRE JARDINE B.Sc, The University of Guelph, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n THE FACULTY OF GRADUATE STUDIES Department of Zoology We accept t h i s thesis as conforming to the required standard THE rJNjATERSITY OF BRITISH COLUMBIA June 1995 © C l a i r e Jardine, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Zoo . lo the University of British Columbia Vancouver, Canada Date AJUTM 1 DE-6 (2788) ABSTRACT In t h i s study I compared a non-cyclic island population of snowshoe hares (Lepus americanus) with a c y c l i c mainland population to determine how the dynamics of the two populations d i f f e r e d following a population decline. Both populations had declined to similar low levels by the f a l l of 1992. During th i s study the Jacquot Island hare population increased 3.5 f o l d from the spring of 1993 to the spring of 1994 while the c y c l i c mainland hare population remained at low densities. I monitored survival and reproduction to determine the proximate causes for the dif f e r e n t dynamics seen. I determined that juvenile survival, overwinter adult survival and reproduction were higher on Jacquot Island than the mainland from the spring of 1993 to the spring of 1994. These were the key proximate factors that allowed the Jacquot Island hare population to increase while the mainland population remained at low densities. There was no difference i n summer adult survival between the two areas i n 1993. I also investigated the ultimate causes for the dif f e r e n t dynamics seen i n the two areas. Higher reproductive output, and higher overwinter survival did not coincide with better hare condition on Jacquot Island. There were, however, fewer mammalian predators on Jacquot Island than the mainland. Higher early juvenile survival on Jacquot Island coincided with lower numbers of small mammal predators. ii TABLE OF CONTENTS Abstract i i Table of Contents i i i L i s t of Tables v L i s t of Figures v i Acknowledgments v i i i Introduction 1 Survival 1 Reproduction 3 The Problem 6 Previous Studies 6 Objectives of Study 9 Study Area 1° Jacquot I s 1 and 10 Mainland 11 Methods 1 1 Hare Trapping 11 Adult Hare Survival 13 Measurement of L i t t e r s 14 Determination of Reproductive Parameters 16 Predator Numbers 16 Juvenile Survival 16 Body Weights and Condition Index 17 Predator Numbers 17 S t a t i s t i c a l Procedures 18 Results 18 Population Density 18 Summer Adult Hare Survival 20 Overwinter Adult Hare Survival 23 Proximate Causes of Mortality of Adult Hares 23 Effect of Maternity Cages on Reproductive Data 23 Pregnancy Rates 33 Timing of L i t t e r s 33 L i t t e r Sizes 33 S t i l l b o r n Rates 37 Newborn Hare Body Weights 37 Sex Ratios of Newborn Hares 41 Juvenile Breeding 41 Total Reproductive Output 41 Evidence of a Fourth L i t t e r 45 Juvenile Survival on Jacquot Island 45 Juvenile Survival on the Mainland 48 Comparison of Juvenile Survival on Jacquot Island and the Mainland i n 1994 49 in Comparison of J u v e n i l e S u r v i v a l over D i f f e r e n t L i t t e r Groups i n 1994 49 O v e r w i n t e r S u r v i v a l o f J u v e n i l e Hares on Jacquot I s l a n d 49 P r oximate Causes of M o r t a l i t y o f J u v e n i l e Hares 54 P r e d a t o r Numbers 54 A d u l t Male S p r i n g Body Weights and c o n d i t i o n I n d i c e s . . 5 7 D i s c u s s i o n 58 S u r v i v a l •. 61 A d u l t S u r v i v a l F o l l o w i n g a P o p u l a t i o n D e c l i n e 61 F a c t o r s I n f l u e n c i n g A d u l t S u r v i v a l 62 J u v e n i l e S u r v i v a l on Jacquot I s l a n d 64 J u v e n i l e S u r v i v a l on Jacquot I s l a n d Compared w i t h t h e M a i n l a n d 65 R e p r o d u c t i o n 68 The E f f e c t o f M a t e r n i t y Cages on R e p r o d u c t i v e data.68 R e p r o d u c t i o n F o l l o w i n g a P o p u l a t i o n D e c l i n e 70 T iming of B r e e d i n g . 72 T o t a l R e p r o d u c t i v e Output 72 L i t t e r S i z e and Neonate Weights 73 C o n c l u s i o n 75 P roximate F a c t o r s I n f l u e n c i n g t h e Jacquot I s l a n d P o p u l a t i o n I n c r e a s e 76 U l t i m a t e Causes o f H i gher S u r v i v a l on Jacquot I s l a n d . . 7 9 U l t i m a t e Causes of H i gher R e p r o d u c t i o n on Jacquot I s l a n d 81 Summary 81 L i t e r a t u r e C i t e d 83 iv LIST OF TABLES Table 1. Predictions of whether Jacquot Island would have lower or higher values than the mainland for each factor examined during this study according to the hypotheses..22 Table 2. 28-day survival rates of radio-collared, adult snowshoe hares (95% confidence l i m i t s i n parentheses and t o t a l hares collared beneath) 24 Table 3. Proximate causes of mortality of radio-collared adult snowshoe hares on Jacquot Island and the mainland.25 Table 4. Summary table of the effect of days a female spent i n a cage prior to parturition on l i t t e r size and mean leveret weight per l i t t e r for 1993 and 1994 34 Table 5. Mean date of b i r t h (± S.D.) for Jacquot Island and mainland study sites i n 1993 and 1994 (number of l i t t e r s i n parentheses) 35 Table 6. Mean l i t t e r Sizes at b i r t h (± S.D.) for Jacquot Island and mainland study sites i n 1993 and 1994 (number of l i t t e r s i n parentheses) 36 Table 7. Mean newborn weights at b i r t h (± S.D.) for Jacquot Island and mainland study sites i n 1993 and 1994 (number of l i t t e r s parentheses) 38 Table 8. Number of female hares born i n maternity cages on Jacquot Island and mainland study sites i n 1993 and 1994 (total number of newborn hares i n parentheses) 44 Table 9. Proximate causes of mortality of radio-tagged, juvenile snowshoe hares on Jacquot Island (JI) and the mainland i n 1994 55 Table 10. Sighting index of snowshoe hare predators on Jacquot Island and the mainland during the summers of 1993 and 1994 56 Table 11. Mean body weights ± S.D. of male snowshoe hares i n spring on study areas (sample size i n parentheses)...59 Table 12. Mean condition indices + S.D. of male snowshoe hares i n spring on study areas (sample size i n parentheses) 60 Table 13. Hypotheses and predictions examined during this study and the results obtained 77 v L I S T OF FIGURES Figure 1. Spring densities (hares/lOOha) and the corresponding potential n a t a l i t y (leverets produced per adult female surviving the breeding season) of an Alberta hare population from 1962 to 1976 (from Carey and Keith 1979) 5 Figure 2. Population estimates (February-April) for snowshoe hares on Jacquot Island (north grid) and the mainland ( s i l v e r control grid) from 1978 to 1992 7 Figure 3. Spring population estimates of snowshoe hares on Jacquot Island (north and south grids) and the mainland (sulphur and s i l v e r grids), 1990-1994, (1991-1992, Zimmerling (1993); 1993-1994, this study) 19 Figure 4. Population estimates with 95% confidence l i m i t s of snowshoe hares on Jacquot Island (north and south grids) and the mainland (sulphur and s i l v e r grids) i n the spring and f a l l from 1993-1994 21 Figure 5(a-b). The effect of time a female spent i n a cage prior to parturition on l i t t e r size and average leveret weight per l i t t e r for a l l l i t t e r s i n 1993 on Jacquot Island 28 Figure 6(a-c). The e f f e c t of time a female spent i n a maternity cage prior to parturition on l i t t e r size for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Island 30 Figure 7(a-c). E f f e c t of time a female spent i n a maternity cage p r i o r to parturition on average leveret weight per l i t t e r , for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Island 32 Figure 8(a-c). Correlation between l i t t e r size and average leveret weight per l i t t e r for l i t t e r groups 1,2, and 3 of 1994 on Jacquot Island 40 Figure 9(a-c). The effect of l i t t e r size on the t o t a l weight of a l i t t e r for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Island 43 Figure 10. Total reproductive output (number of leverets produced per adult female surviving the breeding season) i n r e l a t i o n to population peaks and lows for an Alberta population of snowshoe hares from 1962 to 1976 (Carey and Keith 1979), a Yukon population of snowshoe hares from 1989 to 1990 (O'Donoghue 1991) and 1994, and the Jacquot Island population of snowshoe hares from 1991 to 1992 (Zimmerling 1993) and 1993 to 1994 47 vi Figure 11. Survivorship curves with 95% confidence l i m i t s for f i r s t l i t t e r radio-tagged juveniles on Jacquot Island and the mainland, 1994 50 Figure 12. Survivorship curves with 95% confidence l i m i t s for second l i t t e r radio-tagged juveniles on Jacquot Island and the mainland, 1994 51 Figure 13. Survivorship curves with 95% confidence l i m i t s for t h i r d l i t t e r radio-tagged juveniles on Jacquot Island and the mainland, 1994 52 Figure 14. Survival rates with 95% confidence l i m i t s for the f i r s t 30 days of l i f e for l i t t e r groups 1, 2, and 3 on Jacquot Island and the mainland i n 1994 53 vii ACKNOWLEDGMENTS I would l i k e to thank my supervisor C. J. Krebs for giving me the opportunity to work on Jacquot Island and for his invaluable suggestions throughout t h i s project. I would l i k e to thank my research committee, T.P. Sullivan, J. Myers, and A.R.E. S i n c l a i r for t h e i r advice on a l l aspects of my project. I thank L. Preston, M. Joyce, A. Kalousek, and S. Armstrong for t h e i r help with f i e l d work on the island and M. O'Donoghue, C. Doyle, F. Doyle and C. Olsen for t h e i r advice i n the f i e l d . I would also l i k e to thank C. Doyle for co l l e c t i n g mainland juvenile data and K. Madsen for determining the home ranges of hares on Jacquot Island. M. O'Donoghue helped with bear problems and mainland data analysis. I would l i k e to thank I. Wingate for helping me get organized and A. Williams, base manager of the A r c t i c Institute, for l o g i s t i c a l support. T. Zimmerling provided valuable comments on e a r l i e r drafts of the thesis. I would especially l i k e to thank Brendan Delehanty for the help and encouragement he has given me throughout th i s project. This study was funded by a grant from the Natural Sciences and Engineering Research Council of Canada to C.J. Krebs, grants from the Northern Sciences Training program, and a grant from Sigma X i . viii INTRODUCTION Snowshoe hares (Lepus americanus) exhibit a ch a r a c t e r i s t i c "10-year" population cycle throughout most of t h e i r range (Keith 1990, Krebs et al. 1986). Various hypotheses have been proposed to help explain the cycle including food li m i t a t i o n , predation, and genotypic s h i f t s i n the population (Keith et al. 1984, Trostel et al. 1987, Chitty 1967). Despite numerous studies there i s s t i l l no clear consensus as to the causes of the hare population cycle (Green and Evans 1940, Krebs et al. 1986, Keith et al. 1984). Snowshoe hares do not exhibit population cycles at the southern periphery of th e i r range (Sievert and Keith 1985, Chitty 1950, Dolbeer and Clark 1975). Comparisons of reproduction and survival of c y c l i c and non-cyclic snowshoe hare populations may allow us to determine what proximate factors are necessary for cycles to occur. S u r v i v a l There i s some disagreement about how adult survival changes during the hare cycle. Green and Evans (1940) claimed that adult survival was constant over the cycle. Subsequent studies by Keith and Windberg (1978) showed that winter survival of adults was reduced during the decline and low phases and was correlated with the annual rate of population change (Keith 1990). In the Yukon, adult survival i n the f a l l and winter was correlated with the annual rate of population change, with survival declining gradually over the peak and decline phases (Krebs et al. 1986). Studies of.non-1 c y c l i c populations have a l l been short-term (see Keith 1990), but survival rates have been similar to those found during c y c l i c population declines (Keith 1990, Kuvlesky and Keith 1983). Unlike adult survival, there i s l i t t l e debate that changes in overwinter juvenile survival are closely associated with the population changes seen i n the 10-year cycle of snowshoe hares. Green and Evans (1940) concluded that poor overwinter juvenile survival i n i t i a t e d and sustained the population decline i n a Minnesota hare population. Juvenile survival from f a l l to mid-winter was highly correlated with the rate of population change i n the Yukon (Krebs et al. 1986). In Alberta, Keith