Wes, eee cry shan ok fete ae ihe ess wh bere g bs at hetan eeu buen moa) ‘ “aatted Ss tee Be ferers ae foe Na a 4 Ne Ns Me widen hod ’ She Mex; oy Na eh on Eat it a he Mahe Se ey oe DP odoinn ger ete Benet Arora Soha a, senvrme Veqdtearact Be RES Tah OL is Oo Aah E eben E Caer oak BF e gece Tet "y ae ey ae yaks 2P VARs febis i sore mt ie ae art f eh, a (i Aa ey) f n ee i => The CANADIAN 3IELD-NATURALIST Published by THE OTTAWA FIELD-NATURALISTS’ CLUB, Ottawa, Canada Volume 111, Number 3 July-September 1997 The Ottawa Field-Naturalists’ Club FOUNDED IN 1879 Patron His Excellency The Right Honourable Roméo LeBlanc, P.C., C.C., C.M.M., C.D., Governor General of Canada The objectives of this Club shall be to promote the appreciation, preservation and conservation of Canada’s natural heritage; to encourage investigation and publish the results of research in all fields of natural history and to diffuse infor- mation on these fields as widely as possible; to support and cooperate with organizations engaged in preserving, maintain- ing or restoring environments of high quality for living things. Honorary Members Edward L. Bousfield Anthony J. Erskine Don E. McAllister William O. Pruitt, Jr. Irwin M. Brodo Clarence Frankton Stewart D. MacDonald Hugh M. Raup William J. Cody W. Earl Godfrey Verna Ross McGiffin Loris S. Russell Ellaine Dickson C. Stuart Houston Hue N. MacKenzie Douglas B.O. Savile Bruce Di Labio George F. Ledingham Eugene G. Munroe Pauline Snure R. Yorke Edwards Thomas H. Manning Robert W. Nero Mary E. Stuart Sheila Thomson 1997 Council President: David W. Moore Ronald E. Bedford Philip Martin Vice-Presidents: Michael Murphy Stephen Bridgett Cendrine Huemer Toml@eaue Fenja Brodo Isobel Nicol : : William J. Cody Frank Pope Recording Secretary: David Smythe France Rack Sian Rascnbanal Corresponding Secretary: Lee Cairnie Ellaine Dickson Chris Traynor ecaMaere as Barbara Gaertner Ken Young Alan German Eleanor Zurbrigg Jeff Harrison Those wishing to communicate with the Club should address correspondence to: The Ottawa Field-Naturalists’ Club, Box P.O. Box 35069, Westgate P.O. Ottawa, Canada K1Z 1A2. For information on Club activities telephone (613) 722-3050. The Canadian Field-Naturalist The Canadian Field-Naturalist is published quarterly by The Ottawa Field-Naturalists’ Club. Opinions and ideas expressed in this journal do not necessarily reflect those of The Ottawa Field-Naturalists’ Club or any other agency. Editor: Francis R. Cook, R.R. 3, North Augusta, Ontario KOG 1RO0; (613) 269-3211; e-mail: feook @achilles.net Copy Editor: Wanda J. Cook Business Manager: William J. Cody, P.O. Box 35069, Westgate P.O. Ottawa, Canada K1Z 1A2 (613) 759-1374 Book Review Editor: Dr. J. Wilson Eedy, R.R. 1, Moffat, Ontario LOP 1JO Coordinator, The Biological Flora of Canada: Dr. George H. La Roi, 9302 Edinboro Road, Edmonton, Alberta T6G 2Al Associate Editors: Robert R. Anderson Robert R. Campbell Anthony J. Erskine Warren B. Ballard Brian W. Coad W. Earl Godfrey Charles D. Bird Paul M. Catling William O. Pruitt, Jr. Chairman, Publications Committee: Ronald E. Bedford All manuscripts intended for publication should be addressed to the Editor and sent by postal mail, with the excep- tion of book reviews which should go directly to the Book Review Editor. Subscriptions and Membership Subscription rates for individuals are $23 per calendar year. Libraries and other institutions may subscribe at the rate of $38 per year (volume). The Ottawa Field-Naturalists’ Club annual membership fee of $23 includes a subscription to The Canadian Field-Naturalist. All foreign subscribers (including USA) must add an additional $5.00 to cover postage. Subscriptions, applications for membership, notices of changes of address, and undeliverable copies should be mailed to: The Ottawa Field-Naturalists’ Club, P.O. Box 35069, Westgate P.O. Ottawa, Canada K1Z 1A2. Second Class Mail Registration No. 0527 — Return Postage Guaranteed. Date of this issue: July-September 1997 (July 1997). Back Numbers and Index Most back numbers of this journal and its predecessors, Transactions of The Ottawa Field-Naturalists’ Club, 1879-1886, and The Ottawa Naturalist, 1887-1919, and Transactions of The Ottawa Field-Naturalists’ Club and The Ottawa Naturalist — Index compiled by John M. Gillett, may be purchased from the Business Manager. Cover: Cascade Mantled Ground Squirrel, Spermophilus saturatus, at Cascade Lookout, Manning Provincial Park, British Columbia, 31 May 1990. Photograph courtesy of Maria Leung. See report of the status and distribution of this species in British Columbia by Maria C. Leung and Kimberly M. Cheng, pages 365-375. SN AUG 2 2 1997 The Canadian Field-Naturakst July—September 1997 THSON/AD ~ORARIES Volume 111, Number 3 Fisher, Martes pennanti, Home Range Characteristics in a High Density Untrapped Population in Southern Québec YvEs GARANT!2> and MICHEL CRETE*? ‘Département de Biologie, Pavillon Alexandre-Vachon, Université Laval, Sainte-Foy, Québec G1K 7P4 *Centre d'études nordiques, Pavillon Jean-Charles-Bonenfant, Université Laval, Sainte-Foy, Québec G1K 7P4 Present address: Naturam Environment, 31 Marquette, Baie-Comeau, Québec G4Z 1K4 4Service de la faune terrestre, Ministére de l'Environnement et de la Faune, 150 boul. René Lévesque Est, Se étage, Québec, Québec GIR 4Y1 >Author to whom all correspondence should be sent. Garant, Yves, and Michel Créte. 1997. Fisher, Martes pennanti, home range characteristics in a high density untrapped population in southern Québec. Canadian Field-Naturalist 111(3): 359-364. Most Fisher (Martes pennanti) populations in North America are moderately to heavily trapped. Trapping may reduce density and can indirectly affect spacing patterns of solitary terrestrial carnivores by creating vacant territories. From 1991 to 1993, we studied home ranges of radio-collared Fishers in Gatineau Park (Québec) where trapping had been prohibited for = 20 years. We estimated density using median home range size and a track-count survey technique. Mean home range size deter- mined by minimum convex polygon was 5.4 km? (n = 7) for adult females and 9.2 km? (n = 3) for adult males in the park. The observed home ranges in Gatineau Park were comparable to sizes predicted from body mass/home range size curves among carnivores, but smaller than those measured for trapped populations in northeastern North America. Density was esti- mated at 2.7 Fishers/10 km? using median home range sizes, when assuming saturation of the area by Fishers. An indepen- dent density estimate of 3.0 Fishers/10 km? was obtained using a probability sampling survey based on track counts in the snow. This was one of the highest densities reported in northeastern North America and was mainly attributed to the absence of trapping, a major mortality factor in Fishers. La plupart des populations de pékans (Martes pennanti) de ’ Amérique du Nord sont modérément ou fortement exploitées par le piégeage. Chez les carnivores solitaires, le piégeage peut réduire directement la densité des populations ou affecter indi- rectement |’espacement des individus en créant des territoires vacants. Entre 1991 et 1993, nous avons étudié, par la télémétrie, les domaines vitaux de pékans du Parc de la Gatineau, un site ot le piégeage est interdit depuis plus de 20 ans. Nous avons estimé la densité des pékans a partir de la taille médiane des domaines vitaux et d’un inventaire des pistes en hiver. La taille moyenne du domaine vital, déterminée par le calcul du polygone convexe minimal, était de 5.4 km? (n = 7) pour les femelles adultes et de 9.2 km? (n = 3) pour les males adultes. Les tailles des domaines vitaux mesurées dans le parc étaient comparables a celles pouvant étre prédites a partir de la relation générale reliant la masse corporelle et la taille du domaine vital chez les carnivores. Les aires vitales mesurées dans le parc étaient cependant plus petites que celles mesurées chez des populations piégées du nord-est américain. La densité a été estimée 4 2.7 pékans/10 km? en utilisant la taille médi- ane des domaines vitaux et en assumant que les pékans occupaient la totalité de l'habitat disponible. Une estimation indépen- dante de la densité, par l’échantillonnage des pistes, a indiqué une densité de 3.0 pékans/10 km?. Cette densité correspond a une des valeurs les plus élevées pour le nord-est américain; la forte densité résulterait vraisemblablement de l'absence de piégeage, le principal facteur de mortalité chez le pékan. Key Words: Fisher, Martes pennanti, mustelids, density, home range, intrasexual territoriality, spacing, Québec. Spacing and home range size of terrestrial carni- vores may be affected by factors such as body size, age structure, mortality rates, sex ratio, and by the abundance and productivity of food (Powell 1994). Fishers (Martes pennanti), like many Martes species, generally exhibit intrasexual territoriality (Powell 1979) and show little or no overlap of home ranges between individuals of the same sex. Intrasexual intolerance probably begins when animals reach sex- ual maturity as observed in Pine Marten (Martes martes) (Selas 1990). Fishers attain sexual maturity at one year of age (Wright and Coulter 1967; Leonard 1986; Cherepak and Connor 1992). Home range size was reported in many Fisher studies (Kelly 1977; Raine 1982; Buck et al 1983; Johnson 1984; Arthur et al. 1989; Jones 1991; Kohn et al. 1991), but all northeastern studies concerned har- vested populations. Harvesting affects population size by removing resident animals, which can indirectly influence social structure and spacing patterns of one or both sexes (Litvaitis et al. 1987; Strickland and Douglas © 559 360 1987; Arthur et al. 1989; Powell 1994). Spatial orga- nization of carnivores may remain stable if transient individuals from untrapped areas settle into ranges formerly occupied by residents (Litvaitis et al. 1987), or if territory size is determined primarily by energy requirements (Katnik et al. 1994). However, spatial organization can vary in highly exploited populations or in the absence of refugia (Bailey 1972; Litvaitis et al. 1987; Arthur et al. 1989); har- vested populations do not generally exhibit natural population attributes (Powell 1994). Most Fisher populations outside parks and reserves in the Northeast are moderately to heavily trapped (Douglas and Strickland 1987; Paragi 1990; Krohn et al. 1994). Annual trapping mortality in Maine ranged from 21 to 62%, depending on age and sex classes (Krohn et al. 1994), whereas it averaged 40% in Québec (Garant 1995). Such high trapping rates can maintain Fisher density below carrying capacity and provide vacant territories to transient individuals. Data from lightly harvested and unhar- vested areas are needed for a better understanding of Fisher population dynamics (Krohn et al. 1994). Knowledge of how territoriality affects spacing pat- terns and density in untrapped areas could provide useful information for the management of Fisher populations. Our objectives were to determine Fisher popula- tion characteristics and to estimate home range sizes and density in an untrapped park of southern Québec. Because there was no trapping and there were limited human activities, we postulated that Fisher density would be high and predicted home range size smaller than those in trapped areas. Study Area We conducted the study in Gatineau Park, Québec (45°36'N, 76°05’E), a 350-km? mostly forested area located 30 km northwest of the Ottawa-Hull region (Figure 1). Our main study block was the western half of the park, in the La Péche lake area. The land- scape consisted of rolling hills typical of the -Laurentian shield. Vegetation was representative of the Great Lakes — St.-Lawrence forest region (Rowe 1972); about 45% of the park was composed of hardwoods, and 35% of mixed woods with limited areas of pure conifers. The most common species was Sugar Maple (Acer saccharum), often associated with Yellow Birch (Betula alleghaniensis) and Beech (Fagus grandifolia) (Lopoukhine 1974). Balsam Fir (Abies balsamea), Eastern Hemlock (Tsuga canadensis), White and Red pine (Pinus strobus, P. resinosa), and spruces (Picea spp.) were the most common coniferous species. At the outset of the study, furbearer trapping had been prohibited for 2 20 years in the western part of the park, and for 2 40 years elsewhere, except for the trapping of nui- sance Beavers (Castor canadensis). THE CANADIAN FIELD-NATURALIST Vol iit Methods Fishers were captured in Tomahawk live-traps (Model 107 and 207, Tomahawk Live Trap Co., Tomahawk, Wisconsin) in autumn and winter 1990 to 1992, mostly in the western half of Gatineau Park. Captured Fishers were anesthetized with intramuscu- lar injection of ketamine hydrochloride (Ketaset™, Bristol Lab, Syracuse, New York) at an average dose of 15 mg/kg body mass or gassed with isoflurane (Aeranne™, Anasquest, Mississauga, Ontario) in a Plexiglas box. For most Fishers captured in the park, an upper premolar (PM1) was removed for age esti- mation by cementum annuli (Strickland et al. 1982). Fishers were fitted with ear tags and with radio-col- lars (Model SMRC-3, Lotek Engineering Inc, Newmarket, Ontario, or model 225, Telonics, Mesa, Arizona). Radio-collared Fishers were monitored 2-5 times/month, between December 1990 and December 1993, from a Cessna 172 equipped with 2 Adcock (type H) antennas attached to the aircraft struts. Each location was plotted on a 1:50 000-scale topographic map. To respect independence of obser- vations, only those locations that were separated by > 16h were used for home range size calculation (Arthur et al. 1989). Telemetry accuracy was assessed by comparing known transmitter locations with those determined from the air. In these blind tests (n = 10), aerial locations were < 150 m from the actual locations. Home range size was determined by the minimum convex polygon (MCP) encompassing 100% of locations, using MCPAAL software (M. Stiiwe and C. E. Blohowiak, Smithsonian Institute, Front Royal, Virginia), for ease of comparison with published studies. The harmonic mean method was not used because it was found inappropriate for esti- mating total area used by Fishers (Arthur et al. 1989). We estimated age structure of live-trapped Fishers by calculating frequency distribution of juveniles (< 1 year), yearlings (1.5 year), and adults (& 2.5 years) and compared it with the age structure of Fishers harvested by trappers in the surroundings of the park (Figure 1) using a chi-square test of homo- geneity (Daniel 1978). We used G-statistic analysis, adjusted for our small sample size using Williams’ correction (Sokal and Rohlf 1981), to test if Fisher sex ratio in the park departed from parity. Density was estimated using median range size, assuming saturation of the area by Fishers, and intra- sexual territoriality (Arthur et al. 1989; Thompson and Colgan 1987). We also derived a winter density estimate in late February 1993 using a probability sampling technique (Becker 1991). This sampling design assumes that the number of different Fishers encountered along a set of transects can be deter- mined, and that movement data from a random sam- ple of radio-collared animals can be measured. Three 1997 sets of four transects were randomly distributed in February 1993 in a 58-km? block of Gatineau Park to tally Fisher tracks, 24 hours after a snowfall of 32 cm. Each east-west oriented transect was 4 km long and was walked by an observer on snowshoes. At the same time, nine radio-collared Fishers were monitored from a Hughes-500 helicopter, and their locations were recorded on a 1:20 000-scale topo- graphic map three times/day. These radio telemetry data were used to determine the average distance travelled by Fishers along a north-south axis, referred to as the X axis. The use of a helicopter helped to accurately determine locations and dis- tances travelled by Fishers. We then calculated the probability of encountering a Fisher track in the snow using mean distance travelled by radio-col- lared animals. Using the equations provided in Becker (1991), we estimated the number of animals within our inventory block and calculated confi- dence intervals on this estimate. Results and Discussion Home range Twenty-two Fishers were caught and radio-tagged in Gatineau Park, and home range size was calculat- ed for 10 resident adults, 3 males and 7 females. Data from other Fishers were excluded because they either dispersed from the park, were juveniles, or because of inadequate sample size (radio failure or loss of contact). From 17 to 35 independent locations were necessary before area-observation curves reached an asymptote. Mean home range size (MCP) of females and males in Gatineau Park were 5.4 (range: 2.2-9.6) and 9.2 km? (range: 6.0-12.3) respectively (Table 1), a difference not statistically significant (U = 4.00, P=0.13). Fisher home ranges in Gatineau Park were substantially smaller than those previously reported for harvested Fisher populations of Maine and New Hampshire (Table 1). Mean home ranges of 17 females and 9 males were 8 km” and 23 km’, respectively, in a recent study in Massachusetts (E. York, University of Massachusetts, personal com- GARANT AND CRETE: FISHER IN SOUTHERN QUEBEC 361 munication), and 29.4 km? for 6 males in a trapped area in southeastern Québec (Garant 1995). The ratio of male home range size/female home range size was 1.8 in Gatineau Park, a value of simi- lar magnitude to their body mass ratio of 1.7 (Y. Garant, unpublished data). This home range size ratio was comparable to the value of 1.9 measured in Maine (Arthur et al. 1989) but greater than the ratio of 1.5 estimated in New Hampshire (Kelly 1977). Adult Fishers exhibited a spacing pattern similar to animals from harvested populations and maintained a comparable ratio between male and female home range sizes. No intrasexual territory overlap was observed. However, although we trapped intensively in the western half of the park, some unmarked Fishers may have been present in our main study block. Reduced home range size in mustelids can result from two confounding factors: abundant foods and high density. Thompson and Colgan (1987) found that in American Marten (M. americana), home ranges were the smallest and density the highest when prey was abundant; home ranges grew in size and Marten density decreased when prey was scarce. In a similar way, distribution and abundance of the Least Weasel (Mustela nivalis) in Sweden depended on the abundance of small rodents (Erlinge 1974). Powell (1994) proposed a model for Martes species in which spacing behavior and home range vary with prey abundance. According to this model, spacing should evolve from completely exclusive territories when food resources are very low to ultimately exten- sive home range overlap, with members of both sexes, at high levels of prey. However, this model has not yet been experimentally tested in Fisher. Powell’s model predicts that intrasexual territori- ality would prevail in situations with intermediate levels of prey abundance. Since little or no territorial overlap was observed for adults in most, if not all, studies across Fisher range (Kelly 1977; Buck et al. 1983; Johnson 1984; Arthur et al. 1989), this implies that food base was intermediate everywhere. Food was probably not a limiting factor for Fishers in TABLE |. Home range areas of adult Fishers in untrapped Gatineau Park and trapped areas of New England. Location Sex N Untrapped area Gatineau Park F 7 M 3 Trapped areas Maine F 6 M } New Hampshire F 3) M 6 Home range (km?) xX SE Reference 5.4 0.9 Present study oe 1.8 Present study 16-3 4.7 Arthur et al. 1989 30.9 O53 Arthur et al. 1989 lisa Dal Kelly 1977 22.8 6.4 Kelly 1977 362 ONTARIO THE CANADIAN FIELD-NATURALIST Vol. 111 QUEBEC FicurE 1. Location of Gatineau Park in southern Québec. Gatineau Park since no sign of poor body condition or starvation was observed. Body mass in late autumn-early winter averaged 4.6 kg for males (n = 13) and 2.4 kg (mn = 15) for females, which were among the highest values reported for this species (Powell 1993). Observed home range sizes in Gatineau Park were close to values predicted by the general equation A = 170M! linking body mass (M) and home ranges (A) in carnivores (Lindstedt et al. 1986). Home ranges of Fishers in other studies were consistently larger than those predicted on the basis of body size (Buskirk 1992). In addition to food availability, we believe that spacing behavior and home range size may be influ- enced concomitantly by Fisher density. Fishers are known to defend their territory against intruders (Powell 1993), and instances of fighting have been documented (Arthur et al. 1989). When Fisher den- sity increases after interruption of trapping, one way to minimize social interactions with members of the Same sex and avoid overt aggression might be to contract home range size, as long as food resources remain sufficient to meet metabolic needs. Even if males may adjust their home ranges according to female spacing (Sandell 1989; Powell 1994), home range contraction may not restrict exclusive access to mates in a densily populated area. These three factors (mate access, density, and energy require- ments) that appear to influence home range size in Fisher are not necessarily exclusive. Conversely, when density decreases (e.g., because of exploita- tion) surviving Fishers could extend their territory into vacant ones without any social conflict. Recent studies on Fisher populations at relatively low densi- ties (Jones 1991; Heinemeyer 1993) tend to support this hypothesis. Density Thirteen Fisher tracks were tallied along the tran- sects during the February survey. According to telemetry locations of nine radio-collared animals, the average distance travelled by Fishers along the X axis was 0.90 + 0.09 km during the day of the winter survey. These figures yielded a population estimate of 17.4 + 2.0 Fishers for the 58-km? study block, i.e. 3.0 Fishers/10 km? (+ 11.1%, a = 1997 0.05). Assuming saturation, median home range sizes of female (5.6 km?) and male (9.3 km?) pro- duced a density estimate of 2.7 Fishers/10 km”. Both methods provided comparable density esti- mates, which were among the highest reported in the Northeast. In the White Mountains of New Hampshire, Fisher abundance varied from 1.3 to 2.6 Fishers/10 km? of suitable habitat (Kelly 1977). Arthur et al. (1989) estimated a winter den- sity of 0.5 to 1.2 Fisher/10 km? in Maine. Coulter (1966) and Hamilton and Cook (1955) reported a density of 3.9 Fishers/10 km? based on snow tracking; however, their estimates are at best indices of actual densities (Powell 1994) and are less accurate than density estimates based on telemetry or probability sampling. In the highly trapped Algonquin Park in Ontario, density esti- mates varied between 1.5 and 1.9 Fisher/10 km? (Douglas and Strickland 1987), whereas a maxi- mum density of 2.6 Fisher/10 km? was used in Maine for management purposes (Clark 1986). Age structure and sex ratio Juveniles accounted for 21% of the live-trapped Fishers in the park, whereas yearlings and adults comprised 42% and 37% respectively (n = 24). Age structure of Fisher in Gatineau Park differed signifi- cantly (x? = 40.79, df = 2, P < 0.05) from the age structure of Fishers trapped in the region surround- ing the park, where juveniles comprised 75% of har- vest (n = 92). This contrasted also with other har- vested populations where juveniles usually represent > 60% of the harvest (Douglas and Strickland 1987; Powell 1994). In the lightly trapped Fisher popula- tion of Wisconsin, juveniles comprised 48% of the harvest (Kohn et al. 1993). The Fisher age structure in Gatineau Park was consistent with the survivor- ship of 80% that we estimated during the study (Garant 1995). In Gatineau Park, we captured more females than males but our 0.6 M:F ratio did not statistically depart from parity (G, 4; =inGO. Gf = 1. -P > :0.05) his high proportion of females is nevertheless surprising because live-trapping and kill-trapping tend to be skewed toward males, mainly because adult females are less susceptible to trapping than are adult males (Krohn et al. 1994). Males’ greater vulnerability to trapping was attributed to their presumably greater food requirements (larger size) than females. Krohn et al. (1994) suggested that adult males may be less wary of traps because they are less familiar with their environment as they do not spend their entire adult life in the same home range, as do adult females. Trap spacing can influence rates of capture between sexes since rate of encounter with traps is higher for males if traps are arranged in grids, but higher for females if traps are arranged in a line (Buskirk and Lindstedt 1989). Therefore, because many of our traps GARANT AND CRETE: FISHER IN SOUTHERN QUEBEC 363 were parallel to access roads, this may have increased females’ rate of encountering traps as a result of the higher density of traps in their smaller home ranges. Powell (1993) suspected that true M:F sex ratios for natural Martes populations were close to equali- ty, but that harvested populations might be skewed toward females. Because of Gatineau Park’s relative- ly small size, this Fisher population could be partial- ly affected by kill-trapping on its periphery. By hav- ing larger home range sizes and greater mobility than females, males on the outskirts of the park may suf- fer a greater risk of being trapped. If this were the case, this higher exposure of males to trapping than females could explain to some extent why female Fishers outnumbered males in our study. However, we believe that unharvested populations could be skewed toward females if intrasexual territoriality is to be maintained and if male home range sizes exceed those of females. Such a social structure inevitably leads to a sex ratio skewed toward females because young males are forced to emigrate. This phenomenon was also observed in a well-established but harvested population in Maine: Arthur et al (1989) estimated that three resident adult Fishers, one male and two females, could be present per 25 km? block of habitat, for a M:F ratio of 0.5. Our data suggest that smaller home ranges in Fisher could result from increased social pressure resulting from high density. The untrapped popula- tion of Gatineau Park comprised more adults than in the adjacent harvested area and was skewed toward females. The characteristics of this untrapped Fisher population support Powell's (1994) assumption that trapping affects more than population size; it has repercussions on other population attributes (e.g., age structure, sex ratio) that we should consider for sound management of this furbearer. Acknowledgments This research was funded by the Ministére de l’Environnement et de la Faune du Québec. We acknowledge the contribution of H. Richard and M. Gauthier of Gatineau Park through their cooperation during the project. We thank G. J. Doucet and F. Potvin for their valuable and constructive comments on an earlier draft of the manuscript. J. Berthiaume drew the figure. Literature Cited Arthur, S. M., W. B. Krohn, and J. R. Gilbert. 1989. Home range characteristics of adult Fishers. Journal of Wildlife Management 53: 674-679. Bailey, T. N. 1972. Ecology of bobcats with special refer- ence to social organization. Ph.D. dissertation, University of Idaho, Moscow. 82 pages. Becker, E. F. 1991. A terrestrial furbearer estimator based on probability sampling. Journal of Wildlife Management 55: 730-737. 364 Buck, S., C. Mullis, and A. Mossman. 1979. A radio telemetry study of Fishers in Northwestern California. California-Nevada Wildlife Transactions, 1979: 166-172. Buskirk, S. W. 1992. Conserving circumboreal forests for martens and fishers. Conservation Biology 6: 318-320. Buskirk, S. W., and L. L. McDonald. 1989. Analysis of variability in home-range size of the American marten. Journal of Wildlife Management 53: 997-1004. Cherepak, R. B., and M. L. Connor. 1992. Constantly pregnant... well almost. Reproductive hormone levels of the Fisher (Martes pennanti), a delayed implanter. Norwegian Journal of Agricultural Sciences, Supplement number 9: 150-154. Clark, A. 1986. Fisher assessment—1985. Pages 406-448 in Planning for Maine’s inland fish and wildlife, species assessments and strategic plans, fur and game mammals, 1986-1991. Volume 1: Wildlife, part 1.4. Maine Department of Inland Fish and Wildlife. Augusta, Maine. Coulter, M. W. 1966. Ecology and management of Fishers in Maine. Ph.D. dissertation. State University of New York, Syracuse. 183 pages. Daniel, W. W. 1978. Applied nonparametric statistics. Houghton Mifflin Company, Boston. 503 pages. Douglas, C. W., and M. A. Strickland. 1987. Fisher. Pages 511-529 in Wild furbearer management and con- servation in North America. Edited by M. Novak, J. A. Baker, M. E. Obbard and B. Malloch. Ministry of Natural Resources, Toronto, Ontario. Erlinge, S. 1974. Distribution, territoriality and numbers of the weasel Mustela nivalis in relation to prey abundance. Oikos 25: 308-314. Garant, Y. 1995. Situation du pékan (Martes pennanti) au Québec. Québec, ministére de l'Environnement et de la Faune, Report 95-2630. Hamilton, W. J. Jr., and A. H. Cook. 1955. The biology and management of the Fisher in New York. New York Fish and Game Journal 2: 13-35. Heinemeyer, K. S. 1993. Temporal dynamics in the move- ments, habitat use, activity, and spacing of reintroduced Fishers in northwestern Montana. University of Montana. 158 pages. Johnson, S. A. 1984. Home range, movements, and habitat use of Fishers in Wisconsin. M.S. thesis, University of Wisconsin, Stevens Point. 78 pages. Jones, J. L. 1991. Habitat use of Fishers in northcentral Idaho. M. S. thesis. University of Idaho, Moscow. 147 pages. Katnik, D. D., D. J. Harrison, and T. P. Hodgman. 1994. Spatial relations in a harvested population of marten in Maine. Journal of Wildlife Management 58: 600-607. Kelly, G. M. 1977. Fisher (Martes pennanti) biology in the White Mountain National Forest and adjacent areas. Ph.D. dissertation, University of Massachusetts, Amherst. 178 pages. King, C. M. 1983. Factors regulating mustelid populations. Acta Zoologica Fennica 174: 217-220. Kohn, B. E., N. F. Payne, J. E. Ashbrenner, and W. A. Creed. 1993. The Fisher in Wisconsin. Wisconsin Department of Natural Resources, Technical Bulletin number 183. 24 pages. THE CANADIAN FIELD-NATURALIST Vol. 111 Krohn, W. B., S. M. Arthur, and T. Paragi. 1994. Mortality and vulnerability of a heavily trapped Fisher population. Pages 137-145 in Martens, Sables and Fisher. Biology and Conservation. Edited by S. W. Buskirk, A. S. Harestad, M. G. Raphael and R. A. Powell. Cornell University Press, Ithaca. Leonard, R. D. 1986. Aspects of reproduction of the Fisher, Martes pennanti, in Manitoba. Canadian Field- Naturalist 100: 32-44. Lindstedt, S. L., B. J. Miller, and S. W. Buskirk. 1986. Home range, time, and body size in mammals. Ecology 67: 413-418. Litvaitis, J. A., J. T. Major, and J. A. Sherburne. 1987. Influence of season and human-induced mortality on spatial organization of bobcats (Felis rufus) in Maine. Journal of Mammalogy 68: 100-106. Loupoukhine, N. 1974. The forests and associated vegeta- tion of Gatineau Park, Québec. Forest Management Institute. Information report FMR-X-58. 51 pages. Paragi, T. F. 1990. Reproductive biology of female Fishers in southcentral Maine. M. S. thesis, University of Maine, Orono. 107 pages. Powell, R. A. 1979. Mustelid spacing patterns: Variations on a theme by Mustela. Zeitschrift Tierpsychologica 50: 153-165. Powell, R. A. 1993. The Fisher. Life history, ecology, and behavior. University of Minnesota Press, Minneapolis. 237 pages. Powell, R. A. 1994. Structure and spacing of Martes popu- lations. Pages 101—121 in Martens, sables and fishers: Biology and conservation. Edited by S. W. Buskirk, A. S. Harestad, M. G. Raphael, and R. A. Powell. Cornell University Press, Ithaca. Raine, R. M. 1982. Ranges of juvenile Fisher, Martes pen- nanti, and marten, Martes americana, in southern Manitoba. Canadian Field-Naturalist 96: 432-438. Rowe, J. S. 1972. Forest regions of Canada. Department of Environment, Canadian Forestry Service. Publication number 1300. 172 pages. Sandell, M. 1989. The mating tactics and spacing patterns of solitary carnivores. Pages 164-182 in Carnivore behavior, ecology and evolution. Edited by J. L. Gittleman. Cornell University Press, Ithaca, New York. Selas, V. 1990. Social organization of pine marten. Fauna 44: 214-219. Strickland, M. A., and C. W. Douglas. 1987. Marten. Pages 530-546 in Wild Furbearer management and con- servation in North America. Edited by M. Novak, J. A. Baker, M. E. Obbard and B. Malloch, Ministry of Natural Resources, Toronto, Ontario. Strickland, M. A., C. W. Douglas, M. K. Brown, and G. R. Parsons. 