POPULATION DYNAMICS AND SEASONAL MOVEMENT PATTERNS OF DALL SHEEP IN THE ATIGUN CANYON AREA, BROOKS RANGE, ALASKA

Transcription

1 POPULATION DYNAMICS AND SEASONAL MOVEMENT PATTERNS OF DALL SHEEP IN THE ATIGUN CANYON AREA, BROOKS RANGE, ALASKA By Bob L. Summerfield, B. S. s CAME QL 737.U53 S

2 &l 157. &53 S95 IC1 7~ POPULATION DYNAMICS AND SEASONAL MOVEMENT PAITERNS OF DALL SHEEP IN THE ATIGUN CANYON AREA, BROOKS RANGE, ALASKA A THESIS Presented to the Faculty of the University of Alaska in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE By Bob L. Summerfield, B. S. Fairbanks, Alaska December, 1974 ARLIS Alaska Resources Library & lnfè..~ination Services t\ l r, '-~ <'~ "!.,, L' v... Ctbt..q..t A~~-::t...:'l..d l " "h"r "...,:

3 POPULATION DYNAMICS AND SEASONAL MOVEMENT PAITERNS OF DALL SHEEP IN THE ATIGUN CANYON AREA, BROOKS RANGE, ALASKA REC(M.ŒNDED: / APPROVED: Re sources "Z-1 6c-~ \'f 7-j Date Provost Date /:/~Z/?4-

4 ABSTRACT Population dynamics and seasonal movements of Dall sheep (Ovis dalli) were studied in the Brooks Range in 1973 and Background data were available on the population from 1970 and The population size decreased from over 300 in 1970 to about 275 in Productivity fluctuated from 10 to 59 lambs per 100 ewes, and lamb survival was high. Adult mortality was low during early age and increased rapidly after about age eight. Winter range was a small portion of the total range used in summer. Summer movements began in June and extended up to 25 miles. High use of mineral licks occurred during movements between winter and summer ranges, with an average of 90.0 minutes spent per lick visit. Sheep returned to winter range in August and September. The sizes of ram, ewe, and mixed bands were compared, and seasonal and annual changes in band size were analyzed. iii

5 ACKNOWLEDGMENTS This study was financed by Federal Aid to Wildlife Restoration, Alaska Project Nos. W-17-5 and W-17-6, Job No , through the Alaska Cooperative Wildlife Research Unit, University of Alaska, Fairbanks. 1 extend my sincere thanks to the following people: Dr. David R. Klein, Leader, Alaska Cooperative Wildlife Research Unit, for his advice and encouragement from inception through completion of the study, for critical reading of the manuscript, and for helpful days spent in the field. Mr. David C. Allen, my field assistant, for his dedication and endurance under frequently adverse conditions, for his willingness to always do more than required, and for his companionship during long days away from home. Alyeska Pipeline Service Company, for logistic support and frequent use of their Galbraith Lake facilities. Mr. John Clark, Mr. Ted Hill, and many others stationed at the Galbraith Lake construction camp whose hospitality, encouragement, and assistance will long be remembered. The Bureau of Land Management, for assistance in conducting aerial surveys. Dr. Keith Van Cleve, Associate Professor of Forestry, for analysis of mineral lick soil samples. Dr. Russell D. Guthrie, Professor of Zoology, Dr. Samuel J. Harbo, Jr., Head of the Department of Wildlife and Fisheries, and Dr. Peter C. iv

6 v Lent, Assistant Leader, Alaska Cooperative Wildlife Research Unit, for critical reading of the manuscript and many helpful suggestions for its improvement.

7 .._, TABLE OF CONTENTS Page INTRODUCTION THE STUDY AREA Location. Climate Geology and Physiography. Vegetation... Vertebrate Fauna. STUDY METHODS... POPULATION DYNAMICS. Size and Structure of the Population. Productivity and Lamb Survival. Survival and Mortality of Adult Sheep Accidents..... Disease and Parasites. Hunting. Predation. SEASONAL MOVEMENTS Winter Range. Spring Movements. Mineral Lick Use Summer Range... Fall Movements.. Seasonal and Annual Variation in Band Size and composition. SUMMARY OF CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS. Conclusions.... Management Recommendations. APPENDIX:. LITERATURE CITED v

8 LIST OF TABLES Page Table 1. June sheep classification counts in Atigun Canyon Table 2. Table 3. Table 4. Percent of rams represented by each horn curl class during four years Numbers of sheep counted on wintering areas 2-4 in Productivity and lamb survival in Atigun Canyon during four years Table 5, Dall sheep remains found on the Atigun study area Table 6. Construction of a synthetic cohort from rams found dead in each horn curl class Table 7. Table 8. Table 9. Harvest of Dall sheep rams from the study area and vicinity... Mineral licks used by Dall sheep on the study area Chemical analysis of soils from a lick and a nonlick area Table 10. Duration of mineral lick use by Dall sheep in Atigun Canyon, June-August, Table 11. Percent occurrence of each sheep class in ram, ewe, and mixed bands vi

9 LIST OF FIGURES Figure 1. Map of the Atigun study area... Figure 2. Dall sheep winter range and lambing areas Figure 3. Age distribution of Dall sheep remains found on the Atigun study area. Page Figure 4. Comparison of the percèntage of rams in each horn curl class for the living population and a reconstructed cohort Figure 5. Relationship of horn growth rate to age at death. 40 Figure 6. Sheep feeding on low slopes in the spring 51 Figure 7. Figure 8. Winter range directly across Galbraith Lake from the Alyeska pipeline construction camp Winter range used by rams in a small drainage into the Atigun River Figure 9. The main lambing area in the northeastern portion of Atigun Canyon. 53 Figure 10. Lick A, the most heavily used mineral lick on the study area.. 54 Figure 11. Lick B, a primary lick near the eastern end of Atigun Canyon. 54 Figure 12. Mineral licks and travel routes utilized by Dall sheep on the study area.. 62 Figure 13. Mean duration of lick use for ewes and rams two years of age and older.. 72 Figure 14. Relationship of horn curl to age. 74 Figure 15. Relationship of age to time spent in mineral licks. 75 Figure 16. Mean elevation of Dall sheep bands in and near Atigun Canyon during May-August.. 80 Figure 17. Mean size of ram, ewe, and ali bands during May-August vii

10 viii List of Figures, continued. Figure 18. Percent occurrence of band sizes for ram, ewe, and mixed bands during March-September.. Page. 88 Figure 19. Ram, ewe, and mixed bands as the percent of total identifiable bands recorded in and near Atigun Canyon during May-August

11 INTRODUCTION Mountain sheep are one of the most highly esteemed inhabitants of our wildlands, yet they are among the most sensitive of all mammals to changes in the environment induced by man. Herds of mountain sheep were once widely dispersed throughout western North America from Alaska to Mexico. In most localities, however, numbers of sheep have dwindled steadily since the coming of the white man, and present distribution ineludes only a fraction of the historie range. After years of study, the reasons for this decline -- introduction of disease and parasites, restriction of habitat and competition from domestic livestock, improper game laws and illegal hunting -- are now at least in part understood. In the light of this information, most western states have now undertaken programs to increase the dwindling numbers of mountain sheep on areas which they have historically occupied. This, however, requires both time and money and is usually complicated by complex social and economie conflicts. Alteration of traditional grazing practices or long-established game law policies, for instance, has often been difficult or impossible to accomplish. In Alaska, Dall sheep (Ovis dalli) populations have received less impact from human encroachment than is true ofmountain sheep habitat farther to the south. Dall sheep numbers in Alaska, therefore, are probably not much different today from what they were a century ago. In recent years, however, the increased focus on development of resources in the far north has been strongly felt in Alaska. The results 1

12 1 2 have been increased disturbance of Dall sheep and, in sorne cases, the physical disruption of their habitat. In at least one instance, development activities appear to have directly contributed to the abandonment of portions of traditional sheep range (see Linderman, 1972). Increasing levels of development have also produced increasing human populations and, therefore, greater hunting pressure. This bas led to a decline in the quality of trophy animais collected by hunters and a scarcity of legal rams in the most accessible areas. The future welfare of Dall sheep is, therefore, in question unless wise management policies are initiated to cope with the increasing stresses of development, higher human population levels, and more intensive hunting. This is a challenge which can be met, but only through well planned research contributing to our understanding of mountain sheep ecology and by adjusting our political, economie, and social attitudes. One of the areas of greatest potential development in Alaska is the Brooks Range and northern coastal plain. This is an area richly supplied r with natural resources which have scarcely been tapped. The Brooks Range also contains much excellent Dall sheep habitat, and the potential for conflict between human development in the area and perpetuation of the sheep populations there is great. Development of petroleum and mineral resources, for example, will necessitate greater accessibility. This will not only increase disturbance from aircraft operation and road construction, but will also place the region within easier reach of the public, resulting in still more human disturbance to wildlife and perhaps eventually the loss of critical habitat. This study was undertaken on the northern side of the Brooks Range adjacent to the route of the Trans-Alaska hot oil pipeline, which will