and Windberg (1978) found juvenile survival from August to December was strongly correlated with population changes. For c y c l i c populations overwinter juvenile survival i s low i n the peak and decline years (Keith and Windberg 1978, Krebs et al. 1986). Short term studies of non-cyclic populations indicate that overwinter juvenile survival rates were similar to those found during the decline phase of the cycle (Keith 1990). The effect of early juvenile survival, between b i r t h and weaning at 28 days, on c y c l i c population changes i s less clear. Based on indices of early juvenile survival, Keith and Windberg (1978) concluded that early juvenile survival was not related to rates of population growth while Krebs et al. (1986) concluded, based on a calculated recruitment index, that changes i n early juvenile survival were a cause 2 of the hare cycle. O'Donoghue (1994) measured early juvenile survival d i r e c t l y during a population peak and found survival to 14 days of age was approximately 0.55, 0.19, and 0.47 for f i r s t , second, and t h i r d l i t t e r groups, respectively. Sovell (unpublished) collected survival data for the same population during one year of the decline phase of the cycle and found 14-day survival rates of 0.20 and 0.05 for f i r s t and second l i t t e r s , respectively. Early juvenile survival i n a c y c l i c population appears to change as the population changes, but more data are required to describe the relationship between early juvenile survival and population trends. No studies have examined early juvenile survival i n non-cyclic populations at the southern range boundary (but see Zimmerling 1993). Despite the differences i n adult and juvenile survival rates, the main proximate cause of death i n both c y c l i c and non-cyclic populations i s predation (Keith 1990, Boutin et al. 1986, Sievert and Keith 1985). Poor body condition can increase the r i s k of predation for hares, and factors that reduce body condition, such as food shortage and disease, can ultimately be responsible for poor survival (Keith 1990). The ultimate causes of the differences seen i n the survival rates remain uncertain. Reproduction Changes i n reproductive output are closely associated with the 10 year population cycle of snowshoe hares. During a 16 year study i n Rochester Alberta, Carey and Keith (1979) found 3 that n a t a l i t y (number of young produced per female surviving the breeding season) changed 2.4 f o l d during the cycle. These changes i n n a t a l i t y had a similar p e r i o d i c i t y to the cycle i n hare numbers but the low i n n a t a l i t y preceded the population low by three years (Carey and Keith 1979, Figure 1). Few studies have looked at the reproduction of non-cyclic populations and, of these, none have been more than three years i n duration (see Keith 1990). From these studies i t appears that changes i n the reproductive output of non-cyclic populations have not been so dramatic and have not been useful for predicting subsequent population trends (Keith 1990). Dolbeer and Clark (1975) studied two non-cyclic populations of snowshoe hares. During t h e i r studies the population density i n both areas remained stable. The average annual n a t a l i t y rate (number of young produced per female surviving the breeding season) varied 1.3 f o l d over the three year study i n Colorado and varied only 1.1 f o l d during the two year study i n Utah. Kuvlesky and Keith (1983) recorded n a t a l i t y rates for a non-cyclic population i n central Wisconsin and found that n a t a l i t y rates remained constant while the population density decreased 32%. From these studies i t appears that natality, i n non-cyclic populations located i n the southern part of the hare's geographical range, remains f a i r l y stable compared to the nat a l i t y of c y c l i c populations. 4 •fj* ro c O i-S ° (1) C ro ro to , ^ - WI ro <u cn £ T3 C ro .. a) cn ro r-l to .c *—<>— Q. —-«— I + ro 4-> ro ro t < c ro ,9 °~> ro ' aj cn to «— "a -Q QJ ro ro a) T3 C ro >^  J <4-to • a> to CD ro E CD l i s c I L a> C 3 C T3 Q. O CO jr c o 0) 3 Q. O a. (BM00L/S9JBL|) Aiisusp 6uuds 5 The Problem Most comparisons of c y c l i c and non-cyclic populations have been between southern non-cyclic populations and more northerly c y c l i c populations. Comparisons between these geographically disparate hare populations are d i f f i c u l t to make because of differences i n climate, habitat, vegetation, predators, and potential competitors. Jacquot Island, located i n the northern boreal forest, has been studied since 1977 and does not exhibit c y c l i c fluctuations i n population density (Figure 2, Krebs et al. 1986, Trostel 1986, Zimmerling 1993). Jacquot Island, therefore, provides a rare opportunity to examine a non-cyclic island population of snowshoe hares and compare i t with a nearby c y c l i c mainland population. P r e v i o u s S t u d i e s Two previous studies compared the population dynamics of snowshoe hares on Jacquot Island with the mainland. Trostel (1986) examined trapping data from 1977 to 1984 and did a one year radio-telemetry study on adult hares to determine survival and recruitment rates. During 1991 and 1992 Zimmerling (1993) carried out a detailed study of the demographic factors affecting hares on Jacquot Island. Trostel (1986) found no differences i n recruitment rates between Jacquot Island and the mainland. Recruitment encompasses na t a l i t y and early juvenile survival which Zimmerling (1993) examined separately. During the summer of 1991 juvenile survival was found to be higher on Jacquot 6 TJ H C C To u c o o T3 C ro " ° ro <u ro ro c^ -—> ro^2. (NJ CD CT) O cn CO oo cn to CO cn 00 CT) CNJ CO cn o oo cn + CO cn d q BU /S8JBH ° TJ C CO J2 4-1 O CT O CO - J c o OT 0) i_ CO - C (Nl 0) cn o cn - C (/) o o +J c 00 CO r-s. o cn vi— ' E o CL < vi-CO 0) (0 co 3 L_ "o 1-Cl) +J LL c o (0 o 0) •M TJ CO c E CO c 10 'co <D E c <D o JZ •4-" CO opul and CL ,»•—>, TJ i ' L . (Nl D) Cl) . C L_ 3 t _D) no LL. 7 Island than the mainland (Zimmerling 1993). Lower densities of a main predator of juvenile snowshoe hares, the red squirrel (Tamiasciurus hudsonicus; O'Donoghue 1994), and areas of dense understory cover on Jacquot Island were thought to contribute to the differences i n juvenile survival (Zimmerling 1993). Juvenile survival data were not collected in 1992. It remains to be seen whether juvenile survival i s consistently higher on Jacquot Island and i f i t plays an important role i n determining the population dynamics of snowshoe hares on Jacquot Island. Zimmerling (1993) and Trostel (1986) both concluded that adult hare predation on Jacquot Island and the mainland was different, with predation on the mainland showing a predictable delayed density-dependent pattern not evident on Jacquot Island. Reproductive data, such as l i t t e r size and b i r t h weights, were similar between mainland and island grids i n 1991; however, l i t t l e data were collected on Jacquot Island i n 1992 because both pregnancy rates and survival of females was poor (Zimmerling 1993). Because of low hare numbers on the mainland i n 1992 (Zimmerling 1993) only one f u l l year of reproductive data has been collected concurrently on Jacquot Island and the mainland. Both Jacquot Island and mainland populations experienced declines from the f a l l of 1991 to the f a l l of 1992 and reached similar low l e v e l s . Cyclic populations are characterized by a long decline and period of low density which lasts approximately 4 years despite the apparent 8 abundance of food and scarcity of predators during t h i s time (Keith 1990). Object ives of Study The objective of my study was to compare the population dynamics of hares on Jacquot Island with that of hares on the mainland during the low phase of the population cycle from 1993-1994 and build on the information collected during Zimmerling's (1993) study of Jacquot Island. S p e c i f i c a l l y , I compare adult survival, juvenile survival and reproduction on Jacquot Island with the mainland to determine how these factors d i f f e r between the two areas. I also compare hare body condition and a predator sighting indices on Jacquot Island and the mainland to determine i f these two factors are correlated with any differences i n survival or reproductive output. Three main hypotheses are examined i n thi s study: Hypothesis 1 Differences i n reproduction explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland. Hypothesis 2 Differences i n mortality explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland. Hypothesis 2a Only juvenile mortality d i f f e r s between the island and the mainland. Hypothesis 2b Only adult mortality d i f f e r s between the island and the mainland. 9 Hypothesis 2c Both adult and juvenile mortality d i f f e r between the island and the mainland. Hypothesis 3 Differences i n both reproduction and mortality explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland. The s p e c i f i c predictions associated with the above hypotheses depend on the population trends observed on Jacquot Island during this study and as a result are presented the population density results section. STUDY AREA Jacquot Is land Jacquot Island i s located i n Kluane Lake i n the southwestern Yukon (61°N, 138°W). The island i s approximately 5 km2 and consists of two sections joined by a 30 m wide isthmus. The southern section i s approximately 54 ha and i s separated from the mainland by at least 3 km. The northern section i s much larger (approximately 400 ha) and i s separated from the mainland by a minimum of 1.5 km. From the end of November u n t i l the end of May the island i s connected to the mainland by ic e . During the summer, however, access to and from the island by t e r r e s t r i a l animals i s limited because of the lake. The understory of the i n t e r i o r island i s dominated by willow (Salix spp.) and Shepherdia (Shepherdia canadensis). The canopy i s predominantly white spruce (Picea glauca) with some stands of aspen (Populus tremuloides) and balsam poplar 10 (Populus balsamifera) on h i l l s i d e s . On the outer edges of the island there are a number of windswept bl u f f s where the vegetation consists mainly of herbs and low shrubs. For a more detailed description of the habitat on Jacquot Island see Krebs et al. (1986) and Zimmerling (1993). Mainland The mainland areas used i n t h i s study are the control areas of the Kluane Boreal Forest Ecosystem Project (Krebs et al. 1992), located i n the Shakwak Trench, approximately 60 km southeast of Jacquot Island. Willow (Salix gluaca) and Shepherdia dominate the understory of the mainland grids. Unlike the island there are some mainland areas with a high proportion of shrub birch (Betula glandulosa). White spruce forest, varying from very dense to sparse, i s present on a l l grids. There are also small areas of balsam poplar and aspen. For a more detailed description of the mainland areas see Krebs et al. (1986). METHODS Hare T r a p p i n g This study was conducted from the end of A p r i l to the end of August during 1993 and 1994. Two trapping grids were used on Jacquot Island, one on the north section of the island and one on the south. The o r i g i n a l grids were 10 X 10 with 30 m between stations on the north and 40 m between stations on the south. These grids were spaced d i f f e r e n t l y due to previous studies on the island. I used these trapping grids for the population estimates, except for the f i n a l session of 11 1993 when I used a 12 X 15 grid on the north and a 10 X 16 grid on the south. This was done to cover a larger area for the removal of radio c o l l a r s . A l l grids had a checkerboard arrangement of traps with one Tomahawk li v e t r a p placed at every other station. Despite the differences i n grid spacing between stations on the 10 X 10 north and south grids, the effective trapping areas were quite similar. I estimated the effective trapping area (ETA) of the grid by adding a border equal to the radius of the home range of the hares where possible (Keith 1990). In the summer of 1994 the home range of hares on the north and south islands was 5.43 ha and 3.40 ha respectively (Madsen 1994, Kluane Boreal Forest Ecosystem Project, unpublished). The south island grid was bounded on two opposite sides by water so the ETA of the 10 X 10 south grid was 20.2 ha, and the ETA of the 10 X 15 south grid was 27.4 ha. The north grid was bounded on one side by water so only a 30 m boundary could be added on one side, making the ETA of the 10 X 10 north grid was 22.8 ha and the ETA of the 12 X 15 grid was 31.3 ha. I trapped at the end of A p r i l and at the end of August to get spring and f a l l population estimates. Each session consisted of 3-5 nights of trapping on each grid, alternating between grids. Traps were baited with a l f a l f a cubes and apple. When captured, each hare was marked with a numbered ear tag and had i t s weight, sex, and right hind foot length recorded. The reproductive condition of males was noted as scrota l , testes receding or abdominal, based on testes s i z e . 