1982. Determining the age of Fisher from cementum annuli of the teeth. New York Fish and Game Journal 29: 91-94. Thompson, I. D., and P. W. Colgan. 1987. Numerical responses of marten to a food shortage in north central Ontario. Journal of Wildlife Management 51: 824-835. Wright, P. L., and M. W. Coulter. 1967. Reproduction and growth in Maine Fishers. Journal of Wildlife Management 31: 70-87. Received 6 September 1995 Accepted 1 November 1996 The Distribution of the Cascade Mantled Ground Squirrel, Spermophilus saturatus, in British Columbia Mari C. LEUNG and KIMBERLY M. CHENG! Department of Animal Science, University of British Columbia, Vancouver, British Columbia V6T 1Z4 'To whom correspondence should be addressed. Leung, Maria C., and Kimberly M. Cheng. 1997. The distribution of the Cascade Mantled Ground Squirrel, Spermophilus saturatus, in British Columbia. Canadian Field-Naturalist 111(3): 365-375. The Cascade Mantled Ground Squirrel, Spermophilus saturatus, only occurs in southwestern British Columbia (Canada) and Washington State (United States). Little was known about the squirrel in British Columbia prior to this survey. Range and distribution were determined from museum records, survey forms, opportunistic observations, visual counts, and live- trapping. Visual counts were used most often and were more effective at locations where the relative abundance of S. satu- ratus was high. S. saturatus was found further north and west than previously reported, but not as far east as indicated by museum records. S. saturatus was uncommon and had a discontiguous distribution in British Columbia, which could be a result of frequent extinctions and slow recolonizations at the northern portion of their range. Marginal climatic conditions, drainage systems that delineate the range of S. saturatus, and displacement by the Columbian Ground Squirrel (S. columbianus) at the eastern extent of S. saturatus range in British Columbia, are all probable factors limiting the distribu- tion of S. saturatus. Although the squirrel’s distribution is limited, the population is not likely in danger of extinction at present. Trapping at four locations indicated good overwintering success of the most vulnerable age group, young born the previous year. Key Words: Cascade Mantled Ground Squirrel, Spermophilus saturatus, range, distribution, British Columbia. The Cascade Mantled Ground Squirrel, S. satu- ratus, iS endemic to the Cascade Mountains of southwestern British Columbia and Washington State. Little was known about the squirrel’s north- ern range and distribution in British Columbia (Trombulak 1988) except for data presented in Howell (1938), and Cowan and Guiguet (1965). The survey was conducted in 1989 to gather data on the squirrel’s range and distribution, and to determine if there were any recent changes in the distribution which may have resulted from habitat alteration. Supplementary data were also gathered to outline recruitment success. The species was on the British Columbia provincial red list for “threat- ened or endangered” vertebrate species (Munro 1990) until the report resulting from the survey (Leung and Cheng 1991) was reviewed by the Committee on Status of Endangered Wildlife in Canada (COSEWIC). COSEWIC (1992) decided to let the species remain unlisted. It was also down- listed to the provincial blue list for vulnerable species which are species not considered threatened by imminent extinction or extirpation, but at risk of further disturbance (Harcombe et al. 1994). Methods Before beginning the field survey for S. saturatus, a preliminary range and distribution map was con- structed from data presented in Howell (1938) and museum lists of locations where S. saturatus speci- mens were collected in British Columbia. The muse- ums which contributed included the Royal British Columbia Museum (BCPM;; 6 specimens), the Canadian Museum of Nature (NMC; 19 specimens), the Cowan Vertebrate Museum (CVM, University of British Columbia; 15 specimens), the Royal Ontario Museum (ROM; | specimen), and the [United States] National Museum of Natural History (USNM, Smithsonian Institute; 1 specimen). Locations of other sightings were sought by distributing survey forms to interested individuals, regional naturalist clubs and other outdoors-oriented organizations. The survey form had a picture of S. saturatus and a description on how to distinguish it from chipmunks (Eutamias spp.) and Golden Mantled Ground Squirrels (Spermophilus lateralis). In 1989 and 1990, locations where S. saturatus had previously been recorded from museum collec- tions, were resurveyed. Locations inside and outside the previously delineated range were also checked, particularly in subalpine to alpine habitats. Survey locations were limited by road, trail and terrain access. Field surveys for the ground squirrels were conducted during the active portion of the squirrel’s circannual cycle (Kenagy and Barnes 1988) between May and September in 1989 and 1990. Live-trapping, visual counts, and opportunistic observations (e.g., walking transects and road sight- ings) were used to detect the presence of S. satura- tus and to identify the current range and distribution of the squirrel. When trapping, Tomahawk #201 live traps were baited with peanut butter and oats. Traps were opened at 0800 hours, checked every two hours, and 365 366 closed at 1800 hours. In remote locations, each indi- vidual caught was marked with a metal ear-tag carry- ing a unique three digit number for future identifica- tion. In provincial parks and locations frequented by hikers, black hair dye applied to the pelage was used for marking and identification. For detecting the presence of S. saturatus, a minimum of 25 traps were set for one day. To make relative abundance comparisons, trap- ping was conducted for four consecutive days at four known locations of S. saturatus between 29 May and 8 July 1990. Young of the year were not yet dispers- ing nor trapped during trapping sessions and there- fore did not bias the density estimate upwards at locations checked later. Trapping effort totalled 160 trap-nights at the study areas on Iron Mountain (1650 m; 50°03'N 120°46’W) and Thynne Mountain (1950 m; 49°42’N 120°55'W), and 80 trap-nights at the Cascade Lookout in Manning Park (1750 m; 49°04'N 120°48’W and Brookmere Dump (1000 m; 49°40'N 120°52'W), — study areas which were approximately half the size. Iron Mountain is located in very dry cool montane spruce forest, Thynne Mountain and the Cascade Lookout are located in dry cold Engelmann Spruce-Subalpine Fir forest and the Brookmere Dump lies in dry cool Douglas Fir forest. Although the tree species differed between some sites, all sites were structurally similar in that they were all located in open forest with a developed understorey (Leung 1991). A population index was used to compare the population size of each of the four locations by dividing the total number of indi- vidual squirrels trapped by the trapping effort. — Aside from S. lateralis, S. saturatus is easily dis- tinguished from other small mammals in the region by its pelage colouration and size (Hall 1981) which, together with its diurnal behaviour and high toler- ance to human presence (Banfield 1974), facilitated the use of visual counts surveys to detect its pres- ence. Prior to field surveys, study skins of S. satura- tus and S. lateralis (subspecies tescorum) and associ- ated skulls obtained from the CVM, the Conner Museum (University of Washington), the Burke Museum (Washington State University), and the Slater Museum (University of Puget Sound) were compared so that either species would not be mistak- en for the other when observed in the field. The dark stripes bordering the longitudinal white stripe on each side of the back is well defined in S. lateralis, whereas, on S. saturatus, the median dark stripes are reduced or obsolete and the outer dark stripes are reduced or obscure (Howell 1938). While we did not make an exhaustive examination of all the study skins available at other museums to confirm identification, we are confident that S. satu- ratus specimens have been correctly identified at museums for two reasons: 1) no collection localities for S. saturatus were unusually far from other collec- THE CANADIAN FIELD-NATURALIST Vol. 111 tion localities; and 2) skulls (which often had corre- sponding study skins) identified as S. saturatus or S. lateralis measured at the four museums mentioned were found to be significantly different in size and shape (Leung and Cheng 1994). While misidentifica- tion was unlikely, we were not able to eliminate totally that possibility in the field observations. Field observations were aided by 8 X 21 mm binoculars, a 20 X 60 mm spotting scope, and by intermittently sounding a high pitched whistle to attract the ground squirrels. The minimum observa- tion time, including changing viewpoints at each place, was one hour. The probability of sighting a S. saturatus within a certain time interval was calculated by correlating the cumulative percent sightings at locations of known inhabitation to cumulative time spent observ- ing at these sites. This allowed us to predict the amount of time needed at each location in order to detect S. saturatus if it was there. Results Distribution of S. saturatus Coordinates, elevations, and general habitats for locations where S. saturatus was sighted in 1989 and 1990 are listed in Table 1. Locations where speci- mens of S. saturatus were collected for museums and where S. saturatus were sighted during 1989 and 1990, including reports from survey forms, appear on Figure 1. Most locations associated with the spec- imens are approximate, but all known museum records are listed in Table 2. Locations where we searched for S. saturatus and did not detect any also appear on Figure 1. Coordinates and elevations for these locations are given in Appendix 1. UTM coor- dinates for locations where S. saturatus have been recorded are available from The British Columbia Conservation Data Centre (780 Blanshard St., Victoria, British Columbia, Canada V8V 1X4). Sightings of S. saturatus made in 1989 and 1990 indicate that their range is further north and west than documented by museum specimens. In 1989 and 1990, S. saturatus was recorded at 36 locations. Of these, 16 locations were detected from visual counts, 13 from opportunistic observations, seven from survey forms and none from trap sur- veys. Nine of 35 survey forms distributed were returned, of which seven reported sightings. S. saturatus was trapped where it had already been sighted, and were not trapped in all the attempts at locations where it had not been sighted. Therefore, trap surveys did not detect any new locations of S. saturatus. The results of trapping at the four loca- tions where presence of the ground squirrel was known are shown in Figure 2. Seven squirrels were trapped at Iron Mountain, 11 at Thynne Mountain, 14 at Cascade Lookout and 23 at the Brookmere Dump for a total of 55 squirrels. The mean age and LEUNG AND CHENG: CASCADE MANTLED GROUND SQUIRREL FiGurE 1. Map of survey area. ® = locations where Spermophilus saturatus was sighted during the survey; O = locations surveyed without any detection of S. saturatus; & = locations where S. saturatus has been collected for museum specimens prior to 1989; @ = locations where S. lateralis was sighted during the survey. sex composition of the 55 squirrels (Figure 2) was: adult female (31%), adult male (25%), yearling female (24%), yearling male (21%). Trapping in two consecutive years at Thynne Mountain showed that, of three adult females, four adult males, two young- of-the-year females and two young-of-the-year males (total of 11) marked in 1989, two adult females, two adult males, two yearling females, and one yearling male were recaptured in 1990. In addi- tion, five new individuals (one adult female, one adult male, and three yearling females) were cap- tured for the first time in 1990 at Thynne Mountain (Figure 2). Sightings of S. lateralis During the survey, S. lateralis was sighted at Big White Mountain (2300 m; 49°44’N 118°45’W) and Mount Baldy (2300 m; 49°08'N 119°16'W). Nagorsen (1995) also reported sighting of S. lateralis in Okanagan Mountain Provincial Park (520 m; 49°47'N 119°37'W) (Figure 1). Visual counts survey method For each of the four trapping locations, the pro- portion of total sightings for each time category was calculated (Table 3). The time to first sighting was less for larger populations (Table 3), and the likelihood of seeing a squirrel within the first hour increased with the density of squirrels (Table 3; Figure 3). Although observation periods did not always last until S$. saturatus was sighted, the “60 +” time cate- gory in Table 3 assumes that S. saturatus would have been sighted given enough time. To support this assumption, four artificial points (one for each loca- tion) were created for a cumulative percent sighting value of 100 % at 10 000 cumulative minutes. These points were included in the data for finding the best fit curve, but are not depicted in Figure 3. The aver- age cumulative percent sightings (y) for the four locations were plotted against cumulative time (t). The best fit curve, y = 78(1-e°!") with R? = 0.99, reaches an asymptote at 78% (Figure 3). Although R’ is high, the data points are widely scattered. If the four artificial points are excluded from the data set, the best fit curve is y = 70(1-e°?!') with R* = 0.99. Discussion The shorter time to first sighting and higher likeli- hood of seeing a squirrel within an hour of observa- tion at the Cascade Lookout and Brookmere Dump sites as compared to the Iron Mountain and Thynne Mountain sites (Figure 3) suggests that S. saturatus is more easily detected where it occurs at higher rel- ative abundance. The squirrels at the Cascade THE CANADIAN FIELD-NATURALIST Vol. 111 368 puepyied a4 ourdjeqng-sonidg uuewesugq eapun) ourdye pros Arp AOA O 00SZ M,60.0Z1 N,Z0067 UIeJUNOJ] MITADYL'T ‘OF ye [eIpoyje) puvpyied m4 ourdjeqns-sonidg uuewoelsugq eipun) surdye pros Arp AoA O OSZZ M,T1o0ZT N, £0067 ‘oye’T taddijsApe’'T jo N "Sg Myf ourdyeqng sonidg uuewrslsugq ploo Arp AroA O OOTZ M, ZI 0071 N, V0.6 ye [eIpoyyeD ‘aspoT 1osoy "pe My ourdyeqng sonidg uuewresuq Asp ‘O OSLI M,9b.07I N, 0.6 ye_ suruuryy “Inoxoo] epeoseD "Ee yieg suruueyy Ih{ SPIN JOLI}UI O OST1 A, 870071 N, S006? ‘punoisdure> sutidspjo5 CE diy ourdjeqng-sonidg uuewo)suq O OOVI M€S0CI N,P0.6v yieg suruueyy ‘syepy Auoqmeng ahs qooTWeH UreyuNos] O 0061 M,OfoL CL N,LOo6V Yeo SUITE AA ‘Ot yieg suruuryy puvpyied 117 ourdjeqns-sonidg uuewelsuq pue eipun) ourdye O OSZZ M,9b.071 N,OL 067 ‘UIRJUNOJY JOYjIOIG 18.114 "67 Mf SP[SNOG JOLaUT [ooo Asp ad OSOI M€t.0Cl N, T1060 ye suruuvyy JO | Uy g ‘8C IL SP[SNOG JOLOUT [OOd AIp u 0001 M,€€.071 NiS1 060 yoara Jeddo5 jo q Ne sonidg ourUoyA] pyrut Asp d O0El NM, TVo0CI N,€ 1060 oye'T UOstey) ‘9C yieg suruury puepyied m7 ouidjeqns-sonidg uuewelsuq pur eipun) ourdye d 0061 MN, 6boIZI N, I 106 ‘punoisduies asiop SuLpory Xorg ity ourdjeqng-sonidg uuewopsugq d 0001 M,6SoICI N,€ 1060 yeq suruuryy JO pus MA ‘VC If ourdjeqng-sonidg uurwe)suq d 0071 M,ZO0oITZI N, L060 oye] Aid JO N Ury € "EZ puepyied ij ourdjeqng-sonidg uuewoejsuq pure vrpun) ourdye O OO1C-0081 MOLoTCI N, L160 WwenNnO JUNO] CC a ourdjeqng-sonidg uurwolsug O OOLI NM. €0oI7l N,SCo60 ureJUNO|Y oNsvorL WG puepyred mj ourdjeqng-sonidg uuewelsugq pure vipum ourdye O 00TI MN. LOoIZI N,9Co6V Joptug yUNoW JO HN ‘OC HL Se[sNnog JOLIaUT [ood Arp ad OOL MN, LEOCI N/8Co6V UOoJsOUL JO MMN ‘6l My Se[snog JOLeyUr d 008 M6007 NOt .6v UOJSOUTI Puv JUOW][POD) UseAijoq 81 ML Se[snog JOLaUr [ooo Arp O OSTl NM VVo0CI N,67.60 peoy wingyoel_ YENI sonidg oueUoy| ad 0001 M,6S.0CI N,60067 JOATY UsouVn yt JON ©) sonidg ouvjuoyy pyrut Ap d OSEl M,9P.0CI NO .60 oxe"] GUOJSOpO'T O} prod Cl ML Se[snog JoLajut joy Arp Aroa a 008 M £P0CI N,Cto6V yuoWTeO) JO AN VI My ourdjeqng-sonidg uurwopsug a OOS NM, LOoIZI N,Veo6V WRI 2[P99N JO M tell Mf ourdjeqng-sonidg uuewolsug pyr yom u OSTI M,PS.0ZI NiS€06P urejunoyy Joddoyssern Jo N (li sonidg ourjuoy ad 0001 NM, 90017I N,Sto6V soye’y vpeyinboD Jo MS Wy g ala sonidg our Uo] a 0001 NM, VOoITI N,9€ 067 soye’y vpeyinbod Jo MS Wy p Ol Mf ourdjeqng-sonsdg wurursysug piu yom Pa OOTT M,00017I N,8€.6V soye'y epeyinbod jo § 6 diy ourdjeqng-sonidg uuewospsugq pjoo Arp O OS6l N.SSoOCI N,CV060 ureyunoyy ouuAY], 8 mJ ourdjeqng-sonidg wueursysugq poo Arp d OSV M,SS.0CI N,97067 ureyunoyy ouUAYT, JO N 8 JL Se[snoq JOLaut [ooo Arp O 0001 M,7S.0CI N, 6060 duinp oseqies osouyoo1g ‘9 Mf Se[sNoq JOLIaUr [ood Arp d 0001 NM, VS.0CI N,0S06V oTOUrOOIY JO AA ¢ df eurdjeqng-sonidg uurwejsuq pjoo Ap O OSSI M,6S.07I N,tS060 ureyuNoY] SITE Y sonidg our uoy| a 008 NM, COoICI N,VS$ 060 o4P'T SUITED) JO AN 1g dif Sepsnoq JOLayuI a OS6 M,SS0CI N,VSo6V oe"T SHIED, Cc sonids suvyuoul [OOS Arp AJOA O OS9I M,9¢.00Z1 N,€000S uleyUNO|] UO] ‘I QUOZ OIRUIT|IONSOIg poyloyy (Ur) So}PUIPIOOZD uoneso'T ADAING UONAITA ‘(6S6[) vurlery uo paseq oie souoz ONeUTI[IONSOIg *(CQ) adoos SuIj0ds 10 sre;Nd0UIg Aq UOTRAJASgO pur (yf) SUNYSIS ONstuNj10ddo/peos ‘(J) WO ADAINS :apnyour spoyjou ADAING “OHH PUL G86] Ul Porysis sem snyosnos snpiydousads B19YA SUOTLIO] JO} SyeITQeYy [eIDUSS pUL SUONRAIIO ‘Sa}eUTPIOOS ‘UOTWRIOT “| ATAV]L, 1997 LEUNG AND CHENG: CASCADE MANTLED GROUND SQUIRREL 369 TABLE 2. Locations where S. saturatus specimens were collected in British Columbia Location Princeton Princeton Manning Park, Easy Going Ridge Tulameen Manning Park, Timberline Valley 6500 feet (2000 m) Manning Park, Mount Gordon 5500 feet (1700 m) Ashnola Forks Ashnola Forks Ashnola River Princeton Princeton Princeton Gibson Pass, Manning Park 17 miles (27 km) up Ashnola River Blue River (NNE of Princeton) 12 miles (19 km) East of Princeton Princeton Manning Park Manning Park Manning Park Ashnola Creek Hope-Princeton Summit Pass Hope-Princeton Summit Pass Hope-Princeton Summit Pass Near Princeton Second Summit, Skagit River Second Summit, Skagit River Second Summit, Skagit River Second Summit, Skagit River Second Summit, Skagit River Similkameen River, 2 miles (3 km) below mouth of Ashnola River Similkameen River, 2 miles (3 km) below mouth of Ashnola River Skagit Stirling Creek mouth Stirling Creek mouth Stirling Creek mouth Stirling Creek mouth Stirling Creek mouth Whipsaw Creek Juniper Mountain, Similkameen, 1737 m Lightling Lakes, Manning Park Lookout and Brookmere Dump were also accus- tomed to human activity (i.e., sightseeing and feed- ing at Cascade Lookout and discarding garbage at the Brookmere Dump) and were likely not disturbed into hiding as much by our visual count surveys as the squirrels at the more remote sites, Iron Mountain and Thynne Mountain. Where species occur at low densities, caution should be taken when concluding that the species is absent at a site (Haila and Jarvinen 1983; McArdle 1990). A probability of the species absence at a site is a more realistic representation of data in which species could go undetected despite their presence. The average 78% probability of sight- ing S. saturatus within 60 minutes (Figure 3) gives Museum ID# Date BCPM 4956 1939 BCPM 4955 1939 BCPM 5174 1945 BCPM D5; 1917 BCPM 5175 1945 BCPM 5485 1949 CVM 3554 1951 CVM BOD) 1951 CVM 3556 1951 CVM 3683 1943 CVM 3684 1943 CVM 3685 1943 CVM 4298 1952 CVM 5629 1950 CVM 5630 1947 CVM 5631 1952 CVM 9305 1967 CVM 2966 1949 CVM 2968 1949 CVM 2969 1949 NMC 28847 1950 NMC 7856 1927 NMC 7857 1927 NMC 7858 1927 NMC 28848 1950 NMC 1198 1905 NMC 1199 1905 NMC 1202 1905 NMC 1205 1905 NMC 1207 1905 NMC 8966 1928 NMC 8967 1928 NMC 1165 1905 NMC 8857 1928 NMC 8883 1928 NMC 8884 1928 NMC 8889 1928 NMC 8908 1928 NMC 1163 1905 ROM 28577 1930 USNM 235331 1920 some credence to the conclusion of S. saturatus’ absence on visits of an hour’s duration. Herein, we assume that the survey for S. saturatus, including all methods of detection, give an approximate range and distribution of S. saturatus. Our survey indicated that within the northern part of its range, S$. saturatus is uncommon and its distri- bution fragmented. Three drainage systems surround the range of S. saturatus in British Columbia. To the west is the Fraser River; to the north is the Nicola River system; and to the east is the Okanagan sys- tem. Cowan and Guiguet (1965) state that S. lateralis extends westward only to the east side of the Okanagan Valley, rendering the ranges of the two 370 THE CANADIAN FIELD-NATURALIST Vol. 111 0.3 0.25 0.15 Number trapped / Trap-day 0.05 5 Yearling male Yearling female Adult male | Adult female T 0 hasesrenss Re Cascade Lookout Brookmere Dump Location lron Mountain Thynne Mountain FIGURE 2. Age and sex composition and relative density of Spermophilus saturatus at four locations in British Columbia. sibling species allopatric in Canada. The two species are not known to overlap elsewhere (Hall 1981), but there has not been any recent published survey of S. lateralis. A museum survey of locations of S. later- alis specimens collected from British Columbia (USNM, 36 specimens; CVM, 11 specimens; NMC, 21 specimens; ROM, O specimen; MMMN [Manitoba Museum of Man and Nature] 1 specimen; BCPM 0 specimens: see Appendix 2) did not reveal any location west of the Okanagan drainage and pro- vided some substantiation that the two species do not overlap in the Okanagan Valley. The increased range extending northward and westward as found during this survey probably reflects enhanced human access to these areas in recent years, and not a range expansion by the ground squirrel. Cowan and Guiguet (1965) remarked that the squirrel’s “northern limits are not known’. The apparent shift of the eastern boundary westward in British Columbia (Figure 1) could be due to the competitive exclusion of S. saturatus by S. columbianus. While this hypothesis needs to be test- ed experimentally, circumstantial evidence for com- petitive exclusion includes observations of S. columbianus where S. saturatus used to be, such as the alpine meadows at Blackwall Peak (49°06'N 121°45’W) in Manning Park. S. saturatus has been residing in Manning Provincial Park prior to its des- ignation in 1941 (USNM specimen # 235331 collect- ed in 1920; Carl et al. 1952), but S. columbianus was first observed only after the construction of the high- way traversing the park (Ted Underhill: personal communication) and was not been included in the comprehensive natural history survey of Manning TABLE 3. Percent of total first sightings occurring within consecutive ten minute intervals at four trapping locations Time interval Iron (minutes) Mountain 0-10 20 11-20 0 21-30 0 31-40 13 41-50 0 51-60 13 60+ 54 Number of sightings WS Minimum number of Spermophilus saturatus 7 Thynne Cascade Brookmere Mountain Lookout Dump 23 18 50 Sil! 0 33 iI) 9 17 0 0 0 8 0 0 0 0. 0 28 18 0 13 11 6 11 14 23 n997] LEUNG AND CHENG: CASCADE MANTLED GROUND SQUIRREL Saf Cumulative percent sightings Brookmere Dump Cascade Lookout Thynne Mountain ORs lala are Iron Mountain J Best fit curve T i 20 30 0 10 40 50 60 Cumulative time (min) FiGurE 3. Probability of sighting Spermophilus saturatus within different time limits. Park by Carl et al. (1952). Until 1959, the western- most record of S. columbianus was still east of Manning Park (Sheppe 1959). It is possible for S. saturatus to be outcompeted by S. columbianus because their ecological niches overlap and in some habitats S. columbianus would be dominant. Not only are both squirrels diurnal and semifossorial, but they also have some overlap- ping forage species, including Lupines (Lupinus spp.) and Sunflowers (Taraxacum spp.) (Elliot and Flinders 1985; Andrusiak and Harestad 1987; Trombulak 1988). Since S$. columbianus occurs mainly in meadows, it could only exclude S. satu- ratus from meadow habitats, but not talus slopes or alpine forest where S. saturatus also resides (Reichel 1986). The larger size of S. columbianus (340-812 g) (Burt and Grossenheider 1976) and its cooperative social system (Michener 1983), would probably put S. saturatus (170-276 g) at a major disadvantage because resources would be used faster by S. columbianus and aggressive encounters are usually won by the larger animal. Evidence of competitive exclusion has already been observed between S. lateralis and S. columbianus in Rossland (49°05'N 117°49'W), British Columbia. At one site where all S. columbianus were removed, S. lateralis subsequently reclaimed the area (Leung 1991). The discontinuous distribution of S. saturatus may be due to its poor colonizing ability and to the patchiness of its habitats. Brown (1971) presents data from the Great Basin suggesting that mammals on mountaintops are poor colonizers. Also, because S. saturatus is at the northern limit of its range in British Columbia, it likely experiences local extinc- tions in individual habitat patches and slow recolo- nization (McCauley 1991), which may explain its sparse and unpredictable occurrence in British Columbia. Isozyme analysis of S. saturatus in British Columbia revealed very little genetic varia- tion in the population (Leung and Cheng 1994). This indicates, at least theoretically, that the popu- lation may not adapt to rapid changes in environ- mental conditions. For rodents, mortality is usually highest in the first year of life (Caughley 1977). However, over- winter survival of young born in 1989 at Thynne Mountain was not lower than that of adults. The presence of yearlings at all four trapping sites indi- cate that some individuals born the year before overwintered successfully and were recruited into the local population. The mean age and sex compo- sition of the squirrels trapped at the four sites (Figure 2) was similar to that reported by Trombulak (1987) for S. saturatus populations in forest and meadow habitats in Washington State between 1980 and 1982. Decreased reproduction has been associated with Spermophilus species liv- ing in portions of their range where the growing season for plants is relatively short (Bronson 1979; Philips 1984). The recruitment and survivorship in the locations trapped seem to verify that S. satura- tus does not have a high risk of imminent extinction in British Columbia. Since S. saturatus in British Columbia likely mature slower than their southern counterparts and hence begin breeding after the second winter of hibernation, a longer study is needed to follow individual reproductive success. We recommend long-term monitoring, at least every five years, of S. saturatus to assess changes in population structure and to address any threats to the ground squirrel. 372 Acknowledgments This study was part of a thesis research project carried out by the first author in partial fulfilment of requirements for a M.Sc. degree at University of British Columbia. The research project was support- ed by The Habitat Conservation Fund, World Wildlife Fund (Canada), British Columbia Conservation Foundation, British Columbia Ministry of Environment, British Columbia Ministry of Parks, and the Committee On Status of Endangered Wildlife In Canada (Mammal Sub-committee). We thank Chris Dodd, Don Eastman, Al Harestad, G. J. Kenagy, Bill Munroe, David Nagorsen, David Shackleton, Paul Sherman, Merlin Shoesmith, Jamie Smith, and Tom Sullivan, for ideas and suggestions. Field and technical assistance was provided by W. Goedemondt, R. Altmann, D. Reid, C. Harvey- Clarke, D. Bennett, B. Bennett, T. Wells and C. Nichols. Museum specimens and historical records were provided by D. Nagorsen (Royal British Columbia Museum), C. Ludwig (National Museum of Natural History, Smithsonian Institution), J. Eger ~and S. Woodward (Royal Ontario Museum), R. Cannings and C. Adkins (The University of British Columbia Cowan Vertebrate Museum), K. Pullen (Conner Museum of Washington State University), J. Rodilsky (The Burke Museum of the University of Washington), E. Kritzman and D. Paulson (The Slater Museum of the University of Puget Sound), C. van Zyll de Jong and D. Campbell (Canadian Museum of Nature), and Ted Underhill. Logistical support was also provided by R. Hanah, W. Sobool, D. Crow and T. Haughton (British Columbia Parks staff), R. Moody (British Columbia Conservation Foundation), R. Scheer, R. Simmons, and W. Merillees (British Columbia Ministry of Parks), and The Merritt Snowmobile Club. Literature Cited Andrusiak, L. A., and A. S. Harestad. 1987. Feeding behaviour and distance from burrows of Columbian ground squirrels. Canadian Journal of Zoology 67: 381-384. Banfield, A. W. F. 1974. The mammals of Canada. University of Toronto Press, Toronto. 438 pages. Bronson, M.-T. 1979. Altitudinal variation in the life histo- ry of the golden mantled ground squirrel Spermophilus lateralis. Ecology 60: 272-279. Brown, J. H. 1971. Mammals on mountaintops: nonequi- librium insular biogeography. American Naturalist 105: 467-478. Burt, W. H., and R. P. Grossenheider. 1976. The field guide to the mammals of America north of Mexico. Houghton Mifflin Company, Boston, Massachusetts. Carl, C. G., C. J. Guiguet, and G. A. Hardy. 1952. A nat- ural history survey of the Manning Park area. British Columbia Provincial Museum Occasional Papers 9: 1-130. Caughley, G. 1977. Analysis of vertebrate populations. John Wiley & Sons, New York. 234 pages. THE CANADIAN FIELD-NATURALIST Vol. 111 Cowan, I. McT., and C. J. Guiguet. 1965. The mammals of British Columbia, third printing revised edition. British Columbia Provincial Museum, Victoria. Elliot, C. L., and J. T. Flinders. 1985. Food habits of the Columbian Ground Squirrel, Spermophilus columbianus, in southcentral Idaho. Canadian Field-Naturalist 99: 327-333. Haila, Y., and O. Jarvinen. 1983. Land bird communities on a Finnish island: species impoverishment and abun- dance patterns. Oikos 41: 255-273. Hall, E. D. 1981. The mammals of North America. 2nd edition. John Wiley and Sons, New York. 1181 pages. Harcombe, A., B. Harper, S. Cannings, D. Fraser, and W. T. Munro. 1994. Terms of endangerment. Pages 11—28 in Biodiversity in British Columbia: our changing environment. Edited by L. E. Harding and E. McCullum. Environment Canada, Canadian Wildlife Service, Vancouver. Howell, A. H. 1938. Revision of the North American ground squirrels with a classification of North American Sciuridae. North American Fauna 56: 1—256. Kenagy, G. J., and B. M. Barnes. 1988. Seasonal repro- ductive patterns in four coexisting rodent species from the Cascade Mountains Washington. Journal. of Mammology 69: 274-292. Krajina, V. J. 1959. Biogeoclimatic Zones in British Columbia. University of British Columbia, Vancouver, British Columbia. Leung, M. C. 1991. Status, range, and habitat of the Cascade mantled ground squirrel, Spermophilus saturatus, in British Columbia. M.Sc. thesis, University of British Columbia, Vancouver, British Columbia. 102 pages. Leung, M. C., and K. M. Cheng. 1991. Status report on the Cascade Mantled Ground Squirrel (Spermophilus saturatus) in Canada. The Committee On Status of Endangered Wildlife in Canada. Ottawa. 27 pages. Leung, M. C., and K. M. Cheng. 1994. Genetic variability of the Cascade mantled ground squirrel (Spermophilus saturatus) in British Columbia. Canadian Journal of Zoology 72: 371-374. McArdle, B. H. 1990. When are rare species not there? Oikos 57: 276-277. McCauley D. E. 1991. Genetic consequences of local pop- ulation extinction and recolonization. Trends in Ecology and Evolution 6: 5-8. Michener, G. R. 1983. Kin identification, matriarchies, and the evolution of sociality in ground dwelling sci- urids. Pages 528-572 in Advances in the study of mam- malian behaviour. Edited by J. F. Eisenberg and D. G. Kleiman. American Society of Mammalogists Special Publications Number 7. Munro, W. T. 1990. Committee on the status of endan- gered wildlife in Canada. Bioline 9: 10-12. Nagorsen, D. W. 1995. Status of Western Harvest Mouse in British Columbia. Ministry of Environment, Lands and Parks, Wildlife Working Report Number WR-71. 23 pages. Philips, J. A. 1984. Environmental influences on reproduc- tion in the golden mantled ground squirrel. Pages 108-125 in The biology of ground dwelling squirrels. Edited by J. O. Murie and G. Michener. University of Nebraska, Nebraska. Reichel, J. D. 1986. Habitat use by alpine mammals in the Pacific Northwest USA. Arctic and Alpine Research 18: 111-119. 