13 3 run from Prudhoe Bay to Valdez, Alaska. This 48-inch pipeline and the accompanying gravel road that will ultimately become a part of the state highway system will run through Dall sheep habitat on both sides of the Brooks Range. On the north side of the range, a pumping station, a major construction camp, a staging area and an airstrip will be located within three miles or less of winter range and lambing areas used by the study population. Some sheep within this population are also known to cross the pipeline route during their summer movements. It seems inevitable that this degree of development will affect the adjacent sheep population. To properly assess this effect, we must know something of the natural ecology of the population. The focus of this study, therefore, has been toward an assessment, before substantial development has occurred, of the population dynamics and seasonal movement patterns of these sheep, which live near the northern extremity of mountain sheep distribution in North America. This information should make it possible in future years to better evaluate any man-induced changes that may come about within this mountain sheep population.

14 THE STUDY AREA Location This study was conducted at the juncture of the Endicott and Philip Smith mountains on the northern face of the Brooks Range between the latitudes of 68 10' and 68 35'N and the longitudes of ' and 'W. Topographical features comprising a loose boundary around the study area (see Fig. 1) are the Atigun River and Galbraith Lake on the west, the foothills to the north, the Sagavanirktok River on the east, and the continental divide or crest of the Brooks Range on the south. Along the west boundary, the study area lies adjacent to approximately 25 miles of the route for the Trans-Alaska hot oil pipeline and associated road system as well as the Galbraith Lake construction camp. Within these boundaries, the bulk of study effort was concentrated on the area known as Atigun Canyon, which lies between Galbraith Lake and the Sagavanirktok River where the Atigun River turns from its northward course and begins flowing in a northeasterly direction. At the onset of the study the Trans-Alaska pipeline was scheduled to pass through this canyon, but it has since been rerouted around the canyon on the north. Atigun Canyon lies approximately 30 miles north of Dietrich Pass in the continental divide. Climate Weather information is available from the Galbraith Lake construction.camp on a sporadic basis since October, The nearest climatological station with an extended period of operation is located at the 4

15 5 Figure 1. Map of the Atigun study area (map from U. S. Geological Survey, Philip Smith Mountains Quadrangle, 1956). Legend: Study area boundary..._ Approximate Trans-Alaska pipeline route 1 = Atigun Canyon 2 = Black Mountain 3 = Galbraith construction camp and airfield 4 = Guard House Rock 5 = Pump Station No. 4

16 6

17 7 village of Anaktuvuk Pass, sorne sixty miles west-southwest of Galbraith Lake. Specifie and detailed analysis of the climate of the study ar~a is, therefore, not possible. With the information available, however, it is possible to discuss the climate of the area in broader terms. Temperatures in this section of the Brooks Range may vary from below -50 Fahrenheit in winter to near 80 in summer. The lowest temperature recorded at Galbraith Lake during the winter of was -49 on January 23 and February 28, and the highest temperature recorded by myself on the study area during the summer of 1973 was 75 on July 23. The mean monthly temperature rises above freezing only during the months of June, July, and August, and freezing may occur during any month of the year. Temperature inversions are characteristic of arctic Alaska during both summer and winter (Searby, 1971), a fact which would favor sheep on mountain slopes during the cold months. Strong winds frequently combine with low temperatures in this area to make the climate among the most severe in the state. During the summer of 1973, winds reached 50 m.p.h. on two occasions and m.p.h. severa! times. Wind speeds may also reach m.p.h. in association with winter storms, but fortunately are lighter during periods of extreme cold (Searby, 1968). Still, moderate winds in conjunction with minimum winter temperatures may produce a chili factor lowering temperatures to the equivalent of at least -80. Precipitation on the study area amounts to only 8-10 inches per year, most of which falls during the summer months. Annual snowfall at Anaktuvuk Pass averages 63 inches (Searby, 1968), but is probably a good deal lower in the Atigun and Sagavanirktok drainages as indicated by climatological data from Galbraith Lake and observations made in March and

18 8 April of 1971, 1973 and Light, temporary snows may occur throughout summer, but accumulation in the valley bottoms usually begins in late September. Wind transport of snow and drifting are prevalent, filling small ravines and depressions and exposing ridgetops where sheep are able to forage. Late snowmelt resulting from relatively cold spring temperatures combines with a permafrost layer which halts downward percolation of water to produce an abundance of soil moisture throughout the growing season in what might otherwise be considered an arid environment. A climatic factor which uniquely affects sheep populations above the Arctic Circle is length of day. There is nearly a three-month period of continuous daylight during summer and a period of similar length during winter when darkness predominates. This presents extra opportunities for observation of sheep from May through July but makes observation from November through January impractical. Geology and Physiography The bedrock forming the backbone of the Brooks Range in the region of the study area is derived from Paleozoic marine sediment and consists of limestone, quartzite, shale and schist. Skirting the higher mountains and forming the north front of the range are younger rocks of sandstone, conglomerate, and slate dating from the Permian through early Cretaceous periods (Gryc, 1958; U. S. Dept. of the Interior, 1971). Periodic uplifting and eroding away were characteristic of the area from Jurassic until late Tertiary times, when massive uplifting then gave the range its present form as a high mountainous region. Renewed uplift; erosion and glaciation since that time account for its present appearance (Gryc, 19S8).

19 9 The U-shaped valleys typical of the area were widened, straightened, and scoured by four major glacial advances during the Quaternary period (Detterman, 1953). These glaciers originated in the valley heads, moved downstream, and merged at the mountain front. The highest mountains protruded above the ice even during maximum glacial development (Keller et al., 1961). After recession of the glaciers, moraine and outwash deposits covered the valley floors with a veneer of gravel, and this, along with silt, sand, and localized bedrock, now forms the major streambeds (U. S. Dept. of the Interior, 1971). The elevation of the study area ranges from 2200 feet at the mouth of Atigun Canyon to 7610 feet, with five mountain peaks exceeding the 7000 foot level. Crest lines connecting these peaks are formed by intersecting cirques, aretes, homs and irregular ridges (Keller et al., 1961). Several remnant glaciers remain in the higher north-facing cirques while the cirques at lower levels are now ice free. The north-south oriented valleys of the Atigun and Sagavanirktok rivers are broad and of relatively low gradient, showing the knob and kettle topography typical of glaciated areas as well as polygonal ground, frost boils, mud slumps and other features associated with permafrost regions. Near Galbraith Lake, however, the Atigun River turns sharply in a northeasterly direction and drops 400 feet in approximately 8 miles through Atigun Canyon. Throughout this section the river is normally too swift and contains too many rapids to either wade or navigate safely, and in the lower three miles it frequently cuts through nearly vertical slopes making foot travel adjacent to the river impossible.

20 10 Vegetation Vegetation on the study area may be classified into four plant community types described by Spetzman (1959). These are flood-plain and cutbank vegetation, cottongrass meadows, dry upland meadows, and outcrop and talus vegetation. Wet sedge meadows and aquatic vegetation (also described by Spetzman, 1959) occur in much more limited extent at the lower elevations and are of minor importance to Dall sheep. The most conspicuous species of flood-plain and cutbank communities is feltleaf willow (Salix alaxensis). 1 This is the tallest vegetation of the area, reaching a height of as much as 20 feet. It represents, however, only the second of four successional stages typical of flood-plain and cutbank areas. Associated plants occurring in this community are various species of grasses, sedges, shrubs, herbs, mosses and lichens. Willows and the other associated species are utilized extensively by Dall sheep, as well as by moose and caribou, on localized areas at certain times of the year. Cottongrass meadows occur abundantly in the valley bottoms and on mountain slopes up to about 3600 feet. The cottongrass Eriophorum ~ inatum spissum is the dominant species, forming tussocks 6-12 inches high and equally wide. The tussocks are separated by mossy channels a few inches wide which during summer usually contain standing water. Other species scattered through the relatively closed stands include severa! grasses, sedges, small shrubs and herbs. Sheep frequently graze this community type near the margins of more rugged slopes during both summer and winter. 1 Plant names are from Spetzman (1959).