12 Females were noted as lactating or non-lactating. I also noted whether the females were pregnant or non-pregnant based on weight change and abdominal palpation. I estimated the population sizes on each grid using the jackknife estimator from the CAPTURE program (Otis et al. 1978). This estimator was found to be most robust to the sampling variations found i n snowshoe hare populations i n the Kluane area (Boulanger and Krebs 1994). Population estimates for the mainland grids were obtained during March and November also using the jackknife estimator (Otis et al. 1978). For the population estimates of the mainland I use data from two mainland control grids, S i l v e r and Sulphur. For reproduction and survival information I have combined the mainland controls for increased sample size. In those cases I refer to one mainland control area. A d u l t Hare S u r v i v a l I radio-collared between 10 and 20 hares on each grid during the summers of 1993 and 1994. I used Lotek, model SMRC-3RB, mortality transmitters (mass=40g) attached to either a leather or a nylon webbing c o l l a r . These transmitters doubled t h e i r pulse rate i f they remained s t i l l for four hours and, although monitored every one or two days, needed to be tracked only when the pulse rate doubled and the hare was presumed to be dead. During the winter of 1992-93 there were 7 radio collared adult hares and I used those individuals to estimate overwinter survival. Survival rates for the summer and 13 spring were calculated using the Kaplan-Meier method (Kaplan-Meier 1958, Pollock et al. 1989a) which allows for the censoring of data (owing to loss or f a i l u r e of radios or animals l i v i n g beyond the duration of the study) and the staggered entry of animals (Pollock et al. 1989a, Pollock et al. 1989b). Survival rates were compared using the log-rank test (Savage 1956, Pollock et al. 1989b) Site analysis was used to determine the cause of death for radio-collared hares. The cause of death was determined using the c r i t e r i a set out by the Kluane Boreal Forest Ecosystem Project (C. Doyle, unpublished). The k i l l s i t e was examined for evidence of predators, such as whitewash, scats, or tracks and for the characteristic way some predators deal with th e i r prey. Survival of hares on the mainland was estimated using the same methods as used on Jacquot Island. Measurement of L i t t e r s Female snowshoe hares do not make nests and nurse young only once per day (Severaid 1942, Rongstad and Tester 1971). This makes l i t t e r s very d i f f i c u l t to locate i n the f i e l d and as a result the following method for studying hare l i t t e r s was established by O'Donoghue (1992). In 1993 and 1994 I attempted to catch 10-15 pregnant female hares (on each island grid) shortly before parturition and place each female i n a 60 X 60 X 120 cm, wire mesh cage, at t h e i r point of capture. The back half of each cage was covered i n burlap and spruce bows were woven through the sides and top of the wire mesh to create a concealed area for the hare. Straw was 14 placed on the bottom of the cage and food and water were provided once d a i l y . Each hare received one apple, 4 0-50 willow twigs, 400 ml of water and 100 ml of bunny chow. Upon parturition the cage was opened and the hare was released. After the female l e f t the cage, the l i t t e r was counted, and each leveret was tagged with a numbered eartag, weighed, had i t s right hind foot measured and i t s sex recorded. A maximum of four leverets from a l i t t e r had radio transmitters (Biotrack, model SR-1; mass= 2-3 g) attached for determining subsequent survival (methods described below). The whole l i t t e r was then placed i n a concealed area, usually amongst deadfall, within four meters of the place of b i r t h . The cage was removed from the area to prevent leverets from crawling back i n and to avoid attracting or deterring predators from the area. On the mainland 10-15 females were caught throughout control areas of the study s i t e and were brought to a central area with e l e c t r i c fencing to prevent harassment by predators. Upon parturition the females were removed from the cage, the l i t t e r s were processed as described for the island grids, and then both female and young were transported back to the female's home range. The l i t t e r was then placed in a concealed area and the female was introduced to th i s s i t e and then released. A l l other aspects of the procedure were the same except, on the mainland, a l l leverets from a l i t t e r had radio transmitters attached when possible. 1 5 Determination of Reproductive Parameters To measure the timing of breeding for each area the mean date of par t u r i t i o n of females i n cages was calculated. Testes regression of males was recorded while trapping during the summer to estimate the timing of the end of the breeding period for each area. An estimate of pregnancy rates was made by counting the proportion of pregnant or recently pregnant female hares during a 2 week period around the mean date of par t u r i t i o n for each area. Juveni le S u r v i v a l For determining early juvenile survival radio transmitters (described i n the Measurement of L i t t e r s section) were used. The radio transmitters were glued to a small area of trimmed fur between the shoulder blades of each leveret. This procedure kept the transmitters attached for approximately 3-5 weeks. Radio-tagged juvenile hares were tracked for a visual sighting d a i l y to determine i f they were a l i v e . If a leveret was found dead, the cause of mortality was determined using the same procedure as described for adult hares. I counted a l i t t e r as abandoned when 1) the whole l i t t e r was found dead i n a r e l a t i v e l y open area close to the b i r t h spot, 2) the bodies were whole and there was no evidence of predation and, 3) when autopsied t h e i r stomachs were empty. Survivorship curves for the radio-tagged leverets were calculated using the Kaplan-Meier procedure as described for adult hares. Comparisons of the survivorship curves were made using the log-rank test (Savage 1956, Pollock et a l . 16 1989b). The same procedures were used on the mainland except that a l l leverets from a l i t t e r were radio-tagged when possible. Juvenile hares that were caught i n the August trapping session and weighed more than 1000 g were f i t t e d with adult mortality c o l l a r s . On Jacquot Island I monitored overwinter juvenile survival using trapping data and 3 radio-collared juvenile hares. On the mainland overwinter juvenile survival was measured using 8 radio-collared hares. Body Weights and Condi t ion Index Body condition can be estimated by measuring skeletal size and weight (Bailey 1968). I estimated body condition by measuring right hind foot length (RHF; i n mm) and body mass (in grams) of male snowshoe hares i n A p r i l and early May. The mean right hind foot length for each hare was used i n the calculations to reduce observer error. Only males were used because most females were already pregnant when the measurements were obtained. The following equation was used: Predicted Weight = 738.9 + (RHF) 1 - 3 This equation was calculated by O'Donoghue (1992), and was used to obtain a predicted weight for each i n d i v i d u a l . The ra t i o of observed weight to predicted weight was used as the condition index. Predator Numbers To get an index of predator numbers a sighting index was used on both Jacquot Island and the mainland. Each researcher recorded the number of hours spent i n the f i e l d 17 and the number and type of species seen during t h i s time. An index of predators seen per 100 hrs i n the f i e l d was calculated from th i s information. Red sq u i r r e l numbers were estimated on each of the Jacquot Island grids by active midden counts and vi s u a l sightings of squirrels using middens. A r c t i c ground squirrels (Spermophilus parryii) numbers were estimated by observation of burrows. On the mainland red and ground squirrels numbers were estimated through live-trapping on control grids. S t a t i s t i c a l Procedures I pooled the data from Jacquot Island north and south grids to increase the sample size. To determine i f the data from the two island grids could be pooled I used t - t e s t s . Because of the small sample sizes on Jacquot Island and the lack of data for the mainland i n 1993, comparisons of l i t t e r size and average leveret weight per l i t t e r were made with 1994 data only. A 0.05 l e v e l of significance was used for a l l tests. RESULTS Populat ion Density During t h i s study the hare population on Jacquot Island was higher than the mainland population at a l l times (Figure 3). Jacquot Island reached low population densities i n the spring of 1993 as did the mainland population. The average population density on Jacquot Island increased 3.5 f o l d from 0.4 hares/ha i n the spring of 1993 to 1.4 hares/ha i n the 18 o o OO Z 4-> 4-> o o 3 3 CT CT U CJ (0 CO oo L75 + + H vf cn CD CD C Q . 00 CO cn cn D) c Q . 00 (NJ cn cn c n c *L_ C L OO cn cn c CL 00 o cn cn CD C CL 00 CNJ cn rth cn o • c *— v — ' CT) T J cn C i — CO — ' JO 4-> cn O cn 3 -^CT • O o CO cn - ) cn c v— o es ids CO CD oe Ive ' lO CO ow and c and >, CO hur T J <4— hur 3 o hur *-> Q - CO ates (sul this E T J C 10 (0 cn cu c cn c '(0 >— o E i CO cn CO cu 3 - C cn pop and CO CD /< s cn UjJ sp cn Q . 00 CD cn . C c ro" 4-> 3 CD O CD igun 10 F igun nd iiu; LL CO o d BL| /S9JBH 19 s p r i n g o f 1994, whereas t h e average p o p u l a t i o n d e n s i t y o f t h e mainland c o n t r o l g r i d s remained c o n s t a n t a t 0.08 hares/ha ( F i g u r e 3 ) . Both Jacquot I s l a n d and mainland p o p u l a t i o n s i n c r e a s e d over t h e summer o f 1993; however, Jacquot I s l a n d showed a 7.3 f o l d i n c r e a s e compared t o a 1.8 f o l d i n c r e a s e from A p r i l t o October on t h e mainland. D u r i n g t h e summer o f 1994 t h e Jacquot I s l a n d hare p o p u l a t i o n i n c r e a s e d about 5 f o l d . The mainland c o n t r o l g r i d s i n c r e a s e d between 5 and 23 f o l d from t h e s p r i n g o f 1994 t o t h e f a l l o f 1994 ( F i g u r e 4 ) . Based on t h e s e r e s u l t s , p r e d i c t i o n s f o r each o f t h e hypotheses p r e s e n t e d i n t h e i n t r o d u c t i o n can be made. These are shown i n Table 1. Summer A d u l t Hare S u r v i v a l There were no s i g n i f i c a n t d i f f e r e n c e s i n a d u l t hare s u r v i v a l d u r i n g t h e summer on t h e n o r t h and so u t h g r i d s o f Jacquot I s l a n d i n e i t h e r y e a r (1993, l o g - r a n k t e s t s t a t i s t i c = l . 5 0 , p=0.13; 1994, 100% s u r v i v a l on b o t h g r i d s ) . As a r e s u l t , f o r comparing w i t h t h e mainland, I p o o l e d t h e dat a from both i s l a n d g r i d s t o i n c r e a s e t h e sample s i z e . Twenty-eight day a d u l t s u r v i v a l d i d not d i f f e r s i g n i f i c a n t l y between Jacquot I s l a n d and t h e mainland i n 1993 ( l o g - r a n k t e s t s t a t i s t i c = l . 1 3 , p=0.26) but was 10% h i g h e r on Jacquot I s l a n d t h a n t h e mainland i n 1994 ( l o g - r a n k t e s t s t a t i s t i c = 3 . 0 4 , p=0.002). A d u l t s u r v i v a l on Jacquot I s l a n d was 10% h i g h e r i n 1994 than 1993 ( l o g - r a n k t e s t s t a t i s t i c = 3 . 