1997 LEUNG AND CHENG: CASCADE MANTLED GROUND SQUIRREL 33 Sheppe, W. A. 1959. Notes on the distribution and habitats Trombulak, S. C. 1988. Spermophilus saturatus. of mammals in the Pacific Northwest. Murrelet 40: 1-4. Mammalian Species. Number 322: pages 1- 4. Trombulak, S. C. 1987. Life history of the Cascade man- tled ground squirrel (Spermophilus saturatus). Journal of | Received 27 March 1995 Mammalogy 68: 544-554. Accepted 12 November 1996 Appendix 1. _ Coordinates and elevations for localities surveyed in British Columbia where no Spermophilua saturatus was sighted. Locality Coordinates Elevation (m) 1. Mount Todd 50°55'N 119°56’W 2100 2 Cornwall Hills 50°42'N 121°28’W 2000 a Mount Savona 51°42'N 120°49’W 1500 4. Greenstone Mountain 50°37’N 120°38’W 1800 3: Chuwhels Mountain 50°32’N 120°40’W 1900 6. Tuktakamin Mountain 50°28’N 119°34’W 1800 We N of Devil's Lake 50°25’N 121°53'W 1700-2100 8. Botanie Mountain 50°20’N 121°36'W 1400-1900 9. Silverstar Mountain 50°22'N 119°03'W 1600-1850 10. Swakum Mountain 50°18'N 120°42'’W 1700 ii, Promontory Hills 50°12'N 120°58’W 1000,1700 12: Mount Hamilton 50°08'N 120°21'W 1500 LS. Terrace Mountain 50°06'N 119°37’'W 1600-1900 14. Eureka Mountain 50°05'N 118°23’W 1900 15: Stoyoma Mountain 49°59'N 121°13’W 2200 16. Kane Valley Trail 49°49'N 120°44'W 1350 ie Carrot Mountain 49°45'N 119°39'W 1500 18. E of Boston Bar 49°54'N 121°23'’W 1100 19. W of Gillis Lake 49°55'N 121°05'’W 1300 20. Blue Lake 49°52'N 120°35'W 1100 pM SE of Boston Bar 49°39’'N 121°22'W 800 22: Uztimus Road 49°47'N 121°20'W 500 23% 8 km E of Brookmere 49°49'N 120°40'W 1100 24. Siwash Lake Road 49°43’N 120°21’W 1250 25. Mount Pike 49°43'N 120°40’W 1400 26. SE of Boston Bar 49°40’N 121°24'’W 500 Dap Missezula Mountain 49°40'N 120°32'W 1500 28. Chain Lake 49°44'N 120°19'W 1350 29. Rampart Lake 49°39'N 120°30'W 1400 30. N of Spuzzum Creek 49°38'N 121°34’W 1000 Sis Douglas Lake Trail 49°38’'N 121°08’W 1300 32) Isintok Mountain 49°34'N 119°58’W 1800 3i8), E of Needle Peak 49°34’'N 121°06’W 1500 34. Emancipation Mountain 49°30'N 121°16’W 1100 35% Lodestone Mountain 49°28'N 121°50'W 1850 36. Dog Mountain 49°24'N 121°32’W 1000 S5L- Oglivie Moutain 49°25'N 121°23’W 750-1000 38. Stemwinder Road 49°25'N 120°09'W 1400 39: Apex Mountain 49°28'N 119°55'W 2200 40. Agate Mountain 49°23'N 120°24'W 1600 Al. S of Eureka Creek 49°20'N 121°28'W 500-900 42. Whistle Creek Road 49°18'N 120°06’W 1600 43. NW of IR10 49°16’N 120°09'W 2000 44. Orofino Peak 49°15'N 119°41'W 1500 45. Yola Creek 49°11'’N 121°26’W 1000 46. Crater Mountain 49°11'N 120°05'W 2000-2300 47. Silverhope Creek A49°07'N 121°22’W 1000 48. Shwatum Mountain 49°07'N 121°05’W 650 49. Mount Kobau 49°07'N 119°40'W 1800 50. W of Duruisseau Creek 49°06'N 120°23'W 1300-1700 Shi S of Nepopekum Mountain 49°03'N 121°00'W 600-1500 52° Church Mountain 49°02’N 121°51'W 1000-1250 SiS) Skyline Trail, Manning Park 49°02'N 120°55'’W 1300-1700 54. W of Mount Kruger 49°02'N 119°32'W 1000 S15 Frosty Mountain 49°01'N 120°50’W 1300-2400 THE CANADIAN FIELD-NATURALIST Locations of Spermophilus lateralis collected in British Columbia 374 Appendix 2. Museum Catalog# CVM 1566** CVM 3552" CVM 355935 CVM 447* CVM 1568** CVM S205 CVM 320874 CVM 3686** CVM 6294" CVM 5626** CVM DOTA USNM 106206 USNM 106207 USNM 106208 USNM 106209 USNM 106210 USNM 106211 USNM 106212 USNM 106213 USNM 106214 USNM 106215 USNM 106216 USNM 174129 USNM 174149 USNM 174150 USNM 174151 USNM 174153 USNM 174154 USNM 174155 USNM 174156 USNM 174157 USNM 174159 USNM 174172 USNM 180930 USNM 209374 USNM 209400 USNM 209405 USNM 210222 USNM 222781 USNM 257460 USNM 257461 USNM 257469 USNM 257470 USNM 257471 USNM 257472 USNM 257473 USNM 551546 NMC DUES SE NMC 22845** NMC 22846** NMC DIB aes NMC 0924-7+ NMC ESR NMC 10174*** NMC 47* NMC 10014*** NMC 28840*** NMC 28841 *** NMC 28843*** Subspecies cinerascens cinerascens cinerascens tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum Location Akamina River Mount Rowe, Akamina Pass Mount Rowe, Akamina Pass Creston Toby Creek Goat Peaks Goat Peaks Monashee Pass Pine River Pass Midge Creek Monashee Pass Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Barkerville Yellowhead Pass Moose River, South fork Moose River, North fork Moose River, North fork Moose River, North fork Moose Pass Moose Pass Moose Pass Moose Pass Moose Pass Yellowhead Lake Mount Selwyn Parsnip River, head Parsnip River, upper Parsnip River, middle Head Wapiti River Jarvis Pass Mount Selwyn, Peace River Mount Selwyn, Peace River Sukunka River, head Sukunka River, head Sukunka River, head Sukunka River, head Sukunka River, head Cranebrook Flathead Flathead Flathead Goatfell Goatfell Goatfell Goatfell, near Yahk Golden City Green Mountain, Rossland Grouse Creek, Barkerville Grouse Creek, Barkerville Invermere Collector J. Hatter I. McT. Cowan I. McT. Cowan Hatter and Cowan Cowan . Hollister . Hollister . Hollister . Hollister . Hollister . Hollister . Hollister . Hollister K. Vreeland K. Vreeland K. Vreeland K. Vreeland P. Fay K. Vreeland . G. Sheldon . G. Sheldon . Borden . Borden . Borden . G. Sheldon . G. Sheldon . B. Garrett . Wise . L. Thatcher . L. Thatcher ZABABAAL AZ Z,: tf tot nn eae aca ioel ee E.R. S. Halll John Macoun R. M. Anderson L. Jobin L. Jobin L. Jobin Volo itt (Continued) 1997 APPENDIX 2. Concluded. Museum Catalog# NMC 28844** NMC Vids NMC 9814*** NMC SStoas NMC O8247+* NMC SISSIB ae NMC Diol NMC Soo NMC O93 MMMN 5496*** BCPM 156% BCPM igi BCPM ZA is BCPM PANS A BCPM PAs BCPM ZAZO BCPM DAD IE* BCPM DAD BCPM 212374 BCPM 2124** BCPM 4957* BCPM 4958* BCPM 50347 BCPM Tes BCPM 9641* BCPM 9662* BCPM 9663* BCPM 9664* BCPM 9665* BCPM 18500*** Subspecies tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum tescorum Location Invermere Old Glory Mountain, Rossland Old Glory Mountain, Rossland Old Glory Mountain, Rossland Old Glory Mountain, Rossland Old Glory Mountain, Rossland Rossland Trail Trail British Columbia Indian Point Mountain Indian Point Mountain Eva Lake Eva Lake ’ Eva Lake Eva Lake Mount Revelstoke Monashee Pass Monashee Pass Monashee Pass Slocan Lake, Sandon Ymir Paradise Mine Mount Mobley Barkerville Eigh Mile Lake near Wells Invermere Invermere Invermere LEUNG AND CHENG: CASCADE MANTLED GROUND SQUIRREL Collector W. Spreadborough W. Spreadborough Unknown T. McCabe T. McCabe I. McT. Cowan I. McT. Cowan I. McT. Cowan I. McT. Cowan K. Racey K. Racey I. McT. Cowan K. Racey F. L. Beebe F. L. Beebe C. G. Clifford F. H. H. J. Jobin J. Jobin J. Jobin J. Jobin J. Jobin Okanagan Mountain Provincial Park D. Nagorsen *skin only; ¢skull only; **skin and skull; ***skin, skull, and skeleton. 375 The Recent Spread of Autumn-olive, Elaeagnus umbellata, into Southern Ontario and its Current Status P. M. CATLING!, M. J. OLDHAM2, D. A. SUTHERLAND”, V. R. BROWNELL’, and B. M. H. LARSON* ‘Eastern Cereal and Oilseed Research Centre, Agriculture Canada, William Saunders Building, Central Experimental Farm, Ottawa, Ontario K1A 0C6 *Natural Heritage Information Centre, Ontario Ministry of Natural Resources, P.O. Box 7000, Peterborough, Ontario K9J 8M5 32326 Scrivens Drive, R.R. 3, Metcalfe, Ontario KOA 2P0 4Botany Department, University of Toronto, 25 Willcocks St., Toronto, Ontario M5S 3B2 Catling, P. M., M. J. Oldham, D. A. Sutherland, V. R. Brownell, and B. M. H. Larson. 1997. The recent spread of Autumn-olive, Elaeagnus umbellata, into southern Ontario and its current status. Canadian Field-Naturalist 111(3): 376-380. Autumn-olive (Elaeagnus umbellata Thunb.) was extensively planted during the 1960s and 1970s to provide food and cover for wildlife, and has several other uses including reclamation, erosion control and interplanting in agroforestry. It has been discovered growing outside of cultivation throughout much of southern Ontario over the past 10 years. It has increased rapidly in parts of southern Ontario and the United States and is currently a competitive pest in some areas, hin- dering colonization of native species and competing with native vegetation. Autumn-olive is primarily a problem in open habitats on dry sandy soils. It has the potential to degrade native plant communities of natural sandy openings which con- tain significant native biodiversity. Key Words: Autumn-olive, Elaeagnus umbellata, Elaeagnaceae, alien, weed, distribution, identification, invasion, Ontario. Several alien woody plants have become serious competitive invaders of Canadian ecosystems and they pose major threats to the long term protection of native Canadian biodiversity. Information on the nat- uralization and spread of alien plants in Canada is relevant to the legislation restricting entry and to pre- dictions which may enhance control. Here we report on naturalized occurrences of Autumn-olive (Elaeagnus umbellata Thunb.), also known as Umbellate Oleaster, Japanese Silverberry, or Asiatic Oleaster, a native of eastern Asia. Economic Value Autumn-olive has been available from plant nurseries and landscaping companies in Ontario for at least 30 years, and the cultivar “Cardinal” has been recommended for its attractive scarlet fruit and leaves silvery beneath (Sherk and Buckley 1968). It has also been planted to provide food and cover for wildlife and for erosion control (Henry 1980; Zarger 1980), and has been recommended as a nectar source for honey production (Sternberg 1982*). Autumn-olive is actinorhizal resulting in the improvement of soil quality through nitrogen fixation. Soil improvement capabilities make the plant valuable for reclamation of mine spoil (Brown et al. 1983; Hensley and Carpenter 1986) and in agroforestry. Interplanting with walnut *Unpublished, see Documents Cited section, following Acknowledgments. (Juglans nigra) for example, has improved the pro- duction of walnut in North America, resulting in the the highest yields and the highest return on investment (Schlesinger and Williams 1984; Campbell and Dawson 1989). It has been used extensively in European agroforestry. In Japan Autumn-olive is used for both pickles and the pro- duction of alcoholic beverages (Sakamura and Suga 1987). It is an easy shrub to propagate in nurseries and grows quickly producing visible results in a short time frame (Sternberg 1982). As a result of these qualities, it may may have been more eagerly advertised than it should have, especially as a “wildlife enhancement shrub”. In fact, it ranks rela- tively low on the list of species used by birds as discussed in some depth by Sternberg (1982*). History and Impact A survey of all major Ontario herbaria (CAN, DAO, HAM, OAC, QK, TRT, TRTE, UWO, WAT) suggested that Autumn-olive has only recently escaped from cultivation in Ontario. The first Ontario collection of a possibly naturalized plant (sub E. commutata) was made by Aiken and Darbyshire at Bruce’s Pit near Bells Corners in Ottawa-Carleton in 1983. Very little spreading had occurred from equally spaced original plantings at this site by 1995. The earliest Ontario collection of a definitely naturalized plant was made in Haldimand- Norfolk by Sutherland and Gartshore in 1985, at which time it was reported by Sutherland (1987) as being uncommon in that region. This was the first 376 1997 80° CATLING, OLDHAM, SUTHERLAND, BROWNELL AND LARSON: AUTUMN-OLIVE 377 Elaeagnus wumobellata @ = Escaped (specimen record) @ = Escaped (sight record) ; Sesiyeees i sla a 78° 76° FiGurE 1. Distribution of Autumn-olive in Ontario based on specimens at CAN, DAO, MICH, TRT, TRTE and UWO, as well as sight records of M. J. Oldham and D. A. Sutherland. published report for both Ontario and Canada. In the second published report for Ontario and Canada, Oldham et al. (1992), noted an adventive plant along a railway in Elgin County, but they noted also that it does not often escape cultivation (supporting this view by the observation that it was not listed by Morton and Venn (1990)). Over the past 10 years Autumn-olive has been discovered growing outside of cultivation throughout the southern part of the province (Figure 1). It has increased in both distribu- tion and abundance on the Norfolk Sand Plain (Regional Municipality of Haldimand-Norfolk) to the point where it is now a very serious weed, large- ly as a result of being a significant competitor with native vegetation. Its soil improvement capabilities may lead to a reduction in significant native vegeta- tion which is often associated with nutrient deficien- cy (Eckardt 1987*). The shrub has the potential to degrade natural sandy openings including prairie, savanna, barren, dune and shore communities that have been highly profiled for protection in Canada. Naturalization and impact in other parts of North America appears to be similarly recent. For exam- ple, Autumn-olive was not recognized as adventive in Illinois in 1963, and was first planted in the state during the 1970s. It became a major problem about 10 years later (Ebinger and Lehnen 1981; Sternberg 1982*), as a result of becoming “highly invasive” and forming dense stands at the expense of native vegetation. Ebinger and Lehnen (1981) reported up to 33975 stems/hectare with up to 30% being more than 5 dm high and frequencies of up to 97%. Zimmerman et al. (1993) recently reported it as a problematic invader of reestablished tallgrass prairie in Illinois. Autumn-olive is a detriment to agriculture in some parts of Illinois and in West Virginia (Ebinger 1983). It is difficult to control since burned, mowed, or cut. plants will resprout vigorously (e.g. Kuhns 1986*; Szafoni 1991), making the use of herbicides or hand pulling necessary. Eckardt (1987*) noted that Autumn-olive was only just beginning to be rec- ognized as a potentially serious alien problem in the United States during the late 1980s, at which time it was still being distributed for use in wildlife plant- ings. Some university and museum-based botanists in the US have concluded that further planting of Autumn-olive would be “biologically unsound” or “biologically immoral”. The shrub has also been described as having “the potential of becoming one of the most troublesome adventive shrubs in the cen- tral and eastern United States” (Ebinger and Lehnen 1981; Eckardt 1987*; Sternberg 1982*), based on prolific fruiting, rapid growth, site adaptability and avian dispersal. Few other introduced plants have so quickly become a subject of serious concern. 378 Regional variation in impact in Ontario The escaped occurrences in the Ottawa district (Mountain Provincial Wildlife Area, Leitrim green- belt, Bruce’s Pit) can all be traced to plantings 15 to 25 years old. Spread from these plantings has been minimal (5-8 plants from 10-25 planted shrubs), and the majority of the plants at each site are the original equally spaced plantings. At each of these sites other woody species are aggressively colonizing the open ground including Rhamnus spp., Rhus spp., Populus spp., and Betula spp. The sites included both sandy and clay-loam soils. The situation further south in Ontario is very dif- ferent. In the eastern Lake Ontario region and on the Norfolk sand plain the shrubs were aggressively invading sandy open habitat, and sometimes forming dense stands and evidently competing with native species in adjacent prairie and savanna relicts. Dispersal In many cases escaped plants appear to be a result of spread from nearby cultivated plants. At most of the Ottawa sites the shrubs produced fruit abundant- ly, at least on the lower branches. Younger shrubs appear to produce more fruit than older ones. The seeds, within juicy berries, are probably eaten and dispersed by many different mammals and birds. In the Ottawa area American Robin (Turdus migrato- rius), and Cedar Waxwing (Bombycilla cedrorum) have been observed eating the fruits (P. M. Catling, personal observation). These two birds as well as Eastern Bluebird (Sialia sialis), Swainson’s Thrush (Catharus ustulatus), Hermit Thrush (Catharus gut- tatus), Gray Catbird (Dumetella caroliniensis), Northern Mockingbird (Mimus polyglottos), Bohemian Waxwing (Bombycilla garrulus) and European Starling (Sturnus vulgaris) have been observed eating fruits in other parts of southern Ontario (D. A. Sutherland, personal observation). Seeds regurgitated by, or passing through the diges- tive system of Robins remained viable (Sternberg 1982*). Raccoons (Procyon lotor), Skunks (Mephitis mephitis) and Opossums (Didelphis mar- supialis) are also known to feed on the fruit (Szafoni 1991). Habitat Major infestations of Autumn-olive in Ontario are in areas of dry sandy soils. Although plantings exist on fine-textured, periodically wet soils over lime- stone, spreading has generally been very limited on such sites. In the eastern Lake Ontario area Autumn- olive was associated with Black Oak (Quercus velutina) and/or Trembling Aspen (Populus tremu- loides) in semi-open areas with Canada Bluegrass (Poa compressa). On the Norfolk sand plain, Autumn-olive has been found most frequently in a variety of dry to mesic sandy, forested and open to THE CANADIAN FIELD-NATURALIST Vol. 111 semi-open habitats including: deciduous and mixed forests dominated by Black Oak, White Oak (Quercus alba), White Pine (Pinus strobus) and Red Maple (Acer rubrum); Red Cedar (Juniperus virgini- ana) glades; prairie/savannah relicts dominated by Indian Grass (Sorghastrum nutans); and coniferous plantations. It has also been found in seasonally wet, open floodplain thickets; on open clay hillsides and roadside banks; on gravelly till in openings in White Cedar (Thuja occidentalis) floodplain slope wood- land; on raised sandy knolls in a open to semi-open graminoid fen; and on low sand dunes in Eastern Cottonwood (Populus deltoides) savannah. Soil pH at these sites ranges from 5-7 and a similar pH range is reported from U.S. sites (Eckhardt 1987*). Prospects Although some of the places in the U.S. where it has been reported as a serious pest are distant and have a warmer climate than Ontario, it is a competi- tive invader in the adjacent states of Michigan (Voss 1985), New York (personal observation) and Ohio (Cooperrider 1995). In Michigan, Voss (1985) reported that it is “too freely escaping” and “thor- oughly naturalized as a weed”. Some of the shrubs at eastern Lake Ontario sites had dead upper branches suggesting winter kill, but the plants were still aggressively spreading. Sherk and Buckley (1968) indicate that it can survive up to the limit of hardiness zone 5, which includes the region south of a line from the the lower Ottawa Valley to Georgian Bay. Cultivated plants are known to vary in their hardiness (Sternberg 1982*), but at the very least a large portion of southern Ontario appears to be within the range hardiness. Monitoring and research on reproduction and dispersal rates are necessary to further document the degree of threat (Eckardt 1987*). Research on the effects on the nitrogen cycle in infertility- dependent natural communities is also desirable. The decline of Autumn-olive in plantations in Illinois due to disease (Sternberg 1982*) suggests that natural controls may reduce the competiveness of dense stands. Depending on the effectiveness of natural controls, biological control methods may have to be considered. For a shrub that was not known as a wild plant in Ontario 10 years ago, Autumn-olive has spread very rapidly, but the main problems of competition with native flora may be limited to areas with sandy soils. In particular the sandy shores of the Great Lakes and the sandy areas of southern and eastern Ontario are the most susceptible. These areas are particularly important in terms of native biodiversity and are already seriously impacted. Identification Although this shrub is becoming common over 1997 much of the northeastern United States (eg., Gleason and Cronquist 1991), it has not previously been reported as escaped from cultivation in Ontario in the standard reference texts (eg., Soper and Heimburger 1982; Morton and Venn 1990) or from Canada (eg., Boivin 1966-1967; Scoggan 1979). Consequently it is not expected and is not included in some of the keys frequently utilized to identify southern Ontario plants. Autumn-olive is one of six species in the Elaeagnaceae occurring in Ontario. Soapberry (Shepherdia canadensis (L.) Nutt.) is a widespread IPAM CROP MOSILGia teas 8l ef sata akbar tach aquvacbnsdhebloenieetehdecsaseees CATLING, OLDHAM, SUTHERLAND, BROWNELL AND LARSON: AUTUMN-OLIVE 379 native species. Silverberry (Elaeagnus commutata Bernh.) is native only in northern Ontario. Russian Olive (also called Narrow-leaved Oleaster, E. angustifolia L.) is a cultivated and occasionally escaped in Ontario (but becoming more common). The other three including Sea-buckthorn (Hippophae rhamnoides L.), Thorny Buffalo-berry (Shepherdia argentea Nutt.) and Multi-flowered Oleaster (Elaeagnus multiflora Thunb.), are all cultivated, mostly in southern Ontario. These species can be separated using the following characteristics: 2a. Leaves green and nearly glabrous above; plants not thorny osc cece snide Sin tae au EES Sea A en pene SA Shepherdia canadensis (L.) Nutt. (Soapberry, Canadian Buffalo-berry) 2b. Leaves silvery on both sides; plants thorny ............ MMs AE SEANMENIN AUC ack. cases cunts ttle vob soe soncedewente ensnudicesensp AMIR AVCSBLING Ale csr eee ee eee etetd ine te uwessa Snel ppeleeaves lanceolate, ovate or elliptic .2....0..0..eceeeeceeee Ah ieee Shepherdia argentea Nutt. (Thorny Buffalo-berry) 4a. Branchlets and leaves with only silvery scales; perianth tube as long as the calyx lobe; foamivellow, withisilvery Scales .......ie0i2.2.05--. pudeten Moth, dateecdidn Elaeagnus angustifolia L. (Russian Olive) 4b. Branchlets and leaves with both brown and silvery scales; perianth tube as long as or much longer than the calyx lobe; fruit silvery or red .............. 5a. Leaves silvery on both sides; fruit silvery Pomlcavesisilvery beneath, ereen aAbOVE, Mult DECOMMIME LEG i. -...)cc. 1c. eeiv-cenatns teccissoeceooboetandoscuseveccecssusest 6 6a. Perianth tube as long as calyx lobes; stalk of fruit 1.5-2.5 cm long eee e meee eee sere esses esse ee esessesseseesscsssseesssasesssssoses Patt E. multiflora Thunb. (Multi-flowered Oleaster) 6b. Perianth tube longer than the calyx lobes; stalk of fruit approx. 1 cm long eee come coer ewes ees eases een cessseeassasseeessesseseserssssessese Specimens Examined: ONTARIO: DUNDAS, GLEN- GARRY & STORMONT: 10 km W of Winchester, 45°05'N, 75°29’W, 29 June 1990, Shchepanek & Dugal 8628 (CAN), October 1995, Catling s.n. (DAO). ELcin: Aylmer, Carnation Milk property, 13 Aug. 1991, W. G. Stewart (UWO). Springfield, UTM 054415, 31 July 1992, Oldham 14042 (UWO). 1.5 km NE of Elgin-Kent border, UTM 461035, 9 Aug. 1993, Oldham & Vanderjeugd 15374 (MICH). Essex: Cedar Creek, UTM 537529, Allen & Sutherland 1340 (TRTE). Windsor, UTM 368836, 27 Oct. 1993, Oldham 15956 (DAO). HALDIMAND- NorFo._k: 1.5 km NE of Jacksonburg, UTM 317159, 13 Sept. 1985, Sutherland & Gartshore 6679 (TRTE). Huron: 1 km S Benmiller, UTM 495405, 11 June 1993, Reznicek et al. 9483 (MICH). KENT: Rondeau Prov. Park, South Point trail, 42°15'25’N, 81°51'15”W, 11 Oct. 1994, Larson 3664 (DAO), 30 July 1995, Larson 3908 (TRT). LAMBTON: Walpole Island Indian Reserve, UTM 747167, 8 August 1986, Allen et al. 2744 (TRTE). NORTHUMBERLAND: Brighton Twp., Goodrich-Loomis Conservation Area, 44°07'38"N, 77°49'17"W, June 1994, Brownell s.n. (DAO), 21 May 1995, Catling 20722 (DAO). 4 km N of Smithfield, 44°05’40’N, 77°41'20"W, 18 Aug. 1995, Catling s.n. (DAO). Be ree eRe ee en PR E. umbellata Thunb. (Autumn-olive) Murray Twp., Mayhew Cr. area, 4 km NW of Smithfield, 44°05’58”N, 77°40'11”W, 25 June 1995, Blaney s.n. (DAO). OTTAWA-CARLETON: Bells Corners, Bruce’s Pit, 45°19’45”N, 75°48’20"W, 2 June 1983, Aiken & Darbyshire s.n. (DAO), October 1995, Catling s.n. (DAO). Leitrim, 45°20'50"N, 75°33'50"W, 15 Sept. 1995, Catling s.n. (DAO). Acknowledgments W. J. Cody and J. Cayouette provided comments on the manuscript and S. Porebski assisted with preparation of the map. Documents Cited (marked * in text) Eckardt, N. 1987. Autumn olive: element stewardship abstract. Unpublished report for the Nature Conservancy, Minneapolis, Minnesota. 5 pages. Kuhns, L. J. 1986. Controlling Autumn Olive with herbi- cides. Pages 289-294 in Proceedings of the 40th Annual Meeting, Northeastern Weed Science Society. Sternberg, G. 1982. Autumn Olive in Illinois. Unpub- lished report, Illinois Department of Conservation, Springfield. 12 pages. Literature Cited _ Boivin, B. 1966-1967. Enumération des plantes du Canada. II Lignidées. Le Naturaliste canadien 93: 371-437. 380 Brown, J. E., J. B. Maddox, and W. E. Splittstoesser. 1983. Performance of trees, shrubs, and forbs seeded directly in the fall on mine spoil and silt loam soil. Journal of Environmental Quality 12(4): 523-525. Campbell, G. E., and J. O. Dawson. 1989. Growth, yield, and value projections for black walnut interplant- ings with black alder and autumn olive. Northern Journal of Applied Forestry 6(3): 129-132. Cooperrider, T. S. 1995. The Dicotyledoneae of Ohio. Part 2. Linaceae through Campanulaceae. Ohio State University Press, Columbus, Ohio. 656 pages. Ebinger, J. 1983. Exotic shrubs a potential problem in natural area management in Illinois. Natural Areas Journal 3(1): 3-6. Ebinger, J., and L. Lehnen. 1981. Naturalized autumn olive in Illinois. Transactions of the Illinois State Academy of Science 74(3,4): 83-85. Gleason, H. A., and A. Cronquist. 1991. Manual of the vascular plants of northeastern United States and adja- cent Canada. New York Botanical Garden, New York. Henry, J. 1980. A bonanza for wildlife. Soil Conservation 45(8): 13. Hensley, D. L., and P. L. Carpenter. 1986. Survival and coverage by several N,-fixing trees and shrubs on lime- amended acid mine spoil. Tree Planters Notes 37(3): 27-31. Morton, J. K., and J. M. Venn. 1990. A checklist of the flora of Ontario, Vascular Plants. University of Waterloo. Oldham, M. J., W. G. Stewart, and D. McLeod. 1993. Additions to “A guide to the flora of Elgin County, Ontario” for 1992. The Cardinal 151: 18. Sakamura, F., and T. Suga. 1987. Changes in chemical components of ripening Oleaster fruits. Phytochemistry 26(9): 2481-2484. THE CANADIAN FIELD-NATURALIST Vol. 111 Schlesinger, R. C., and R. D. Williams. 1984. Growth response of black walnut to interplanted trees. Forest Ecology and Management 9(3): 235-243. Scoggan, H. J. 1979. The flora of Canada, part 4, Dicotyledoneae (Loasaceae to Compositae). National Museum of Natural Sciences Publications in Botany Number 7(4): 1117-1711. Sherk, L. C., and A. R. Buckley. 1968. Ornamental shrubs for Canada. Canada Department of Agriculture, Publication Number 1286. 187 pages. Soper, J. H., and M. L. Heimburger. 1982. Shrubs of Ontario. Royal Ontario Museum, Toronto. Sutherland, D. A. 1987. Annotated checklist of the plants of Haldimand-Norfolk. 152 pages + appendices in The natural areas inventory of the Regional Municipality of Haldimand-Norfolk. Volume II. Norfolk Field- Naturalists, Box 995, Simcoe, Ontario. Szafoni, R. E. 1991. Vegetation management guidelines: Autumn Olive, Elaeagnus umbellata Thunb. Natural Areas Journal 11(2): 121-122. Voss, E. G. 1985. Michigan Flora. Part II. Cranbrook Institute of Science, Bloomfield Hills, Michigan. Zarger, T. G. 1980. Trees for reclamation in the eastern United States. Forestation of surface mines for wildlife. General Technical Report, Northeastern Forest Experiment Station. USDA Forest Service NE-61: 71—74. Zimmerman, V. D., J. E. Ebinger, and K. C. Diekroeger. 1993. Alien and native woody species invasion of aban- doned crop land and reestablished tallgrass prairie in East-Central Illinois. Transactions of Illinois Academy of Science 86(3-4): 111-118. Received 12 March 1996 Accepted 31 October 1996 The Distribution and Numbers of Bowhead Whales, Balaena mysticetus, in Northern Foxe Basin in 1994 S. E. Cosens!, T. QAMUKAQ’, B. PARKER?®, L. P. DUECK! AND B. ANARDJUAK* ‘Department of Fisheries and Oceans, 501 University Crescent, Winnipeg, Manitoba R3T 2N6 *Igloolik Hunters and Trappers Organization, Igloolik, Northwest Territories XOA OLO 3Department of Renewable Resources, Government of the Northwest Territories, Igloolik, Northwest Territories XOA OLO 4Hall Beach Hunters and Trappers Organization, Hall Beach, Northwest Territories XOA OKO Cosens, S. E., T. Qamukagq, B. Parker, L. P. Dueck, and B. Anardjuak. 1997. The distribution and numbers of Bowhead Whales, Balaena mysticetus, in northern Foxe Basin in 1994. Canadian Field-Naturalist 111(3): 381-388. Aerial reconnaissance surveys were conducted in June, 1994, to document the distribution of Bowhead Whales in the coastal waters of Foxe Basin and Roes Welcome Sound during the spring migration. The summer distribution of Bowheads in northern Foxe Basin was examined using both aerial surveys and boat-based observers. Results indicated that there are potentially three migration routes used by Bowheads to enter Foxe Basin and that once in northern Foxe Basin, Bowheads aggregate in a well-defined area north of Igloolik Island. Three systematic aerial strip surveys were done in August to esti- mate the number of surface whales present in northern Foxe Basin. Estimates of 256 + 31.3 and 284 + 48.6 Bowheads were derived from two useable surveys. Key Words: Bowhead Whale, Balaena mysticetus, eastern Canadian Arctic, Foxe Basin, distribution, numbers. By the early 20th century, Bowhead Whale, Balaena mysticetus Linnaeus 1758, numbers in the eastern Canadian Arctic were depleted by commer- cial whaling (de Jong 1983; Reeves et al. 1983). Other than reports by local residents that stocks are recovering (Anonymous 1995*), little is known about the current status of these whales. The most recent published estimate of Bowhead numbers in northern Hudson Bay and Foxe Basin, is “at least a few tens” of animals (Reeves and Mitchell 1990) based on examination of opportunistic sightings. There has been no systematic attempt to document Bowhead numbers in this area. Spring and fall migration routes are not well documented (Moore and Reeves 1993) and there has been only one sur- vey (McLaren and Davis 1982*) to locate the winter- ing area used by these Bowheads. In addition, the summer distribution of animals in this area, although generally known from whaling records and recent sightings (Reeves et al. 1983), has not been system- atically studied. In 1994, we began a study of Bowhead whales in Foxe Basin to gather information on spring and sum- mer distribution and on the numbers of whales using Foxe Basin during the summer open water period. Aerial reconnaissance surveys were conducted in late June to gather information on spring distribu- tion. Additional aerial reconnaissance surveys were flown in August to document summer distribution. Sightings from boats also provided information on both spring and summer distribution. Systematic aerial strip surveys were flown on three days in mid- *See Documents Cited section between Acknowledgments and Literature Cited. August to estimate how many Bowheads were pre- sent in northern Foxe Basin. Methods June reconnaissance surveys focussed on the coastal waters of Melville Peninsula, Roes Welcome Sound and the east and south east portions of Southampton Island (Figure 1). Fury and Hecla Strait was surveyed on 24 June. A flight on 28 June along the Igloolik ice edge was conducted in response to a hunter report of Bowheads having been seen along the ice. On this day, the survey aircraft departed from Hall Beach. On all other days, the air- craft left from Igloolik. Wind and ice conditions were noted during each survey. In August, reconnaissance surveys were flown only in northern Foxe Basin, including Fury and Hecla Strait, Gifford Fiord and Steensby Inlet (Figure 2). Transects were also flown between Hall Beach and Rowley Island. Wind and ice conditions were recorded during each survey. All reconnaissance surveys were flown with a deHavilland Twin Otter equipped with a Global Positioning System (GPS), at an altitude of 338 m at 204 to 222 km/h (110 to 120 knots). All windows were flat. In June, one observer sat in the co-pilot’s seat and three observers (two on the pilot’s side and one on the co-pilot’s side) sat in the passenger sec- tion of the plane. For the surveys in August, one observer sat in the co-pilot’s seat and two observers sat, one on either side, in the passenger section. Although the surveys were primarily for Bowhead Whales, incidental sightings of other species were also recorded. Belugas (Delphinapterus leucas) and Walrus (Odobenus rosmarus) made up the majority of sightings of other marine mammals. 381 382 THE CANADIAN FIELD-NATURALIST Vol. 111 28 June — Me eae Ener
am
*\ Frozen Strait
FiGurE 1. Aerial reconnaissance survey routes flown and Bowheads seen in June, 1994. Repulse Bay was used as a
refuelling location. -
Survey routes were documented by the observer
in the co-pilot seat by recording the time of day and
GPS location every 3 to 10 min or when the plane
changed bearing. The GPS locations of marine
mammal sightings were also recorded. Observers in
the passenger section of the plane recorded the time
of day of each sighting. The latitude and longitude
of these sightings were later determined by match-
ing time of day with the GPS data used to plot the
survey routes.