21 11 Dry upland meadows are found on well-drained coarse mineral soil from about 2500 to 4000 feet in elevation. Vegetation is usually only a few inches high and somewhat sparse, and it may vary considerably in species composition from one site to another. Dryas octopetala and lichens, however, normally are common to all sites. The associated species of grasses, sedges, shrubs, herbs and mosses and their elevational position on mountain slopes make this community type an important one to Dall sheep. Plant communities among rock outcrops and on talus slopes are perhaps the most important vegetation in Dall sheep habitat, since these provide available forage near escape terrain on winter range and are abundant throughout the higher summer range. Vegetation is found in these sites up to 6000 feet in elevation, although it may become quite sparsè above 4000 feet. Even at lower levels plants are scattered on shallow soil between areas of exposed rock. Species composition may vary considerably, depending on rock type, but generally consists of a mixture of scattered saxifrages, ferns, grasses, dwarf herbs and mat shrubs. Vertebrate Fauna The study area is inhabited by a variety of vertebrate wildlife species including fish, birds and mammals. A portion of these are resident to the area throughout the year, and others only pass through during the course of their seasonal movements. Sorne of the mammalian species affect Dall sheep directly as predators, and this relationship will be discussed in a later section. Many of the nonpredators also play a role in sheep ecology, however, as competitors for food. Most of the bird species occurring on the study area visit for only

22 12 a short period during the summer months. As Murie (1944) noted, however, sorne of these, such as the snow hunting (Plectrophenax nivalis), may cornpete with Dall sheep for food. A total of 72 bird species were identified in in the Atigun and Sagavanirktok valleys by Sage (ILS. Dept. of the Interior, 1971), and 52 identified species plus several unidentified species were observed by D. Allen and myself in Listed in the Appendix are the scientific and coddilon names of mammals identified on the study area and those few birds which remain there the year round.

23 STUDY METHODS The initial studies of Dall sheep in the Atigun Canyon area were conducted by Ron Andersen during the summer of 1970 under the project title "Effects of Human Disturbance on Dall Sheep." This project was continued by Ray Priee and assistant James Male through the 1971 field season. During both years, valuable baseline information was collected on population parameters and movement patterns in addition to observations of human disturbance. Their work allowed me to anticipate what to expect when I began my study and has made analysis of population ecology over a four year period possible. Field observations for the present study were conducted during March 28-April 2, 1973; May 23-September S, 1973; April 9-12, 1974; and June 3-13, A total of 127 days were spent observing sheep on the study area. During the May 23-September S, 1973 period, field work was conducted with the aid of assistant Dave Allen from a base camp established near the center of Atigun Canyon. Backpacking trips of up to four days were frequently made from this camp. The remaining periods of investigation were conducted solely by backpacking. Three basic techniques were used in making observations. Reconnaissance flights were conducted with fixed-wing aircraft at irregular intervals to determine population size, distribution and movements. Ground surveys were made during ail periods of investigation and were conducted on an approximately weekly basis during the May 23-September S, 13

24 period for more frequent and detailed information on population characteristics and movement patterns. During these surveys the size, composition, activity and elevation of ali sheep bands observed were recorded on prepared forms, and the location of ali bands was keyed to topographie maps. Sheep were classified according to age as lambs, yearlings, and adults. Only adults were classified according to sex as no lambs and only a few yearlings could be sexed at a distance. Adult rams were further classified by horn size to the nearest one-quarter curl using the technique described by Murphy (1974:18). Daily observation of the population was used to gain an understanding of various other aspects of sheep behavior and ecology which were useful in interpreting population ecology and movement information. A 15-60X spotting scope and 7x35 binoculars were used as an aid in ali ground surveys and daily observations.

25 POPULATION DYNAMICS There are four major sheep wintering areas within the study area (see Fig. 2). Sorne of these areas are used primarily by ewes and young sheep, others primarily by rams. The sheep within each of these areas are isolated from other sheep groups for several months of the year. Mixing of these groups occurs, however, during the rut when rams visit the ewe wintering areas and during the summer when sheep congregate at mineral licks and then disperse over summer range. Since all sheep found on the study area could potentially interbreed and come into contact on summer range, they should all be considered as one population. It should also be noted, however, that sorne sheep move out of and into the study area during summer and probably during the rut, and therefore, the study population may be considered a subunit of a much larger population. Size and Structure of the Population The analysis of the size and structure of the study population is facilitated by considering separately each wintering area, in view of the fact that these sheep are concentrated on winter range for a large portion of the year. Area 1, Atigun Canyon, (Fig. 2) is the major winter range unit located within the study area bath in terms of size and the number of sheep present. Table 1 lists the results of population composition counts made in this area during June of It is apparent that Atigun Canyon is used primarily by ewes and young sheep, with legally huntable rams (3/4-curl or larger) averaging only 5.2% of 15

26 r 16 Figure 2. Dall sheep winter range and lambing areas (map from U. S. Geological Survey, Philip Smith Mountains Quadrangle, 1956). Legend: approximate winter range boundary areas receiving occasional winter use lambing area

27

28 ~ Table 1. June sheep classification counts in Atigun Canyon (18-20 June) (8-9 June) (13 June) (8-9 June) No. % No. % No. % No. % Lambs Yearlings 19* Ewes 121* /4-curl rams /2-curl rams /4-curl rams , Full curl rams Total rams Unidentified Total *Estimated figure: count included 91 ewes, 14 yearlings and 35 unclassified ewes and yearlings, hence (14/105)X35 5 sheep were added to the yearling count and (91/105)X35=30 sheep were added to the ewe count.,_. 00

29 19 the total. Also apparent is a continuous decline in the adult population, exclusive of lambs and yearlings, since The number of adults counted was 165 in 1970, 159 in 1971, 141 in 1973, and 122 in For this four-year period, an average decrease of 10.8 adult sheep occurred each year. Considering this decline, the percentage of ewes in the population has remained remarkably stable since 1970, varying around 48%. Age structure changes among the rams in Atigun Canyon have also occurred during this period. Table 2 shows each ram class as the percent of total rams seen on Area 1 each year. A slight shift from the younger age classes (1/4- and 1/2-curl) to the older age classes (3/4 and full curl) occurred between the June, 1970 and June, 1971 composition surveys, a large increase in the younger age classes occurred between the June, 1971 and June, 1973 surveys, and another shift towards the older age classes occurred between the June, 1973 and June, 1974 surveys. Such age structure changes are probably due to large variations in initial cohort size from year to year and the effect of their carryover into the older age classes. The production in 1970 illustrates such effects. A large lamb crop was produced in 1970 with a lamb:ewe ratio of 54:100. By 1973 these sheep were three years old and most of the rams were small 1/2-curls, producing the increase in the younger age classes in The shift back towards the older age classes in 1974 was due to smaller lamb crops in 1971 and 1972 and probably also to the movement of sorne young rams out of the area to join ram bands at other locations. Age structure changes are more difficult to detect among ewes than among rams. However, the sex ratio in the genus Ovis is near unity at birth (Geist, 197lb), and the large fluctuations in the size of the lamb crop which have been observed probably also have resulted in changes in

30 20 Table 2. Percent of rams represented by each horn curl class during four years. Ram Year Classification /4-curl 43% 39% 47% 23% 1/2-curl /4-curl full curl Total

31 21 in the ewe structure. Age structure changes may also be influenced by differentiai survivorship and mortality rates among cohorts w~ich, as noted by Murphy (1974), may occur at any age throughout life as a result of annually variable weather conditions, changes in food availability, disease, predation, intense social interaction, or any other factors which affect the energy intake and expenditure of cohorts. A further trend evident from the annual composition counts in Atigun Canyon is a changing sex ratio. During 1970, 1971, and 1974, the ratio of adult males to adult females (two years and older) was 36, 38, and 39 rams per 100 ewes respectively. In 1973, however, the proportion of rams to ewes was about doubled (74:100). The large number of rams that year, as mentioned before, resulted from the large lamb crop in However, one would have expected an equally large increase in the number of ewes in 1973, which would have stabilized the ram:ewe ratio somewhat. Since there was actually a decrease in the proportion of ewes seen in 1973, a lower birth rate and/or subsequent survival rate for ewes than for rams is indicated between 1970 and The Atigun sheep population reached its greatest size and density in 1970 and thereafter began to decline. Female lambs characteristically show lower survival rates in declining populations than do rams (Geist, 197lb). Adverse forage conditions, which may be associated with declining populations, may also result in poorer intrauterine growth for females than for males (Robinson et al., 1961), and presumably could affect sex ratios at birth. Sex ratios on ewe wintering areas are not only a product of different natality and mortality rates for ewes and rams, but are also influenced by the social structure of mountain sheep. The large increase in the proportion of rams to ewes from 1973 to 1974 is a case in point.