1 1 , p=0.002). On t h e mainland t h e r e was no 20 3 % o o CO Z o o ^ CT CT Q . <D O O -=• > w ro .5 —) —) CO CO cn CD CD Dl C O . CO CO CD CD CO CD CD D) C "i— Q . CO 4- .2 5" S JS cn ^ c o a. CO <" d) d> CO 4_* c <u — 10 T 3 § ' C 0 RA C i-to a) V4- > In "5 1 2 — 3 C C3 c o ^.S ro CD ^ I ± ! ±3 5 -D s—' c to CO 1-2 \p O) CO CO "O 3 C CD Q . CD CD o x: — ° - % CO 2 CD . CD 5 1 E D5 J5 O LT J2 it o d EL|/SaJBH 21 T J C CO c o •+J u 3 T J - I cu co t. o or* E CO LU OO LU X h-o C L > -T J i £ i ! re 3 t! "8 5 SI C _ cu ^ > c\j .5, c o 3 T J CO CNJl c o '4-1 u 3 T J O L. Q. CU CO t C\J E M o cu > ~ 3 ro <3 i H o CtC O o < cu - G CD CU E CO 10 cu E ro co cu E ro co cu CD CU > U 43 -2 a. Q. * £ O cu o cu <: o cu o cu o cu E ro co cu o cu o cu E CO CO cu E re co cu E CO CO (0 f*> t o E _0) 0) > 3 CO £ o E 3 T J CO 22 s i g n i f i c a n t difference i n survival rates between 1993 and 1994 (log-rank test statistic=0.80, p=0.42; Table 2). Overwinter Adult Hare S u r v i v a l During the winter of 1993-1994 (September to Ap r i l ) there were 7 radio-collared adult hares on Jacquot Island (approximately 54% of the adults i n the population). Two of the hares died, one due to an unknown predator, the other due to unknown causes. There was no dispersal of radio-collared hares off the island during t h i s period. Based on trapping results, survival of adult hares from September 1993 to May 1994 was 83% over eight months. On the mainland 38 hares were collared from September to A p r i l . The survival rate for thi s period was 12.6% over eight months (95% confidence interval 8-17%). Proximate Causes of M o r t a l i t y of Adult Hares During the summers of 1993 and 1994 on Jacquot Island there were only 6 mortalities out of 55 collared hares. The majority of deaths were due to predation. On the mainland there were 19 mortalities out of 54 collared hares and the majority of deaths were due to unknown causes and unknown predators (Table 3). E f f e c t of Materni ty Cages on Reproductive Data Females were kept i n maternity cages and supplied with food prior to part u r i t i o n . To determine the possible impact of t h i s procedure on the young I did regressions of days a female spent i n the maternity cage prior to pa r t u r i t i o n on l i t t e r size and average leveret weight per l i t t e r . In 1993 23 Table 2. 28-day survival rates of radio-collared, adult snowshoe hares (95% confidence limits in parentheses and total hares collared beneath). Summer Summer 1993 1994 Jacquot Island 0.90 1.00 * (0.83-0.96) (1.00-1.00) 23 32 Mainland 0.85 0.90 (0.68-0.94) (0.85-0.95) 12 42 * significantly different from Jacquot Island summer 1993 (log rank test statistic=3.11, p=0.002); and mainland 1994 (logranktest statistic=3.04, p=0.002). 24 Table 3. Proximate causes of mortality of radio-collared adult snowshoe hares on Jacquot Island and the mainland. Percentage of Mortalities Mortality Factors Island Mainland Island Mainland Lynx 0.0 0.0 0.0 7.7 Coyote 0.0 0.0 0.0 0.0 Unknown Mammal 16.7 16.7 0.0 0.0 Goshawk 0.0 0.0 0.0 15.4 Great-Horned Owl 0.0 0.0 0.0 0.0 Unknown Avian 0.0 16.7 0.0 0.0 Unknown Predator 66.7 16.7 0.0 30.7 Unknown 16.7 50.0 0.0 46.2 Total Killed 6 6 0 13 Total Radio-Collared 23 12 32 42 25 females were held i n cages an average of 5.8 days (standard deviation of ± 2.9 days) prior to part u r i t i o n . The time a female spent i n a maternity cage prior to pa r t u r i t i o n had no sig n i f i c a n t e ffect on l i t t e r size (Figure 5a; regression, Y=0.10X + 5.59, R2=0.04, p=0.53) or average leveret weight per l i t t e r (Figure 5b; regression, Y=-0.36X + 60.03, R2=0.02, p=0.63). For the f i r s t l i t t e r of 1994 females were held i n maternity cages an average of 5.2 days (standard deviation of ±3 . 7 days). The time a female spent i n a maternity cage prior to par t u r i t i o n was s i g n i f i c a n t l y correlated to l i t t e r size (Figure 6a; regression, Y=0.14X + 3.36, R2=0.2, p=0.05) and s i g n i f i c a n t l y , negatively correlated to average leveret weight per l i t t e r (Figure 7a; regression, Y=-1.32X + 67.49, R2=0.22, p=0.04). For the second l i t t e r of 1994 females were held i n cages and average of 7.2 days (standard deviation of ± 3.9 days). The time a female spent i n a maternity cage p r i o r to parturition was s i g n i f i c a n t l y correlated to l i t t e r size (Figure 6b; regression, Y=0.18X +5.3, R2=0.27, p=0.05) but not to average leveret weight per l i t t e r (Figure 7b; regression, Y=-0.58X + 71.47, R2=0.10, p=0.27). For the t h i r d l i t t e r of 1994 females were held i n maternity cages an average of 5.0 days (standard deviation of ± 2 . 8 days). The time a female spent i n a maternity cage prior to par t u r i t i o n was not s i g n i f i c a n t l y correlated to l i t t e r size (Figure 6a; regression, Y=0.09X + 7.21, R2=0.01, 26 F i g u r e 5(a-b). The ef f e c t of time a female spent i n a cage prior to p a r t u r i t i o n on l i t t e r size and average leveret weight per l i t t e r for a l l l i t t e r s i n 1993 on Jacquot Island. 27 (a) 1993 all litters Litter size 8 7 + 6 5 4 3 2 + 1 0 R = 0.04 p=0.53 0 10 15 Average leveret weight per litter (9) (b) 1 993 all litters 80 70 60 50 40 30 20 + 10 0 • • 1 » » R2 = 0.02 p=0.63 0 10 15 Time in cage prior to parturition (days) 28 F i g u r e 6 ( a - c ) . The effect of time a female spent i n a maternity cage prior to parturition on l i t t e r size for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Island. 29 (a) 10 8 6 4 2 0 1994 first litter 10 litter size = 0.1423 Days + 3.36 R2 = 0.20 p=0.05 15 20 (b) 1994 second litter 10 8 6 4 2 0 litter size = 0.1 848Days+ 5.3358 R2 = 0.27 p=0.05 10 15 20 (C) 14 -r 12 10 + 8 6 4 2 0 1994 third litter • 10 R2 = 0.01 p=0.69 15 20 Time in cage prior to parturition (days) 30 F i g u r e 7 ( a - c ) . Ef f e c t of time a female spent i n a maternity cage pr i o r to parturition on average leveret weight per l i t t e r , for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Island. 31 (a) 1994 first litter 100 80 60 40 20 0 10 Weight = -1.33Days +67.49 R 2=0.22 p=0.04 15 20 (b) 1994 second litter 100 80 60 40 20 0 I t 10 • R 2=0.10 p=0.27 15 20 (C) 100 80 60 40 20 0 1994 third litter 10 R2=0.01 p=0.68 15 20 Time in cage prior to parturition (days) 32 p=0.68) or l i t t e r weights (Figure 7c; regression, Y=-0.41X + 74.99, R2=0.01, p=0.68). The effect of time a female spent i n a maternity cage prior to parturition on l i t t e r size and average leveret weight per l i t t e r i s summarized i n Table 4. Pregnancy Rates Pregnancy rates (proportion of females pregnant during each l i t t e r group) varied between 89% and 94% with an average of 92% (standard deviation of ± 1.7%) on Jacquot Island during 1993 and 1994. The mainland hares that were trapped for maternity cages were a l l pregnant or had recently given b i r t h . Timing of L i t t e r s The mean dates of b i r t h were similar between years and between the mainland and the island for each l i t t e r group (Table 5). In 1994 the int e r v a l between l i t t e r s , based on mean par t u r i t i o n dates, varied between 36 and 38 days, with an average of 37 days. L i t t e r S izes L i t t e r size data were not analyzed i n 1993 due to small sample sizes and lack of mainland data. In 1994 mean l i t t e r sizes were 17% larger on Jacquot Island than the mainland (Table 6; two-way ANOVA, p=0.008). There was no interaction effect between location (Jacquot Island and the mainland) and l i t t e r group (two-way ANOVA, p=0.89). For both the island and the mainland i n 1994 the f i r s t l i t t e r was s i g n i f i c a n t l y smaller than second and t h i r d l i t t e r s (Tukey test, p<0.001 i n each case) and there was no difference between l i t t e r s two 33 Table 4. Summary table of the effect of days a female spent in a cage prior to parturition on litter size and mean leveret weight per litter for 1993 and 1994. The effect of days in cage prior to parturition on: Litter size Mean newborn weights Year litter group Days in cage prior to parturition + (S.D.) R 2 value p value R 2 value p value 1993 all 5.8 (2.9) 0.04 0.53 0.02 0.63 1994 1 5.2 (3.7) 0.20 0.05* 0.22 0.04* 1994 2 7.2 (3.9) 0.27 0 .05* 0.10 0.27 1994 3 5.0 (2.8) 0.01 0.69 0.01 0.68 * significant effect of days a female spent in cage prior to parturition 34 Table 5. Mean date of birth (+ S.D.) for Jacquot Island and mainland study sites in 1993 and 1994 (number of litters in parentheses). Mean date of birth 1993 1994 Litter Jacquot Island Mainland 1 May 17 (D 2 June 18+ 2.6 (4) 3 July 22 ±2.8 (7) Jacquot Island Mainland May 12 + 3.3 May 13 +2.6 (20) (11) June 19 + 5.5 (15) July 26 ±7.3 (14) June 19 ±3.4 (9) July 25 ±3.8 (10) 35 Table 6. Mean litter sizes at birth (+ S.D.) for Jacquot Island and mainland Study Sites in 1993 and 1994 (number of litters in parentheses). Mean litter Size 1993 1994 Litter Jacquot Island Mainland Jacquot Island1 Mainland 1 4.0 - 4.1 + 1.2 2 3 . 2 ± 0 . 8 2 (D (20) (11) 2 7.3 + 0.5 - 6.7 + 1.4 5.8 ± 1 . 1 (4) (14) (9) 3 5.9+1.5 - 7.6 + 2.1 6 .4+1 .2 (7) (14) (10) 1 significantly larger litter sizes on Jacquot Island (2-way ANOVA p=0.008) 2 significantly larger than litter groups 2 and 3 (2-way ANOVA, Tukey test p<0.001) 36 and one (Tukey test p=0.502). During the t h i r d l i t t e r one female on Jacquot Island had 14 leverets which i s larger than any l i t t e r size I have found recorded i n the l i t e r a t u r e . S t i l l b o r n Rates S t i l l b o r n rates were very low on both the island and the mainland for 1993 and 1994. On Jacquot island there was no incidence of s t i l l births i n 1993 and the f i r s t l i t t e r of 1994 and s t i l l b o r n rates were only 1.0% for second and t h i r d l i t t e r s i n 1994. On the mainland the s t i l l b o r n rates were 2.9% for f i r s t l i t t e r , 0.0% for second l i t t e r , and 4.7% for th i r d l i t t e r of 1994. Newborn Hare Body Weights Mean newborn weight data were not analyzed i n 1993 due to small sample sizes and the lack of data from the mainland. In 1994 the average weights of newborn hares were 12% greater on Jacquot Island than the mainland (Table 7; two way ANOVA, p=0.02). There was no si g n i f i c a n t interaction between location (Jacquot Island and the mainland) and l i t t e r group (two-way ANOVA, p=0.96). On both the island and the mainland the newborn weights for f i r s t l i t t e r s were s i g n i f i c a n t l y smaller than t h i r d l i t t e r s (Tukey test p<0.001), however there were no s i g n i f i c a n t differences between the weights of l i t t e r groups one and two (Tukey test p=0.08) and l i t t e r groups two and three (Tukey test p=0.25). For the f i r s t and t h i r d l i t t e r s of 1994 the average leveret weight per l i t t e r was s i g n i f i c a n t l y negatively correlated with l i t t e r size ( F i r s t l i t t e r ; Figure 8a, R2=0.30, 37 Table 7. Mean newborn weights at birth (+ S.D.) for Jacquot Island and mainland study sites in 1993 and 1 994 (number of litters in parentheses). Mean newborn weights (g) 1993 1994 Litter Jacquot Island Mainland Jacquot Island1 Mainland 1 47.4 (D. - 60.6 +10.32 (20) 54.3 + 7.02 (9) 2 56.4 + 5.0 (4) - 67.3 + 7.5 (14) 60.4 ± 4 . 8 (9) 3 60.3 + 6.7 (7) - 72.9 ± 9 . 8 (14) 64.5 + 4.6 (10) 1 significanlty heavier mean newborn weights on Jacquot Island (2-way ANOVA, p=0.02) 2 significantly lighter than litter 3 (2-way ANOVA, Tukey test, p=0.001) 38 Figure 8 ( a - c ) . Correlation between l i t t e r size and average leveret weight per l i t t e r for l i t t e r groups 1,2, and 3 of 1994 on Jacquot Island. 39 (b) 90 80 70 60 50 40 30 20 10 0 1994 second litter R2 = 0.20 p=0.08 10 (C) 1994 third litter 100 90 + 80 70 60 50 40 30 20 10 0 R2 = 0.60 p=0.001 10 15 Lit ter size 40 p=0.01; Third l i t t e r ; Figure 8c, R2=0.60, p=0.001). No si g n i f i c a n t correlation was found for second l i t t e r (Figure 8b; R2=0.20, p=0.08). 