Local hunters reported the date, location and num-
ber of Bowheads they sighted opportunistically, dur-
ing their travels by boat in northern Foxe Basin. The
authors (BP and TQ) recorded the date, location and
number of whales seen during boat-based surveys to
photograph naturally marked Bowheads.
Both boat-based sightings and aerial reconnais-
sance data indicated that, in August, Bowheads were
aggregated in a relatively well-defined area north of
Igloolik Island. Strip surveys were therefore flown to
estimate the numbers of Bowheads using the area.
The survey block was located between 82°15’ and
80°35’W and extended approximately from the
south shore of Richards Bay or the north side of
Igloolik Island to the south coast of Baffin Island
(Figure 3). Survey transects were north-south and
spaced about.6.5 km (10 min of longitude) apart
(Table 1). The complete survey block was sampled
on each of three days: 11, 13 and 15 August. The air-
craft remained on transect at all times and did not
break off from the flight path to examine individual
groups. The entire survey block was sampled in 2 to
2.5 hours.
We used the same deHavilland Twin Otter for the
strip surveys as was used for reconnaissance surveys
and, again, flew at an altitude of 338 m at 204 to 222
km/h (110 to 120 knots). Aircraft position was moni-
tored using the Global Positioning System in the
plane; the observer in the co-pilot’s seat recorded
both time of day and GPS location at the beginning
1997 COSENS, QAMUKAQ, PARKER, DUECK AND ANARDJUAK: BOWHEAD WHALES 383
ON ae ° ° ‘Yo 70°3 U
10 August de Calg ihe Stee sby {
11 August - - - Wy
s Inlet
2g
‘
i]
1
1
CaM
xy
68°00
FiGurE 2. Aerial reconnaissance survey routes flown and Bowheads seen in August, 1994.
and end of each transect. This observer also recorded
Bowhead and other marine mammal sightings along
with their GPS locations. The two observers in the
rear of the aircraft recorded the time of day, species
and group size of each sighting. The location of
these sightings was determined by matching time of
day with the GPS data recorded by the observer in
the co-pilot’s seat.
All sightings, including those of the pilot, were
used to tabulate numbers seen. Inclusion of the
sightings made by the pilot and the observer in the
co-pilot’s seat resulted in sampling of the strip
under the plane because animals seen ahead of the
aircraft were counted. Data were checked for dupli-
cate sightings to ensure that Bowheads seen by
more than one observer were counted only once in
the analysis. Observers remained in the same seats
throughout the survey.
Using an inclinometer, the survey strip width
was estimated to be 600 m on either side of the air-
craft. Animals estimated to be within 600 m of the
aircraft were counted. Identification of species and
accurate estimation of group size was found to be
difficult for whales more than about 600 m from
the aircraft; however, sightability of whales with
distance from the aircraft was not systematically
documented. Inclinometer readings were periodi-
cally taken during the survey to confirm that sight-
ings were within 600 m.
Data from the 13 August survey were not used
because Beaufort 4-5 wind conditions reduced visi-
bility of the whales. The number of Bowhead sight-
ings made in this survey was about half that made
in the other two and most of these sightings were
made along the flight line of the plane thus the strip
width appeared to be reduced from what it was on
calm days.
Using the method outlined by Kingsley (1993),
the total number of whales sighted on each day was
multiplied by an expansion factor k , calculated as
k=S/W (1)
where S = transect spacing and W = strip width.
Transect spacing averaged 6.54 on 11 August and
6.42 on 15 August. Slight differences in transect
placement from one survey to the next resulted in
384
THE CANADIAN FIELD-NATURALIST
Voll a1
TABLE |. Transect locations, transect lengths and Bowhead counts from systematic aerial surveys conducted on
11 and 15 August in northern Foxe Basin.
Transect Longitude Length km
Number on 11 August
1 822157 27.4
2 $2205" 34.1
3 SIS55" 35.9
4 81° 45’ 46.3
5 S135! 52.0
6 lea 49.6
7 Sie a, 40.0
8 81° 05’ 43.3
9 80555" 32.6
10 80° 45’ 27.4
11 80° 35’ 25.4
variation in average transect spacing. Strip width
was estimated to be 1200 m. The derived estimate of
numbers
(2)
where J = the number of transects and x. = the num-
ber of whales counted on the j" transect is applicable
only to the survey block in northern Foxe Basin and
cannot be applied to the whole geographic range of
the stock.
The variance of the estimate of numbers was also
calculated following the method used by Kingsley
(1993) where
J
i
De Cease.
pa J J
AT kel) 4
Vie DMN)
2
+v
(3)
Standard errors of the estimates of Bowhead
numbers are reported.
Results and Discussion
In June, Fury and Hecla Strait was blocked by
land-fast ice (Figure 4). The northern portion of
Foxe Basin was open and a shore lead extended
south along the coast of Melville Peninsula. The
southern portion of Foxe Basin was covered with
unconsolidated ice. Cloud-free satellite images of
Roes Welcome Sound and Frozen Strait were not
available for the spring survey period, however, we
saw from the aircraft that Frozen Strait was also
blocked by land-fast ice. Land-fast ice was also pre-
sent in Roes Welcome Sound from just south of
Wager Bay at about 65°07’'N to about 64°32’N.
North of Wager Bay, a shore lead extended along the
mainland coast to about 66°11'N. By August, ice in
Fury and Hecla Strait had broken up and Foxe Basin
was relatively ice-free.
Reconnaissance surveys flown on 22, 24 and 26
June found few Bowheads (Figure 1). We saw no
Count on Length km Count on
11 August on 15 August 15 August
0 22.4 0
2 27.6 0
10 34.1 0
3 41.9 3
21 44.8 3
) 49.4 12
6 43.0 39
0 32.0 0
0 DSA 0
0 22.0 0
0 20.7 0
Bowheads along the coast of Melville Peninsula and
only four in the vicinity of Southampton Island.
Hunter reports of Bowheads at the Igloolik ice edge
were received on 28 June and an aerial count, made
the same day, indicated that at least 20 whales were
present (Figure 1) but, on 29 June, we counted at
least 30 whales from a boat. With the exception of
two possible sub-adults, all whales seen appeared to
be large adults. Observations from both the plane
and boat indicated that the whales were probably
feeding. They dived repeatedly, often raising their
flukes out of the water and remained submerged for
up to 20 min. Boat-based observations made during
the break-up period indicated that the whales were
seen in the same general location until the ice began
to disintegrate. Whales then moved northward
through the deteriorating ice, using melt holes for
breathing.
Early whalers believed that Bowheads migrated
from the high arctic through Fury and Hecla Strait
into Foxe Basin (Reeves et al. 1983). Early whaling
reports (Reeves et al. 1983) also indicated that bow-
heads in Roes Welcome Sound were thought to
migrate through Frozen Strait into Foxe Basin. Both
Frozen Strait and Fury and Hecla Strait were blocked
with land-fast ice during the June reconnaisance sur-
veys so whales found at the ice edge had to have
arrived through Foxe Basin which was open. Up to
three migration routes may, therefore, be used by
bowheads to move into northern Foxe Basin.
Local hunters from both Igloolik and Hall Beach
indicated that Bowheads are commonly seen in and
around Fury and Hecla Strait after the ice has broken
up. Two sightings of Bowheads (see Figures | and 3)
were reported in the vicinity of Fury and Hecla Strait
in 1994, although movement through the channels
was not documented.
Bowheads were seen along the coast of Melville
Peninsula, in the vicinity of Hall Beach, in July
(Figure 5). Their distribution in August was centred in
an area north of Igloolik Island (Figures 2 and 5)
COSENS, QAMUKAQ, PARKER, DUECK AND ANARDJUAK: BOWHEAD WHALES 385
1997
Murray Maxwell Bay
Jens Munk Island
akto
ao)
e
S
2
i
?
‘
‘
@
:
“
Murray Maxwell Bay
ao)
r=
&
fou
x
c
=]
=
”
o
o
=
i: OSewee ae cae ba
a a >
v
a a a ye en ~q
at af ee ee f>
ithe’ 3
*@ O00: 5-~
rAd
ee @ 00-5 >
deo
Beene eens ee. @ 3
ik Island
FIGURE 3. Locations of survey transects and Bowhead sightings during systematic strip surveys on (a) 11 August and (b)
15 August, 1994.
386
Battin
= 7
THE CANADIAN FIELD-NATURALIST
totand ,
| b, Paes 1 a weirs
_ Fury and Hecla Strait ~~
= i |
2 Ee Le i ae
Vol. 111
AES ARMA
NOAA-11 VISIBLE BAND
27-JUNE-1994 20:17GMT
SA]
FiGurE 4. Satellite image of ice conditions during reconnaisance surveys in northern Foxe Basin, including Fury and Hecla
Strait in June, 1994.
where they were consistently found throughout the
summer, associated with loose pack ice. Both fluke-
up diving and skim feeding (see Wursig et al. 1989)
were observed in this area. Other behaviours such as
tail-slapping, rolling and, apparently, sleeping were
also observed. Two calves were seen from the boat
and two were also seen from the aircraft. It is not
known whether these were the same or different
calves.
Based on observations from reconnaissance sur-
veys, the area between 82°15’ and 80°35’W was
selected for systematic surveys. From these surveys,
we estimated 256 + 31.3 Bowheads in the survey
block on 11 August and 284 + 48.6 Bowheads on 15
August (Figure 5), based on sightings of 47 and 53
individuals respectively (Table 1). The distribution
of whales was somewhat more clumped on 15
August, thus the Standard Error is wider than the one
calculated for the survey on 11 August.
Underestimation of strip width would result in an
overestimation of Bowhead numbers. Harwood and
Borstad (1985), flying at an altitude of 305 m, used a
2 km wide strip for Bowhead whale surveys in the
southeast Beaufort Sea. Their distribution of sighting
distances indicated that Bowheads were visible up to
a distance of 1100 m. If a strip width of 2 km were to
be assumed in the present study, estimates of surface
Bowheads would reduce to 154 + 17.3 (11 August)
and 170 + 26.8 (15 August).
The surveys conducted in 1994 were intended to
provide a minimum estimate of visible Bowhead
present in Foxe Basin. Future surveys would bene-
fit from using line transect methods where sighta-
bility as a function of distance from the aircraft,
group size and behaviour can be quantified and
appropriate correction factors applied to the final
estimate (Cooke 1985). Regardless of the inadequa-
cies of the 1994 surveys, results indicate that the
number of Bowheads present in northern Hudson
Bay/Foxe Basin is larger than a few tens of ani-
mals. A more rigorous survey is needed to more
accurately estimate numbers using this area during
the summer.—
Bowheads in northern Foxe Basin appear to
aggregate in a relatively well defined area. Both aeri-
al surveys and boat-based observations suggested
that Bowheads of all age classes were present
because calves as well as juveniles and adults were
19977
COSENS, QAMUKAQ, PARKER, DUECK AND ANARDJUAK: BOWHEAD WHALES
387
Oo June, July
27| @ August, September 0
ic.
@
8
BO
°*
8
ob
QO é O O
Richards ®% ae Nee =
Bay 2 lonakto Island
Cu
. Murray Maxwell Bay
ei
Jens Munk
Island
aS
10 km
0 tenes
FIGURE 5. Boat-based sightings of Bowhead Whales made by local hunters and the authors. The Bowhead on the west
side of Fury and Hecla Strait, at 69°52’N 83°18'W, was seen by a helicopter crew en route to Igloolik.
seen. Information on the age structure of this stock
should be collected to determine whether all compo-
nents of the stock are present in northern Foxe Basin
during the summer.
Bowheads were also present in northern Hudson
Bay at the time of our surveys in Foxe Basin (Robert
Luke, personal communication), thus our estimate
did not include all animals present in northern
Hudson Bay and Foxe Basin. Although northern
Hudson Bay and Foxe Basin Bowheads are consid-
ered to be one stock, (see Reeves and Mitchell
1990), additional research is required to clarify stock
relationships so that census results can be properly
interpreted and management units identified.
Acknowledgments
This study was supported by Tungavik Federation
of Nunavut Implementation Funds and by the
Community Resource Management Program. Polar
Continental Shelf Project supplied aircraft support.
The Science Institute of the Northwest Territories
provided accommodations and laboratory space that
was greatly appreciated. We thank the pilots of
Bradley Air for their patience and help during the sur-
veys. We also thank the hunters of Igloolik and Hall
Beach for their sighting records and discussions about
migration routes that were instrumental in directing
survey efforts. The project was conducted under
Scientific Research Licence Number 12690N from
the Science Institute of the Northwest Territories. Stu
Innes and Keith Hay provided valuable comments on
an earlier version of this manuscript. The manuscript
was greatly improved thanks to the comments made
by three anonymous reviewers.
Documents Cited
Anonymous. 1995. Nunavut bowhead traditional knowl-
edge study: Preliminary report based on interviews
conducted during 1995. Report presented at the meet-
ing of the Nunavut Wildlife Management Board,
Rankin Inlet. 37 pages.
McLaren, P. L., and R. A. Davis. 1982. Winter distribu-
tion of arctic marine mammals in ice-covered waters
of eastern North America. Unpublished report by LGL
for Petro-Canada, Calgary.
388
Literature Cited
Anonymous. 1993. Agreement between the Inuit of the
Nunavut Settlement Area and Her Majesty the Queen in
right of Canada. Tungavik Federation of Nunavut and
Department of Indian Affairs and Northern Development,
Ottawa, Ontario. 279 pages.
Cooke, J. G. 1985. Notes on the estimation of whale densi-
ty from line transects. Report of the International
Whaling Commission 35: 319-323.
Davis, R. A., and W. R. Koski. 1980. Recent observations
of the Bowhead whale in the eastern Canadian high arc-
tic. Report of the International Whaling Commission 30:
439-444.
de Jong, C. 1983. The hunt of the Greenland whale: A
short history and statistical sources. Report of the
International Whaling Commission Special Issue 5:
83-106.
Harwood, L. A., and G. A. Borstad. 1985. Bowhead
whale monitoring study in the southeast Beaufort Sea,
July-September 1984. Environmental Studies
Revolving Funds Report Number 9. 99 pages.
Kingsley, M. C. S. 1993. Census, trend, and status of the St
THE CANADIAN FIELD-NATURALIST
Vol. 111
Lawrence Beluga population in 1992. Canadian
Technical Report of Fisheries and Aquatic Sciences
Number 1938. 33 pages.
Moore, S. E., and R. R. Reeves. 1993. Distribution and
movement. Pages 313-386, in The Bowhead Whale.
Edited by J.J. Burns, J. J. Montague and C. J. Cowles.
The Society of Marine Mammalogy Special Publication
Number 2. 787 pages.
Reeves, R. R., and E. Mitchell. 1990. Bowhead whales in
Hudson Bay, Hudson Strait, and Foxe Basin: A review.
Le Naturalist canadien 117: 25-43.
Reeves, R., E. Mitchell, A. Mansfield, and M.
McLaughlin. 1983. Distribution and migration of the
Bowhead whale, Balaena mySticetus, in the eastern
North American arctic. Arctic 36: 5—64.
Wirsig, B., E. M. Dorsey, W. J. Richardson, and R. S.
Wells. 1984. Feeding, aerial and play behaviour of the
Bowhead whale, Balaena mysticetus, summering in the
Beaufort Sea. Aquatic Mammals 15: 27-37.
Received 30 October 1995
Accepted 18 November 1996
Wolf, Canis lupus, Predation and Maternal Defensive Behavior in
Mountain Goats, Oreamnos americanus
STEEVE D. CoTE, ALBERTO PERACINO, and GENEVIEVE SIMARD
Groupe de Recherche en Ecologie, Nutrition et Energétique, Département de Biologie, Université de Sherbrooke,
Sherbrooke, Québec JIK 2R1
Cété, Steeve D., Alberto Peracino, and Geneviéve Simard. 1997. Wolf, Canis lupus, predation and maternal defensive
behavior in Mountain Goats, Oreamnos americanus. Canadian Field-Naturalist 111(3): 389-392.
Four attacks by single Wolves on Mountain Goats were observed at Caw Ridge, Alberta, during July and August 1995.
One Wolf killed a yearling female and an adult female successfully defended her kid against an adult Wolf. In 206 agonis-
tic encounters between kids and older Goats, mothers defended their kids only five times. These results do not support the
hypothesis that maternal defense against conspecifics is common in Mountain Goats but demonstrate that females can
defend kids against predators.
Key Words: Wolf, Canis lupus, Mountain Goat, Oreamnos americanus, predation, maternal defensive behavior, Alberta.
Wolves (Canis lupus) and Mountain Goats
(Oreamnos americanus) both inhabit the Rocky
Mountains in Alberta, Canada (Mech 1970; Rideout
1978). Wolves are known to prey upon Mountain
Goats opportunistically but goats are usually a rare
prey (Smith 1986; Huggard 1993; Festa-Bianchet et
al. 1994; but see Fox and Streveler 1986). Huggard
(1993) observed that Mountain Goats represented
<2% of the biomass consumed by two Wolf packs in
summer and 0.2% in winter in Banff National Park,
Alberta. Peterson et al. (1984) reported that,
although Mountain Goats were present in their study
area on the Kenai Peninsula (Alaska), they were
absent from the Wolf diet. Inaccessibility of goat
range (Rideout 1978) and the potential risk presented
by the very sharp horns of Mountain Goats (Geist
1967; Nelson and Mech 1985) could explain the
scarcity of reports of Wolf predation on goats.
Offspring maternal defense against predators has
been reported for several large ungulate species
[Bighorn Sheep (Ovis canadensis): Hornocker 1969,
Berger 1978; Buffalo (Syncerus caffer): Schaller
1972; Chamois (Rupicapra pyrenaica): Locati 1990;
Moose (Alces alces): Stephenson and Van
Ballenberghe 1995; Mule Deer (Odocoileus
hemionus): Hamlin and Schweitzer 1979; Muskoxen
(Ovibos moschatus): Gray 1987; Pronghorn
(Antilocapra americana): Lipetz and Bekoff 1980;
Thomson’s Gazelle (Gazella thomsonii): Estes 1991;
White-tailed Deer (Odocoileus virginianus): Smith
1987; Wildebeest (Connochaetes taurinus): Kruuk
1972; Zebra (Equus burchelli): Schaller 1972]. For
Mountain Goats, Brandborg (1955) first suggested
that maternal defensive behavior was important in
defense against conspecifics, and could also be
against predators. To our knowledge, however, there
are no direct reports of maternal defense against
predators by Mountain Goats. Geist (1971, 1974)
stated that maternal protection of kids against con-
specifics is common and necessary in Mountain
Goats because of frequent juvenile and adult aggres-
sion but no study has presented quantitative evidence
of such behavior. Here we examine maternal defen-
sive behavior of Mountain Goats against con-
specifics and against predators and report a direct
observation of predation of a Wolf on a goat.
Interspecific observations
During a study of Mountain Goats in west-central
Alberta, Canada, we observed a female goat defend
her 4-month-old kid against an adult Wolf. We also
observed a Wolf kill a 15-month-old female. The
events reported here occurred on Caw Ridge
(54°04’N, 119°25’W), a gently rolling mountain
complex in the front range of the Rocky Mountains.
On 30 August 1995, we observed a group of 40
goats (38 were marked) and 12 kids foraging in an
open slope at about 2010 m altitude. They were
approximately 100 m from timberline when, at
12:55, two adult Wolves (one gray and one com-
pletely black) ran out of the forest and chased the
goats uphill for 300 m to the closest rocky cliff. The
Wolves did not get closer than 40-50 m from the
goats before they reached the cliff. At 13:02, the
gray Wolf approached the goats at the bottom of the
cliff and, after a few attempts, grabbed goat Number
166, a 3-month-old male kid of 23 kg marked 2
weeks before. As soon as the Wolf pulled the kid
down the rocky ledge, the kid’s mother (Number 23,
a 7-year-old first marked as a kid) jumped down and
charged the Wolf. She hit it twice on the rump and
missed it on an other attempt. The Wolf released
Number 166 and both mother and kid fled to the cliff
to join the other goats. Three other adult goats then
charged the Wolf and forced it to retreat. Number 23
apparently did not injure the Wolf which returned to
join the other Wolf about 150 m away. The goats
then disappeared to the other side of the escape ter-
rain followed at about 200 m by the two Wolves that
skirted round the cliff.
389
390
At 16:15, the group of goats came back to feed on
the same slope they had used in the early afternoon.
At 17:16, the same gray Wolf (as determined by its
coloration) appeared alone at the top of the ridge and
started pursuing the goats that ran toward a rocky
cliff. As the Wolf approached the base of the cliff,
the last three goats changed direction and started to
run toward the forest. The Wolf caught up to the
goats and grabbed the smallest one (a marked year-
ling female) by a hindleg but the goat escaped and
kept running towards the forest. The Wolf recaptured
the goat by the same hindleg while running downhill
and they rolled together 15 m downslope. The goat
got up again but was quickly caught at the throat and
knocked down by the Wolf. The goat managed to
stand and escape once again but was again recap-
tured, bitten at the throat, and died in <3 min. The
Wolf then disappeared in the forest (<20 m away)
for 5 min. It came back to the carcass at 17:36 and
dragged it into the forest out of sight. At 17:44, the
other goats started to bed in the cliff. Goat Number
75 (the mother of the yearling which had been
killed) looked for several minutes at the site where
the Wolf had disappeared and was the last goat to
bed. She had not attempted to defend the yearling.
We documented two other Wolf attacks on goats
in 1995. On 11 July, an adult Wolf attacked a group
of 63 goats including 16 kids feeding in an open for-
est at 1920 m but was unsuccessful. On 20 August, a
juvenile Wolf chased a group of 84 goats including
20 kids that were foraging at about 400 m from a
steep rock face but the goats ran to the cliff and the
Wolf never got closer than 30 m to them.
Another case of antipredator defense behavior was
observed on 9 June 1994 in the same goat popula-
tion. M. Festa-Bianchet and S.D.C. observed a
Wolverine (Gulo luscus) near a group of 15 goats
feeding below a ledge close to timberline. At this
time, kids were only 1 to 2-weeks-old and therefore
vulnerable to predators. Number 35, a 5-year-old
female, ran towards the Wolverine and drove it
away. This female did not have a kid.
Intraspecific observations
We sampled agonistic encounters involving kids
between May and September 1995 and 1996. In
249.9 h of focal observation periods (Altmann
1974), we observed kids 3591 times within <4 m of
goats other than their mothers or other kids. Kids
were displaced by older goats in 206 (5.7%) of these
encounters. Adult females, subadult females (1 and
2-year-old), and subadult males (1 and 2-year-old)
were responsible for 63.9%, 30.3%, and 5.8% of
these agonistic interactions, respectively. The mother
intervened and defended her kid only five times
(once against an adult female, once against a 2-year-
old female, once against a yearling female which
was the kid’s sister, and twice against unclassified
THE CANADIAN FIELD-NATURALIST
Vol. 111
individuals), suggesting that defense against con-
specifics is rare.
Discussion
Our observations suggest that female Mountain
Goats can protect their young against large predators
but that goats may be vulnerable when far from
escape terrain (Geist 1971; Rideout 1978; Smith
1983). Since 1989, 19 marked goats were known to
have been killed by predators at Caw Ridge, includ-
ing six taken by Wolves (Festa-Bianchet et al. 1994;
Coté et al., unpublished data). Even if, in general,
Mountain Goats appear not to be a common prey of
Wolves (Smith 1986; Huggard 1993), Fox and
Streveler (1986) reported that 62% of 124 Wolf scats
collected in southeastern Alaska contained goat
remains. Most goat populations in west-central
Alberta are small (40-50 individuals [x = 44] inhab-
iting defined home range with very little immigra-
tion and emigration) and have a very slow rate of
increase (Smith 1988). It may only takes one pack
that specializes on goat predation to cause an impor-
tant increase in the yearly mortality of a goat popula-
tion. If a Wolf pack killed ten goats during a winter,
it could reduce population numbers by as much as
20-25%, as has been suggested for Cougar (Felis
concolor) predation on Bighorn Sheep (Wehausen
1996; Ross et al., submitted). Therefore, it appears
important to consider the possibility of individual
Wolf packs specializing on goat predation when
managing Mountain Goat populations.
The role of maternal defensive behaviour against
conspecifics appeared much less important in our
study population of Mountain Goats than suggested
by Geist (1971, 1974). During 206 encounters, the
female defended her young only five times. We
observed just four cases where older goats used horn
contact to displace a kid and each time the goat
pushed the kid harmlessly instead of rushing it, sug-
gesting little need for maternal defense. Furthermore,
other researchers observed that kids were less com-
monly attacked by other goats than adult females and
juveniles (Chadwick 1977; Dane 1977). In a winter
study of goats at a baiting site, Masteller and Bailey
(1988) observed that an orphan kid received much
aggression, while kids with their mothers received
few threats. Therefore, it appears that kids may be
sheltered from many interactions by their mother
(Singer 1977). This suggests that, even if kids are
rarely defended against aggressors, they receive
some protection just by being close to their dam
(Chadwick 1977).
In our study, subadult males performed only 5.8%
of the agonistic interactions received by kids.
Similarly, Chadwick (1977) found that interactions
with 2-year-old males accounted for only 13% of the
kids’ total number of social interactions. These
results do not support Geist’s (1974) hypothesis that
1997
female Mountain Goats protect their kids against con-
specifics and especially against subadult males.
Therefore, there appears to be little selective pressure
for maternal defense against conspecifics in goats
(see Maestripieri 1992 for a review of maternal
defense in mammals). No maternal defense against
conspecifics was observed in studies of Bighorn
Sheep (Geist 1971), Apeninne Chamois (Locati and
Lovari 1990) and Muskoxen (Gray 1987). However,
all our observations were conducted in summer when
resources were abundant, and we might have reached
different conclusions if it would have been possible
to observe the goats during winter. When snow is
deep, resources are scarce, rates of aggression
increase (Petocz 1973), and kids feed in craters dug
by their mothers (Chadwick 1977). They may at this
time receive more protection from their dam.
Maternal defense in Mountain Goats seems to be
used only in extreme situations such as defense
against predators (Holroyd 1967; this study). Such
behavior has been reported in several species of
large ungulates (Packer 1983) and is likely to be
selected even if it prevents offspring predation only a
few times during the lifetime of a female.
Acknowledgments
Financial support was provided by the Rocky
Mountain Goat Foundation (grant to $.D.C. and
Marco Festa-Bianchet), Alberta Natural Resources
Service, the Alberta Sport, Recreation, Parks and
Wildlife Foundation, and the Natural Sciences and
Engineering Research Council of Canada (operating
grant to M. Festa-Bianchet and scholarships to
S.D.C.). We thank the Alberta Natural Resources
Service staff in Edson and Grande Cache for logistic
support and Chantal Beaudoin for help in the field.