32 22 Geist (197lb) has shown that young rams remain attached to ewe bands until 2-4 years of age at which time they join ram bands and establish permanent home range patterns. Many of the young rams which wintered in Atigun Canyon in 1973 did not return in 1974 but probably remained with the ram bands on Areas 2 and 3 (see Fig. 2). Population counts taken in Area 2 in 1973 and 1974 show evidence of this. Twenty-one rams were counted there in June, 1973, but four or five mature rams were killed from this area in August by hunters working out of a nearby guide's camp (D. Keyes, Bureau of Land Management, pers. comm.). In June, 1974, however, 21 rams were counted there again and the average age (as indicated by horn curl) had decreased somewhat. The above discussion of spring sex ratios in Atigun Canyon was not intended to present a picture of the actual sex ratio of the study population but was given only to illustrate the changes in sex and age composition which have occurred on this one wintering area since Geist (197lb:288) noted that "sex ratios can be considered valid only if taken within two weeks prior to the ewes entering estrus," and, I might add, if taken over a sufficiently large area to include both ewe and ram home ranges. The entire study area is probably large enough to give an accurate indication of the true sex ratio from composition counts made in the spring as both ram and ewe ranges are included. This information is available from April and early-june surveys of wintering areas 2-4, the results of which are shown in Table 3. Area 2 is an area used almost exclusively by rams. Only one and two ewes were counted there in 1973 and 1974 respectively, while the total estimated number of sheep using this range is 25. Areas 3 and 4 are also used mainly by rams but contain a few ewes as weil. A total of 30 sheep is estimated to winter

33 Table 3. Numbers of sheep counted on wintering areas 2-4 in Area and Date Classification 2: (June 7) 3: (June 5) 4: (April 10) Lambs Yearlings Ewes /4-curl rams /2-curl rams /4-curl rams Full curl and larger rams Total rams Total Total N...

34 24 on these areas. Over the entire study area, an actual 228 sheep were counted in the spring of Of the 167 sheep which were two years of age or older, 72 were rarns and 95 were ewes, giving a sex ratio of 76 rarns per 100 ewes. This figure is intermediate in comparison to sex ratios given in the literature for nonexploited and lightly exploited mountain sheep populations. An estimate of the total study population size can be calculated from the June, 1974 composition counts in wintering Areas 1-4. A total of 182 sheep were counted on Area 1, which is probably close to the actual number present. Combined with the 55 total sheep estimated on Areas 2-4, a minimum of 237 sheep were present. By early June, however, sorne sheep may have already begun to wander from the wintering areas. It is also likely that sorne sheep spent the winter of at locations other than wintering areas 1-4 since snow accumulation was unusually light (see discussion of winter range in Seasonal Movements section). With these considerations, a fair estimate of the total size of the wintering population on the entire study area in 1974 would seem to be approximately 275 animais. Productivity and Lamb Survival Table 4 summarizes lamb production and survival in Atigun Canyon for the four years in which classification counts were made. It is evident that productivity fluctuates between wide extremes and, therefore, results in large differences in initial cohort size. These fluctuations might occur as a result of annually variable pregnancy rates, re~orption of fetuses, stillbirths, or neonatal mortality. In the latter case it must be remembered that June classification counts represent

35 ,...---,----~- -"-~ ~ - Table 4, Productivity and lamb survival in Atigun Canyon during four years. Year No. lambs % survival No. yearlings ~ 18~ No. ewes Lam:ewe ratio 54:100 20:100 10:100 59:100 Yearling: ewe ratio 16:100 39:100 22:100 9:100 Average :100 22:100 N "'

36 26 only apparent natality since many lambs probably die within a few days after birth and are never counted. The number of ewes which become pregnant in a mountain sheep population depends on at least two things, range quality and the age structure of the population. The age at which ewes reach sepual maturity is closely related to their nutritional state. Under conditions of average range quality, most females probably breed first at 2-1/2 years of age, but on exceptionally good range sexual maturity may be achieved at 1-1/2 years (Geist, 197lb; Streeter, 1970). Hence, on high quality ranges the number of potentially reproductive females may be higher than on ranges of lower quality, even if the actual number of ewes is the same. Also, it is known from domestic sheep that ewes achieve their highest reproductive output during middle age. The highest conception rates, highest number of lambs born per ewe, and highest survival of lambs occurs among ewes 5-7 years of age (Turner and Dolling, 1965). On the other hand, reproductive performance in female ibex (Capra ibex), a close relative of mountain sheep, has been shawn to decline after about 10 years of age, while young ibex and domestic sheep ewes lose or desert a higher percentage of their lambs than older females (Nievergelt, 1966; Alexander, 1960). Populations of middle-aged sheep, therefore, probably demonstrate the greatest production, while reproductive output in populations with large young- and old-aged segments would likely be somewhat lower. In applying this information to the study population, one might speculate that the large lamb crop of 1970 was a product of a healthy mi"ddle-aged ewe population. In 1973, however, these ewes had passed their reproductive prime and the large crop of lambs they produced in

37 had not yet reached their highest reproductive potential. Consequently, production in 1973 was much reduced. An excellent lamb çrop was produced in 1974, but it is doubtful if one year's increase in the age of the ewes born in 1970 could have made such a large improvement in the reproductive performance of the population. It seems likely, therefore, that other factors have also contributed to the wide fluetuations in productivity. Geist (197lb) compared the number of bighorn ewes seen from a fire lookout in Banff National Park during eleven years to the lamb:ewe ratios in the summers following and concluded that productivity is inversely related to population density (r=-0.75, t=3.4, p<:o.ol). As further evidence of this, Murphy (1974) showed that the lamb:ewe ratio in the area of Mt. McKinley National Park where population density was highest was significantly lower (X 2 =6.43, p<:o.ol) than for the park as a whole. Finally, Woodgerd (1964) demonstrated that among the Wildhorse Island population of bighorns, productivity dropped as the population expanded and density increased to a stable level. Composition counts from Atigun Canyon substantiate the inverse relationship of population density and productivity. The peak population level, 250 sheep in 1970, was followed by a low lamb:ewe ratio (20:100) in 1971 while the smallest population count, 167 sheep in 1973, was followed by a high lamb:ewe ratio (59:100) in Weather during both gestation and the lambing season also has profound influences on lambing success. Murphy (1974) showed a significant negative correlation (y= lx, r=-0.81, F(l,lO)=l7.6, p<o.oi) between snowfall during winter and the lamb:ewe ratio in the summer following for Dall sheep in Mt. McKinley National Park. Good

38 l 28 nutrition is important to "every phase of the reproductive process from conception to parturition and post-partum vitality," (Pitzman, 1970:37). However, during winters of deep snow sheep are restricted to small portions of their habitat, which increases population density and reduces the amount of forage available per sheep. Pawing for food through deep snow also increases the energy expenditure for sheep. Normal winter weight losses of 15% and 13% have been reported for Dall sheep from the Alaska Range and the Kenai Peninsula in Alaska respectively (Heimer, 1973). Among reindeer, weight losses of 17-24% early in gestation have been shown to result in the frequent resorption of fetuses (Preobrazhenskii, 1961). Assuming a similar condition exists for Dall sheep, it seems almost certain that during severe winters many ewes are unable to provide enough energy for optimum fetal growth. Even if a fetus is successfully formed, it is known from experience with domestic sheep that lambs born after severe winters are smaller and less viable than those produced in mild years (Robinson et al., 1961). Small, weak lambs born to nutritionally stressed ewes are especia11y susceptible to immediate postnatal death caused by desertion, failure to initiate nursing, or hypothermia. If they survive, they are subject to several physio1- ogical and behavioral disorders which may be retained throughout 1ife (see Geist, 1971b: ). The winters of and , which according tou. S. Weather Bureau statistics appear to have been relatively mild throughout northern Alaska, were followed by large lamb crops on the study area (54 and 59 lambs: 100 ewes respectively). The poor lamb crop of 1971 (20 lambs:loo ewes) followed a severe winter; however, the virtual failure of the lamb crop in 1973 followed a winter which was not