9 For a l l l i t t e r groups the t o t a l weight of a l i t t e r increased with l i t t e r size (Figure 9a-c; F i r s t l i t t e r , R2=0.67, p<0.001; Second l i t t e r , R2=0.72, p<0.001; Third l i t t e r , R2=0.86, p<0.001). Sex Ratios of Newborn Hares Sex ratios were s t a t i s t i c a l l y d ifferent than 1:1 i n two cases (Table 8; G-test). On the north grid of Jacquot Island for f i r s t l i t t e r there was a disproportionate number of female newborns and on the mainland for t h i r d l i t t e r there was a disproportionate number of male newborns. Juveni le Breeding On Jacquot Island i n 1993 there were 2 scrotal f i r s t l i t t e r males caught i n mid-July. In 1994 there were 3 scrotal f i r s t l i t t e r males caught on the mainland and 4 caught on the island during mid-July. There was no evidence of any juvenile females being pregnant or lac t a t i n g during either year of the study. T o t a l Reproductive Output Total reproductive output per female surviving the breeding season can be estimated by multiplying pregnancy rates by mean l i t t e r size and (1 - s t i l l b o r n rates) and adding these over the three l i t t e r groups (O'Donoghue 1991). This value corresponds to potential n a t a l i t y (Carey and Keith 1979, O'Donoghue 1992). No s t a t i s t i c a l tests could be done 41 F i g u r e 9 ( a - c ) . The effect of l i t t e r size on the t o t a l weight of a l i t t e r for l i t t e r groups 1, 2, and 3 of 1994 on Jacquot Is1and. 42 (b) 1994 second litter 600 T 500 400 300 200 100 0 y = 47.7x + 126.8 R2 = 0.72 p<0.001 10 (C) 900 800 700 600 500 400 300 200 100 0 1994 third litter y = 37.2x + 258.2 R2 = 0.86 p<0.001 10 15 Litter size 43 Table 8. Number of female hares born in maternity cages on Jacquot Island and mainland study sites in 1993 and 1994 (total number of newborn hares in parentheses). Number of females 1993 1994 Litter Jacquot Island Mainland Jacquot Island Mainland North South 1 3 28 * 22 15 (4) (36) (44) (34) 2 16 51 21 (29) (91) (48) 3 20 54 24* (41) (103) (64) * significantly different than 1:1 (G-test, df=1) 44 because only one estimate i s made per year, but i n 1994 Jacquot Island had a higher reproductive output compared to the mainland: 16.8 juveniles/female on Jacquot Island and 15.2 juveniles/female on the mainland. Total reproductive output of Jacquot Island, the mainland and another c y c l i c population of hares i s shown i n Figure 10. Evidence of a Fourth L i t t e r There i s evidence that a fourth l i t t e r was produced i n 1994. On the island 75% (n = 16) of adult males had scrotal testes i n the l a s t two weeks of July when breeding for a fourth l i t t e r would be occurring. It i s doubtful that the females with l a t e r t h i r d l i t t e r s would have a fourth l i t t e r , but some of those with early l i t t e r s may have. On the mainland i n 1994, a juvenile hare was caught i n September that, due to i t s small size, was probably a fourth l i t t e r leveret. Unfortunately Jacquot Island was not trapped during September to determine i f a fourth l i t t e r occurred; however, during the f i n a l trapping session of August, two adult females were judged to be pregnant. There was no sign from the August trapping that a fourth l i t t e r occurred on the island i n 1993, however the p o s s i b i l i t y cannot be ruled out. Juveni le S u r v i v a l on Jacquot Is land Only two leverets were radio-tagged from the f i r s t l i t t e r on the south grid during 1993. Ten and 13 leverets were radio-tagged on both grids during second and t h i r d l i t t e r s respectively. Because sample sizes were small the data from both grids were pooled. None of the 25 leverets that were 45 F i g u r e 10. Total reproductive output (number of leverets produced per adult female surviving the breeding season) i n rela t i o n to population peaks and lows for an Alberta population of snowshoe hares from 1962 to 1976 (Carey and Keith 1979), a Yukon population of snowshoe hares from 1989 to 1990 (O'Donoghue 1991) and 1994, and the Jacquot Island population of snowshoe hares from 1991 to 1992 (Zimmerling 1993) and 1993 to 1994. 46 47 radio-tagged were known to have died. Five of the radio-transmitters disappeared before 30 days of age. It i s not possible to know whether the radios malfunctioned or whether the leverets were preyed upon and the radios destroyed. Because the fate of these f i v e leverets i s unknown they were treated as censored individuals i n the survival calculations. Juvenile survival data for Jacquot Island i n 1994 were collected from 94 radio-tagged leverets. Of the 94 leverets that were radio-tagged, there were 16 known deaths. Thirteen radio transmitters disappeared before 30 days and were not recovered. Because the fate of these 13 leverets i s unknown, they were treated as censored individuals i n the survival calculations. There were no s i g n i f i c a n t differences i n the survival rates of juvenile hares on the two island grids for l i t t e r one or two ( l i t t e r one, log-rank test statistic=0.25, p=0.803; l i t t e r two, log-rank test statistic=l.60, p=0.110) so I pooled the data to increase the sample s i z e . For the th i r d l i t t e r , there were only 4 radio tagged leverets on the north grid so I pooled the data from both grids. Juveni le S u r v i v a l on the Mainland No survival data were collected i n 1993. During 1994, 123 leverets were radio-tagged and there were 47 known deaths. Fifteen of the 123 radio transmitters disappeared before 30 days and were not recovered. Because the fate of these 15 leverets i s unknown they were treated as censored individuals for the survival calculations. 48 Comparison o f J u v e n i l e S u r v i v a l on Jacquot I s l a n d and the Mainland i n 1994 For the f i r s t l i t t e r , early juvenile survival on Jacquot Island was not s t a t i s t i c a l l y d i fferent from the mainland (Figure 11; log-rank test statistic=0.94, p=0.35). However survival was higher on Jacquot Island for second and t h i r d l i t t e r s (Figure 12 and Figure 13; l i t t e r 2, log-rank test statistic=3.90, p<0.001; l i t t e r three, log-rank test statistic=2.18, p=0.03). The average survival rate over a l l three l i t t e r s was 0.79 on Jacquot Island and 0.52 on the mainland. Comparison of J u v e n i l e S u r v i v a l Over D i f f e r e n t L i t t e r Groups i n 1994 Early juvenile survival of the f i r s t l i t t e r group on Jacquot Island was lower than second l i t t e r and s i g n i f i c a n t l y lower than t h i r d l i t t e r (Figure 14; f i r s t and t h i r d l i t t e r s log-rank test statistic=2.39, p=0.02; f i r s t and second l i t t e r s log-rank test statistic=l.76, p=0.08). On the mainland second l i t t e r survival was s i g n i f i c a n t l y lower than f i r s t and t h i r d l i t t e r s ( f i r s t and second l i t t e r , log-rank test statistic=3.06, p=0.002; f i r s t and t h i r d l i t t e r , log-rank test statistic=2.73, p=0.006; Figure 14). Overwinter S u r v i v a l of J u v e n i l e Hares on Jacquot I s l a n d Overwinter juvenile survival on Jacquot Island from August 1993 to spring 1994, based on trapping was 48%. Only three radio-collared juvenile hares were monitored overwinter. One 49 50 o CO LO (NJ O (NJ CO CO LD TJ CU CD < LO C D d CO d d CO d d vT d CO d (NJ T— O d d T 3 c o o CO XT L _ CD O CD 14— , CO ±L T 3 i (J (0 5 E TJ CU Ci= -C c ^ O TJ O c (0 LO Y CD jo - C CO 4-» £ o "> 3. 0 c 9- 0 !c CO CO _0) .> cu M 00 -o cu (NJ 2 _ CD ^ (O JJ V 3 ° O) TJ iZ 2 fiujAjAjns uoi+jodojd 5 1 52 TJ C ro (/> uot 1 land cr _c o 'ca HJa E CM CL 3 o D) L— 1 1 •t 1 1 1 1 1 1 1 1 1 1 1 1 CT1 d oo d d CD d d d d CM d «- o d 10 ay X i X5 o C CO re 4-1 c </) i _ ' r e VI— E OJ <D -C -C 4-< 4-" L— o V*— nd LO ro +J X ! jUlj an o 4J c o a> 3 T3 CT VI— O c ro o -> o c o LO CO CT) TJ -C 4-" an CM" to a) 1— 4-1 CO re Q. 3 O > i -rvi O) L _ 3 <D to *-> 4J *r L_ o VI— ure 0) VI— CD CT) M— CT) LL. o «— ajEJ |BAiAjns A"ep-rj£ 5 3 of these died due to coyote predation and the other two survived to the following spring. Eight radio collared juveniles were monitored on the mainland over the winter. Survival was 11.3% (95% confidence interval 0-32%) from August 1993 to A p r i l 1994. Proximate Causes of M o r t a l i t y of Juven i l e Hares On the mainland red squirrels, unknown predators and weasels were the main proximate causes of mortality. The main proximate causes of mortality on Jacquot Island were non-predator/abandoned and unknown predator. Predation accounted for 62.5% of deaths on Jacquot Island while on the mainland i t accounted for 97.9% of deaths (Table 9). Abandonment accounted for 6 out of the 16 mortalities (2 of the 49 l i t t e r s ) on Jacquot Island whereas on the mainland no deaths were attributed to abandonment. Predator Numbers Fewer predators were seen on Jacquot Island than on the mainland i n 1993 and 1994; however, i n 1993 the sighting index for coyotes (Canis latrans) on Jacquot Island was similar to that on the mainland. In 1994 there were no coyote sightings on Jacquot Island while those on the mainland had increased from 1993 (Table 10). The other main mammalian predator, the lynx (Lynx canadensis), was not seen at a l l on Jacquot Island i n either year and was seen only during 1993 on the mainland. Fewer avian predators were sighted on Jacquot Island than on the mainland i n both years. Sightings of individual 54 Table 9. Proximate causes of mortality of radio-tagged, juvenile snowshoe hares on Jacquot Island (J I) and the mainland in 1994. Percentage of mortalities Litter 1 Litter 2 Litter 3 Mortality factor Jl Mainland J l Mainland J l Mainland unknown predator 44.4 60.0 20.0 15.4 50.5 25.0 red squirrel 33.3 20.0 0.0 38.5 0.0 12.5 small mammal 0.0 0.0 0.0 7.7 0.0 6.2 weasel 0.0 0.0 0.0 7.7 0.0 50.0 great-horned owl 0.0 20.0 0.0 0.0 0.0 0.0 boreal owl 0.0 0.0 0.0 11.5 0.0 0.0 red-tailed hawk 0.0 0.0 0.0 0.0 50.5 6.2 unknown avian 0.0 0.0 0.0 15.4 0.0 0.0 abandoned/non-predator 22.2 0.0 80.0 3.8 0.0 0.0 total mortalities 9 5 5 26 2 16 total radio-tagged 30 27 34 46 30 50 55 Table 10. Sighting index of hare predators on Jacquot Island and mainland grids during the summers of 1993 and 1 994. Sightings/100 hrs 1993 1994 Predator type Predator Jacquot * Mainland Jacquot Mainland Mammalian coyote 0.14 0.13 0.00 0.23 lynx 0.00 0.14 0.00 0.00 weasel 0,00 0.05 0.00 0.04 Total mammalian 0.14 0.32 0.00 0.27 Avian great horned owl 0.00 0.07 0.00 0.02 hawk owl 0.00 0.47 0.00 0.00 red-tailed hawk 0.00 2.52 0.50 2.01 goshawk 0.00 0.91 0.00 0.42 marsh hawk 0.00 1.02 0.06 0.96 unidentified raptor 0,70 0.52 0.26 0.46 Total avian 0.70 5.51 0.82 3.87 *August not included 56 raptor species were also higher on the mainland than on Jacquot Island during both years; however, i n 1993 there was a higher incidence of unidentified raptor sightings on Jacquot Is1and. S l i g h t l y more raptors were seen on Jacquot Island i n 1994 than i n 1993. During the summer of 1994 a great-horned owl was heard hooting on the island i n August which i s not reflected i n the sighting index observations. This i s not unexpected since the owls are nocturnal and most of the f i e l d work occurred during the day. The mainland had fewer mammalian and raptor sightings i n 1994 compared to 1993. The sighting indices of a l l individual species, except the coyote, were lower i n 1994 than 1993 (Table 10). Red s q u i r r e l densities were estimated at 0.2/ha for both Jacquot Island north and south grids. Densities of red squirrels did not vary between years. On the mainland red squ i r r e l densities ranged between 1.75/ha and 3.83/ha (Kluane Boreal Ecosystem Project, unpublished). Ground squirrels were not observed on Jacquot Island during t h i s study nor i n previous studies on Jacquot Island. Adult Male Spring Body Weights and Condi t ion Indices There was no s i g n i f i c a n t difference i n male spring body weights or condition indices i n 1993 or 1994 between north and south Jacquot Island grids (1993, weights, t=0.87, p=0.21, condition index t=0.61, p=0.28; 1994, weights, t=0.77, p=0.23, condition index, t=0.68, p=0.26). Therefore 57 I pooled Jacquot north and south grids i n both years to increase sample size. The body weights and condition indices of snowshoe hares on Jacquot Island and the mainland could not be compared i n 1993 because only two hares were caught on control areas on the mainland i n A p r i l and May of 1993. There were no si g n i f i c a n t differences i n the body weights and the condition indices of hares on Jacquot Island and the mainland i n 1994 (weights, t=1.27, p=0.11; condition index, t=1.52, p=0.07) and there were no differences between years on Jacquot Island (weight, t=1.29, p=0.11; condition index, t=1.25, p=0.11; Tables 11 and 12). DISCUSSION During a previous study by Zimmerling (1993) both Jacquot Island and the mainland underwent population declines reaching low levels by the f a l l of 1992. The hare population decline on Jacquot Island was less dramatic and of shorter duration than that of the mainland population. The Jacquot Island population reached peak spring densities i n 1991 and declined 71% over the next 2 years, reaching low densities i n the spring of 1993, whereas, the mainland population declined 94% over 4 years, from the spring of 1990 to the spring of 1994. The Jacquot Island hare population declined one year la t e r than the mainland population and never reached the low levels seen on the mainland. Unlike the c y c l i c population on the mainland, which showed no