M. Festa-Bianchet and James A. Schaefer kindly
reviewed an earlier draft of this manuscript. This is
contribution number 106 of the Groupe de recherche
en écologie, nutrition et énergétique, Université de
Sherbrooke.
Literature Cited
Altmann, J. 1974. Observational study of behavior: sam-
pling methods. Behaviour 49: 227-267.
Berger, J. 1978. Maternal defensive behavior in bighorn
sheep. Journal of Mammalogy 59: 620-621.
Brandborg, S. M. 1955. Life history and management of
the mountain goat in Idaho. Idaho Department of Fish
and Game Wildlife Bulletin 2. 142 pages.
Chadwick, D. H. 1977. The influence of mountain goat
social relationships on population size and distribution.
Pages 74—91 in Proceedings of the First International
Mountain Goat Symposium. Edited by W. Samuel and
W. G. MacGregor.
Dane, B. 1977. Mountain goat social behavior: social
structure and “play” behavior as affected by dominance.
Pages 92-106 in Proceedings of the First International
Mountain Goat Symposium. Edited by W. Samuel and
W. G. MacGregor.
CoTE, PERACINO, AND SIMARD: WOLF PREDATION AND MOUNTAIN GOATS
391
Estes, R. D. 1991. The significance of horns and other
male secondary sexual characters in female bovids.
Applied Animal Behaviour Science 29: 403-451.
Festa-Bianchet, M., M. Urquhart, and K. G. Smith.
1994. Mountain goat recruitment: kid production and
survival to breeding age. Canadian Journal of Zoology
72: 22-27.
Fox, J. L., and G. P. Streveler. 1986. Wolf predation on
mountain goats in Southeastern Alaska. Journal of
Mammalogy 67: 192-195.
Geist, V. 1967. On fighting injuries and dermal shields of
mountain goats. Journal of Wildlife Management 31:
192-194.
Geist, V. 1971. Mountain sheep. The University of
Chicago Press, Chicago. 383 pages.
Geist, V. 1974. On the relationship of social evolution and
ecology in ungulates. American Zoologist 14: 205-220.
Gray, D. R. 1987. The muskoxen of Polar bear Pass.
Fitzhenry and Whiteside, Markham, Ontario. 191 pages.
Hamlin, K. L., and L. L. Schweitzer. 1979. Cooperation
by coyote pairs attacking mule deer fawns. Journal of
Mammalogy 60: 849-850.
Holroyd, J.C. 1967. Observations of Rocky mountain
goats on Mount Wardle, Kootenay National Park, British
Columbia. Canadian Field-Naturalist 81: 1—22.
Hornocker, M. G. 1969. Defensive behavior in female
bighorn sheep. Journal of Mammalogy 50: 128.
Huggard, D. J. 1993. Prey selectivity of wolves in Banff
National Park. I. Prey species. Canadian Journal of
Zoology 71: 130-139.
Kruuk, H. 1972. The spotted hyena. University of
Chicago Press, Chicago. 335 pages.
Lipetz, V. E., and M. Bekoff. 1980. Possible functions of
predator harassment in pronghorn antelopes. Journal of
Mammalogy 61: 741-743.
Locati, M. 1990. Female chamois defends kids from
eagle attack. Mammalia 54: 155-156.
Locati, M., and S. Lovari. 1990. Sexual differences in
aggressive behaviour of the Apennine chamois.
Ethology 84: 295-306.
Maestripieri, D. 1992. Functional aspects of maternal
aggression in mammals. Canadian Journal of Zoology
70: 1069-1077.
Masteller, M.A., and J. A. Bailey. 1988. Agonistic
behavior among mountain goats foraging in winter.
Canadian Journal of Zoology 66: 2585-2588.
Mech, L. D. 1970. The wolf: the ecology and behavior of
an endangered species. Natural History Press, Garden
City, New York. 384 pages.
Nelson, M. E., and L. D. Mech. 1985. Observation of a
wolf killed by a deer. Journal of Mammalogy 66:
187-188.
Packer, C. 1983. Sexual dimorphism: the horns of african
antelopes. Science 221: 1191-1193.
Peterson, R. O., J. D. Wollington, and T. N. Bailey.
1984. Wolves of the Kenai Peninsula, Alaska. Wildlife
Monographs Number 88. 52 pages.
Petocz, R. G. 1973. The effect of snow cover on the
social behaviour of bighorn rams and mountain goats.
Canadian Journal of Zoology 51: 987-993.
Rideout, C. B. 1978. Mountain goat. Pages 149-159 in
Big game of North America. Edited by J. L. Schmidt and
D. L. Gilbert. Stackpole Books, Harrisburg.
Schaller, G. B. 1972. The Serengeti lion. University of
Chicago Press, Chicago. 480 pages.
392.
Singer, F. J. 1977. Dominance, leadership and group
cohesion of mountain goats at a natural lick, Glacier
National Park, Montana. Pages 107—113 in Proceedings
of the First International Mountain Goat Symposium.
Edited by W. Samuel and W. G. MacGregor.
Smith, C. A. 1983. Responses of two groups of Mountain
Goats, Oreamnos americanus, to a wolf, Canis lupus.
Canadian Field-Naturalist 97: 110.
Smith, C. A. 1986. Rates and causes of mortality in
mountain goats in southeast Alaska. Journal of Wildlife
Management 50: 743-746.
Smith, K. G. 1988. Factors affecting the population
dynamics of mountain goats in west-central Alberta.
Proceedings of the Biennial Symposium of Northern
Wild Sheep and Goat Council 6: 308-329.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Smith, W. P. 1987. Maternal defense in Columbian
white-tailed deer: when is it worth it? American
Naturalist 130: 310-316.
Stephenson, T.R., and V. V. Ballenberghe. 1995.
Defense of one twin calf against wolves, Canis lupus, by
a female moose, Alces alces. Canadian Field-Naturalist
109: 251-253.
Wehausen, J. D. 1996. Effects of mountain lion predation
on bighorn sheep in the Sierra Nevada and Granite
Mountains of California. Wildlife Society Bulletin 24:
471-479.
Received 22 December 1995
Accepted 1 November 1996
New and Noteworthy Records from the Vascular Flora of Nova Scotia
MICHAEL J. OLDHAM! and MARIAN ZINCK2
\Natural Heritage Information Centre, Ontario Ministry of Natural Resources, P.O. Box 7000, Peterborough, Ontario
K9J 8M5
2Nova Scotia Museum of Natural History, 1747 Summer Street, Halifax, Nova Scotia B3H 3A6
Oldham, Michael J., and Marian Zinck. 1997. New and noteworthy records from the vascular flora of Nova Scotia.
Canadian Field-Naturalist 111(3): 393-398.
Eleven taxa, ten species and one hybrid, are reported new to the vascular flora of Nova Scotia. Three species are rare native
plants (Carex alopecoidea, Carex digitalis, Carex tincta), two appear to be adventive from elsewhere in North America
(Bulbostylis capillaris, Hypericum gentianoides), and five are introductions from Europe (Cerastium pumilum, Cerastium
semidecandrum, Erophila verna, Myosotis stricta, Veronica verna). The hybrid sedge, Carex flava x Carex viridula ssp.
oedocarpa, has not previously been reported from North America. Notes are provided on four other plants rarely reported
from Nova Scotia (Aira caryophyllea, Aira praecox, Carex wiegandii, Veronica peregrina).
Key Words: Nova Scotia, vascular plants, floristics, new records.
We report ten species and one hybrid new to the
vascular flora of Nova Scotia: Bulbostylis capillaris,
Carex alopecoidea, Carex digitalis, Carex flava x
Carex viridula ssp. oedocarpa, Carex tincta,
Cerastium pumilum, Cerastium semidecandrum,
Erophila verna, Hypericum gentianoides, Myosotis
stricta, and Veronica verna. These taxa are not listed
in the most recent flora of Nova Scotia (Roland and
Smith 1969) nor are they listed for Nova Scotia in
the most recent flora of Canada (Scoggan 1978-
1979) or in the bibliography of Catling et al. (1995).
Of the additions, the three Carex species appear to
be overlooked rare native species. Bulbostylis capil-
laris and Hypericum gentianoides are probably
adventive in Nova Scotia, although native to North
America. The sedge hybrid Carex flava x Carex
viridula ssp. oedocarpa has not previously been
reported from North America. The remaining five
species are European weeds found in campgrounds
where they appear to be at least in part dispersed by
camping equipment.
In addition, we present information on four
species which are rare or infrequently reported in
Nova Scotia: Aira caryophyllea, Aira praecox,
Carex wiegandii, and Veronica peregrina. Following
specimen citations we discuss the distribution, habi-
tat and status of each species in Nova Scotia. Taxa
are arranged alphabetically, with those new to Nova
Scotia indicated by *. Herbarium acronyms follow
Holmgren et al. (1981) and Boivin (1980), except
where otherwise indicated.
Aira caryophyllea L. (Poaceae) Silvery Hairgrass
Annapolis County, Kejimkujik National Park,
main campground, rare on campsites; 25 June
1992; M. J. Oldham 13868 (DAO, MICH,
NSMP, mjo - personal herbarium of M. J.
Oldham). Inverness County, Cape Breton
Highlands National Park, Chetticamp Camp-
ground, rare in open gravel of campsites, with
Cerastium fontanum; 21 June 1992; M. J.
Oldham 13847 (DAO, MICH). Shelburne
County, Islands Provincial Park, near
Shelburne, rare campground weed, with Aira
praecox; 26 June 1992; M. J. Oldham 13884
(DAO, MICH). Victoria County, Cape Breton
Highlands National Park, Broad Cove Camp-
ground, campsites; 22 June 1992; M. J. Oldham
13855 (DAO, NSPM).
Catling et al. (1984) found Silvery Hairgrass on
Seal Island, Nova Scotia, and reported it new to east-
ern Canada. This species and the related Early
Hairgrass (Aira praecox) are tiny, annual grasses
introduced into North America from Europe. Aira
caryophyllea was previously known in Canada from
Yukon and British Columbia (Boivin 1969; Scoggan
1978-1979). Elsewhere in eastern North America it
is known from dry, open areas near the coast from
Vermont and Massachusetts south to Texas and
Florida (Gleason and Cronquist 1991).
Aira praecox L. (Poaceae) Early Hairgrass
Shelburne County, Islands Provincial Park, near
Shelburne, uncommon campground weed, with
Aira caryophyllea; 26 June 1992; M. J. Oldham
13883 (DAO, MICH, NSPM). Yarmouth
County, Ellenwood Provincial Park, near
Yarmouth, campground; 28 June 1992; M. J.
Oldham 13896 (DAO, MICH, NSPM).
Early Hairgrass was previously reported in eastern
Canada only from Seal Island and Mud Island, both
in Yarmouth County, Nova Scotia (Erskine 1958;
Roland and Smith 1969; Catling et al. 1984). Erskine
(1958) speculated that its presence on Seal Island
resulted from seeds coming ashore in the sand-bal-
last of ships driven up on the beach. Elsewhere in
Canada Aira praecox is known from southern British
Columbia (Hubbard 1969; Scoggan 1978-1979). In
393
394
the eastern United States it occurs from southern
New York to Virginia in dry sandy soil near the
coast (Gleason and Cronquist 1991). Both hairgrass
species are small and inconspicuous, and their pres-
ence in widely separated park campgrounds in Nova
Scotia suggests they should be looked for elsewhere
in dry, sandy open areas in the Maritime provinces
and New England. The two Aira species grew inter-
mixed in campsites in Islands Provincial Park,
Shelburne County.
*Bulbostylis capillaris (L.) C.B. Clarke (Cyperaceae)
sedge
Annapolis County, Jerry Lake, 44°43’ North
Latitude, 65°26’ West Longitude; locally abun-
dant on sandy roadside; 1 July 1993; M. J.
Oldham 15113 (MICH, NSPM, mjo).
Bulbostylis capillaris is a small annual sedge
which is widespread in eastern North America from
Maine to Minnesota, south to Florida, Texas, and
Arizona (Kral 1971; Gleason and Cronquist 1991).
In Canada it was previously known from southwest-
ern Quebec where it is doubtfully native (Bouchard
et al. 1983), and southern Ontario where it occurs
both as a rare native and as an introduction along
railways and roadsides (Argus et al. 1982-1987;
Oldham 1994). Bulbostylis capillaris is apparently
adventive in Nova Scotia, where it grows along rail-
way tracks and roadsides. This sedge was first col-
lected in Nova Scotia at Halifax in 1967, but was not
reported (Taschereau and Chute, NSPM).
Subsequent collections exist from Wolfville, Kings
County (Vanderkloet in 1972, ACAD); Annapolis
Royal, Annapolis County (Johnson in 1971, ACAD);
Shelburne County (Newell and Newell in 1981;
ACAD); and near Bridgewater, Lunenburg County
(Reznicek in 1990; MICH). At its most recent locali-
ty it was found with another annual, Radiola
linoides, a European species known in northeastern
North America only from Nova Scotia (Gleason and
Cronquist 1991).
*Carex alopecoidea Tuckerman (Cyperaceae) sedge
Antigonish County, St. Georges Bay, west of
Antigonish, 45°38’ North Latitude, 61°47’
West Longitude; local in moist, overgrown
clear-cut area near shore, with Carex stipata; 17
June 1992; M. J. Oldham 13813 (CAN, MICH,
NSPM, mjo). 24 June 1993; M. J. Oldham
14984 (DAO, TRTE).
This northeastern North American sedge is dis-
tributed from Maine to Minnesota, south to New
Jersey, Indiana, and Missouri (Gleason and
Cronquist 1991). In Canada it ocurrs in southern
Saskatchewan, Manitoba, Ontario, and Quebec, but
*new to Nova Scotia
THE CANADIAN FIELD-NATURALIST
Vol. 111
has not previously been reported from any of the
Maritime Provinces (Scoggan 1978-1979; Boivin
1992). Carex alopecoidea is rare in New England
(Seymour 1982), and is known in Maine only from
Oxford and Kennebec Counties (Richards et al.
1983).
The Antigonish County colony grew in an open,
moist, cut-over area with the superficially similar
sedge, Carex stipata. Carex alopecoidea appears
native to Nova Scotia and should be added to the
province’s rare plant list (Maher et al. 1978).
Carex digitalis Willd. (Cyperaceae) sedge
Annapolis County, Kejimkujik National Park,
local in sandy woods adjacent to main camp-
ground; 25 June 1992; M. J. Oldham 13873
(MICH, NSPM, ctb - personal herbarium of C.
T. Bryson, mjo).
Carex digitalis is a woodland sedge ranging from
Maine to Florida, west to Wisconsin, Illinois,
Missouri, and eastern Texas (Bryson 1980; Gleason
and Cronquist 1991). Elsewhere in Canada it is
known only from southern Ontario and southwestern
Quebec (Scoggan 1978-1979; Boivin 1992).
Bouchard et al. (1983) consider it rare in Quebec.
Although considered frequent in New England by
Seymour (1982), C. digitalis has not previously been
reported from the Canadian Maritime Provinces.
Nova Scotia material of C. digitalis belongs to var.
digitalis (Fernald 1950; Bryson 1980). In Kejimkujik
National Park Carex digitalis grows in dry, sandy
woods where it is undoubtedly native.
*Carex flava L. x Carex viridula Michx. ssp. oedo-
carpa (N. J. Andersson) B. Schmid (Cyperaceae)
hybrid sedge
Cumberland County, Cape d’Or, north shore of
Minas Channel, locally common in moist, open
seepage on exposed headland; 25 June 1993; M.
J. Oldham 15001 (MICH, NSPM, mjo).
One of the parents of this hybrid, Carex flava, is
common and widespread in northeastern North
America, while the other, C. viridula ssp. oedocarpa
(= C. demissa of Roland and Smith 1969), is rare in
North America and known from a few scattered sites
along the east coast from Newfoundland south to
New Jersey (Crins and Ball 1989). Crins and Ball
(1989) suggest that C. viridula ssp. oedocarpa, a
widespread European sedge, is most likely adventive
in North America. The hybrid between these two
sedges is known from Europe (Jermy et al. 1982),
but has not previously been reported from North
America (Cayouette and Catling 1992). There is one
previous Nova Scotia collection from Digby County,
“in a swale near the lake, west of Centreville”. It was
initially identified as Carex flava var. fertilis, and
was annotated by W. J. Crins in 1985 as this hybrid
(E. C. Smith et al. in 1954, ACAD).
1997
At Cape d’Or in Nova Scotia, the hybrid was
locally common in an open seepage slope with
Carex flava and Carex viridula ssp. oedocarpa.
*Carex tincta Fern. (Cyperaceae) sedge
Antigonish County, St. Georges Bay,
Antigonish Harbour, west of Bayfield, 45°38’
North Latitude, 61°47’ West Longitude, com-
mon at edge of mixed woods along dirt road; 20
June 1992: M. J. Oldham 13835 (MICH,
NSPM, mjo).
This poorly understood member of the difficult
Ovales Section of Carex (Reznicek 1989) is rare
throughout its range in Canada (Boivin 1992; A. A.
Reznicek, personal communication). Although most
recent authors recognize Carex tincta as a species
(e.g., Scoggan 1978-1979; Seymour 1982; Boivin
1992), some mention the possibility of it being a
hybrid (e.g., Scoggan 1978-1979; Gleason and
Cronquist 1991; Boivin 1992). In Canada it ranges
from southern Alberta (Fernald 1950; Scoggan 1978-
1979; Boivin 1992), Ontario (A. A. Reznicek, per-
sonal communication), Quebec (Scoggan 1978-1979;
Boivin 1992), New Brunswick (Scoggan 1978-1979;
Hinds 1986; Boivin 1992), and Prince Edward Island
(Scoggan 1978-1979; Erskine et al. 1985; Boivin
1992). Previous reports from British Columbia,
Saskatchewan, and Newfoundland (e.g., Fernald
1913; Scoggan 1978-1979) have been more recently
discounted (Taylor 1983; Boivin 1992). Hinds (1986)
considers Carex tincta locally common in moist
meadows, roadside ditches, woodland borders and
clearings in New Brunswick, especially in the Saint
John River drainage system, and it is locally frequent
in large areas of Maine and New Hampshire (A. A.
Reznicek, personal communication).
In Nova Scotia Carex tincta was found to be
locally common at the edge of a mixed woodland
where it is undoubtedly native.
Carex wiegandii Mackenzie (Cyperaceae) sedge
Shelburne County, Port la Tour bog, between
Port la Tour and Cape Negro, uncommon and
local at edges of extensive bog; 30 June 1993;
M. J. Oldham 15054 (MICH, mjo).
Reznicek and Ball (1980) map this rare sedge only
from Cape Breton Island in Nova Scotia. This col-
lection extends its range to southern Nova Scotia.
The range of Carex wiegandii is centred on the Gulf
of St. Lawrence, extending inland to Ontario, and
south, very sparingly to Maine, northern
Massachusetts and Pennsylvania (Rothrock 1978;
Reznicek and Ball 1980).
Cerastium pumilum Curtis (Caryophyllaceae) Curtis’
Mouse-ear Chickweed
Annapolis County, Kejimkujik National Park,
uncommon weed on campsites in main camp-
OLDHAM AND ZINCK: VASCULAR FLORA OF NOVA SCOTIA
395
ground; 25 June 1992; M. J. Oldham 13869
(CAN, MICH, mjo). Shelburne County, Islands
Provincial Park, near Shelburne, uncommon
campground weed; 26 June 1992; M. J. Oldham
13878 (MICH, mjo); 26 June 1992; M. J.
Oldham 13885c (NSPM). Yarmouth County,
Ellenwood Lake Provincial Park, rare on camp-
sites with Aira praecox and Cerastium
fontanum; 30 June 1993; M. J. Oldham 15079
(MICH).
This small, early flowering, annual European
weed has been greatly overlooked in the North
American flora. It has been found in most states in
eastern North America, although many of these are
very recent reports (Rabeler and Cusick 1994).
Cerastium pumilum is not reported from Canada by
Boivin (1969) or Scoggan (1978-1979), although it
is now known from several southern Ontario sites
(Morton and Venn 1984; Morton and Venn 1990; M.
J. Oldham collections at MICH and WAT).
Collections in 1992 from New Brunswick (Charlotte
County, M. J. Oldham 13913, CAN, MICH, mjo)
and Maine (Hancock County, M. J. Oldham 13920,
MICH, NEBC, mjo) appear to be first records
(Richards et al. 1983; Hinds 1986; Seymour 1982).
In Nova Scotia Cerastium pumilum was collected
from campgrounds in three widely separated parks.
This chickweed is part of a characteristic early
spring weedy flora of open, usually sandy campsite
tent-pads. Several other species discussed in this
paper are also components of this flora: Aira
caryophyllea, A. praecox, Cerastium semidecan-
drum, Erophila verna, Myosotis stricta, Veronica
peregrina, and Veronica verna. In Nova Scotia,
other species growing in this habitat include
Cerastium fontanum, Poa annua, Sagina procum-
bens, and Veronica arvensis. This habitat is greatly
overlooked by botanists, as evidenced by the fact
that five species new to the Nova Scotia flora were
discovered in park campgrounds in 1992 and 1993.
These species are best sought in the early spring,
since most are small, early flowering annuals, and
they wither up and disappear later in the season. In
addition, heavy trampling by humans makes them
much more difficult to detect in summer. Some of
these species may be dispersed by camping equip-
ment, as was speculated for Veronica verna in
Ontario (Crins et al. 1987).
Cerastium semidecandrum L. (Caryophyllaceae)
Small Mouse-ear Chickweed
Inverness County, Cape Breton Highlands
National Park, Chetticamp Campground, locally
common on campsites with Cerastium
fontanum; 21 June 1992; M. J. Oldham 13848
(MICH, WAT, mjo). Inverness County, Cape
Breton Highlands National Park, Corney Brook
Campground, rare on campsites; 21 June 1992;
396
M. J. Oldham 13853 (CAN, MICH). Shelburne
County, Islands Provincial Park, near Shelburne,
rare campground weed; 26 June 1992; M. J.
Oldham sight record. Victoria County, Cape
Breton Highlands National Park, Broad Cove
Campground, campsites; 22 June 1992; M. J.
Oldham 13858 (CAN, MICH, NSPM, WAT,
mjo). Yarmouth County, Ellenwood Provincial
Park, near Yarmouth, campsites; 28 June 1992;
M. J. Oldham sight record.
Scoggan (1978-1979) reports this small,
European, weedy annual only from southwestern
British Columbia and southern Ontario in Canada. It
is not listed for New Brunswick by Hinds (1986) or
Prince Edward Island by Erskine et al. (1985). The
discovery of Cerastium semidecandrum in five dif-
ferent campgrounds suggests that this plant has been
overlooked in Nova Scotia and probably other
Maritime provinces, as it has been elsewhere in east-
ern North America (Rabeler and Cusick 1994).
Erophila verna (L.) Chevall. (Brassicaceae)
Whitlow-grass
Inverness County, Cape Breton Highlands
National Park, Corney Brook Campground, rare
on campsites; 21 June 1992; M. J. Oldham
13852 (CAN, MICH, NSPM). Shelburne
County, Islands Provincial Park, near
Shelburne, rare campground weed; 26 June
1992; M. J. Oldham 13885b (NSPM).
Yarmouth County, Ellenwood Provincial Park,
near Yarmouth, rare campground weed; 28 June
1992; M. J. Oldham 13989 (MICH, NSPM).
Whitlow-grass is a small annual or winter-annual
European plant widely naturalized in North America.
It is treated by some authors (e.g., Scoggan 1978-
1979; Gleason and Cronquist 1991) as Draba verna.
In Canada, Scoggan (1978-1979) reports it from
British Columbia, Ontario, and Quebec. The only
published report from the Canadian Maritime
Provices is from Albert County, New Brunswick,
where it was found in a Fundy National Park camp-
ground (Hinds 1986). Specimens at NSPM show this
early-spring ephemeral as being in Nova Scotia since
at least 1983, when it was collected at Smiley’s
Provincial Park, Hants County (Butters and Moores,
NSPM). Since that time, collections have been made
from Annapolis Royal, Annapolis County (Wilson in
1989, NSPM) and Kentville Research Station, Kings
County (Zinck in 1992; NSPM). At each site it is
locally common in disturbed soil. This Whitlow-
grass has been seen at Golden Arm Campground,
near Sydney, Cape Breton County (Sullivan-Fraser,
personal communication 1995); a documenting spec-
imen was lost in transit.
Hypericum gentianoides (L.) Britton, Sterns & Pogg.
(Guttiferae) Orange-grass
Lunenberg County, southeast side of Hwy. 103
THE CANADIAN FIELD-NATURALIST
Voli i
between the highway and Fancy Lake, along
railtracks, ca. 5 km southwest of Bridgewater,
dry sand and gravel along railway tracks,
frequent in a very small area with Bulbostylis
capillaris; 16 September 1990; A. A. Reznicek
8732 (MICH).
Orange-grass or pineweed is a diminutive St.
John’s-wort previously known in Canada only from
the Windsor area of southwestern Ontario (Gillett
and Robson 1981; Argus et al. 1982-1987). A pub-
lished report from southwestern Quebec (Scoggan
1978-1979) is apparently based on a label error
(National Museum of Canada Rare Plant Project file,
copies on file at Natural Heritage Information
Centre). Argus and Pryer (1990) consider the species
a rare plant in Canada.
The Nova Scotia population is probably adventive
since suitable habitat for the species was not present
in the vicinity of the collection, which was made
along a railway (A. A. Reznicek, personal communi-
cation). Hypericum gentianoides does however range
into New England as a native plant (Seymour 1982;
Richards et al. 1983). See addendum page 398.
Myosotis stricta Link (Boraginaceae) Blue Scorpion-
grass
Shelburne County, Islands Provincial Park, near
Shelburne, rare campground weed; 26 June
1992; M. J. Oldham 13882 (MICH, NSPM).
Victoria County, Cape Breton Highlands
National Park, Broad Cove Campground, camp-
sites; 22 June 1992; M. J. Oldham 13857
(CAN, MICH, NSPM, mjo). Yarmouth County,
Ellenwood Provincial Park, near Yarmouth,
rare campground weed; 28 June 1992; M. J.
Oldham 13897 (MICH).
Blue Scorpion-grass is another small European
annual or winter-annual which is locally introduced
in dry waste places in northeastern North America
(Gleason and Cronquist 1991). Some authors call
this plant Myosotis micrantha (e.g., Scoggan 1978-
1979; Gleason and Cronquist 1991). Previous
Canadian reports are from southern British
Columbia, Alberta, Ontario, Quebec (Scoggan 1978-
1979), and New Brunswick (Hinds 1986). In New
England, Seymour (1982) reports it only from
Massachusetts. A 1992 Maine collection (Hancock
County, M. J. Oldham 13925, MICH) appears to be
the first from that state (Seymour 1983; Richards et
al. 1983).
Veronica peregrina L. (Scrophulariaceae) Purselane
Speedwell
Victoria County, Cape Breton Highlands
National Park, Broad Cove Campground, camp-
sites; 22 June 1992; M. J. Oldham 13861
(CAN, MICH).
Roland and Smith (1969) mention only a single
1997
Nova Scotia record of Purslane Speedwell from
Nova Scotia, citing the report of Smith and Erskine
(1954) from Point Pleasant Park in Halifax. The
specimen cited above extends the known Nova
Scotia range north to Cape Breton Island. Veronica
peregrina is widespread in North America. It is
known throughout the southern part of Canada, from
coast to coast (Scoggan 1978-1979).
Veronica verna L. (Scrophulariaceae) Spring
Speedwell
Shelburne County, Islands Provincial Park, near
Shelburne, rare campground weed; 26 June
1992; M. J. Oldham 13881 (MICH, NSPM).
Victoria County, Cape Breton Highlands
National Park, campsites; 22 June 1992; M. J.
Oldham 13856 (CAN, MICH, NSPM).
Gleason and Cronquist (1991) mention the occur-
rence of this small European weed only from south-
ern Ontario and Michigan in northeastern North
America. Hinds (1986) has reported it from road-
sides in New Brunswick, and it was collected in
1992 from a New Brunswick campground (Charlotte
County, M. J. Oldham 13914, CAN, MICH). A 1992
Lewis County, New York collection (M. J. Oldham
13787, MICH, NYSM) appears to be the first from
New York State (Mitchell 1986). Crins et al. (1987)
found Veronica verna to be quite common and
widespread in southern Ontario, particularly in
provincial park campgrounds.
Acknowledgments
We thank Tony Reznicek (University of Michigan,
Ann Arbor) for verifying or identifying several of the
collections reported here, allowing us to report his
discovery of Hypericum gentianoides and providing
encouragement and useful comments on the
manuscript. Stephen Darbyshire (Agriculture Canada,
Ottawa) identified or verified numerous Nova Scotia
grass collections, including the Aira collections
reported here. Rich Rabeler (University of Michigan,
Ann Arbor) verified the Cerastium collections.
Literature Cited
Argus, G. W., and K. M. Pryer. 1990. Rare vascular
plants in Canada: our natural heritage. Canadian
Museum of Nature, Ottawa, Ontario. 191 pages.
Argus, G. W., K. M. Pryer, D. J. White, and C. J.
Keddy. Editors. 1982-1987. Atlas of the rare vascular
plants of Ontario. Four parts. National Museum of
Natural Sciences, Ottawa, Ontario. (looseleaf).
Boivin, B. 1969. Enumération des plantes du Canada.
Provancheria Number 6, Université Laval, Québec. 185
pages.
Boivin, B. 1980. Survey of Canadian herbaria.
Provancheria Number 10, Université Laval, Québec.
187 pages.
Boivin, B. 1992. Les Cypéracées de l’est du Canada.
Provancheria Number 25, Université Laval, Québec.
230 pages.
OLDHAM AND ZINCK: VASCULAR FLORA OF NOVA SCOTIA
2h
Bouchard, A., D. Barabé, M. Dumais, and S. Hay. 1983.
The Rare Vascular Plants of Quebec. Syllogeus Number
48, National Museum of Natural Sciences, Ottawa,
Ontario. 75 pages.
Bryson, C. T. 1980. A Revision of the North American
Carex Section Laxiflorae (Cyperaceae). Ph. D. disserta-
tion, Mississippi State University. 191 pages.
Catling, P. M., V. R. Brownell, and B. Freedman. 1984.
Silver Hairgrass, Aira caryophyllea, new to eastern
Canada, and other notable records from Seal Island,
Nova Scotia. Canadian Field-Naturalist 98: 248-249.
Catling, P. M., S. Porebski, and B.S. Brookes. 1995.
Plants of the Maritimes: a bibliography for agriculture,
resource management, landscape planning and biologi-
cal research. The CanaColl Foundation, K. W. Neatby
Building, Central Experimental Farm, Ottawa.
Cayouette, J., and P. M. Catling. 1992. Hybridization in
the genus Carex with special reference to North
America. Botanical Review 58: 351-440.
Crins, W. J., and P. W. Ball. 1989. Taxonomy of the
Carex flava complex (Cyperaceae) in North America
and northern Eurasia. II. Taxonomic treatment. Canadian
Journal of Botany 67: 1048-1065.
Crins, W. J., D. A. Sutherland, and M. J. Oldham. 1987.
Veronica verna (Scrophulariaceae), an overlooked ele-
ment of the naturalized flora of Ontario. Michigan
Botanist 26: 161-165.
Erskine, D.S., P. M. Catling, and R. B. MacLaren.
1985. The Plants of Prince Edward Island with new
records, nomenclatural changes, and corrections and
deletions. Publication 1798, Research Branch,
Agriculture Canada, Ottawa, Ontario. xxii + 272 pages.