39 29 especially severe. The most reasonable explanation for low production in 1973 is the cold, wet weather which occurred during the lambing season and probably resulted in a high rate of neonatal mortality by hypothermia. Predation may influence lamb production in either a positive or a negative manner, depending on the relative numbers of predators and prey and the size of the prey population in relation to the quality of its habitat. Moderate levels of predation may reduce competition for available food in high density sheep populations, thereby placing ewes on a higher nutritional plane and increasing productivity. A high ratio of predators to prey, however, could concentrate sheep within the more rugged portions of their winter range where vegetation is often less abundant. This would not only lower the nutritional plane of ewes and place them under.physiological stress, but if harassment were frequent enough might cause psychological stress as well. Experimentally-induced stress led to neurosis in domestic sheep (Liddell, 1954) and produced behavioral abnormalities in the offspring of pregnant female rats (Thompson, 1957). Nutritional and psychological stress ind~ced by predators might, therefore, contribute to the loss of embryos or the production of smaller, less viable, or behaviorally affected lambs. Predation pressure may vary annually as the numbers of predators and prey change, as alternate prey species become available and, because predation is greater during winters of heavy snowfall (Murphy, 1974), as the weather fluctuates. Evidence from Dall sheep studies in Alaska indicates that lambs experience little mortality for the remainder of the summer after they reach about a week of age (Murie, 1944; Murphy, 1974).. During their

40 30 first winter, however, they frequently suffer high mortality, usually at a greater rate than other age classes (Murie, 1944; Murphy, 1974; Nichols, 1973). Lambs born in Atigun Canyon probably follow a similar pattern, although mortality data for this age class are sparse. Remains of only six lambs were found on the study area in 1970 and 1971, and no lamb carcasses were found in 1973 or The small number of lamb carcasses found is not surprising in view of their high perishability and the rugged areas in which lambing occurs. Low summer mortality among lambs in this area is suggested by the 1973 lamb crop. Eight lambs were counted on the lambing areas in Jqne and at!east seven of these were present when the sheep returned to Atigun Canyon in late August. Information on survival through the first winter is available for the 1970" and 1973 cohorts (see Table 4). Of 66 lambs counted in 1970, 42, or 64%, survived to be counted as yearlings in Of the eight lambs counted in 1973, all survived until the 1974 count. Survival rates of 64% and 100% are probably not typical, however, but were more likely the result of favorable circumstances in those years. Conditions which produced a large lamb crop in 1970 may also have favored good survival of those lambs. For instance, Murphy (1974) found a strong negative correlation between snowfall in the year before birth and survival to yearling age among sheep in Mt. McKinley National Park during recent years. Wolf densities were also apparently lower on the study area in 1970 than more recently. The eight lambs which were counted in June, 1973 probably represènted the largest and strongest of their cohort since high neonatal!osses seem to have occurred that year. Snowfall was exceptionally light

41 31 during the winter of , and because the sheep population was at its lowest level since at least 1970, more food was apparently avaifable per sheep. In addition, wolf densities apparently dropped somewhat from a high the previous summer and caribou were present in the area most of the winter and were likely the focus of wolf predation. The combination of all these factors suggests highly favorable survival conditions for the 1973 lamb crop. Survival and Mortality of Adult Sheep From this and other studies it is apparent that patterns of survivorship and mortality among adult mountain sheep almost always follow a similar form. Sheep which live past yearling age show high survival during the early years of life and then accelerated mortality beginning about midway through the maximum life expectancy. This pattern may vary slightly between populations or between cohorts within the same population, but the general characteristics remain the same for both rams and ewes in all populations. Geist (197lb:294) has noted three means for estimating the agespecifie survival and mortality of adult mountain sheep. These are: (1) aging skeletal remains found in the field by the horn annuli technique (Geist, 1966); (2) comparing the size of successive age (or horn size) classes observed in the living population; and (3) observing the return of individually identifiable sheep to a given seasonal home range in successive years. Since no sheep in the study population were marked and few could be recognized individually, the first two methods must be used in this analysis. A total of 86 sheep remains were found on the study area between

42 r and A breakdowp of the number found for each sex and horn curl class is shown in Table S. As the majority of remains were collected in 1973, it is believed that most of these sheep died during or after The homs on sorne skulls were quite decayed, however, indicating that a few remains may have accumulated as early as Unfortunately, mortality data compiled from remains has been shown to be subject to severa! biases. These include variation in the perishability of carcasses among the various age and sex classes and the tendency of finding relatively more of certain sex and age classes depending on whether ram or ewe wintering areas are more closely searched for remains (Geist, 197lb). Neither of these biases appears serious for the ram remains data collected on the Atigun study area, as will be shown later. Age structure changes within a population also may lead to unrealistic conclusions concerning the age-specifie mortality patterns of that population. Murphy (1974) hypothesized that large fluctuations in initial cohort size and variable conditions (e.g., weather) which influence subsequent survival lead to large differences in age-specifie mortality patterns among cohorts. These differences may be so great that remains which accumulate over a period of severa! years will not reflect the true mortality patterns of the population. Such appears to have been the case in Mt. McKinley National Park when Murie (1944) collected his sample of over 800 skulls. There have also been age structure changes and large variations in initial cohort size in the Atigun population. The remains found' during this study, however, indicate mortality patterns which are not inconsistent with those that would be ex~ected from the age structure of the living population (see p. 36 and Fig. 4).

43 ... Table S. Dall sheep remains found on the Atigun study area. Year Classification Total Lamb Yearling Ewe l/4.:curl ram /2-curl ram /4-curl ram Full curl ram Full curl plus ram Large ram, class unknown Adult, class unknown Unidentifiable Total "'

44 34 Of the ram remains found, 24 had homs in good enough condition to allow confident aging by horn annuli while 32 could be classed according to size to the nearest one-quarter curl. Fig. 3 shows the age distribution of the 24 aged ram remains. The mean age at death was 9.00 years with a standard deviation of 2.67 years and a standard error of the mean of 0.54 years. The minimum life expectancy of adult rams may be defined as the age at which 95% of those surviving to yearling age are still alive while the maximum life expectancy is the age at which 95% are dead (Geist, 197lb). For this population, the calculated minimum and maximum life expectancies are 3.5 years and 14.5 years respectively (for details of calculations see Geist, 197lb:294). These figures are close to the ages of the youngest and oldest remains found, which were four and thirteen years of age respectively. For purposes of analysis of the data from the ram remains, the 24 known-aged rams can be considered as members of the same cohort. Average mortality for any age interval may then be calculated as the sum of the number found dead each year in that interval over the sum of the number alive in each previous year (Geist, 197lb:295). This method gives an average mortality between the ages of two and eight years of 5.2% and between the ages of nine and thirteen years of 38.1%, clearly illustrating the two-phase survivorship pattern noted earlier. For all rams greater than two years of age, the average annual mortality rate is 12.9\. A second synthetic cohort may be constructed from the 32 ram remains classed to the nearest quarter curl. Such a cohort, shown in Table 6, contains 41\ 1/4-curl rams, 34\ 1/2-curl rams, 21% 3/4-curl rams, and 4\ ful1 curl rams. These percentages were compared to the percentage composition of living rams recorded in classified counts of the study area

45 'E u tio ~ - """" - ~ l""" - - r Il Age at De ath (Yeora) Figure 3. Age distribution of Dall sheep ram remains found on the Atigun study area

46 36 during and the two were found to be very similar (Table 6 and Fig. 4). The 1/2- and full curl classes are slightly over-represented in the remains data while the 1/4- and 3/4-curl classes show slight under-representation. These differences in percentage composition may result from one or more of the biases suggested by Geist (197lb) and Murphy (1974), but since the discrepancies are of small magnitude, these biases apparently are not large in the Atigun carcass sample. The causal faètdrs behind the two phase mortality pattern exhibited by rams greater than two years of age have been described by Geist (197lb :295): The ages of low mortality in rams coïncide with their low dominance status and near exclusion from breeding by largerhorned, older rams. Conversely, when rams reach near ultimate body and horn size, and become dominant breeding rams during rut, their mortality increases. Large, dominant rams, notes Geist, return from the rut in an exhausted condition because they '~irtually do not feed while guarding ewes and.they.fight extensively and do mut:h running and chasing when following the estrous ewe and discouraging competitors." Having lost most of their fat deposits, these dominant. rams are vulnerable to mortality during the severe winter months following the rut. ' Rams which are on high quality range and thus grow fast and achieve dominance status relatively early in life will be subjected to the physiological stresses imposed by the rut at an earlier age than slowgrowing, late maturing rams on poor range. One would expect, then, that on the average, rams reaching sexual maturity on good range would have a shorter life span than those on range of lower quality. This principle has peen demonstrated for both mountain sheep and European ibex by cornparing longevity in introduced, expanding populations (where forage

47 r- A Table 6. Construction of a synthetic cohort from rams found dead in each horn curl class. Horn Curl Class 1/4 1/2 3/4 Full Full+ Total No. remains found Reconstructed cohort No % % observed in population "' "

48 Observed Population -- Reconstructed Cohort -c u ~ n. 1/4 1/2 3/4 4/4 4/4+ Horn Curl C lass Figure 4. The percentage of rams in each horn curl class in the living population compared to a reconstructed cohort.