change i n population density from the spring of 1993 to the spring of 1994, the Jacquot 58 Table 11. Mean body weights ± S.D. of male snowshoe hares in spring on study areas (sample size in parentheses). No significant differences were found between years on Jacquot Island or between areas in 1994. Male body weights Spring Spring 1993 1994 Jacquot Island 1217 + 121 1271 + 109 (9) (33) Mainland - 1230+ 140 (26) 59 Table 12. Mean condition indices ± S.D. of male snowshoe hares in spring on study areas (sample size in parentheses). No significant differences were found between years on Jacquot Island or between areas in 1994. Condition Indices Spring Spring 1993 1994 Jacquot Island 0.937 + 0.08 0.976+0.08 (9) (33) Mainland - 0.941 +0.09 (26) 60 Island hare population increased 3.5 f o l d . This immediate rebound i n the population after a decline i s not characteristic of c y c l i c populations which have a decline and low period which can l a s t 4 years (Keith 1990). Consequently either reproduction or survival were higher on Jacquot Island than the mainland from the spring of 1993 to the spring of 1994. S u r v i v a l In t h i s section I discuss how overwinter adult survival, overwinter juvenile survival, early juvenile survival (birth to 30 days), and summer adult survival d i f f e r between Jacquot Island and the mainland. I also discuss possible causes of these differences. Adult S u r v i v a l Fol lowing a Populat ion Decl ine There was no difference i n adult survival between the mainland and Jacquot Island i n the summer of 1993. Therefore, poorer adult survival during the summer on the mainland compared to Jacquot Island does not appear to be responsible for keeping the mainland hare population at low spring densities from 1993 to 1994. Overwinter survival of hares on Jacquot Island was higher than on the mainland i n the 1993-1994 winter. Overwinter survival on Jacquot based oh trapping data was 83% (equivalent to 98% per 30 days) and based on 7 radio-collared hares was 71% (equivalent to 96% per 30 days). This i s higher than the 12.6% (equivalent to 77.2% per 30 days) survival seen on the mainland during t h i s period (based on 61 radio-collared hares). Therefore higher overwinter survival may be an important factor that allowed Jacquot Island to increase so dramatically. F a c t o r s I n f l u e n c i n g A d u l t S u r v i v a l Predation has been found to be the main proximate cause of death for both c y c l i c and non-cyclic snowshoe hare populations (Keith 1990, Boutin et al. 1986, Sievert and Keith 1985). Similar results were found i n t h i s study with 60% of the mortalities being attributed to predation (Table 3). Food shortage, poor weather and disease may increase the ris k of predation for hares because these factors tend to cause hares to become malnourished which reduces body condition (Keith 1990). L i t t l e empirical evidence i s available, but Sievert and Keith (1985) found that hares with below average body condition indices suffered higher predation rates than hares i n good condition. In 1994 Jacquot Island hares had higher summer survival than the mainland and than Jacquot Island i n 1993. In both these cases there were no differences i n the condition indices of male snowshoe hares i n the spring indicating that these factors do not provide a l i k e l y explanation for the lower survival rates seen. Differences i n predator numbers seem a more l i k e l y explanation of the patterns of adult survival, as there were differences i n predator numbers between areas. Adult survival on Jacquot Island was higher i n 1994 than 1993 and, 62 although s l i g h t l y more avian predators were sighted i n 1994 than i n 1993, mammalian predator sightings dropped to zero. There was no s i g n i f i c a n t difference i n adult survival on Jacquot island and the mainland i n 1993. In general, fewer predators were sighted on Jacquot Island compared to the mainland during the summer of 1993, however the sighting index for coyotes was similar. In the summer of 1994 adult survival was higher on Jacquot Island than the mainland and the sighting index for a l l predators was lower on Jacquot Island. The main difference between 1993 and 1994 on Jacquot Island was that sightings of mammalian predators dropped to zero and survival increased. Higher summer survival of hares on Jacquot Island i n 1994 compared to 1993, and on Jacquot Island compared to the mainland i n 1994, was not associated with fewer predator numbers i n general, but with the presence of fewer mammalian predators. If mammalian predators have the largest impact on survival rates of adult hares, the majority of k i l l s i n 1993 should be due to coyotes. In fact, the major mortality factor i n the summer of 1993 was unknown predators due to the lack of signs l e f t at k i l l s s ites i n summer. Radio-collaring more hares i n future studies may help increase the number of k i l l s for which the cause of death i s known. Other studies have also suggested that mammalian predators have the most impact on hare population densities and adult survival. A r c t i c hares (Lepus arcticus) that were introduced onto islands reached high densities on 2 islands where no 63 mammalian predators were present (Mercer et al. 1981). Angerbjorn (1986) found that adult survival of a population of mountain hares decreased on an island i n Sweden when a fox remained on an island i n the summer. When a lynx remained on Jacquot Island during the summer of 1992 there was a sig n i f i c a n t drop i n survival rates on the part of the island where the animal concentrated i t s hunting e f f o r t s (Zimmerling 1993). Juveni le S u r v i v a l on Jacquot Is land Comparisons of early juvenile survival between years are important for determining i f the results seen are part of a consistent pattern. In 1991 Zimmerling (1993) found that hares on the Jacquot Island south grid had s i g n i f i c a n t l y higher survival than those on the north grid for f i r s t and second l i t t e r s . The main cause of mortality on the north grid was coyote predation while no deaths on the south grid were attributed to coyote predation. If there had been no coyote predation on the north grid and a l l those juveniles survived, survival on both grids would have been similar. This suggests that coyote predation was responsible for the differences i n survival on the two grids. Zimmerling (1993) hypothesized that higher survival on the south grid was related to dense cover at 10 cm above the ground. When coyotes are present, survival may d i f f e r between grids on Jacquot Island because cover density d i f f e r s between the two grids. An alternative to the cover hypothesis i s that the coyote concentrated i t s hunting on the 64 north grid during 1991, and that differences i n predator pressure caused the differences i n survival. These two hypotheses can be examined when coyotes are present on the island. Radio-collaring the coyote(s) would allow us to determine where they concentrate t h e i r hunting e f f o r t s . An examination of cover at sites where hares were k i l l e d would allow us to determine i f poor cover i s associated with higher predation rates. During 1994 no juvenile deaths were attributed to coyote predation and there was no difference i n early juvenile survival between the Jacquot Island grids. Exposure was an important mortality factor for juveniles in Zimmerling's (1993) study, but not i n this study indicating that i t i s not a regular occurrence. There were, however, i n my study two instances of l i t t e r s being abandoned in an open area where they may have been more vulnerable to exposure and predation. It i s possible that l i t t e r s were abandoned i n other studies and that the deaths were attributed to other factors. More intensive studies (monitoring the leverets twice da i l y for example) of early juvenile survival w i l l help us to determine i f abandonment i s a regular occurrence. Juveni le S u r v i v a l on Jacquot Is land Compared with the Mainland The pattern of survival of l i t t e r s d i f f e r s between Jacquot Island and the mainland. On the mainland the second l i t t e r had lower survival than f i r s t and t h i r d l i t t e r s during the peak and the low phase (this study, O'Donoghue 1994) while on 65 Jacquot Island the f i r s t l i t t e r has had the lowest survival (this study and Zimmerling 1993). O'Donoghue (1994) hypothesized that small mammal predators actively foraged for the newborn hares and that the number of potential predators of leverets during the second l i t t e r was greater than for f i r s t l i t t e r on the mainland because independence of juvenile squirrels corresponds to the second l i t t e r of hares. One hypothesis for the different patterns of survival seen on Jacquot Island and the mainland i s that lower densities of small mammal predators and better cover on Jacquot Island during second and t h i r d l i t t e r s result i n higher survival for those l i t t e r s . F i r s t l i t t e r s are usually born just as the vegetation i s emerging and cover may be reduced making them more susceptible to predation than are second and t h i r d l i t t e r s . Trostel (1986) examined recruitment, which encompasses early juvenile survival, reproduction, and dispersal, through live-trapping, and found no differences between Jacquot Island and the mainland from 1977-1984. However, Zimmerling (1993) d i r e c t l y examined early juvenile survival (using radio-telemetry) on Jacquot Island over one year and found that survival on the island was higher than the mainland during the f i r s t year of the decline phase on the mainland. Early juvenile survival during both my and Zimmerling's (1993) studies was higher than that measured for the mainland population during the peak (O'Donoghue 1994). During t h i s study juvenile survival was similar on Jacquot Island and the 66 mainland for f i r s t l i t t e r s of 1994. However, survival was higher for second and t h i r d l i t t e r s on Jacquot Island resulting i n overall higher survival on Jacquot Island. Jacquot Island appears to have consistently higher early juvenile survival (during the f i r s t 30-days of l i f e ) than the mainland (Zimmerling 1993, this study). Although mainland data are lacking for 1993 i t appears that high early juvenile survival allowed Jacquot Island to recover quickly after the population decline. The higher survival seen on Jacquot Island may be due to fewer small mammal predators and better cover on Jacquot Island. Long term monitoring of early juvenile survival on Jacquot Island and the mainland should be done to determine i f early juvenile survival i s consistently higher than the mainland. Experimental studies of early juvenile survival w i l l allow us to determine i f better cover and fewer small mammal predators are responsible for the higher survival seen on Jacquot Island. Survival of juveniles from September 1993 to May 1994 based on trapping was 48% on Jacquot Island while on the mainland survival of juveniles over the same period was 11.3% based on telemetry. Trapping usually underestimates survival because animals that have dispersed are counted as mortalities (Boutin and Krebs 1986) so Jacquot Island may have had even higher overwinter survival. High overwinter juvenile survival appears to have contributed to the 1993 population recovery on Jacquot Island; however, overwinter survival of juveniles does not appear to be consistently high 67 on Jacquot Island. Zimmerling (1993) found overwinter survival of juveniles (based on 11 radio-collared juveniles) to be 36% during the winter of 1991-1992 when the population was undergoing a decline. Overwinter survival should be studied on Jacquot Island using radio-telemetry to determine how overwinter juvenile survival varies from year to year. Using radio-telemetry to study survival allows us to dif f e r e n t i a t e between mortality, emigration, and immigration and would allow us to determine i f dispersal was an important factor influencing the population dynamics of Jacquot Island. Reproduction In t h i s section I discuss the impact of the length of time a female was caged prior to parturition on reproductive data collected on Jacquot Island. I also discuss how reproduction differed between Jacquot Island and a c y c l i c population and suggest reasons for the differences found. The E f f e c t of M a t e r n i t y Cages on Reproductive Data Females were kept i n maternity cages and supplied