Erskine, J.S. 1958. A study of the Tusket Islands.
Proceedings of the Nova Scotia Institute of Science 24:
271-296.
Fernald, M.L. 1913. Carex tincta a valid species.
Rhodora 15: 186-187.
Fernald, M. L. 1950. Gray’s Manual of Botany. Eighth
edition. D. Van Nostrand Company, New York. 1632
pages.
Gillett, J. M., and N. K. B. Robson. 1981. The St.
John’s-worts of Canada (Guttiferae). Publications in
Botany 11, National Museum of Natural Sciences,
Ottawa, Ontario. 44 pages.
Gleason, H. A., and A. Cronquist. 1991. Manual of
Vascular Plants of Northeastern United States and
Adjacent Canada. Second edition. New York Botanical
Garden, Bronx, New York. 910 pages.
Hinds, H. R. 1986. Flora of New Brunswick. Primrose
Press, Fredericton, New Brunswick. 885 pages.
Holmgren, P. K., W. Keuken, and E. K. Schofield. 1981.
Index Herbariorum. Part I. The herbaria of the world.
Seventh edition. Bohn, Scheltema and Holkema,
Utrecht. 452 pages.
Hubbard, W. A. 1969. The Grasses of British Columbia.
Handbook Number 9, British Columbia Provincial
Museum, Victoria, British Columbia. 205 pages.
Jermy, A. C., A. O. Chater, and R. W. David. 1982.
Sedges of the British Isles. Handbook Number 1,
Botanical Society of the British Isles, London, England.
268 pages.
Kral, R. 1971. A treatment of Abilgaardia, Bulbostylis
and Fimbristylis (Cyperaceae) for North America. Sida
4: 57-227.
Maher, R. V., D. J. White, G. W. Argus, and P. A.
Keddy. 1978. The rare vascular plants of Nova Scotia.
398
Syllogeus Number 18, National Museum of Natural
Sciences, Ottawa, Ontario. 37 pages.
Mitchell, R.S. 1986. A checklist of New York State
plants. Bulletin Number 458, New York State Museum,
Albany, New York. 272 pages.
Morton, J. K., and J. M. Venn. 1984. The flora of
Manitoulin Island. Second revised edition. University of
Waterloo Biology Series Number 28, Waterloo, Ontario.
106 pages.
Morton, J. K., and J. M. Venn. 1990. A checklist of the
flora of Ontario vascular plants. University of Waterloo
Biology Series Number 34, Waterloo, Ontario. 218
pages.
Oldham, M. J. 1994. Natural heritage resources of
Ontario: Rare vascular plants. Natural Heritage
Information Centre, Peterborough, Ontario. 48 pages.
Rabeler, R. K., and A. W. Cusick. 1994. Comments on
some introduced Caryophyllaceae of Ohio and nearby
states. Michigan Botanist 33: 95-108.
Reznicek, A. A. 1989. New England Carex (Cyperaceae):
taxonomic problems and phytogeographical considera-
tions. Rhodora 91: 144-152.
Reznicek, A. A., and P. W. Ball. 1980. The taxonomy of
Carex Section Stellulatae in North America north of
Mexico. Contributions from the University of Michigan
Herbarium 14: 153-203.
Addendum
THE CANADIAN FIELD-NATURALIST
Vol. 111
Richards, C. D., F. Hyland, and L. M. Eastman. 1983.
Revised check-list of the vascular plants of Maine.
Bulletin Number 11, Josselyn Botanical Society. 73
pages.
Roland, A. E., and E. C. Smith. 1969. The Flora of Nova
Scotia. Nova Scotia Museum, Halifax, Nova Scotia. 746
pages.
Rothrock, P. E. 1978. Carex wiegandii Mackenzie
(Cyperaceae): new for Pennsylvania. Rhodora 80: 451.
Scoggan, H. J. 1978-1979. The flora of Canada. Four
parts. Publications in Botany Number 7, National
Museum of Natural Sciences, Ottawa, Ontario. 1711
pages.
Seymour, F.C. 1982. The flora of New England. Second
edition. Phytologia Memoirs V, Plainfield, New Jersey.
611 pages.
Smith, E. C., and J. S. Erskine. 1954. Contributions to
the flora of Nova Scotia. Rhodora 56: 242-252.
Taylor, T.M.C. 1983. The sedge family of British
Columbia. British Columbia Provincial Museum
Handbook Number 43, Victoria, British Columbia. 375
pages.
Received 12 March 1996
Accepted 22 November 1996
We recently became aware of the following two
additional Nova Scotia records of Hypericum gen-
tianoides. Lunenburg County, railway path at Italy
Cross; 25 September 1996; M. Zinck (NSPM).
Shelburne County, west side of Hwy. 103, 1.6 km
southwest of Granite, roadside ditch with Hypericum
canadense and Radiola linoides, 27 September 1979,
D. F. Brunton 2093 and H. L. Dickson (MICH,
mixed collection with Hypericum canadense).
17 February 1997 MICHAEL J. OLDHAM
Vegetative Concealment, Proximity to Trails, and Predator Activity as
Relative Factors Affecting Nest Success and Egg Loss in Spruce
Grouse, Dendragapus canadensis
ROBERT G. D’ EON
Kokanee Forests Consulting Ltd., 201-625 Front Street, Nelson, British Columbia V1L 4B6
D’Eon, Robert G. 1997. Vegetative concealment, proximity to trails, and predator activity as relative factors affecting nest
success and egg loss in Spruce Grouse, Dendragapus canadensis. Canadian Field-Naturalist 111(3): 399-402.
Sixty-seven nests of Spruce Grouse (Dendragapus canadensis) were located during five nesting seasons in north-central
Ontario. There was no difference in success of nests > 25 m from a trail than those < 25 m away. Nests with high predator
activity indices had lower success than those with low indices. There was no correlation between nest concealment and
success. I suggest that the importance of concealment to nest predation diminishes as vegetative cover increases from
patchy and open to uniformly dense. Further, I submit different types of predators attracted to varying cover types, special-
izing in either visual or olfactory cues, as the proximal cause.
Key Words: Spruce Grouse, Dendragapus canadensis, nest success, predator activity.
Nest destruction by predators is the primary
source of production loss in grouse (Tetraoninae)
(Johnsgard 1973; Zwickel 1975; Boag et al. 1979;
Bergerud and Gratson 1988). Further, Boag et al.
(1979) indicated that most deaths of young Spruce
Grouse (Dendragapus canadensis) occur at the egg
state and are caused primarily by predators. Some
believe that breeding success and fledging survival
determine spring population numbers (Bergerud
1970; Myrberget 1972). Despite this, the relative
importance of factors influencing nest vulnerability
is not well understood.
Boag et al. (1979) suggested that productivity loss
due to predation in Spruce Grouse was inversely
related to population density. Several studies have
since refuted this hypothesis by showing no relation-
ship between egg loss and female density (Redmond
et al. 1982; Smyth and Boag 1984; Boag et al. 1984).
Compelling evidence has shown nest success to be
directly proportional to vegetative concealment of
nests (Wallestad and Pyrah 1974; Keppie and
Herzog 1978; Redmond et al. 1982). However,
Owing to inconsistencies in their own data (i.e., eggs
in well-concealed nests that did not hatch and vice-
versa), Redmond et al. (1982) concluded that some
factor(s) other than concealment alone affects nest
success. Boag et al. (1984) used artificial nests to
investigate egg loss among Spruce Grouse; Red
Squirrels (Tamiasciurus hudsonicus) were the princi-
pal predator destroying clutches. The authors
hypothesized that the probability of nest destruction
is a function of nest placement relative to activity
centers of predators.
These studies concentrated largely on either nest
concealment or predator activity as a factor affecting
egg loss, neglecting the possibility of a combined
effect varying with habitat conditions. Furthermore,
the validity of conclusions based on studies using
artificial nests has been questioned (Roper 1992).
The objectives of my study were to identify the rela-
tive importance of vegetative concealment, predator
activity in the immediate vicinity of nests, and prox-
imity of nests to trails, to nest success and egg loss
among natural nests of Spruce Grouse.
Study Area
Spruce Grouse nests were located in a 266-ha area
approximately 148 km northwest of Sudbury, Ontario
(47°36'N, 82°45’W). The area consisted primarily of
11 to 23 year-old Jack Pine (Pinus banksiana) hand-
planted or aerially-seeded plantations. Jack Pine
heights varied from 2-3 to 7-9 m in the younger and
older plantations, respectively. Ground vegetation in
high-density Jack Pine areas was primarily uniformly
dense blueberry (Vaccinium spp.), Sweet-Fern
(Comptonia peregrina), Trailing Arbutus (Epigaea
repens), Wintergreen (Gaultheria procumbens), and
lichens (Cladina spp.). In clearings and low-density
Jack Pine areas, ground vegetation was primarily
dense grasses, sedges (Carex spp.), raspberry (Rubus
spp.), and lichens. There were numerous natural and
human-made trails varying in prominence and width
from 0.5 to 3.0 m. Topography was flat with a total
elevational range of approximately 10 m.
Methods
Sixty-seven nests were located with the aid of
pointing dogs during five nesting seasons (1985-
1988, 1990) prior to and during incubation and post-
hatch. Nests were found in the course of on-going
Spruce Grouse population studies (e.g., Keppie
1992). Field procedures were similar to those used
earlier in New Brunswick and Alberta (e.g., Keppie
1987), with 2-4 people searching for grouse daily
with trained dogs from April through to at least
October. The study area was searched repeatedly on
399
400
a sectional rotation basis. Population counts were
considered complete (Keppie 1992).
Nests were monitored (usually visited every sec-
ond day) until empty (i.e., eggs hatched, nest aban-
doned by female, or destroyed by predator). At each
nest, vegetative concealment of the nest was mea-
sured, using a systematized visual estimate (Keppie
and Herzog 1978) once the nest was empty, to reduce
site disturbance during incubation. Concealment was
measured horizontally and vertically. For horizontal
concealment, the nest was viewed in each cardinal
direction from 10 m at 0.5 m above ground level.
From each vantage, an estimate of horizontal con-
cealment was scored to a maximum of 25. If a nest
was directly at the base of a tree, it was automatically
scored 25 points for the effect of the tree bole. The
four scores were summed to a maximum of 100.
Vertical concealment was estimated from 1.8 m
directly above the nest for a single score to a maxi-
mum of 100. Horizontal and vertical scores were
combined for a concealment index to a maximum of
200. Redmond et al. (1982) used several objective
and subjective methods to estimate nest concealment
and concluded that this visual technique is of similar
accuracy to any other method they evaluated.
An index of mammalian activity around each of 13
nests was obtained in 1986. A pair of sand transects
was established, once nests were empty, radiating
north and west from the nests. Transects were con-
structed by cutting out existing ground vegetation and
scraping out 2-3 cm of soil within a rectangular area
of 30 cm by 6 m. Sand was imported and placed
within the transect until flush with ground level. Care
was taken to minimize disturbance of vegetation
around the nest and transects. The number and
species of mammalian tracks (Murie 1974) were
recorded every second day for one month, and tran-
sects were swept clear on each visit. In 1985 and
1986, distances from nests to the nearest natural or
man-made trail were recorded.
The fate of each nest was determined by examining
egg shell fragments and other evidence in and around
the nest. A successfully hatched egg was one which
had one half of the egg shell cupped inside the other
half in a relatively symmetrical fashion; thus allowing
an accurate count of eggs hatched. Abandoned nests
were identified by the presence of cold eggs at con-
secutive visits without sighting the female on or near
the nest, and were treated as unsuccessful nests. For
depredated nests, the type of predator was determined
by evidence of shell breakage and nest disturbance
(Bump et al. 1947; Boag et al. 1984).
Both nest success (defined as = 1 egg hatched
[Keppie 1982]) and egg loss (individual eggs that
fail to hatch) were used as measures of nest success.
This was done to account for Red Squirrels, a major
nest predator on the study area, which tend to prey
on individual eggs rather than entire nests (Boag et
al. 1984), thus accounting in the analysis for depre-
THE CANADIAN FIELD-NATURALIST
Vol. 111
dated eggs that belonged to successful nests.
Statistical tests of independence for comparing rates
of egg and nest success were made with log-likeli-
hood non-parametric tests using William's correction
(G,,,) (Zar 1984). Student's-t parametric tests were
used to compare average values of nest parameters
(all tests of homogeneity P > 0.05).
Results
A total of 67 nests was found over the five years.
In 37 (55%) of these, eggs were depredated; in 27
(40%), eggs were not depredated; in 2 (3%), eggs
were destroyed by research dogs; and in 1 (2%), the
nest contained eggs of undetermined fate. Of nests
containing depredated eggs, 16/37 (43%) were
preyed upon by a large mammal (e.g., Red Fox
[Vulpes vulpes], Black Bear [Ursus americanus],
Striped Skunk [Mephitis mephitis]), 9 (24%) by Red
Squirrels, 2 (5%) by another rodent, 1 (3%) by a
bird, and 9 (24%) by unknown animals. For those
nests where the outcome (successful or not) was
known, 51% (33/65) were successful.
Concealment
The average nest concealment rating was 104 (n =
67, SD = 41.2, range = 35-191). In nests rated < 104
(low concealment) 44% (74/169) of eggs laid
hatched; in nests rated > 104, 59% (90/153) of eggs
hatched, a similar proportion in both concealment
ratings (G,,; = 2:4, df = 1, P = O.93)0 NEst Saeecss
was also similar between low and high concealment
ratings (< 104 = 54% [19/35 nests], > 104 = 57%
[17/30 nests]; G4 = (0.3, df =1, P= 062): There was
also no difference between average concealment rat-
ings for successful nests (x = 109, n = 33, SD = 39.7)
and unsuccessful nests (x = 102, n = 32, SD = 43.0)
(t = 0.73, df = 62, P = 0.47). Further, when classed
by concealment ratings (Table 1), there were no dif-
ferences between proportions of successful and
unsuccessful nests in each category (all G, aj < 0.44,
df = P= 025):
Predator Activity Around Nests
Mammalian tracks (n = 208) were recorded in 12
visits to each nest; of these, 191 (92%) were attribut-
ed to Red Squirrels, 10 (5%) to other small rodents,
4 (2%) to skunks, and 3 (1%) to Red Fox. The aver-
age total track count per nest was 17.3 tracks (n =
12, SD = 16.8, range = 2-37). This value was used to
differentiate between low and high predator-activity
nests. In nests with < 17.3 track counts (1.e., low
activity), 82% (31/38) of eggs laid hatched, signifi-
cantly more than the 23% (6/26) that hatched in
nests with > 17.3 track counts (G,,, = 6.8, df =1,
P=0.01). Similarly, 6 of 7 nests with low predator
activity were successful, whereas 3 of 5 nests with
high predator activity were successful (G, dj = 0.6,
df= 1, P=0.46).
Proximity to Trails
In nests < 25 m from a trail, 63% (26/41) of eggs
hatched, similar to nests > 25 m from a trail (42%
1997 D’ Eon: NEST SUCCESS AND EGG Loss IN SPRUCE GROUSE 401
[39/92]) (G,,, = 1.60, df = 1, P = 0.21). Nor was
there any difference in success rates between nests
< 25 m from a trail (6/9 [67%]) and > 25 m from a
trail (7/16 [44%]) (G,4, = 0:35; dik= P= 055)iihe
average distance from a trail for depredated nests
(66 m , n = 10, SD = 65.0) was greater than that for
other nests (26.8 m, n = 16, SD = 24.60) (t = 2.18,
df = 25,0P =0:04):
Unsuccessful
4
11
10
20
45
(n = 64)
Discussion
Proximity of nests to trails did not increase the
chance of a nest being depredated, contrary to the
intuitive assumption that mammals that use trails
would more likely discover nests close to trails. Scat
collections confirmed that mammals used trails on
the study area during the nesting season (D. M.
Keppie, unpublished data), supporting the hypothesis
that predators often use trails for reasons other than
foraging for prey as suggested by Boag et al. (1984).
Sand transects were an easy and effective way to
index mammalian activity around nests and warrant
further use. Results of sand-transect experiments
indicated that high predator activity in the vicinity of
nests increases the probability of nest failure and egg
loss, supporting Boag et al.’s (1984) hypothesis. Red
Squirrels accounted for the majority of tracks, are
probably opportunistic predators, and will revisit a
nest once they have discovered it (R. D’Eon and D.
Keppie, personal observations). This behaviour pat-
tern would likely compound the effects of nest place-
ment in a Red Squirrel activity centre by increasing
the probability of all eggs in a nest being preyed
upon once discovered. Therefore, in these situations,
predator abundance (i.e., Red Squirrels) and disper-
sion of activity centers would likely be the most
important factors influencing nest success.
A regional comparison (Table 1) reveals a con-
vincing correlation between magnitude of visual
concealment and nest success of Spruce Grouse
among studies. New Brunswick (N.B.) had the high-
est proportion of nests rated >100 points for conceal-
ment (72%, n = 36) and proportion of successful
nests (81%, n = 36), followed by Ontario (48% and
51% respectively, n = 65), and Alberta (19% and
30% respectively, n = 64). Despite this correlation
however, concealment of nests in Ontario was unre-
lated to nest success and egg loss. The disparity
between these results and past studies is likely due to
area differences in vegetative cover. All nests found
by Keppie and Herzog (1978) in Alberta were in
Lodgepole Pine (Pinus contorta) forests character-
ized by sparse, low ground vegetation, whereas
Redmond et al. (1982) found most nests in open
spruce —pine forests with moderately dense and
patchy ground cover. However, the Young Jack Pine
forests of the present study typically had dense, uni-
form ground vegetation and contrasted greatly with
the earlier studies.
Redmond et al. (1982) suggested that differences
in mechanisms of locating prey by predators, a func-
D.c. franklinni (Alta.)*
oO SF SO GF | ON
—
Successful
66)
Unsuccessful
13
14
5
(0)
3)
D.c. canace (Ont.)°
(n
Successful
18
11
4
0)
33
VA Aa oy i>
Unsuccessful
D.c. canace (N.B.)°
(re 36)
TABLE |. Concealment and success for 166 Spruce Grouse nests of Dendragapus canadensis canace and D.c. franklinii from New Brunswick (N.B.), Ontario (Ont.), and Alberta
Successful
2
(Alta.), 1970-1990, using a subjective visual concealment method after Keppie and Herzog (1978)*.
’One additional nest in N.B. had "very good" concealment but success was unknown; three unsuccessful nests that had “very good” concealment were deserted just after discovery,
prior to start of incubation.
°One additional nest in Ont. had “poor” concealment, another had “good” concealment, but both were destroyed by a dog.
‘One additional nest in Alta. had “good” concealment and two had “very poor” concealment but success was unknown.
aN.B. and Alta. data from Redmond et al. (1982)
(200 points maximum)
Very good (>100)
Good (51 - 100)
Concealment rating
Poor (26 - 50)
Very poor (< 25)
Totals
402
tion of habitat structure, could result in different suc-
cess rates for Spruce Grouse nests. Most nests in the
present study were depredated by large mammals
(Red Fox, Black Bear, Striped Skunk) that likely pri-
marily use auditory and olfactory (in this case pri-
marily olfactory) senses to locate prey (Banfield
1974), and Red Squirrels, a predator that appears to
be largely opportunistic in its search pattern. The
presence of these types of predators is likely a func-
tion of the uniformly dense ground vegetation, that
tends to conceal nests well but does little to deter
predators using non-visual cues. I suggest, that
although concealment generally contributes to nest
success, concealment in uniformly dense vegetative
cover is not as important in determining egg loss and
nest success as in areas of sparse and/or patchy
ground vegetation where predators may be primarily
using visual cues to locate nests (e.g., Keppie and
Herzog 1978). Further, I submit the importance of
nest concealment in relation to nest predation dimin-
ishes as the overall vegetative ground cover of an
area increases from patchy and open to uniformly
dense, by attracting different types of predators that
hunt either visually (as in patchy and open areas) or
by olfactory cues (as in uniformly dense areas).
The findings in this study rest on assumptions that
require validation. I urge caution in their interpreta-
tion. Activities of field workers and dogs around
nests may be a cause of unrepresentative nest infor-
mation in studies of this kind. Some have suggested
field workers at nests could be a partial cause of high
nest failure in birds (e.g., Dwernychuk and Boag
1972; Zwickel 1975). Keppie and Herzog (1978)
investigated this concern and concluded that dogs
did not promote a greater rate of Spruce Grouse nest
failure, nor was there conclusive evidence that find-
ing a nest by any means lessened its chances of
being successful.
Despite some advancement, a particularly persis-
tent question remains: how does a female grouse
choose her nest site? Bergerud and Gratson (1988)
stated that the most important reproductive decision
a female must make is where to locate her nest. This
decision should relate primarily to locating a nest
that will best avoid detection by predators. Further
work should attempt to determine the influence of
vegetative cover on the types of predators preying on
nests and how these factors influence a female’s
choice of nest sites.
Acknowledgments
I thank D. M. Keppie for his insightful criticisms
of previous drafts and permission to use data collect-
ed by himself and assistants as part of ongoing
research funded by grants from the Natural Science
and Engineering Research Council of Canada.
Literature Cited
Banfield, A.W. F. 1974. The mammals of Canada.
THE CANADIAN FIELD-NATURALIST
Vol. 111
University of Toronto Press, Toronto, Ontario. 438
pages.
Bergerud, A. T. 1970. Population dynamics of the willow
ptarmigan Lagopus lagopus alleni L. in Newfoundland
1955 to 1965. Oikos 21: 299-325.
Bergerud, A. T., and M. W. Gratson. 1988. Survival and
breeding strategies of grouse: female nesting strategies.
Pages 503-533 in Adaptive strategies and population
ecology of northern grouse. Volume II. Theory and
Synthesis. Edited by A. T. Bergerud and M. W. Gratson.
University of Minnesota Press, Minneapolis, Minnesota.
Boag, D. A., S. G. Reebs, and M. A. Schroeder. 1984.
Egg loss among Spruce grouse inhabiting lodgepole pine
forests. Canadian Journal of Zoology 62: 1034-1037.
Boag, D. A., D. H. McCourt, P. W. Herzog, and J. H.
Alway. 1979. Population regulation in spruce grouse: a
working hypothesis. Canadian Journal of Zoology 57:
2275-2284.
Bump, G., R. W. Darrow, F. C. Edminster, and W. F.
Crissey. 1947. The ruffed grouse: life history, propaga-
tion, management. Telegraph Press, Harrisburg,
Pennsylvania. 915 pages.
Dwernychuk, L. W., and D. A. Boag. 1972. How vegeta-
tive cover protects duck nests from egg-eating birds.
Journal of of Wildlife Management 37: 955-958.
Johnsgard, P. A. 1973. Grouse and quails of North
America. University of Nebraska Press, Lincoln,
Nebraska.
Keppie, D. M. 1982. A difference in production and asso-
ciated events in two races of spruce grouse. Canadian
Journal of Zoology 60: 2116-2123.
Keppie, D. M. 1987. Impact of demographic parameters
upon a population of spruce grouse in New Brunswick.
Journal of Wildlife Management 51: 771-777.
Keppie, D. M. 1992. An audio index for male spruce
grouse. Canadian Journal of Zoology 70: 307-313.
Keppie, D. M., and P. W. Herzog. 1978. Nest site charac-
teristics and nest success of spruce grouse. Journal of
Wildlife Management 42(3): 628-632.
Murie, O. J. 1974. A field guide to animal tracks. Peterson
field guide series. Houghton Mifflin Company. Boston,
Massachusetts. 375 pages.
Myrberget, S. 1972. Fluctuations in a north Norwegian
population of willow grouse. Proceedings of the
International Ornithological Congress 15: 107—120.
Redmond, G. W., D. M. Keppie, and P. W. Herzog.
1982. Vegetative structure, concealment, and nest suc-
cess at nests of two races of spruce grouse. Canadian
Journal of Zoology 60: 670-675.
Roper, J. J. 1992. Nest predation experiments with quail
eggs: too much to swallow? Oikos 65: 528-530.
Symth, K. E., and D. A. Boag. 1984. Production in spruce
grouse and its relationship to environmental factors and
population parameters. Canadian Journal of Zoology 62:
2250-2257.
Wallestad, R., and D. Pyrah. 1974. Movement and nesting
of sage grouse hens in central Montana. Journal of
Wildlife Management 38: 630-633.
Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall Inc.,
Englewood Cliffs, New Jersey. 718 pages.
Zwickel, F. C. 1975. Nesting parameters of blue grouse
and their relevance to populations. Condor 77: 423-430.
Received 18 March 1996
Accepted 5 December 1996
Cadmium, Copper, and Lead in Fish from the Lower Nelson River
System in Northern Manitoba
EVA Pip and JEFFRAY STEPANIUK
Department of Biology, University of Winnipeg, Winnipeg, Manitoba R3B 2E9
Pip, Eva, and Jeffray Stepaniuk. 1997. Cadmium, copper, and lead in fish from the Lower Nelson River system in north-
ern Manitoba. Canadian Field-Naturalist 111(3): 403-406.
Cadmium, copper and lead were examined in a catch of fishes from the Lower Nelson River system. In Lake Whitefish,
(Coregonus clupeaformis) cadmium and copper concentrations in skeletal muscle decreased in larger fish. In Northern
Pike, (Esox lucius), lead concentrations in muscle were inversely correlated with fish weight, and fish downstream of the
Limestone Dam showed higher copper concentrations than those upstream, reflecting a parallel difference in environmental
copper levels. Interspecific differences in metal concentrations were not significant. In Brook Trout (Salvelinus fontinalis),
heart tissue showed the highest cadmium concentrations, while lead was highest in the liver and ovary.
Key Words: Cadmium, copper, lead, fishes, Lake Whitefish, Stizostedion vitreum, Northern Pike, Esox lucius, Lake
Whitefish, Coregonus clupeaformis, Salvelinus fontinalis, Goldeye, Hiodon alosoides, Longnose Sucker,
Catostomus catostomus, Nelson River, Manitoba.
Heavy metal contamination of aquatic environ-
ments is of continually increasing concern as anthro-
pogenic loadings are becoming more widespread.
Since fish are the trophic components of the greatest
economic significance, published research has con-
centrated on them. While species such as Walleye
(Stizostedion vitreum), Northern Pike (Esox lucius)
and Lake Whitefish (Coregonus clupeaformis) can
survive in rather high environmental concentrations
of heavy metals (van Loon and Beamish 1977), fish
may bioaccumulate metals even at low levels
through long-term exposure via the gills and food
chain. The Nelson River system in northern
Manitoba has been intensively developed for hydro-
electric power, and many local residents, as well as
tourists, consume fish from this waterway. The
objective of the present study was to examine levels
of cadmium, copper and lead in a sample catch of
fishes from the study area.
Study Area
Sampling was conducted on the Nelson River at
two sites: above (56° 24’ 57”N, 94° 11’ 11”W) and
below (56° 31’ 59”N, 94° 03’ 22”W) the Limestone
Dam. Brook Trout (Salvelinus fontinalis) were
obtained at McMillan Creek (56° 39’56”N, 94° 31’
03”W), a tributary of the Limestone River which
drains into the Nelson River. In addition to the
Limestone Dam, two additional major hydroelectric
developments are located nearby on the Nelson
River: Long Spruce, 23 km upriver from the
Limestone Dam, and Kettle Generating Station,
16 km upstream of Long Spruce. A map of the study
area and description of the developments are given
by Pip and Stepaniuk (1992).
Materials and Methods
Walleye, Northern Pike, Lake Whitefish, Goldeye
(Hiodon alosoides) and Longnose Sucker
(Catostomus catostomus) were obtained from the last
week of May to the first week of August, 1988, using
a series of three gillnets, of stretched-mesh size rang-
ing from 3.9-12.7 cm. Each series was set 2-4 m deep
and allowed to remain overnight. Brook Trout were
obtained by angling. The total harvest consisted of: 28
Northern Pike, 14 Lake Whitefish, 3 Goldeye, 2
Brook Trout, and 1 each of Longnose Sucker and
Walleye. All of these individuals were analyzed.
Fresh weight (nearest g) and fork length
(nearest mm) were recorded. Samples of skeletal
muscle were removed below the dorsal fin (first dor-
sal fin in the case of Walleye), and frozen. Otoliths,
cleithra, and scales were collected as appropriate for
age determinations. The age of the oldest fish in the
sample (18 y) coincided with the start of construc-
tion of Limestone Dam (1975), although the generat-
ing station was still not in full operation at the time
of sampling.
Water samples were collected 10-20 cm below the
surface in acid-washed polypropylene bottles.
Surface sediments were obtained with a plastic shov-
el. All samples were frozen within 5 h of collection.
Frozen tissue and sediment samples were freeze-
dried. Particles larger than 4 mm were excluded
from sediment samples. Each sample was extracted
as three separate | g replicates. Extraction was car-
ried out by heating for | h at 60°C. with 7.5 mL con-
centrated nitric acid and 1.5 mL 70% perchloric acid.
After cooling to room temperature, 10 mL of 1%
nitric acid were added, followed by filtration through
Whatman #541 hardened ashless filter paper. The fil-
trate and two 2 mL rinses of 1% nitric acid were
403
404 THE CANADIAN FIELD-NATURALIST
Vol. 111
TABLE 1. Weight, length, age, and metal concentrations in muscle tissue of fish at the study sites. Values in parentheses are
standard errors.
Parameter Northern Lake Brook Longnose
Pike Whitefish Goldeye Trout Sucker Walleye
Wet weight (g)
x 996 (77) 951129) 2a Si(28) 1210 (240) 400 570
range 400-1900 400-2100 230-300 971-1450 — =
Fork length (mm)
x 516 (9) SISA (CID) 259)(5) 440 (10) 390 370
range 442-610 292-500 252-268 430-450 — -
Age (yr)
X 8.3 (0.5) 10.2 (1.0) - OEE) 6 9
range 5-13 5-18 — 8-11 — —
Cd (ug/g dry wt)
x 2.9 (0.2) 2.5 (0.3) Sl (OST) 3.0 (0.2) ted 2.6
range 0.4 — 5.0 0.44.2 2.8-5.1 2.8-3.1 — _
Cu (ug/g dry wt)
Xe 1.7 (0.3) te (3) 1.4 (1.1) 3.0 (0.2) 0.4 0.5
range <0.1-4.9 <0.1-2.8 0.3-3.7 2.8-3.1 - =
Pb (ug/g dry wt)
4.8 (1.2) 4.2 (1.2) 6.5 (2.4) Cl) 8.7 0.1
range <0.1-20.5 <0.1—16.4 1.9-10.2 <0.1-2.4 — —
N 28 14 3 D l 1
pooled and diluted to 30 mL with 1% nitric acid.
This extract was aspirated into an IL-151 atomic
absorption spectrophotometer (Instrumentation
Laboratory Inc., Wilmingtom, Massachusetts).
Water samples were acidified and treated with
ammonium pyrrolidinedithiocarbamate and methyl
isobutyl ketone (American Public Health
Association 1985), and the organic extract was aspi-
rated for spectrophotometry as above.
All glassware was acid-leached prior to use.
Blanks consisted of all reagents and steps in the pro-
cedure, less sample material. The standard additions
method was used to compensate for matrix absorp-
tion effects, using certified atomic absorption stan-
dards (Fisher Scientific Co., Fairlawn, New Jersey).
Statistical tests were carried out using SPSS
(SPSS, Inc., Chicago, Illinois). The critical signifi-
cance level was p = 0.05.