49 39 conditions are good and individual growth rates are expected to be high) to longevity in stable or declining populations on poorer range (whe~e individual growth rates should be somewhat lower; see Geist, 197lb:296). There is also considerable difference in growth rates among rams from the same population, however, and evidence from the Atigun study population indicates that the inverse relationship of growth rate and longevity is equally true for these rams. Individual horn segments were measured for twelve of the ram remains found on the study area and the total length of horn segments 3-5 (produced between the ages of 1-1/2 and 4-1/2 years) was used as an indication of growth rate. There are several reasons for choosing these segments. The first and second segments (lamb and yearling) are almost always partially worn away and thus would not give an accurate indication of growth. The third through fifth segments are usually the largest and most easily discernible and would, therefore, give the most precise index of growth. Since these are the years of fastest growth, any subtle differences in growth rates would show up here best. After age four, the fastest growing rams may begin taking part in the rut, which could affect their horn growth and complicate any comparison based on later segments. By regression analysis a negative correlation was found between the total length of horn segments 3-5 in millimeters and age at death in years (y= x, r=-0.693). This relationship is illustrated in Figure S. Within the same population, therefore, there exist faster growing individuals of relatively short life span and slower growing individuals of relatively long life span. Since fast growers are known to predominate in expanding (i.e., high quality) populations and slow

50 40 ~ E s 10 1,., c E 11: U) c ~ 0 :z:... 0 z:.. 11:11 c.j ~ '350 y= x r= Age at Death (Years) Figure S. Relationship of horn growth rate ta age at death.

51 41 growers are known to predominate in stable or declining (i.e., lower quality) populations, the existence of both types within the same population may be due simply to individual nutritional variation or possibly occurs as a genetic adaptation evolved to cope with a periodically changing environment (e.g., fluctuations in weather, density, predation, disease) which sometimes produces sudden changes in population quality. Mortality and survival data for ewes in the Atigun population are sparse, however, better information is available from other mountain sheep populations. Large carcass samples of Dall sheep ewes collected in Mt. McKinley National Park (Murie, 1944) and Nelson's bighorn (Ovis canadensis nelsoni) collected on the Desert National Wildlife Range (Hansen, 1967) indicate a two phase survivorship pattern for ewes similar to that exhibited by rams. As with rams, mortality rates of ewes seem to be tied closely to the energy expended on reproduction. Young ewes probably can maximize lifetime reproductive output by channelling relatively more energy into maintenance and growth than into early reproduction (Murphy, 1974). As a consequence, survivorship of ewes in the young age classes is enhanced. It has been inferred from knowledge of domestic sheep and European ibex that reproductive output for mountain sheep ewes probably peaks at about six years of age (Geist, 197lb). Middle-aged ewes should, therefore, channel all available energy into reproduction if maximum lifetime reproductive output is to be achieved (Murphy, 1974). As age continues to increase, however, recovery from the physiological stresses of reproduction becomes more difficult each year. Consequently, one would expect increasing mortality rates and/or decreasing productivity in the older age classes. The available evidence suggests that both occur after about age 8-10.

52 42 Mortality data from both the Nelson's bighorn and Dall sheep populations noted previously indicates that age-specifie mortality rates begin to increase when ewes first reproduce, increase at a greater rate through middle age, and achieve their highest rate of increase bètween 9 and 10 years of age (for life tables of these two populations see Bradley and Baker, 1967). The reproductive output of female ibex was shown by Nievergelt (1966) to decline after about 10 years of age. If a similar decline takes place for mountain sheep, then it seems certain that the mortality rate increases and reproductive performance wanes as ewes enter the older age classes. Survivorship may actually be increased as a consequence of decreased reproductive success if barren females or those which have lost their young have a better chance of survival than those which are pregnant or lactating (Geist, 197lb). On the other hand, the available evidence for caribou and muskox (~moschatus) indicates that pregnancy may have certain hormonal and behavioral effects which increase chances for winter survival (P. Lent, Alaska Cooperative Wildlife Research Unit, pers. comm.). While the physiological stresses incurred as a result of reproductive effort probably contribute to increased mortality rates among adult mountain sheep of both sexes, these stresses are usually not the ultimate causes of death. Severa! direct causes of mortality and their importance to the study population will be discussed in the following paragraphs. Accidents Accidents occur in ali animal populations and populations of mountain sheep are no exception. Although the percentage of the population affected

53 L 43 is usually thought to be low, nearly every published report concerned with the life history and population dynamics of mountain sheep mentions the occurrence of accidentai injury or death. While accidents reported from different localities have been attributed to such events as poisoning, puncture wounds, drowning, hanging, collision with vehicles, avalanches, and freezing in overflow ice OMUrie, 1944; Smith, 1954; Spalding, 1966; Geist, 197lb), by far the greatest number of accidents appears to involve falls. Lambs are particularly vulnerable to falls during the first month of life because, although they may be more agile than older sheep, they lack experience in traversing the precipitous terrain. Among adult sheep, most falls probably occur during the rut when aggressive behavior and chasing are common. Also, males travel extensively between bands at this time when ice and snow reduce footing to its poorest. Evaluation of the importance of accidents as a mortality factor from remains found in the field is difficult since accidentai death is easily confused with death from other sources. Sorne falls, for instance, may be caused by weakness or dizziness resulting from pathological conditions (Smith, 1954). Sick sheep also tend to bed at the base of a cliff where the cause of death may be interpreted as a fall (Welles and Welles, 1961). Other sheep which do die from accidents may later be fed upon by predators and the results interpreted as death from predation. Only three instances of accidentai injury or death have been observed in the Atigun area, all apparently involving falls. Andersen (1971) discovered a dead ewe on July 3, 1970 lying on a steep grassy slope with its head turned back in an abnormal position. He concluded it had probably fallen while running and had broken its neck. Andersen also saw a lamb

54 44 become injured in 1970 while fleeing from a helicopter (pers. comm. to D. Klein, Alaska Cooperative Wildlife Research Unit). On June 27, 1973 D. Allen observed a lamb with a bad limp in one rear leg which was apparently sustained in a fall. It was seen only once thereafter and probably either died or recovered from its injury and became indistinguishable from other lambs. Disease and Parasites Although a great deal of information has been compiled on the diseases and parasites of mountain sheep, most of it pertains to Ovis ~ densis. The pathology of Ovis dalli has, by comparison, received very little study. Severa! diseases have been detected in Aiaskan Dall sheep (Ericson and Neiland, 1973), but no reports of diseased sheep have come from the Brooks Range. Rausch (1951) found no sign of disease or helminth parasites among six sheep examined from Anaktuvuk Pass region and no instances of gross pathological conditions were observed during the present study. The Atigun population thus appears to be in generally good health with the exception of a few minor anomalies. The most common anomaly noted was broken homs. This was observed among severa! ewes and rams but most homs break above the terminus of the horn core and do not cause further complications. One ewe was also seen with a horn which grew almost lateral! y from the top of her head. The left jaw of a mature ewe was found with the third premolar missing. The tooth may never have been present or was lost at an early age as the cavity in which it should have fitted was not fully developed and the first two premolars were slanted backwards leaving no gap between them and the first molar. The teeth and jaw otherwise appeared in good