with food prior to par t u r i t i o n . It i s possible that t h i s could either negatively (increased stress) or p o s i t i v e l y (additional food) affect the female and the young. I t r i e d to l i m i t the impact of the cages by keeping the females i n them for as short a period as possible and by l i m i t i n g my contact with them to once da i l y . For l i t t e r groups one and two of 1994 the time a female spent i n a maternity cage prior to p a r t u r i t i o n was s i g n i f i c a n t l y related to l i t t e r size; however, l i t t e r size could not increase during the l a s t week 68 of pregnancy. I think the positive " e f f e c t " of the maternity cages on l i t t e r size was a consequence of putting the females i n maternity cages at similar weights, but dif f e r e n t stages of pregnancy. Because larger l i t t e r s have a heavier t o t a l weight (Figure 9a-c) i t would follow that females with larger l i t t e r s would themselves weigh more at the time of b i r t h than females with smaller l i t t e r s . Since a l l females were put i n maternity cages at similar weights, the females with larger l i t t e r s would be put i n the cages at an e a r l i e r stage of pregnancy than the females with small l i t t e r s and would therefore remain i n maternity cages longer than females with small l i t t e r s . Extra food, i n the form of commercial rabbit chow, was added to hare grids i n the Yukon and O'Donoghue (1992) did not find any difference i n the l i t t e r sizes of hares on these grids compared to control areas. For the f i r s t l i t t e r group i n my study a s i g n i f i c a n t effect of days caged on average leveret weight per l i t t e r was found. It i s conceivable that newborn weights were negatively affected by the time spent i n the maternity cage; however, this can more e a s i l y be explained by the relationship between l i t t e r size and newborn weights for the f i r s t l i t t e r (Figure 8a). In the second l i t t e r where there was no s i g n i f i c a n t relationship between l i t t e r size and weight, the length of time spent i n maternity cages was related to l i t t e r size only. I think that the apparent effect of maternity cages on average leveret weight per l i t t e r i s a consequence of the 69 negative relationship between average leveret weight and l i t t e r size found i n the f i r s t l i t t e r . I do not believe that the time a female spent i n the cage prior to pa r t u r i t i o n affected reproduction on Jacquot Island. The results that were obtained can more eas i l y be explained by the fact that females with larger l i t t e r s were put i n cages at an e a r l i e r stage of pregnancy. The apparent negative ef f e c t of time spent i n a cage pr i o r to p a r t u r i t i o n on the average leveret weight per l i t t e r seen i n the f i r s t l i t t e r group of 1994 was most l i k e l y a result of the negative relationship between l i t t e r size and l i t t e r weights. A s i g n i f i c a n t effect of time spent i n maternity cage on l i t t e r size and l i t t e r weight has not been found i n other studies using t h i s same technique (O'Donoghue 1991, Zimmerling 1993). Reproduction Fol lowing a Populat ion Decl ine Unfortunately no reproductive data were collected on the mainland during 1993 so i t i s impossible to d i r e c t l y determine i f differences i n reproduction or survival are responsible for the different dynamics seen on Jacquot Island and the mainland. It i s possible, however, to compare the reproductive output found on Jacquot Island with the pattern that has been seen i n other c y c l i c populations. The reproductive output of a c y c l i c population i n Alberta was monitored over 15 years and encompassed two population lows (Carey and Keith 1979; Figure 10). During t h i s study reproductive output was at i t s lowest l e v e l two summers pri o r to the lowest population density. The year before the 70 population low reproductive output increased 1.5 f o l d from i t s lowest levels but the population continued to decline. This one year delay before the population began to recover was not evident on Jacquot Island. Reproductive output on Jacquot Island increased 2.3 fo l d the summer after reproduction was at the lowest l e v e l and the population density increased. Higher reproductive output may have played a part i n allowing the Jacquot Island hare population to increase while the mainland remained at low densities over the summer of 1993; however, i t i s not the only factor. Carey and Keith's (1979) population had s u f f i c i e n t l y high reproduction to increase i n the year prior to the population low. I estimated the expected f a l l population density (EFPD) for the years p r i o r to the population lows of the Alberta population, taking only reproduction and adult survival into account. EFPD=(RN *SPD/2)+(SPD*SAS) Where RN i s realized n a t a l i t y (number of leverets produced per female ali v e at the start of the breeding season), SPD i s spring population density and SAS i s summer adult survival. Based on t h i s calculation I predicted a 3 to 4 fo l d increase i n population density over the summer. Therefore juvenile survival or overwinter adult survival must be important factors i n causing the c y c l i c population to remain 1 low because reproductive output i s high enough to allow the population to increase. 7 1 The reproductive output of c y c l i c populations appears to be s u f f i c i e n t l y high to allow the population to increase the summer before the population low; however, the population does not increase, indicating that other factors, l i k e adult or juvenile survival, must be keeping the population at low numbers. Nonetheless, the reproductive output on Jacquot Island was at least 1.5 times higher than the reproductive output of a c y c l i c population i n Alberta the year before the low indicating that reproduction of c y c l i c populations may i n some way be r e s t r i c t e d during t h i s period. Timing of Breeding The timing of breeding of the island population i n 1993 and of both the island and mainland populations i n 1994 was advanced by almost two weeks compared to a 1991 study on the mainland and a 1992 study on Jacquot Island (O'Donoghue 1992, Zimmerling 1993). This allowed for the production of a fourth l i t t e r on the mainland which has not been recorded during previous studies i n the Kluane area. There i s also evidence that there was a fourth l i t t e r on the island, however, no detailed study of fourth l i t t e r was completed i n either location. T o t a l R e p r o d u c t i v e Output The t o t a l reproductive output (number of leverets produced per adult female surviving the breeding season) on Jacquot Island varied 2.5 fold from 6.8 i n 1992 (Zimmerling 1993) to 16.8 i n 1994. This i s similar to the 2.4 fold v a r i a t i o n i n reproductive output seen i n a c y c l i c population i n Alberta 72 (Carey and Keith 1979). It i s important to note that although non-cyclic populations do not show predictable population fluctuations they do go through periods of increase and decline. The 2.5 fold variation i n t o t a l reproductive output of hares on Jacquot Island was recorded during a population increase. No other studies of non-cyclic populations have recorded t o t a l reproductive output during a population increase and the t o t a l reproductive output has always been constant (Dolbeer and Clark 1975, Kuvlesky and Keith 1983). More long term studies of reproduction of non-c y c l i c populations w i l l allow us to determine i f reproduction on Jacquot Island i s similar to other non-cyclic populations. L i t t e r S i z e and Neonate Weights Through l i v e trapping, Trostel (1986) found that juvenile recruitment, which encompasses reproduction, immigration, emigration, and early juvenile survival, was similar between the non-cyclic Jacquot Island hare population and the c y c l i c mainland population between 1977 and 1984. In a more detailed study Zimmerling (1993) compared reproduction i n one year on Jacquot Island with the mainland and found no differences. During t h i s study, however, both l i t t e r sizes and newborn weights were larger on Jacquot Island than on the mainland i n 1994, indicating that females were able to produce more offspring. Large l i t t e r s and heavier neonate weights have been recorded for other species l i v i n g on islands. Ebenhard (1990) found that an insular population of f i e l d voles 73 (Microtus agrestis) had larger l i t t e r and neonate masses than mainland populations, however the body masses of the adult island voles were also greater. No evidence of differences in the size of hares on Jacquot Island and the mainland were found so i t appears that Jacquot Island hares were able to allocate more resources for reproduction than mainland hares i n 1994. A possible explanation for the larger l i t t e r size and newborn weights found on Jacquot Island was that hares on the Island were i n better physical condition than mainland hares. Boonstra and Singleton (1993) concluded, by examining the stress response of male snowshoe hares, that control hares i n a declining population were i n poor condition compared to well-fed hares. However, a comparison of condition indices of male snowshoe hares trapped i n the spring on the island and the mainland indicated that there was no s i g n i f i c a n t differences i n t h e i r condition (Table 12). Another possible explanation could be the lack of predators on Jacquot Island compared with the mainland. Ylonen (1989) showed that reproduction i n bank voles (Clethrionomys glareolus) could be suppressed when they were exposed to weasel scent. A sighting index of predator numbers indicated that there were fewer predators on the island and adult survival rates on the island were also higher. It i s possible that reproduction on the mainland was lower than the island because there were more predators on 74 the mainland. The possible impact of predators on snowshoe hare reproduction should be examined experimentally. CONCLUSION Jacquot Island and the mainland both underwent population declines reaching similar low levels i n the f a l l of 1992 (Zimmerling 1993). During my study the mainland population remained at low densities from the spring of 1993 to the spring of 1994 while Jacquot Island increased 3.5 f o l d over the same period. Three main hypotheses were put forward to explain the different population trends seen during t h i s period. Hypothesis 1. Differences i n reproduction explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland. Hypothesis 2. Differences i n mortality explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland Hypothesis 2a. Only juvenile mortality d i f f e r s between the island and the mainland. Hypothesis 2b. Only adult mortality d i f f e r s between the island and the mainland. Hypothesis 2c. Both adult and juvenile mortality d i f f e r between the island and the mainland. Hypothesis 3. Differences i n both reproduction and mortality explain the differences seen i n the population dynamics of snowshoe hares on Jacquot Island and the mainland. 