Results
Metal levels in whole water samples were low,
with concentrations of approximately 0.01 ug/L for
cadmium, and up to maxima of 7 ug/L for copper
and lead. Bottom sediment concentrations were
higher, with mean values (+ S.E.) of 1.85 (0.25), 6.3
(0.8) and 30 (3) ug/g dry weight for cadmium
(n=23), copper (n=23) and lead (n=13) respectively.
All metals showed great variation within individu-
al species Table 1), and ANOVA showed no signifi-
cant interspecific differences in skeletal muscle con-
centration among Northern Pike, Lake Whitefish and
Goldeye. Unpaired T-tests showed that concentra-
tions in muscle tissue were significantly (p<0.001)
higher than in sediments for cadmium, but consis-
tently (p<0.001) lower than in sediments for copper
and lead. In Northern Pike, cadmium and copper
concentrations were linearly correlated with each
other (p = 0.018).
Northern Pike and Lake Whitefish were examined
in greater detail. Lead in muscle was inversely corre-
lated with body wet weight in Northern Pike (p=
0.007). In Lake Whitefish, copper was inversely cor-
related with weight, fork length (p = 0.02) and age
(0.005), while cadmium was inversely correlated
with fork length (0.001). Correlations for lead and
copper were most pronounced when age, weight and
length were log transformed.
T-tests comparing Northern Pike upstream and
downstream of Limestone Dam showed no significant
differences with respect to weight, length or age of
fish in the samples, and no significant differences for
cadmium or lead concentrations in muscle. However
copper concentrations were significantly greater in the
downstream catch (p<0.001). Lake Whitefish could
not be compared, as this species was present only in
samples taken downstream of the dam.
The two Brook Trout from McMillan Creek were
dissected and individual organs were examined
(Table 2). In both specimens, heart tissue showed the
highest cadmium levels, but the greatest lead con-
centrations were observed in liver and ovaries; cop-
per was also high in these organs in the old female.
Discussion
Intraspecific variation of metal concentrations in
muscle tissue was quite large; similar variability has
1997
PIP AND STEPANIUK: CADMIUM, COPPER, AND LEAD IN FISH
405
TABLE 2. Metal concentrations in organs of Brook Trout, as ug/g dry weight of tissue. A = Female, 11 years old, 1450 g
wet weight, 45 cm fork length. B = Male, 8 years old, 971 g wet weight, 43 cm fork length.
Tissue Individual Cd
Skeletal muscle A 2.8
B 31
Bone A Sell
B 6.6
Scales A 1.6
B DD
Brain A 35)
B 2.9
Liver A 3.0
B Dal
Gills A 3.0
B 3.9
Heart A 6.7
B 1
Reproductive organs A Dol
B 43
been observed by Pagenkopf and Neuman (1974),
Mathis et al. (1979) and Wren et al. (1983).
Concentration factors compared to sediment were
greatest for cadmium, and least for lead. Similar
trends have been reported by Enk and Mathis (1977)
for a stream ecosystem. Johnson (1987) found that
lead showed the lowest biomagnification potential of
nine trace metals examined in whole fish.
No significant differences in muscle metal levels
were seen between carnivorous Northern Pike and
omnivorous Lake Whitefish. This is consistent with
the results of other workers (e.g., Mathis and
Cummings 1973; Hutchinson et al. 1975; Johnson,
1987) who reported that cadmium, copper and lead
levels in muscle are not greater in carnivorous fishes.
In Lake Whitefish, copper concentrations per unit
muscle weight decreased in older, heavier and larger
fish, and cadmium decreased with size as well. In
Northern Pike, lead decreased with body weight, but
copper showed no significant relationship with age
or weight in this species. This finding supported that
of Wren et al. (1983) for Northern Pike. In other
species, Mathis and Kevern (1975) and Mathis et al.
(1979) found no correlation between muscle cadmi-
um or lead and fish length and weight. Thus relation-
ships between metal uptake and growth appear to
vary greatly with species. Cadmium and lead con-
centrations have been shown to decrease with
increasing age in aquatic arthropods (Jop and Wojtan
1982) and freshwater gastropods (Pip 1992) and
mussels (Pip 1995).
Northern Pike muscle tissue showed a substantial
difference for copper above and below Limestone
Dam, and appeared to reflect differences in environ-
mental copper concentrations. Pip and Stepaniuk
Cu Pb
Sal 2.4
DES <().1
By) 30
DD Di
9.9 69
Wail 38°
6.3 30
2.6 2.4
107 203
7.9 13)7/
4.7 24
6.6 28
MBS 52
18.1 <0)
71 123
14.4 39
(1992) found significantly higher copper levels in sed-
iments downstream, compared to upstream, of
Limestone Dam. Macrophytes also showed higher
copper concentrations downstream of the dam. The
reason for this effect is not known and was not related
to differences in sediment particle size distribution.
In Brook Trout, muscle tissue showed compara-
tively low levels of copper and lead, compared to
other organs. Mathis et al. (1979) have reported large
variation in lead concentrations for the same organs
among individual fish. High levels of cadmium and
copper have been found in liver and kidney tissue of
some species (Hutchinson et al. 1975; Badsha and
Goldspink 1982; Bendell-Young et al. 1986), but
cadmium and lead have been reported to be greatest
in heart tissue in fish studied by Mathis et al. (1979).
Heart tissue showed the highest cadmium concentra-
tions in the Brook Trout in the present study.
Concentrations of cadmium and copper in the
liver have been reported to increase with age
(McFarlane and Franzin 1980), and accumulation is
proportional to rate of fish growth (Bendell- Young
et al. 1986). McFarlane and Franzin (1980) suggest-
ed that accumulation in the liver may be associated
with metalothioneins, metal-binding proteins that act
as sequestering agents. These workers also reported
high copper levels in female gonads, which was
observed in the present study as well.
In many pollution monitoring programs, skeletal
muscle is the only fish tissue examined, but this tis-
sue contains the lowest metal levels. Whole fish in
turn tend to have much lower metal body burdens
than plankton or bottom dwelling organisms (Mathis
and Cummings 1973; Namminga et al. 1974). Fish
skeletal muscle, therefore, may not be a good indica-
406
tor of ecosystem health. Similarly, prediction of
metal concentrations in fish is difficult because of
the large number of variables which may influence
metal uptake, for example growth rates, environmen-
tal calcium concentrations (McFarlane and Franzin
1980), and possible effects of stress conditions such
as parasites, disturbance, competition, and presence
of other pollutants.
Literature Cited
American Public Health Association. 1985. Standard
methods for the examination of water and wastewater.
Washington, D.C. 1268 pages.
Badsha, K. S., and C. R. Goldspink. 1982. Preliminary
observations on the heavy metal content of four species
of freshwater fish in NW England. Journal of Fisheries
Biology 21: 251-267.
Bendell-Young, L. I., H. H. Harvey, and J. F. Young.
1986. Accumulation of cadmium by white suckers
(Catostomus commersoni) in relation to fish growth and
lake acidification. Canadian Journal of Fisheries and
Aquatic Sciences 43: 806-811.
Enk, D., and B. J. Mathis. 1977. Distribution of cadmium
and lead in a stream ecosystem. Hydrobiologia 52:
153-158.
Hutchinson, T. C., A. Fedorenko, J. Fitchko, A. Kuja, J.
van Loon, and J. Lichwa. 1975. Movement and com-
partmentation of nickel and copper in an aquatic ecosys-
tem. Pages 89-105 in Trace Substances in
Environmental Health, Volume IX. Edited by D.D.
Hemphill, University of Missouri, Columbia.
Johnson, M. G. 1987. Trace element loadings to sedi-
ments of fourteen Ontario lakes and correlations with
concentrations in fish. Canadian Journal of Fisheries and
Aquatic Sciences 44: 3-13.
Jop, K., and K. Wojtan. 1982. Concentrations of cadmi-
um and lead in the body of some macrobenthos species
from five streams of southern Poland. Acta
Hydrobiologica 24: 197-210.
Mathis, J., and T. F. Cummings. 1973. Selected metals
in sediments, water, and biota in the Illinois River.
Journal of the Water Pollution Control Federation 45:
1573-1583.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Mathis, J., and N. R. Kevern. 1975. Distribution of mer-
cury, cadmium, lead and thallium in a eutrophic lake.
Hydrobiologia 46: 207-222.
Mathis, J., T. F. Cummings, M. Gower, M. Taylor, and
C. King. 1979. Dynamics of manganese, cadmium, and
lead in experimental power plant ponds. Hydrobiologia
67: 197-206.
McFarlane, G. A., and W. G. Franzin. 1980. An exami-
nation of Cd, Cu, and Hg concentrations in livers of
northern pike, Esox lucius, and white sucker,
Catostomus commersoni, from five lakes near a base
metal smelter at Flin Flon, Manitoba. Canadian Journal
of Fisheries and Aquatic Sciences 37: 1573—1578.
Namminga, H.E., J. E. Scott, and S. L. Burks. 1974.
Distribution of copper, lead and zinc in selected compo-
nents of a pond ecosystem. Proceedings of the
Oklahoma Academy of Sciences 54: 62-64.
Pagenkopf, G. K., and D. R. Neuman. 1974. Lead con-
centrations in native trout. Bulletin of Environmental
Contamination and Toxicology 12: 70-75.
Pip, E. 1992. Cadmium, copper and lead in gastropods of
the Lower Nelson River system, Manitoba, Canada.
Journal of Molluscan Studies 58: 199-205.
Pip, E. 1995. Cadmium, lead and copper in freshwater
mussels from the Assiniboine River, Manitoba, Canada.
Journal of Molluscan Studies 61: 295-302.
Pip, E., and J. Stepaniuk. 1992. Cadmium, copper and
lead in sediments and aquatic macrophytes in the Lower
Nelson River system, Manitoba, Canada. II. Metal con-
centrations in relation to hydroelectric development.
Archiv fiir Hydrobiologie 124: 451-458.
Van Loon, J. C., and R. J. Beamish. 1977. Heavy-metal
contamination by atmospheric fallout of several Flin Flon
area lakes and the relation to fish populations. Journal of
the Fisheries Research Board of Canada 34: 899-906.
Wren, C. D., H. R. Maccrimon, and B. R. Loescher.
1983. Examination of bioaccumulation and biomagnifi-
cation of metals in a Precambrian Shield lake. Water,
Air, and Soil Pollution 19: 277-291.
Received 1 April 1996
Accepted 15 November 1996
Comparison of Plant and Animal Diversity on New Reservoir Islands
and Established Lake Islands in the Northern Boreal Forest of Québec
MICHEL CRETE! ?, JEAN Huot?, MARIE-JOSEE ForTIN*? and G. JEAN DoucetT*
'Ministére de l'Environnement et de la Faune, Service de la faune terrestre, 150 Boul. René-Lévesque Est, 5e étage, Québec,
Québec GIR 4Y1
Centre d'études nordiques, Université Laval, Sainte-Foy, Québec G1K 7P4
3Département de biologie, Université de Sherbrooke, Sherbrooke, Québec JIK 2R1
4Vice-présidence Environnement, Hydro-Québec, 75 boul. René-Lévesque Ouest, Montréal, Québec H2Z 1A4
Créte, Michel, Jean Huot, Marie-Josée Fortin, and G. Jean Doucet. 1997. Comparison of plant and animal diversity on new
reservoir islands and established lake islands in the northern boreal forest of Québec. Canadian Field-Naturalist
111(3): 407-416.
We determined species diversity of passerines, small mammals, and woody plants on new islands of the La Grande-3 hydro-
electric reservoir, in the northern boreal forest of Québec, and compared it with that on natural islands of two large neigh-
bouring lakes. We predicted that species diversity on reservoir islands would be greater than that on lake islands because
relaxation (progressive loss of species caused by an excess of extinction over immigration) had not fully operated on reser-
voir islands due to their relatively recent creation, only 11 years before the study. Animal diversity, richness and species
composition were similar in both island groups. Exceptions were a greater abundance of Northern Waterthrush (Seiurus
noveboracencis) and Meadow Vole (Microtus pennsylvanicus) on lake islands. However, richness of woody plants on reser-
voir islands exceeded that on lake islands. We did not find four shrub species in our samples on lake islands which were
recorded on reservoir islands. The mean age of trees on lake islands (192 y) significantly exceeded that of trees on reservoir
islands (112 yr). In contrast to the mainland, where young stands regenerating following fire abounded, mature forest stands
strongly dominated on lake islands. We speculated that fire cycles might be longer on natural islands than on the mainland,
and that reservoir islands could become similar to lake islands with time. Difference in plant diversity between island groups
may have resulted from differences in fire cycle rather than from time since isolation alone. We discuss the future of reser-
voir islands with respect to plant and animal diversity.
Nous avons déterminé la diversité spécifique des passereaux, des petits mammiféres et des plantes ligneuses sur de nouvelles
jles du réservoir hydroélectrique La Grande-3, au nord de la forét boréale québécoise, et nous l'avons comparée 4 celle d'iles
naturelles de deux grands lacs voisins. Nous avons prédit que la diversité en espéces des les du réservoir dépasserait celle
des iles naturelles parce que la relaxation (perte progressive d'espéces résultant d'un surplus d'extinctions par rapport aux
immigrations) des espéces ne se serait pas encore totalement manifestée sur les iles du réservoir, compte tenu de leur créa-
tion récente, 11 ans avant l'étude. La diversité animale, la richesse et la composition en espéces étaient semblables pour les
deux groupes d'iles, sauf pour la Paruline des ruisseaux (Seiurus noveboracencis) et le Campagnol des champs (Microtus
pennsylvanicus) qui étaient tous deux plus abondants sur les iles des lacs. Cependant, la richesse des plantes ligneuses des
iles du réservoir surpassait celle des iles des lacs. Nous n'avons pas détecté la présence de quatre espéces d’arbuste sur les
iles des lacs, qui furent rencontrées sur les iles du réservoir. Nous avons observé que l'age moyen des arbres était plus grand
sur les iles des lacs (192 ans) que sur celles du réservoir (112 ans). De plus, les peuplements forestiers matures dominaient
fortement sur les iles des lacs, contrairement a la terre ferme entourant le réservoir. II se pourrait que le cycle des feux ait été
plus long sur les iles naturelles que sur la terre ferme, et que les iles du réservoirs puissent ressembler a celles des lacs avec
le temps. Les différences de diversité végétale entre les deux groupes d'iles pourraient donc avoir résulté d'un cycle des feux
différent plutot que du temps d'isolement en soi. Nous discutons du futur des fles du réservoir du point de vue de la diversité
animale et végétale.
Key Words: Small mammals, passerines, woody plants, boreal forest, diversity, islands, isolation, Québec.
Island biogeography was initially developed in an
effort to explain species richness on oceanic islands
(Darlington 1957; MacArthur and Wilson 1967).
There has been a renewed interest in the subject
given the accelerated rate of extinction of species
due to ecosystem fragmentation (Saunders et al.
1991). Species extinction related to loss of wilder-
ness areas, most notably due to agriculture and
urbanization, is not a novel problem for ecologists
(Curtis 1956) but the desire to maintain biodiversity,
despite an increasing number of fragmented habitats,
is a more recent concern (McNeely et al. 1990).
Scandinavian countries have become increasingly
concerned about changes to forest composition by
intensive forestry and consequent habitat simplifica-
tion (Enoksson et al.1995; Angelstam 1996).
However, the boreal forest of North America has not
been yet subjected to such intensive silviculture, and
many parts still remain relatively natural (Haila and
Jarvinen 1990). In northern boreal forests, logging
has not been introduced and forest fires are uncon-
trolled. In northern Québec, natural fires often
destroy large areas of forest (Payette et al. 1989;
Hunter 1993) which can become vast homogeneous
407
408
THE CANADIAN FIELD-NATURALIST
Vol. 111
TABLE 1. Number of passerines observed or heard per census point, richness indices and diversity indices (x +S.E.) in
mature forest stands of 10 lake islands and 10 reservoir islands in the northern boreal forest of Québec, 17 to 24 June,
1993. Variables from both island groups were compared using the Kruskal-Wallis test.
Islands
Lakes Reservoir K-W
Species
Yellow-rumped Warbler, Dendroica coronata 20 O02 23 0)3
Ruby-crowned Kinglet, Regulus calendula Loe 0S MA z= (03
Dark-eyed Junco, Junco hyemalis l.3)2e OL (laihea()D
Northern Waterthrush, Seiurus novaboracensis 0.9+0.2 03 = 02 xf
Boreal Chickadee, Parus hudsonicus Ca O02 0.4+0.2
Blackpoll Warbler, Dendroica striata 0:32: 0:2 O02
Gray Jay, Perisoreus canadensis 0.4 + 0.3 O27 0M
Fox Sparrow, Passerculus iliacea 0.1+0.1 Onn
Hermit Thrush, Catharus guttatus - 032102
White-winged Crossbill, Loxia leucoptera 0303 -
Swainson’s Thrush, Catharus ustulatus 02220" -
Pine Grosbeak, Pinicola enucleator O77 -S0n -
White-crowned Sparrow, Zonotrichia leucophrys - O05
White-throated Sparrow, Zonotrichia albicollis - 0.1+0.1
Lincoln’s Sparrow, Melospiza lincolnii - One Oat
Common Redpoll, Carduelis flammea Oh 051 -
Yellow Warbler, Dendroica petechia - OOF
Cape May Warbler, Dendroica tigrina - On Ou
Winter Wren, Troglodytes troglodytes - Ole Ot
Richness
Total species 12 is
Mean richness/census point 4.9+0.5 4.9+0.6
Diversity
Log-Alpha index 4.918 + 0.584 5.347 + 0.830
Shannon index 1.496 + 0.115 1.488 + 0.087
Berger-Parker index 0.283 + 0.028 0.345 + 0.025
Simpson Index 0.246 + 0.033 0.251 + 0.020
2s (NS)
stands (Heinselman 1981). Natural fire periodicity
covers about 100 years around the James Bay area in
eastern Canada. The interval between fire distur-
bance lengthens progressively towards the Atlantic
coast to the east, and towards the forest-tundra and
tundra to the north (Foster 1983; Payette et al. 1989).
In southern boreal forests, Bergeron (1991) observed
that fire cycles covered similar period on lakeshore
and island stands; i.e., 99 vs. 112 years during the
last century. Fires on islands occurred more fre-
quently but were smaller than fires on the mainland.
In southern boreal forests, fires are suppressed
because logging is an important economic activity;
large clearcuts are common. There, the preservation
of biodiversity is becoming a major concern for
wildlife managers, in particular locally in remnant
mature stands surrounded by young cutovers.
We studied an extreme case of isolation: conver-
sion of hilltops to islands after flooding of the La
Grande-3 hydroelectric reservoir (LG-3). We mea-
sured plant and animal diversity on these newly
created islands and compared them with those on nat-
ural islands of two large adjacent lakes. We predicted
that species diversity on reservoir islands would be
greater than that on lake islands because relaxation
(progressive loss of species caused by an excess of
extinction over immigration: Diamond 1975) had not
fully operated on reservoir islands due to their recent
creation, 11 years before the study. We selected only
mature stands as defined by Morneau and Payette
(1988) in order to eliminate the effect of stand age
that influences species composition in northern bore-
al forests (Morneau and Payette 1988; Créte et al.
1995); we expected identical fire cycles on islands
and on the mainland (Bergeron 1991). We wanted to
predict the future of reservoir islands in terms of bio-
diversity and we expected that our findings might be
extended to exploited boreal forests for remnant
uncut patches of comparable size.
Study area
The study area was in northwestern Québec,
approximately 200 km east of James Bay, near the
LG-3 hydroelectric reservoir (53°-54°N, 75°-76°W)
(Figure 1). The area belongs to the Canadian Pre-
cambrian Shield dominated by Archean granitic and
1997
gneissic rock, and covered with glacial deposits
(Rowe 1972). The annual temperature averages
2.5°C and precipitation reaches 650 mm, with 40%
falling as snow. The growing season lasts 130 days,
with a frost-free period of 60-80 days (Wilson 1971;
Environment Canada 1986).
Vegetation is typical of the northern boreal forest
and consists of open woodlands, dominated by Black
Spruce (Picea mariana) and Jack Pine (Pinus
banksiana), with a lichen carpet (Gérardin 1980).
Natural fires have occurred, on average, at 100-year
intervals during recent centuries (Payette et al.
1989). In well drained areas on the mainland around
the LG-3 reservoir, 17 species of woody plants, 8
species of small mammals and 26 species of nesting
passerines were found in forest stands of varying
ages (Créte et al. 1995). However, when considering
CRETE, HUOT, FORTIN, AND DOUCET: DIVERSITY ON NEW RESERVOIR ISLANDS
409
only mature stands, richness decreased to 14 species
of woody plants, 5 of small mammals and 10 of nest-
ing passerines.
Only islands covered with mature forests were
chosen to eliminate variability from fire succession.
We selected 10 reservoir islands and matched them
to 10 comparable islands within two large natural
lakes: six in Lac de la Montagne du Pin and four in
Lac Patukami (Figure 1). Islands selected in LG-3
reservoir were former hilltops isolated from the
mainland when the reservoir was flooded in 1982.
Reservoir islands were chosen so that their size and
distance from the mainland could be matched with
those of lake islands. Reservoir islands covered from
9 to 82 ha (average (x) = 28 ha + standard error
(S.E.) = 7) as compared to 10 to 62 ha (x= 28 + 6)
for lake islands. Minimum distance from the shore
FicurE 1. Location of the study area in northwestern Québec and of the 20 islands where passerines, small mammals and
woody plants were censused.
410
THE CANADIAN FIELD-NATURALIST
Vol. 11
TABLE 2. Number of small mammals captured per 100 trap-nights, richness indices and diversity indices (x + S.E.) in
mature forest stands of 10 lake islands and 10 reservoir islands in the northern boreal forest of Québec, 4 to 17 August
1993. Variables from both island groups were compared using the Kruskal-Wallis test.
Rodents
Southern Red-backed Vole, Clethrionomys gapperi
Meadow Vole, Microtus pennsylvanicus
Heather Vole, Phenacomys intermedius
Northern Bog Lemming, Synaptomys borealis
Meadow Jumping Mouse, Zapus hudsonius
Insectivores
Masked Shrew, Sorex cinereus
Richness
Total species
Mean richness/census grid
Diversity
Log-Alpha index
Shannon index
Berger-Parker index
Simpson index
= P'S0:05
ranged from 100 to 1130 m (x = 548 + 125) for lake
islands, and from 180 to 1950 m (x = 530 + 164) for
reservoir islands. Distance between islands varied
from 90 to 2400 m (x = 956 + 192) on Lac de la
Montagne du Pin, from 170 to 620 m (x = 347 + 61)
on Lac Patukami, and from 460 to 4500 m
(x = 2220 + 172) on the LG-3 reservoir. Elevation
ranged from 8 to 23 m (x= 14+ 2) on lake islands
and from 8 to 38 m (x= 15 + 3) on reservoir islands.
Annual variations in water level (+ 12 m) create a
“tidal” range on reservoir islands, which has been sub-
jected to strong disturbances (flooding, ice erosion
and forest debris accumulation). During periods of
low water level, a beach of variable width (5-175 m)
is created, which separates the treeline from the
water's edge and where some herbaceous plants grow;
this is less pronounced on lake islands. Forest stands
on selected islands were exclusively mature wood-
lands (> 75 years) with relatively homogeneous struc-
ture and plant composition, and without wetlands or
anthropogenic disturbances. Areas surrounding the
LG-3 reservoir and the study lakes were covered by
forest stands of varying ages (Créte et al. 1995).
Methods
Species census
Three groups of organisms were selected to esti-
mate species diversity: nesting birds of the order
Passeriformes, small mammals of the orders
Rodentia and Insectivora, and woody plants; nomen-
clature follows Peterson (1980), Banfield (1977) and
Marie-Victorin (1964). Breeding passerines were
counted from June 17 to 24, 1993, using the fixed-
Islands
Lakes Reservoir K-W
11.56 + 2.72 11.01 +2.91
0.52 + 0.29 0.14 + 0.12 *
0.10 + 0.05 0.32 + 0.29
0.10 + 0.06 0.02 + 0.02
= 0.02 + 0.03
0.60 + 0.30 0.36 + 0.20
5 6
2.7+0.3 2.0 + 0.3
0.736 + 0.111 0.478 + 0.103
0.425 + 0.083 0.258 + 0.119
0.840 + 0.046 0.895 + 0.056
0.759 + 0.057 0.855 + 0.072
radius point-count method (Ralph et al. 1993). A
census point was located in the middle of each island
and the circular plot had a radius of 50 m. The
perimeter of each plot was at least 50 m away from
the shore to minimize edge effect. Each census point
was visited twice on different days by different
observers; the census, which lasted 15 minutes, was
conducted between 04:30 and 09:30 on clear calm
days (no rain and winds < 20 km/h). Abundance of
individual species was determined from the maxi-
mum number of birds observed or heard during both
visits for each census point.
On each island, 12 50-m transects, running paral-
lel to the shore and covering the bird census plot,
were used to sample small mammals. Trapping was
carried out from 4 to 17 August 1993. One “Victor”
number 0 mousetrap and one “Museum Special” trap
(both Ekco Canada Inc., Niagara Falls, Ontario)
were set side-by-side at every 10 m along each tran-
sect; in all, 72 traps of each type were used per
island. For three consecutive nights, traps were bait-
ed with apple cubes and peanut butter. Dental struc-
ture [based on Van Zyll de Jong (1983) for insecti-
vores and on Banfield (1977) for rodents] was used
to identify all captured specimens. Abundance was
expressed as the number of trapped specimens per
100 trap-nights per island. The number of captured
specimens was divided by the number of functional
traps, which was the number of traps set minus the
number of traps accidentally triggered. That number
was then multiplied by 100.
Forest characteristics
Nine transects served to determine forest structure
#997
CRETE, HuoT, FORTIN, AND DOUCET: DIVERSITY ON NEW RESERVOIR ISLANDS
411
TABLE 3. Frequency occurrence of woody plant species per height stratum, richness indices and diversity indices (x + S.E.)
in mature forest stands of 10 lake islands and 10 reservoir islands in the northern boreal forest of Québec. Fisher's exact test
was used to compare qualitative variables, whereas the Kruskal-Wallis test was used to compare quantitative variables of
both groups.
Low shrubs
Labrador Tea, Ledum groenlandicum
Sour-top Blueberry, Vaccinium myrtilloides
Leather Leaf, Chamaedaphne calyculata
Sheep Laurel, Kalmia angustifolia
Swamp Laurel, Kalmia polifolia
High shrubs
Willows, Salix spp.
Speckled Alder, Alnus rugosa
Dwarf Birch, Betula glandulosa
Bartram’s Serviceberry, Amelanchier batramiana
Pin Cherry, Prunus pensylvanica
Trees
Black Spruce, Picea mariana
Jack Pine, Pinus banksiana
Tamarack, Larix laricina
Richness
Total species
Mean richness/quadrat
Diversity*
Shannon Index
Simpson Index
P= 0.05; ~* P= 0.01
“Diversity indices were calculated using presence-absence data.
and abundance of woody plants on each island. Eight
of them, whose projection crossed the centre of the
bird census plot, started on the perimeter of a 40-m
square, one transect on each conner and one in the
middle of each side; transects followed the cardinal
directions. The 40-m square was centered on the bird
census point where was located the last transect.
Transect length varied according to plant stratum:
1 m for the ground stratum, 2 m for herbaceous and
low shrub (< 0.5 m) strata and 10 m for high shrub
(0.5 m < height <2 m) and tree (> 2 m) strata.
Coverage of woody plants was recorded by species
only over 10-m transects.
Density and basal area of trees were estimated
with a point-centered quarter method (Cottam and
Curtis 1956) at one end of all nine transects; two
classes of diameter at breast height (d.b.h.) were
selected: 2.5 cm < d.b.h. <6 cm, and d.b.h. > 6.0 cm.
Abundance of fallen dead trees was estimated using
the number of trunks bisecting 10-m transects.
Density of standing dead trees was determined by
counting the number of trunks in a 2 X 10-m sam-
pling plot, which was parallel to the central 10-m
Islands
Lakes Reservoir Fisher
1.0+0 1.0+0
OF -EAOD 1.0+0
0.7+=0.2 (OS7/ 25 (05)
0B O2 1.0+0 ee
- 0.2+0.1
0.3 + 0.2 0.5 = 0.2
= 0:7=02 ot
- 0.5+0.2 *
- 0.2 + 0.1
0.1+0.1 =
1.0+0 1.0+0
0.6 + 0.2 O7-S02
0.5 + 0.2 OS Ost
K-W
9 12
SEZ EEIES Tea) se lei at
2.054 2.336
0.139 0.425
transect. A clinometer served to determine canopy
height in the centre of the quadrat. The age of the
five largest trees found around each plot was deter-
mined by counting growth annuli.
Age distribution of forest stands
We observed, during reconnaissance flights, that
mature forest stands were very common on lake
islands and we estimated their ages to compare with
those on the mainland. Forest stands growing on
mesic sites were divided into four post-fire succes-
sional stages, according to composition and distribu-
tion of lichens (Morneau and Payette 1988): (1)
recent burns (0-20 y) with dead lichens or bare soil;
(2) shrubs (21-40 y), characterized by a discontinu-
ous cover of Cladonia species; (3) young forests (41-
60 y), distinguished by a continuous cover of
Cladina mitis and C. rangiferina; and (4) mature
woodlands (> 60 y), with a ground stratum dominat-
ed by C. stellaris. The frequency distribution of for-
est stands per post-fire successional stage was com-
puted for all islands on Lac de la Montagne du Pin
and Lac Patukami, and compared with that of the
mainland, along the road surrounding the LG-3
412 THE CANADIAN FIELD-NATURALIST
Vol. 111
TABLE 4. Age of dominant trees, percentage of plant cover per height stratum and dendrometric characteristics (x + S.E.)
of mature forest stands of 10 lake islands and 10 reservoir islands in the northern boreal forest of Québec. Variables from
both island groups were compared using the Kruskal-Wallis test.
Age of trees (y)
Cover(%)
Ground stratum
Lichens
Bryophytes
Bare ground and litter
Herbaceous stratum
Shrub stratum
Low shrubs (< 0.5 m)
High shrubs (0.5 > m < 2)
Tree stratum
Small timber (2 I 6 CT “91 9 qniog MoT uedO ANOTTEAA
a © 8 (¢ C cl CGMNG UGS 2.
= .+, ‘WodIed jsasojo 0} popunol Joquinyy ,
i eee eee
——— a — —— —————_—_—_—________________________________ ninnnnnnnnEnEEDEEEEREEEEEEEEEEEEUEUEEEU EEE EEUU
+ JOE
a I jbuars LC $3140) “[PAkIH
0+ 001 9+ C6 ot I * 66 [+8 8 + 6C i {IIL oT TOW]
+ 9FLP L¥SE CFIL OS punoln seg
Ssollosajyea 10yIO
OF 1 + + LI OF 1 SSO
Ol 55 1h suoyory]
L+Z OFT OFI THT [FZ Joey] punoiyH
er teil CFC CRE 9 + 8E + L=I sseispsoy julofany,g ‘sisuappUDI SYSOASDUDIDID
See will CCl CFEC 9 + 8E (ee se © (0) II SOMT]-SSBID)
Let Seu & "ir UO -YOLNSC ‘siajdoiyjnajs DIDINAYNO
Lev Gates UId4-yeoO ‘siaidokup wnidip2ouudkyH
Orta OF I WIJ -Pjerys esojnurds ‘ninyvj1p si4ajdoKsq
+ + + + MBIISPIg UIOYWON ‘ajvasog wnyvyH
+ + + JOMO]J-UIM J, “Vavdosna sypuai J,
+ young eyig ‘ojvjndys pqsosinsuvs
sonids Je|dog Jejdog re|dog
OU AA - wes[eg wes[eg JOpIV [IejosIOH MOTTE wes[eg seAIq
yortg Jodeg PIO suno X oytusane
yey o}PIPSWII}U] qniyg Ayreq
‘(panuyjuoy)) *7 ATV IL
558
THE CANADIAN FIELD-NATURALIST
Vol. 111
TABLE 3. Characteristics of low shrub-sized plants in different successional stages, summer 1981, Susitna River.?