55 condition. A similar condition, but involving the second premolar, has been reported in a significant proportion of Ovis canadensis nelson~ and is more prevalent in ewes than in rams (Bradley and Allred, 1966). Hemming (1969) also found the second premolar missing in 5 of 63 Dall sheep examined. It should be noted here that examination of many sheep jaws from the Atigun area has revealed no trace of bene necrosis which is common among sheep in certain areas of Alaska. One of the eight ewes which was followed by a lamb in 1973 still retained most of her winter coat after all ether sheep had nearly completed molting. Her delay in shedding could possibly have been due to pathological conditions. Hunting The number of sheep hunters in Alaska and the harvest of Dall sheep rams have risen annually for several years. This increase in hunting pressure has prompted many hunters to search for less accessible sheep ranges where greater numbers of harvestable rams remain. The Brooks Range has traditiona1ly been one of these areas of 1ow accessibi1ity, and during past years the sheep hunting season has begun ear1ier there and bag limits have at certain times been greater than in the rest of the state. Probably as a result of this, hunting pressure has increased at a greater rate in the Brooks Range recent1y than over the state as a whole. The Atigun study area lies in Game Management Unit 26 where the harvest of 3/4-cur1 or Iarger rams has risen from 47 in 1970 to 149 in 1973, a 217% increase. The exact kil1 of rams from the study area cannet be determined, but can be estimated from harvest tickets returned by hunters to the Alaska Department of Fish and Game. The reported numbers of sheep killed during in subunits 2606 (Atigun River drainage) and

56 (Sagavanirktok, Ribdon, and Lupine River drainages) are shown in Table 7. From these statistics, an annual harvest since 1970 of S-8 sheep from the study area seems probable. With the beginning of major construction activities for the Trans Alaska pipeline in sight, the Commissioner of Fish and Game, at the request of the Alaska State Legislature, issued an emergency order on February 21, 1974 closing all big game hunting for five miles on each side of the pipeline corridor north of the Yukon River. This closure is aimed at avoiding undue impact on big game species during pipeline construction and will continue indefinitely. Approximately one-half of the Atigun study area has been affected, hence the sheep harvest from this area will probably be sharply reduced. Most hunters have in the past entered the study area from the Atigun valley where access by float plane is good. Hunting will still be permitted in the eastern portion of this area, but access in the upper Sagavanirktok drainage is limited to two lakes of marginal size for float planes. Predation Potential Dall sheep predators in the Brooks Range include the golden eagle, raven, wolverine, grizzly bear, red fox and wolf. Of these, the eagle, raven and fox, because of their small size, would be largely restricted to attacking young lambs. The lambing season and several months following, however, is a time of food abundance for these three predators, with ground squirrels, other rodents, and carrion making up the bulk of their diet. Because of the relative difficulty of finding and. killing lambs as compared to finding other food items, predation by these species must be very insignificant. There have been documented

57 .loo.. 47 Table 7. Harvest of Dall sheep rams from the study area and vicinity. Year Subunit Refers to the Atigun drainage. 2 Refers to the Sagavanirktok, Ribdon, and Lupine drainages

58 48 cases of ali three species attacking ungulates, but in no instance have they become an important source of mortality. The wolverine is a scavenger-predator weil known for its ferocity. While there are no documented accounts of wolverine attacking mountain sheep, Guiguet (1951) observed an unsuccessful attack on an adult mountain goat. Considering the low population density characteristic of the species, however, it seems unlikely that wolverine could be an important predator of Dall sheep. Only one wolverine sighting occurred during the present study. Grizzly bears are the largest carnivores in the area, and if they could catch sheep they could easily kill them. Healthy sheep of all ages, however, can probably easily avoid bears. The behavior of both species reflects this and they usually act with indifference towards each ether (Sheldon, 1930; Pitzman, 1970). At!east nine, grizzlies frequented the study area during the present study but no grizzly-sheep interactions were noted. The wolf, in contrast to ether predators, is capable of having major influence on sheep populations. predators is weil documented. The effectiveness of wolves as sheep In Arctic Alaska they usually hunt in packs averaging about 4-6 individuals (Stephenson, 1974). but single wolves have been observed to kil! seemingly healthy shee~ in their normal habitat with apparent ease (Heimer, 1972). Between 1970 and 1973, wolf numbers on the Atigun study area appear to have increased dramatically, probably as a result of the prohibition of aerial wolf hunting in Wolves were seen only once each by R. Andersen in 1970 and R. Priee in 1971, but were observed frequently in During the summer of 1973, two wolf dens in the vicinity of the

59 49 study area were occupied. One was near the west bank of the Atigun River two miles south of Galbraith Lake and contained five pups. The other, on the east bank of the Sagavanirktok River four miles downstream from Atigun Canyon, contained the litters of two females with a total of 15 pups (T. Hill, Galbraith Lake camp, pers. comm.). In addition to these, six other wolves which were identifiable by their coloration were observed several times. The local wolf population at this time, therefore, numbered at least 31 (20 pups + 3 adult females + 2 adult males + 6 other wolves). The diet of wolves seems to be governed largely by the availability of their various prey species. In the Atigun area during winter, sheep are apparently taken often, moose occasionally, and caribou when available. During summer, ground squirrels and other small mammals are abundant and easier to catch than larger species and thus make up a major portion of the diet. Wolves were frequently sighted capturing ground squirrels during the summer of 1973, but no attempts at killing large prey were seen. This same pattern of seasonality in the diet was observed by Murie (1944) in Mt. McKinley National Park. The wolves studied by Kelsall (1968) in northern Canada traveled with the caribou herds in winter, probably because this was their only major food source. On the Atigun study area, caribou are present sorne winters and absent others. Wolves, however, remain in the area each winter, although their numbers may not be as high in winter as in summer. Their year-round existence there probably depends on the availability of alternate prey species when caribou are absent. Sheep and a few moose are always present on the study area and constitute acceptable alternate prey species. l

60 50 Since moose densities north of the Brooks Range are comparatively low (about 10 winter on the study area), fewer are killed by wolves than in areas where moose densities are higher. The fresh remains of a cow and calf were found in Atigun Canyon during April, 1973 and the very old remains of two others were found later that summer. Sheep remains, while not ali wolf kills, were ~ch aore abundant; a total of 86 were found between 1970 and The ultiaate cause of death of an animal is almost impossible to establish with certainty when only its skull and homs, a few bones, or hair are found. It may have been killed by predators or it may equally well have died from other causes before being eaten by predators. Severa! points suggest, however, that wolves kill a high percentage of the sheep which die on the study area. These include the apparent low level of disease in the population, the generally good condition of the dentition in skulls found, the high level of the wolf population, and the scarcity of other major prey species during most winters. A high level of wolf predation on sheep in the Atigun area is further suggested by the location at which the sheep remains were found. Most remains were discovered around the periphery of winter range. More precisely, most sheep die or are killed near mountain bases on relatively gentle slopes away from escape terrain. Escape terrain is an essential element of sheep habitat, but sheep frequently move some distance from the cliffs and rock outcroppings during winter as grazing may be better around the margins of the steeper slopes (see Fig. 6). lt is here that sheep are most easily caught when surprised by wolves. If wolf predation were not a major cause of mortality among these sheep, most remains would probably be found in more rugged terrain.

61 51 The wolf population seemed markedly lower during field trips to the study area in April and June, 1974 than during the previous sudder. It is not known, however, whether this was a reflection of aortality or of movement out of the area. Several instances of rabies were discovered among the local fox population by construction personnel during the spring and some wolf mortality from the disease aay also have occurred. Increased construction activities associated with the Trans-Alaska pipeline, however. may equally well have caused sœe of the wolves to aove to less disturbed areas. In any case. predation on sheep seeaed less during the winter of than during the previous winter. Only three fresh sheep remains were found during the two weeks spent on the study area in 1974 while eight were found during a coaparable period in 1973.

62 52 Figure 6. Sheep feeding on low slopes in the spring. Most wolf predation occurs in such locations. Figure 7. Winter range directly across Galbraith Lake from the Alyeska pipeline construction camp. Note the scarcity of snow except in draws and ~epressions (April 2, 1973).

63 .s;; r Figure 8. l'l'inter range us.ed by rruns. in a small drainage into the Atigun River (Area 2) six miles south of Galbraith Lake. (April 11, 1974) Figure 9. The main lrunbing area in the northeastern portion of Atigun Canyon.

64 54 Figure 10. Lick A, the eroding bluff just a.bove the stream bed, is the most heavily used mineral lick on the study area. Figure 11. Lick B, also a stream-eroded bluff, is a primary lick ne ar the eastern end of Atigun Canyon.