75 My results support hypothesis 3 and Table 13 outlines the predictions for each hypothesis and the results obtained i n this study. Proximate F a c t o r s I n f l u e n c i n g the Jacquot I s l a n d P o p u l a t i o n Increase During the summer of 1993 28-day survival rates for adult hares on Jacquot Island and the mainland were similar. However, overwinter survival of adults, from September to May, was six times higher on Jacquot Island than the mainland. Juvenile survival u n t i l weaning was not monitored on the mainland i n 1993 and only 25 leverets were marked on Jacquot Island. None of these 25 leverets were known to have died, which suggests that early juvenile survival was very high on Jacquot Island i n 1993. Overwinter survival of juveniles, from September 1993 to May 1994, was higher on Jacquot Island than the mainland. These results eliminate hypotheses 2a and 2b since both adult and juvenile mortality were lower on Jacquot Island than the mainland (see Table 13). Reproduction was not monitored on the mainland during 1993. Comparisons with reproductive data from a di f f e r e n t c y c l i c population of hares (Carey and Keith 1979) indicate that reproductive output was higher for Jacquot Island than for a c y c l i c population during the low phase of the hare cycle. Reproduction, therefore, may be an important factor i n allowing Jacquot Island to increase while c y c l i c populations remain at low densities. Since reproduction on 76 CO ZD CO or CO LU CO h-O D_ >-T J C co c o 4-> o T J O CD ro| o E T J ±i 75 21 - Q C _ 0) CM .2, c o C\J E ro t o CM M o CD 4 ^ > i= 3 CO •r t -c? g CM E g •4-" o 3 T J O k— Q . CD CC o h-o < 0> O) CD E CO CO CD E ro CO CD E ro co CD D) CD > •3 s 2 Q -Q . ^ 2 O CD O CD O CD o CD E ro 00 CD o CD E ro co c: CD > 3 CD O <D O CD O CD O CD E ro to CD E ro co ro t o E CO t o E 3 T J CO 77 Jacquot Island was higher than for a c y c l i c population during the low phase of the population cycle, hypotheses 2a-c, that lower mortality on i t s own allowed Jacquot Island to increase while the mainland was at low densities, were tentatively rejected (see Table 13). Reproductive data should be collected on both Jacquot Island and the mainland during the population low to be able to test these hypotheses. These results indicate that both higher reproductive output and lower mortality played a role i n allowing the Jacquot Island hare population to increase while the mainland remained at low densities (hypotheses 3). While reproduction on Jacquot Island appears to be higher than i n c y c l i c populations during the low, the reproductive output of c y c l i c populations i s s u f f i c i e n t l y high that the population could increase during t h i s period. Despite t h i s , mainland populations, remained at low densities suggesting that poor survival must also be a factor keeping the mainland population at low densities during 1994. Four components of survival were considered: early juvenile survival (to 30 days), summer adult survival, and overwinter adult and juvenile survival. On Jacquot Island overwinter survival of juveniles and adults and early juvenile survival was higher than the mainland. In thi s study adult survival over the summer was not di f f e r e n t between Jacquot Island and the mainland indicating that i t was not a factor i n the Jacquot Island population increase 78 between 1993 and 1994 and that i t was not involved i n keeping mainland densities low. Higher reproductive output, higher early juvenile survival, and higher overwinter adult and juvenile survival on Jacquot Island are the proximate factors that allowed the Jacquot Island population to increase while the mainland remained at low densities from 1993 to 1994. The ultimate causes of the differences observed i n the two areas are harder to determine. Various hypotheses have been proposed to explain the observed differences i n survival and reproduction between the mainland and Jacquot Island. U l t i m a t e Causes of Higher S u r v i v a l on Jacquot I s l a n d Zimmerling (1993) hypothesized that fewer small mammal predators and more dense cover at 10 cm above the ground were responsible for the higher early juvenile survival found on Jacquot Island i n 1991. The only other year i n which mainland and Jacquot Island data were collected concurrently was 1994 and survival on Jacquot Island was higher i n this year also. During this study, as i n Zimmerling's (1993) study, higher juvenile survival on Jacquot Island coincided with lower numbers of small mammal predators ( s q u i r r e l s ) . The importance of cover as protection against predation could not be assessed i n 1994 since only 10 leverets were depredated. Juvenile survival may be consistently higher on Jacquot Island than the mainland; however, we need to obtain additional data where juvenile survival i s collected on Jacquot Island and the mainland during the same years. 79 Experimental and observational studies w i l l allow us to determine i f differences i n predator numbers and cover could account for the di f f e r e n t survival rates and d i f f e r e n t pattern of survival of the l i t t e r groups seen on Jacquot Island and the mainland. Overwinter survival of both juvenile and adult hares was higher on Jacquot Island than on the mainland i n 1993-1994. There have been no detailed studies comparing overwinter survival on the two areas so i t i s not known i f overwinter survival of adult and juveniles i s consistently higher on Jacquot Island. In the spring of 1994 there were no differences i n condition between hares i n the two areas indicating that i t i s not poor condition of mainland hares that caused them to have higher mortality rates. One hypothesis that could explain the different survival rates i s differences i n predator numbers affecting the two populations. A detailed radio-telemetry study of overwinter survival on Jacquot Island and the mainland would allow us to determine i f survival i s consistently higher on Jacquot Island and comparisons of food a v a i l a b i l i t y and predator numbers may allow for the determination of the ultimate causes of any differences found. Summer adult survival did not d i f f e r between Jacquot Island and the mainland i n 1993, however, i t was higher than the mainland i n 1994. Zimmerling (1993) found that survival on Jacquot Island south grid was lower than on the mainland i n 1992. This study concurs with Zimmerling's hypothesis 80 that summer adult survival varies i n a stochastic manner depending on predator numbers. On the mainland predation follows a predictable delayed density dependent pattern (Trostel 1987). U l t i m a t e Causes of Higher Reproduction on Jacquot I s l a n d Reproduction was studied concurrently on Jacquot Island and the mainland i n 1991 (Zimmerling 1993) and no s i g n i f i c a n t differences were found. However, i n 1994 Jacquot Island hares had higher reproductive output than the mainland. Condition indices were not different between Jacquot Island and the mainland, and as a result lower reproduction on the mainland does not appear to be a result of poor condition. There were, however, large differences i n predator numbers and the influence of predators on snowshoe hare reproduction should be examined experimentally. Summary Higher reproduction, higher early juvenile survival, and higher overwinter adult and juvenile survival were a l l key proximate factors allowing the Jacquot Island hare population to increase while the mainland population remained at low densities from the spring of 1993 to the spring of 1994. These results support hypothesis 3, that a combination of higher reproductive output and lower mortality accounted for the dif f e r e n t dynamics seen on Jacquot Island compared to the mainland. More research i s required to determine the ultimate causes of these differences. This study and e a r l i e r 81 studies (Zimmerling 1993, Trostel 1986) have found some evidence to suggest that differences i n predator numbers and habitat are important areas requiring future work. One of the limitations of t h i s study i s the lack of r e p l i c a t i o n . Only one set of c y c l i c and non-cyclic populations was compared i n th i s study and consequently the conclusions are limited to Jacquot Island and the mainland. It may be useful to sample other northern island hare populations to determine i f they, too, are non-cyclic. If we are able to f i n d other areas with similar dynamics to Jacquot Island i t w i l l be possible to make more far reaching conclusions. 82 LITERATURE CITED Angerbjorn, A. 1986. Population dynamics of mountain hares Lepus timidus L. on islands. Doctoral dissertation. University of Stockholm, Department of Zoology, Stockholm 1986. Bailey, J.A. 1968. A weight-length relationship for evaluating physical condition of co t t o n t a i l s . J. Wildl. Manage. 3 2:835-841. Boonstra, R., and Singleton, G.R. 1993. Population declines i n the snowshoe hare and the role of stress. General and Comparative Endocrinology, 91:126-143. Boulanger, J., and Krebs, C.J. 1994. Comparison of capture-recapture estimators of snowshoe hare populations. Can. J. Zool. 7 2:1800-1807. Boutin, S., and Krebs, C.J. 1986. Estimating survival rates of snowshoe hares. J. Wildl. Manage. 50:592-594. Boutin, S., Krebs, C.J., S i n c l a i r , A.R.E., and Smith, J.N.M. 1986. Proximate causes of losses i n a snowshoe hare population. Can. J. of Zool. 6 4:606-610. Carey, J.R., and Keith, L.B. 1979. Reproductive change i n the 10-year cycle of snowshoe hares. Can. J. Zool. 57:375-390. Chitty, D. 1967. The natural selection of s e l f regulatory behaviour i n animal populations. Proceedings of the Ecological Society of Australia, 2:51-78. Chitty, H. 1950. The snowshoe rabbit enquiry, 1946-1948. J. Anim. Ecol. 19:15-20. Dolbeer, R.A., and Clark, W.R. 1975. Population ecology of snowshoe hares i n the central rocky mountains. J. Wildl. Manage. 3 9:535-549. Ebenhard, T. 1990. A colonization strategy i n f i e l d voles (Microtis agrestis)i reproductive t r a i t s and body si z e . Ecology, 7 1:1833-1848. Green, R.G., and Evans, CA. 1940. Studies on a population cycle of snowshoe hares on the Lake Alexander area. I. Gross annual census 1932-1939. J. Wildl. Manage. 4:220-238. Kaplan, E.L. and Meier, P. 1958. Nonparametric estimation from incomplete observations. Journal of the American S t a t i s t i c a l Association, 5 3:457-481. 83 Keith, L.B. 1990. Dynamics of snowshoe hare populations. In Current Mammalogy 2. Edited by H.H Genoways. Plenum Press,New York. pp. 119-195. Keith, L.B., Cary, J.R., Rongstaad, O.J., and Brittingham, M.C. 1984. Demography and ecology of a declining snowshoe hare population. Wildl. Mon. 90:1-43. Keith, L.B. and Windberg, L.A. 1978. A demographic analysis of the snowshoe hare cycle. Wildl. Mon. 5 8:1-70. Krebs, C.J., Boonstra, R., Boutin, S., Dale, M., Hannon, S., Martin, K., S i n c l a i r , A.R.E., Smith, J.N.M., and Turkington, R. 1992. What drives the snowshoe hare cycle i n Canada's Yukon? In W i l d l i f e 2001: Populations. Edited by McCullough, D.M. and Barrett, R. pp. 886-896. Krebs, C.J., Gilbert, B.S., Boutin, S. S i n c l a i r , A.R.E., and Smith, J.N.M. 1986. Population biology of snowshoe hares. I. Demography of populations i n the southern Yukon, 1976-84. J. Anim. Ecol. 55:963-982. Kuvlesky, W.P. and Keith, L.B. 1983. Demography of snowshoe hare populations i n Wisconsin. J. Mammal. 64:233-244. Mercer, W.E., Hearn, B.J., and Finlay, C. 1981. A r c t i c hare populations i n insular Newfoundland. In Proc. of World Lagomorph Conf. Edited by Myers, K. and Mac Innes, CD. Guelph, Ontario, 1979. pp. 478-485. O'Donoghue, M. 1991. Reproduction, juvenile survival and movements of snowshoe hares at a c y c l i c population peak. M.Sc. Thesis, University of B r i t i s h Columbia, Vancouver, B.C. O'Donoghue, M. 1994. Early survival of juvenile snowshoe hares. Ecology, 7 5:1582-1592. O'Donoghue, M., and Krebs, C.J. 1992. Effects of supplemental food on snowshoe hare reproduction and juvenile growth at a c y c l i c population peak. J. Anim. Ecol. 61:631-641. Otis, D., Burnham, K.P., White, G.C., and Anderson, D.R. 1978. S t a t i s t i c a l inference from capture data on closed animal populations. Wildl. Mon. 62:1-135. Pollock, K.H., Winter stein, S.R., Bunck, CM., and Curtis, P.D. 1989b. Survival analysis i n telemetry studies: The staggered entry design. J. Wildl. Manage. 53:7-15. Pollock, K.H., Winterstein, S.R., and Conroy, M.J. 1989a. Estimation and analysis of survival distributions for radio-tagged animals. Biometrics, 45:99-109. 84 Rongstad, O.J., and Tester, J.R. 1971. Behaviour and maternal relations of young snowshoe hares. J. of Wildl. Manage. 35:338-346. Savage, I.R. 1956. Contributions to the theory of order s t a t i s t i c s - t h e two sample case. Ann. Math. Stat. 2 7:590-615. Severaid, J.H. 1942. The snowshoe hare. Its l i f e history and a r t i f i c i a l propagation. Maine Department of Inland Fisheries and W i l d l i f e , Augusta, Maine. Sievert, P.R., and Keith, L.B. 1985. Survival of snowshoe hares at a geographic range boundary. J. Wildl. Manage. 46:854-866. Trostel, K. 1986. Investigation of causes of the 10-year hare cycle. M.Sc. Thesis, University of B r i t i s h Columbia, Vancouver, B.C. Trostel, K., S i n c l a i r , A.R.E., Walters, C.J., and Krebs, C.J. Can predation cause the 10-year cycle? Oecologia, 74:185-192. Ylonen, H. 1989. Weasels Mustela nivalis supress reproduction i n c y c l i c bank voles Clethrionomys glareolus. Oikos, 55:138-140. Zimmerling, T. 1993. A study of factors influencing a non-c y c l i c , island population of snowshoe hares. M.Sc. Thesis, University of B r i t i s h Columbia, Vancouver, B.C. 85 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099008/manifest

Comment

Related Items