65 SEASONAL MOVEMENTS Mountain sheep in northern climates generally do not remain in one location throughout the year but seasonally move between various segments of their habitat, thus optimizing their chances for survival and reproduction. Geist (197Ib) has shown that in some localities, particularly the bighorn and Stone sheep ranges of Canada, rams may have as many as seven seasonal home ranges and ewes may have up to four, while a minority of sheep have only two home ranges, summer and winter. In the less diverse sheep habitat of the Brooks Range with its long winters and short summers, the home range pattern of Dall sheep approaches the simplest of these forms. Both ewes and rams have summer and winter home ranges, and rams additionally visit the ewe wintering areas during the rut. Both sexes concentrate at mineral licks during spring and summer while moving between seasonal range units, but since the mineral licks on the study area lie between winter and summer ranges, they have not been considered as separate range units. Winter Range Although snowfall on the north side of the Brooks Range is normally light, it is great enough to concentrate sheep within a small portion of their total range for severa! months of the year. The se areas of winter concentration have certain characteristics in common which permit sheep to survive the cold, dark winters. 55

66 56 First, winter range must occur in areas of light or no snow accumulation so that sheep can obtain forage with a minimum expenditure of energy. Evidence indicates that vegetation covered by a thin layer of soft snow may be preferred to exposed vegetation since the former is protected from desiccation and nutrient leaching by the snowcover (Nichols, 1973). The combination of snow depth and snow hardness impose certain limits, however, beyond which pawing for vegetation becomes energetically uneconomical. Secondly, assuming snow depth is not limiting, winter range must show an adequate growth of vegetation during spring and summer to meet the forage requirements of sheep from at least September through May. And thirdly, terrain within the winter range must be sufficiently rugged to permit escape from predators. Several factors may contribute to the existence of snow free areas in the Brooks Range during winter. Direction of exposure is important since insolation on south-facing slopes delays snow accumulation in fall and hastens melting and regrowth of vegetation in spring. Furthermore, sunlight on southern slopes in early spring significantly increases temperatures near ground level, benefiting sheep at a time when survival may be most difficult, i.e., when they have lost their fat reserves and when range conditions are at their worst. A second important factor in limiting snow accumulation is steepness of terrain, Snowfall is less per unit of ground surface on steep slopes and also may cling to these slopes less readily. This factor may be quite significant in the rugged cliff terrain in which sheep are often found and may increase the amount of available escape terrain on winter range. A final and probably the most important factor contributing to the

67 57 absence of snow on winter range is exposure to wind. The continuously cold temperatures of the area result in snow which is extremely light and easily blown about. Ridgetops and sideslopes which are repeatedly blown free of snow may provide good grazing for sheep throughout the winter months. Where drifting snow tends to collect, however, wind contributes to snow compaction, making these areas of little use to sheep. Fortunately, the factors contributing to an absence of snow often exist simultaneously at the same locations. That is, south-facing slopes on the study area generally receive the most wind in winter, and the steeper of these receive greater insolation due to a more nearly perpendicular angle of incidence at that time of year. There is, therefore, a considerable amount of exposed grourid with drifted snow prevalent in draws, depressions and canyon bottoms (see Figs. 7 and 8). Exposed ground, however, does not imply an abundance of suitable winter range because adequate vegetation is frequently lacking. This is especially true on talus and seree slopes and on ridgetops above 5,000 feet where vegetation is extremely sparse. Escape terrain is also generally abundant throughout the mountains, but the cliffs and rock outcroppings which provide the best escape cover usually contain less vegetation than the more gentle slopes. Figure 2 depicts the locations within the study area where concentrations of sheep were observed in winter. These areas also represent the major locations meeting the qualifications of limited snow accumulation, adequate vegetation, and good escape terrain. Area 1, Atigun Canyon, is the major wintering area on the study area and is used primarily by ewes and young sheep. This range presently supports a population of about 175 sheep, most of which winter on the north side of

68 58 the canyon. Area 2 is a wintering area used almost exclusively by rams and contains about 25 animais. Areas 3 and 4 are used primarily by rams but also contain ewes. The sexes generally remain separated, however, except during the rut. An estimated 30 sheep currently winter here. In addition to these areas, similar populations of sheep exist outside the study area across the Atigun and Sagavanirktok valleys. There are also patches of habitat scattered throughout the study area which are not choice winter range but which receive occasional winter use by small numbers of sheep, especially during winters of light snowfall. Three such areas where sheep have been observed in winter are indicated in Figure 2 by dashed lines. However, there are undoubtedly many other areas which receive occasional winter use. Spring Movements The coming of spring brings dramatic changes to the Brooks Range. During May, the daylight period increases by about eight minutes each day, temperatures rise above freezing for the first time in severa! months, and the snowcover begins to recede. At this time the sheep move down from the windblown ridges to feed on vegetation which is made available by melting snow. Although plants do not produce new growth for a few more weeks, the vegetation which has remained frozen under the snow all winter is of higher quality than that which has been exposed to the desiccating and nutrient-leaching effects of dry winds, greater temperature fluctuations, and intense light conditions. Near the end of May the vegetation on the low south-facing hills above the Atigun River beg~ns to show new green growth and offers choice grazing. Within a few days, however, vegetation at the higher elevations on the north side of the canyon also resumes growth and sheep move back up the slopes where

69 59 escape terrain is more abundant and where the ewes can find isolation for lambing. Published lambing dates for Dall sheep fall between May 5 and June 28 (Hemming, 1969) with the greatest frequency of births usually occurring in the last week of May. On May 20-22, 1970, R. Andersen counted 19 lambs in a ground survey of Atigun Canyon while R. Priee counted 11 lambs in a similar su~ey June 2-3, The first lambs seen in 1973 were two 1 found on May 28 which were judged to be less than 24 hours and 48 hours old. However, only eight lambs were seen during the entire summer. In 1974, lambing seems to have occurred earlier than usual with one lamb reported as early as May 7 (E. Ludlow, Bureau of Land Management, pers. comm. to D. Klein, Alaska Cooperative Wildlife Research Unit). When 1 conducted a survey of the area on June 8 the lambs appeared quite large and many had moved well away from the lambing areas. The lambing areas in Atigun Canyon are shown in Figure 2. The boundaries of these areas were established by plotting all lamb sightings through the first week of June in the years and outlining those areas which showed definite concentrations. Some movement of lambs away from lambing areas occurs before June 7, but outlining only the areas where sightings were concentrated probably minimizes the error created by these movements. It is apparent that there are two major lambing areas, both within winter range on the north side of the canyon. One of these is the southwest slope of Black Mountain directly across Galbraith Lake from the pipeline construction camp. This area consists of severa! shallow draws with numerous rock outcroppings which provide good escape cover. Sheep can be found there throughout each winter, but lambing apparently occurs on this area only during good lambing years. New lambs

70 60 were observed there during 1970 and 1974 when large lamb crops were produced, but no lambs were seen there in 1971 or 1973 which were poor lambing years. The second lambing area is located in more rugged terrain in the lower half of the canyon (Fig. 9). New lambs were observed there every year of the study and in greater numbers than on Black Mountain; thus this seems to be the most important lambing site on the study area. The area around Guard House Rock (see Fig. 1) has been reported by construction personnel as a good lambing area (Andersen, 1971), but no new lambs were recorded there in 1970, 1971, or during the present study. While a few lambs may be born at locations throughout winter range, the majority of lambing occurs on the two areas just described. Mineral Lick Use Concurrent with the lambing season, sheep begin to show increasing interest in the severa! natural mineral licks which are located on the study area. The location and a description of these licks are given in Figure 12 and Table 8 respectively. Two basic categories of mineral licks exist in this area. Those which receive the greatest volume of use and which seem to be preferred because the sheep travel relatively long distances to reach them have been designated "primary" licks (A and B in Fig. 12; see also Figs. 10 and 11). Licks which receive only occasional light use or licks which are used as substitutes when the primary licks cannat be reached have been designated "secondary" licks (C-H in Fig. 12). Mineral licks are apparently used by sheep to satisfy bath nutritional and social needs. It has been generally assumed that licks

71 61 Figure 12. Mineral licks and travel routes utilized by Dall sheep on the study area (map from U. S. Geological Survey, Philip Smith Mountains Quadrangle, 1956). Legend: A, B C,D,E,F,G,H primary mineral lick location* secondary mineral lick location* sheep,trail other major travel route *Letters refer to lick descriptions in Table 8 and the text.

72 62