Early Intermediate
Juvenile Young
Dryas Balsam Willow Horsetail Alder Balsam
Poplar Poplar
Salix alaxensis
Height (m) 033027 0:51 05 0.79 + 0.10 0.55 + 0.07 55) = 016
Crown length (cm) 0.22 + 0.13 0.3 + 0.0 0.31 + 0.03 0.16 + 0.02 0.55 + 0.08
Age Bes) Sena o= I S10) S21
Crown dominance? 2 2 2 2 6
n D 25 18 Ds 19
Salix novae-angliae
Height (m) 0.81 + 0.0 0.83 + 0.17 0.37 + 0.04 ESE (0) 1.25+0.14
Crown length (cm) 0.6 + 0.0 0.30 + 0.04 0.09 + 0.01 O51 = 0m 0.53 + 0.07
Age 4 4+0 Hae) oj ee 927
Crown dominance D 2 2 6 6
n 1 il 16 4 8
Populus balsamifera
Height (m) 0.19 + 0.06 0.32 + 0.06 0.49 + 0.10 0.61 + 0.15 2.48 + 0.59
Crown length (cm) 0.19 + 0.02 0.22 + 0.03 0:22 + 0:03 0:32 20.12 120 = 0:37,
Age B) a=) ez )| 7+0 DEsu| 16+3
Crown dominance 1 2 2 2 6
n 8 30 18 iS 7
Alnus tenuifolia
Height (m) 1.67 + 0.92 0.54 + 0.09 1.19+0.18 £39 = 0:12
Crown length (cm) 0:62-0:29 0.22 + 0.03 0.43 + 0.05 0.73 + 0.09
Age a 0 3410 4+] 6+ 1
Crown dominance 3 2 6 6
n 2 2 q 32
Alnus sinuata
Height (m) 2.46 + 0.46
Basal diameter (cm) 2.23 + 0.66
Age Val
Crown dominance 5
n 4
Picea glauca
Height (m) 0.71
Crown length (cm) 0.34
Age 10
Crown dominance 6
n
“Numbers are rounded to nearest unit of measurement. n=number of stems sampled.
>Crown dominance: 1=open grown, 2=dominant, 3=codominant, 4=intermediate, 5=overtopped, 6=subordinate, 7=ground.
measured tree or shrub was determined by counting
growth rings on cross sectional cuttings or cores
taken as near the ground as possible. Young woody
stems were excavated on some sites to determine the
total age of stem, part of which had been buried by
sediment. Ages analyzed included only the above-
ground age. Densities were also recorded by size
class: < 0.4 m, 0.4-2 m, 2 m - 4 m tall and <4 cm
DBH (diameter-at-breast-height),<4 m tall and >4
cm DBH, and > 4 m tall. Stems >0.4 m tall were
considered available browse even though snow could
reduce that availability in deep snow years. Moose
could reach stems up to 4 m tall or > 4 m tall if DBH
was < 4 cm. These categories are related to the
amount of browsing disturbance possible and to the
vegetation succession. Crown dominance was also
reported for each species: 1 open grown, 2 dominant
(received sunlight from above and sides), 3 codomi-
nant (received sunlight from above but not the
sides), 4 intermediate (barely reaching main canopy),
5 overtopped (below general level of canopy), 6 sub-
ordinate (under overtopped), and 7 ground - lowest
level. Two soil pits were dug in each site to evaluate
depositional profiles.
To estimate the proportion of land area covered by
each of the vegetation types at a given stream flow,
97]
Intermediate
Young Poplar
Early Shrub
Alder
pad >16
= aig) eee
a es
1 Sn ee
ee ae
> 1b
@ 04108 |
0.004, J
HELM AND COLLINS: VEGETATION SUCCESSION AND DISTURBANCE
559
Late
Old Poplar Birch - Spruce
0 10 20 30 40 50 0 10 20 30 40 500 10 20 30 40 50 0 10 20 30 40 500 10 20 30 40 50
MH Poba_ [Il Pig!
Bepa E=Salix
Alte &8Late Shrub
[__]Forbs &&8Grass
Cover (%)
FiGuRE 2. Scaled relative cover (%) (relative cover x total vascular plant cover) by major species or herba-
ceous life form and height class in each successional stage. Shrubs between 0.4 and 4 m tall are sub-
ject to Moose browsing disturbance. Poba = Populus balsamifera, Bepa = Betula papyrifera, Pigl =
Picea glauca, Salix = Salix alaxensis + Salix novae- angliae, Alte = Alnus tenuifolia, Late shrub =
Rosa acicularis + Viburnum edule + Ribes triste.
transects were plotted approximately every 4 cm (2
km) on aerial photographs (1:48 000 black- and-
white taken in 1980). Points were systematically
plotted approximately every 7 mm (350 m) along
each transect from the base of slope on one side of
the river to the base of first hill on the other side,
which was considered the limit of the floodplain.
The points were classified according to successional
stage during a helicopter survey in June 1981.
Results
General
Early Shrub sites were most common on the two
lower reaches where the floodplain was wide with
Intermediate
Early Shrub Alder
>4m
<4m,>4cm
2-4m,<4cm
0.4-2.0m}
0.0-0.4m
Size Class
split and braided channels. Approximately 20 to 30%
of available land in this reach was occupied by
Barren or Early Shrub stages (Table 1). Above the
Chulitna River confluence, the river was narrower
with well-defined channels. Here the mature forests
predominated and occupied almost three-fourths of
the available land area. Similarly the intermediate
stage only occupied 6% of the area compared to
>17% in the lower two reaches.
Some Early Shrub sites have remained in early
succession from 1951 (aerial photographs) to 1995
although they are periodically flooded. While some
Early Shrub sites eroded between 1981 and 1995,
others advanced successionally. Part of a Birch -
Late
Young Poplar Old Poplar Birch - Spruce
+ oooe
eee
PPeTy
}eeoe
$630000000000055 ee
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Mm Poba [ill Pigl
Bepa E=Salix
Alte Late Shrub
Density (1000 stems/ha)
FIGURE 3. Density (stems ha™') of major woody species by size class in each successional stage. Size
classes are by height (first number) and DBH (second number, if present). Stems sized 0.4 to 2.0
m tall and 2 to 4 m with DBH < 4 cm are subject to Moose browsing disturbance. Species codes
are explained in Figure 2.
560 THE CANADIAN FIELD-NATURALIST
Vol. 111
TABLE 4. Characteristics of trees and tall shrubs in different successional stages, summer 1981,
Susitna River.
Intermediate Late
Alder Young Balsam Old Balsam Paper Birch-
Poplar Poplar White Spruce
Populus balsamifera
Height (m) G2 05 14.0+1.1 PAS
DBH (cm) 7.2+0.4 24.8 + 1.9 332 =A
Age 19 1 44+3 98 +6
Crown dominance 2 Bs 2
n 28 36 40
Alnus tenuifolia
Height (m) 5-302 49+0.2 3:6 +03 3.905
DBH (cm) 7.3+0.4 6.9+0.4 19-09 3.4+0.5
Age 20 +3 22 28+2 Pires
Crown dominance 2 4 >) 4
n 40 32 36 12
Alnus sinuata
Height (m) DDS SS yeaa I 2.4 3.4+0.8
DBH (cm) 54 = 210 2.6 + 0.4 DA 13s
Age 17+4 We 38 50 + 12
Crown dominance 4 5 5 5
n 4 3 1 3
Picea glauca
Height (m) 22407, 2.6 10.9 + 1.0 13.8 + 1.0
DBH (cm) S Vee) - 21.4+1.7 26.3 + 1.8
Age [Dee 13 100+8 90+5
Crown dominance 4 6 3 Z
n 6 1 16 38
Betula papyrifera
Height (m) 3.7+0.8 46+4.0 12.8 + 0.6
DBH (cm) - - PA Wz XS)
Age [Pee 26 17 70 +3
Crown dominance 4 4 2
n 7 2 35
“Numbers are rounded to nearest measurement unit. n=number of stems sampled.
>Crown dominance: 1=open grown, 2=dominant, 3=codominant, 4=intermediate, S5=overtopped,
6=subordinate, 7=ground.
Spruce forest along a cut bank was eliminated by
flooding and erosion between 1981 and 1984.
Several minor types were also found along the
helicopter transect: Bog, Wet - Sedge Grass
Meadow, and other forests. The Bog and Wet Sedge
Grass Meadow had developed on poorly drained
soils rather than on the well-drained sites supporting
successional pathways described in this paper.
Early Shrub Stage
Early Shrub communities had the least cover of
any successional stage, and most plants were <0.4 m
tall (Table 2, Figure 2). Dominant plant species
included Yellow Dryas (Dryas drummondii), Balsam
Poplar, Feltleaf Willow (Salix alaxensis), Variegated
Horsetail (Equisetum variegatum), or combinations
of these. Willows were the only species in this stage
with mean height > 0.40 m (Table 3). Heights of
Balsam Poplar and Feltleaf Willow were suppressed
by browsing and flooding. Age variation in most
Early Shrub stands was small, suggesting that
recruitment was simultaneous.
The least vegetation cover occurred on the harsh,
cobbly Yellow Dryas sites and was not readily
detectable on the black-and-white photographs or
from the helicopter survey. Overstory development
on Yellow Dryas sites was stunted, and Balsam
Poplar remained sapling sized for 20 to 40 years on
one stand that had been examined but not quantita-
tively sampled. Balsam Poplar communities occu-
pied dry, nutrient-poor sites with sand content near
90% in many cases. Willow sites had intermediate-
textured soils and ground cover. Greatest vegetation
cover and finest-textured soils, with silt >60%,
among the Early Shrub sites occurred on Horsetail
sites, which also included traces of sedges (Carex
spp.) and cottongrasses (Eriophorum spp.).
1997
0-15 yr 20-50 yr
wTICECSCOUT 200 yr
* Beaver
Meese Logging
FIGURE 4. Conceptual model of successional pathways along the Susitna River and their controlling
factors. Flooding includes erosion and sedimentation. Years above diagram represent general-
izations of when types may dominate. Width of arrows represents relative importance of the
pathway.
Thinleaf Alder grew more rapidly than other
shrubs, 5-year-old alders being approximately 1.5 m
tall while same-aged Balsam Poplars averaged
approximately 0.5 m in height. Willow stands had
the best-developed shrub community within the
Early Shrub stage (Tables 2, 3). These shrubs formed
the overstory (crown dominance = 2, Table 3).
Many Early Shrub sites changed considerably
between 1981 and 1984. One stand (20) was suffi-
ciently eroded that it could not be resampled. Two
Juvenile Balsam Poplar stands (21 and 25) had
almost identical cover values but stand 21 had more
litter in 1984. One horsetail stand (16) had more
horsetail and more bare ground in 1984 compared to
1981. Two horsetail stands (8, 9) had more Balsam
Poplar in 1984. Willows in another Horsetail stand
(13) increased from 1% cover to almost 20% cover.
During the same period horsetail cover in a Willow
stand (22) increased 22% to 53% while woody cover
decreased from 18% to 5%. These different respons-
es were functions of varied disturbance regimes and
changes in water levels associated with dissimilar
channels.
Intermediate Stage
The intermediate stage of succession accounted
for 6 to 18% of vegetated land along the three reach-
es of the floodplain, and most of this was Closed
Tall Scrub Alder (Table 1). These sites were charac-
terized by Thinleaf Alder or Balsam Poplar which
had developed into tall shrubs or trees. These would
be classified as Open Tall Scrub Alnus tenuifolia /
Calamagrostis canadensis and Closed Broadleaf
Forest Balsam Poplar Populus balsamifera / Alnus
tenuifolia | Calamagrostis canadensis, respectively
at Level V of Viereck et al. (1992). Low shrub-sized
plants (<1.5 m tall) were rare in the understory
(crown dominance = 6) beneath a Thinleaf Alder or
Balsam Poplar overstory (Figure 3, Tables 3, 4).
Total vegetation cover in Alder stands averaged
87% across all height classes (Table 2). Thinleaf
Alder provided 59% cover, whereas Balsam Poplar
provided only 13% cover, but was present in all
stands. Bluejoint Reedgrass (Calamagrostis
canadensis) produced a dense understory with 38%
cover. Average ages of tall shrub-sized Thinleaf
Alder and Balsam Poplar in Alder sites were approx-
imately 20 years (Table 4). Mean Balsam Poplar and
Thinleaf Alder heights were 6.2 and 5.3 m, respec-
tively, in the Alder sites, but Thinleaf Alder provided
more cover especially in the taller classes (Tables 2,
4, Figure 2). Little browse was available for Moose
because of the short stature and low cover values of
the other woody species (Figures 2, 3). This was the
first stage to have a well-developed litter layer that
could contribute to nutrient cycling. Multiple buried
organic layers were found where sedimentation
buried old litter layers.
Vegetation developed into Young Poplar Forests
when Balsam Poplar overtopped Thinleaf Alder
(crown dominance shifted from 2 to 4) and dominat-
ed the overstory with 62% cover (Figure 2, Tables 3,
4). These Balsam Poplar trees averaged 44 years of
age and 14 m in height, which was more than double
their height in the Alder stage and twice as tall as
Thinleaf Alder in this stage. Thinleaf Alder stems
had similar ages in both the Alder and Young Poplar
Forests, suggesting that individual Thinleaf Alder
562
stems may live only 20 years, although parent root
systems might live longer. Both Alder and Young
Poplar Forests were similar except for the age and
species structure of the overstory and increase in
Prickly Rose (Rosa acicularis) and High Bush-
cranberry (Viburnum edule) (Table 2, Figure 2).
Most Thinleaf Alder cover was limited to the 2 - 4
and 4 - 8 mclasses, and Bluejoint still dominated the
understory < | m tall (Figure 2).
Late Stage
Late stage of succession contained Old Balsam
Poplar Forests and Paper Birch - White Spruce
Forests which occupied over half of the vegetated
land surveyed (Table 1). It also contained transition-
al Balsam Poplar - White Spruce Forests. Old Poplar
and Birch - Spruce stages would be identified as
Open Broadleaf Forest Balsam Poplar Populus bal-
samifera / Alnus tenuifolia | Calamagrostis canaden-
sis - Rosa acicularis - Viburnum edule and Open
Mixed Forest Spruce - Paper Birch Picea glauca -
Betula papyrifera | Calamagrostis canadensis - Rosa
acicularis - Viburnum edule types, respectively, at
Level V of the Alaskan Vegetation Classification
(Viereck et al. 1992). Old Balsam Poplar Forests
characterized 25 to 40% of the vegetated floodplain
while mixed stands of Paper Birch and White Spruce
occupied 23 to 32% of the area (Table 1).
Although Old Balsam Poplar sites averaged 90%
total vegetation cover, the overstory had thinned rel-
ative to the Young Poplar Forest (Figure 2). Shrub
cover and density, especially of browse species,
increased substantially (Figure 2). The oldest Balsam
Poplar tree measured in 1984 was approximately 170
years, but many were 110 to 140 years of age. One
(stand 3) was an Old Poplar site that had been flood-
ed between 1981 and 1984, and litter was reduced
from 100% to 50%. White Spruce was not detected
along the sampling transects in 1981, but 9% cover
was reported in 1984. In 1981 several minor forbs
were found [Tall Bluebell (Mertensia paniculata),
Pyrola (Pyrola spp.), American Red Currant (Ribes
triste), Northern Bedstraw (Galium boreale), ferns],
but these were negligible in 1984. The site had about
twice as much Prickly Rose, High Bushcranberry,
and Bluejoint Reedgrass in 1984 as it did in 1981.
Similarly, another stand (17) went from 67% to 1%
litter and many small forbs were no longer present.
Prickly Rose and High Bushcranberry tripled
between 1981 and 1984.
Flooding was still an important factor affecting
understory species in Old Poplar Forests as evi-
denced by the preceding observations and numerous
buried horizons, indicating that a forest floor was
present during the flooding (Viereck et al. (1993).
Changes in vegetation types occurred most often
when banks were eroded or plants were physically
removed, rather than from these depositional events
that temporarily altered understories.
THE CANADIAN FIELD-NATURALIST
Vol. 111
As individual Balsam Poplar trees and understory
Thinleaf Alder matured and died, White Spruce trees
were released (crown dominance shifted from 3 to 2,
Table 4), and Paper Birch became established on soil
clinging to the roots of uprooted Balsam Poplar
trees. Paper Birch - White Spruce Forests consisted
of a mosaic of well-developed treed areas with gaps
dominated by shrubs where older trees had died.
Herbaceous and low shrub species became more
important in the gaps and understory of Birch -
Spruce while cover in intermediate layers decreased
(Figure 2). Densities of browsable shrubs increased
substantially (Figure 3) although many of these
stems were in open areas. In one site where Paper
Birch had been logged, most stumps had sprouted
and associated vegetation had developed the appear-
ance of open patches within other Birch - Spruce
stands. The oldest cored Paper Birch in a Birch -
Spruce site was almost 170 years although ages of
most mature trees were between 100 and 140 years
while the oldest measured White Spruce tree was
124 years.
Some abandoned oxbows or otherwise poorly-
drained sites developed into wetlands. These were
dominated by Thinleaf Alder, Dwarf Arctic Birch
(Betula nana), Buffaloberry (Shepherdia cana-
densis), Bluejoint Reedgrass, and sphagnum moss
(Sphagnum spp.).
Discussion
Plant Species Establishment
The first plant species, such as Balsam Poplar and
willows, to colonize a site had light, nondormant,
wind- or water-dispersed seeds that germinated soon
after landing on a moist, suitable substrate (Viereck
1970; Densmore and Zasada 1983; Walker et al.
1986; Helm and Allen 1995). Species that survived on
these new surfaces had to survive flooding and sedi-
mentation. Fall- and winter-dispersed seeds (Thinleaf
Alder, Paper Birch, White Spruce) usually lagged
summer dispersers by a year before they colonized.
Colonization appeared to be a stochastic event that
did not occur each year (Walker et al. 1986).
Survivability of new seedlings that germinated along
high water lines depended on minimal flooding for
the remainder of the summer (Sigafoos 1964; Fenner
et al. 1985; Bradley and Smith 1986). Helm and
Allen (1995) observed that Balsam Poplar seedling
survival on a glacial floodplain may depend on dis-
persal close to summer rains, but early enough so
that seedlings could survive flooding. Successful
colonization along the Susitna might not occur each
year since late summer flows were usually higher
than spring flows. Two- to five-year old sites were
rare both in 1981 and 1984 which suggested that
new sites had not become available or were not colo-
nized during the preceding few years. Bradley and
Smith (1986) reported that suitable conditions for
1997
recruitment occurred once every five years on the
Milk River, which would be a reasonable estimate
for the Susitna River.
Thinleaf Alder colonized where Balsam Poplar,
willows, or horsetails were already growing. This
delay might have occurred because its seeds were
dispersed in the fall rather than during the growing
season. Appropriate mycorrhizal fungi may not be
present on primary successional sites for Thinleaf
Alder to colonize (Helm and Carling 1993; Helm et
al. 1996). Partial shade was important for White
Spruce seedlings (Safford 1974) while Paper Birch
seedlings required light shade for 2 to 3 months
(Brinkman 1974), thus reducing the likelihood of
successful colonization on barren sand or silt bars.
However, Paper Birch and White Spruce established
more readily on mineral soils with little or no flood-
ing (Clautice 1974; Youngblood and Zasada 1991).
Most White Spruce and Paper Birch seed travel <
100 m (Zasada 1986), making distance from seed
source a possible limiting factor.
Paper Birch generally became established after
White Spruce, and seedlings occurred only on miner-
al soil on rootballs of trees uprooted by wind or
gravity. This provided mineral soil and elevated the
seedling above competition on the forest floor
(Beatty and Stone 1986; Jonsson and Dynesius
1993). Old Paper Birch were primarily observed
straddling rootballs from uprooted trees that were
partially decomposed, indicating that the seedlings
we observed on rootballs could indeed mature in
those microsites. Gaps and windthrow microsites
have been shown to be important for regeneration of
other tree species at northern latitudes (Deal et al.
1991; Jonsson and Dynesius 1993).
Mechanisms of Successional Pathway Changes
Figure 4 depicts major successional pathways, rel-
ative importance (width of arrows) of each pathway,
causes of changes, and approximate “time after sta-
bilization” that various stages occurred. Burial of 5
to 10 years of plant growth with 0.5 m or more of
sediment and rotten centers on older trees made
aging only approximate. The Water compartment
represented surface areas beneath water. Barren rep-
resented the first surfaces above water with little
plant colonization (<2% cover). At any stage, ero-
sion by flooding could cut banks, remove the sub-
strate and vegetation, and return the location to
Water. Older sites were found along cut banks while
young sites were more often found on areas of recent
deposition rather than along cut banks. Other distur-
bances usually prevented succession from advancing
(arrow pointing to same cell) or regressed the site by
removing the overstory.
The several phases within the Early Shrub stage
could develop into each other (arrow cycling within
that compartment), but substrate differences associ-
ated with the various phases likely resulted from
HELM AND COLLINS: VEGETATION SUCCESSION AND DISTURBANCE
563
different intensities of flooding and soil deposition
(Figure 4). A Willow stand with willows and poplar
present in 1981 was dominated by Variegated
Horsetail with an understory of new Balsam Poplar
seedlings in 1984. In contrast, Feltleaf Willow and
Variegated Horsetail both became more abundant
and larger between 1981 and 1984 on a hummocky,
ice-affected site where the hummocks formed by ice
action. In another site Variegated Horsetail, but not
Feltleaf Willow, was reduced by flooding some time
between July and September 1984.
Ice scour usually bent and scraped willows and
juvenile and sapling Balsam Poplar in Early Shrub
and Alder stages but did not change the vegetation
type (Figure 4). Bent stems sprouted from the hori-
zontal or diagonal stem, substantially increasing the
number of browsable stems. Larger woody trunks of
Alder communities protected understory plants from
ice damage, but in younger sites ice often scraped
the substrate, removing both the substrate and plants
growing there. These became Barren sites. Ice also
transported sediments to some sites.
Browsing by Moose and Arctic Hares slowed veg-
etation development by reducing heights on many
shrubs in earlier stages. This allowed Thinleaf Alder
to dominate more rapidly. Beavers were very active
on some sites and removed most Balsam Poplar
stems in a Young Poplar Forest, reverting it to an
Alder site between 1981 and 1984 (Figure 4).
Beavers also removed full-sized trees in Old Poplar
Forests but did not alter the vegetation type. A hypo-
thetical pathway exists for the formation of White
Spruce stands if Beavers removed sufficient poplar
from an Old Poplar Forest or Poplar-Spruce Forest.
Shade and lack of mineral soil would probably pre-
clude Paper Birch from growing on the site, thus
favoring spruce establishment. Although other stud-
ies have noted substantial changes in vegetation as a
result of tree-cutting by Beaver, the trees were
Trembling Aspen (Populus tremuloides) (Johnston
and Naiman 1990; Moen et al. 1990).
Disturbances could also slow vegetation develop-
ment indefinitely. Some Juvenile Poplar sites
appeared to be on 30-year old surfaces based on
comparison of aerial photographs from 1951 and
1981, but above-ground ages were < 10 years.
Underground portions of stems extended about 0.5 m
below the present soil surface and accounted for 5 -
10 years of additional growth on deeply-sedimented
sites. All of our Alder sites had Balsam Poplar as a
codominant and would be expected to progress to
Young Poplar, rather than remaining in Alder stage
indefinitely.
One Old Poplar site had trunks broken 2 to 3 m
above the ground, apparently by winter winds since
no trees were uprooted from the frozen ground, and
most trees fell in the same direction. Some White
Spruce regeneration was evident, but the herbaceous
564
understory was too dense and the soils too organic
for hardwood seedlings, especially Paper Birch, to
become established. We believe that this site may
eventually become a White Spruce site, but could
become an herbaceous, low or tall shrub site and
remain in this stage indefinitely if other trees do not
become established. This is indicated by wind arrow
in Figure 4. Logging and firewood cutting in one
Birch - Spruce site produced a community similar to
the gap portion of the Birch - Spruce Forest although
herbaceous competition could sometimes hinder
woody regeneration from seed.
Comparisons with Other Northern Rivers
Vegetation succession along the Susitna River was
similar to other boreal forest rivers in terms of the
early shrub development (Feltleaf Willow, Balsam
Poplar) and establishment of Old Poplar Forests, but
differed primarily in the effects of substrate on initial
community composition and in the importance of
Birch - Spruce stage in late succession. Because of
the greater presence of deciduous trees, mosses and
lichens were only a minor component of the ground
layer. Similarities to other northern rivers included
the importance of Feltleaf Willow and the rareness
of Balsam Poplar - White Spruce sites.
More Feltleaf Willow was found on gravelly
alluvium than on sandy or silty materials on the
Colville River on the Arctic slope (Bliss and
Cantlon 1957) in contrast with the Susitna River.
Gill (1972) also found Feltleaf Willow up to 6 m
tall on the Mackenzie River delta. Feltleaf Willow
almost never reaches these sizes in southcentral
Alaska because of intense moose browsing and
competition from tree species. These factors gener-
ally precluded Feltleaf Willow from reaching even
2 m height along the Susitna River. Observations of
heavy browsing slowing growth of palatable
species is consistent with observations on the
Tanana River in interior Alaska (Wolff 1976;
Bryant and Chapin 1986).
Vegetation along the Susitna River differed from
other northern rivers in the greater variety of early
seral communities resulting from variable river
dynamics and substrate textures. The earliest stage
described by Viereck (1970) along the Chena River
in Interior Alaska was a 15-year-old Feltleaf Willow
site. It established on coarser soils than Willow sites
in the Susitna River floodplain, and it was much
older than most of our Willow. However, Gill (1972)
and Teversham and Slaymaker (1976) reported
Balsam Poplar species on their coarser sites similar
to the Susitna River. The greater variety of substrate
textures on the Susitna River resulted from different
fluvial dynamics. Sands were frequently deposited
by intense summer floods, particularly below the
Chulitna confluence, while silts and fine sands were
deposited by milder floods and in backwaters behind
ice jams (Helm et al. 1985).
THE CANADIAN FIELD-NATURALIST
Vol. 111
Willow was reduced from Alder and later stages
of succession by its shorter life span, browsing, and
shade intolerance (Walker et al. 1986; Bryant 1987;
Viereck et al. 1993). Van Cleve et al. (1980)
observed that most willows died under closed
canopy, although Feltleaf Willow and Thinleaf Alder
survived in a state of reduced vigor. Prickly Rose
and High Bushcranberry increased in understories
because they were more shade tolerant (Van Cleve et
al. 1980; Viereck et al. 1993).
Nanson and Beach (1977) observed that Balsam
Poplar transitioned rapidly to White Spruce with no
mixed stage in British Columbia. Their White
Spruce seedlings were released when Balsam Poplar
died on 100- to 150-year-old surfaces, similar to our
transitional sites of Poplar - Spruce. White Spruce
would normally be expected to follow Balsam
Poplar in the successional sequence since it is longer
lived and more shade tolerant than Paper Birch
(Reed and Harms 1956). However, Paper Birch is
able to reproduce on mineral soil in gaps left by fall-
en trees, and recruitment is sufficient to maintain the
Paper Birch - White Spruce stage in the Anchorage -
Matanuska Valley adjacent to our study area. This
strong deciduous component with associated leaf fall
probably limits the development of the moss and
lichens in the ground layer. Along the Tanana River
moss development did not exceed minor cover until
the Balsam Poplar - White Spruce stage when White
Spruce became codominant or dominant (Viereck et
al. 1993).
Vegetation establishment appeared slower in the
Susitna River floodplain compared to the Chena and
Tanana Rivers in interior Alaska (Viereck 1970;
Viereck et al. 1993). Early Shrub communities could
persist in the Susitna River floodplain for 15 years or
more while Alder tended to dominate 20 to 50 year
old sites. In contrast Closed Alder and Willow stages
dominated 5- to 10- year-old sites in the Tanana
floodplain, and Young Balsam Poplar dominated 20-
to 40-year-old sites (Viereck et al. 1993). Balsam
Poplar dominated the canopy in Chena River sites
that were approximately 50 years old, then White
Spruce became dominant by 120 years (Viereck
1970). However, in the Tanana River floodplain,
White Spruce was not dominant until about years
200-300 (Viereck et al. 1993). Our Old Balsam
Poplar Forests were over 100 years old, and White
Spruce was just becoming important in the understo-
ry, which was more similar to the Tanana River than
to the Chena River. However, Birch - Spruce stage
dominated sites over 200 years old in the Susitna
River floodplain and appeared to be self-replacing.
Conclusion
Disturbances caused by flooding, including both
sedimentation and erosion, and wildlife herbivory
were major factors regulating vegetation succession
1997
along the Susitna River. Vegetation appeared to
establish only in certain years, perhaps in response
to rainfall and flood regimes. Most disturbances by
flooding or ice caused sedimentation of the site and
only caused a retrogression to bare ground or water
if the substrate itself was eroded beneath the plants.
Uprooted trees appeared necessary to produce safe
sites with mineral soils and increased sunlight for
Paper Birch establishment in forest understories.
Small-scale disturbances caused by tree-falls created
openings in the canopy and mineral soil for seedling
establishment and development of shrub mosaics.
Moose restricted the height growth of Balsam Poplar
and willows in early stages. Beavers, however,
could remove entire trees and force vegetation suc-
cession to an earlier stage. Vegetation patterns in the
Susitna River floodplain thus resulted from distur-
bances by flooding, ice, wind, and browsing as well
as stochastic events associated with seed dispersal
and establishment.
Acknowledgments
Funding was provided by the Alaska Power
Authority through subcontracts with Terrestrial
Environmental Specialists in 1981-1982 and Harza-
Ebasco Susitna Joint Venture in 1985, Federal Aid
Wildlife Restoration Projects W-24-3 and W-24-4,
and by the University of Alaska Fairbanks Agri-
cultural and Forestry Experiment Station in 1984.
Additional thanks go to J. Moor, N. Hull, L. Werner
for help with field and laboratory work and to Larry
Erie for the location map. J. LaBelle of Arctic
Environmental Information Data Collection, Uni-
versity of Alaska, was instrumental with the geomor-
phological interpretations on the original project. We
would like to thank A. Jubenville and F.S. Chapin
Ill for reviews of early versions of the manuscript.
Literature Cited
Beatty, S. W., and E. L. Stone. 1986. The variety of soil
microsites created by tree falls. Canadian Journal of
Forest Research 16: 539-548.
Bliss, L. C., and J. E. Cantlon. 1957. Succession in river
alluvium in Northern Alaska. American Midland
Naturalist 58: 452-469.
Bormann, B. T., H. Spaltenstein, M. H. McClellan, F. C.
Ugolini, K. Cromack, and S. M. Nay. 1995. Rapid
soil development after windthrow disturbance in pristine
forests. Journal of Ecology 83: 747-757.
Bradley, C. E., and D. G. Smith. 1986. Plains cottonwood
recruitment and survival on a prairie meandering river
floodplain, Milk River, southern Alberta and northern
Montana. Canadian Journal of Botany 64: 1433-1442.
Brinkman, K. A. 1974. Salix L. Willow. Pages 746-750
in Seeds of woody plants in the United States. Edited by
Schopmeyer, C. S. United States Department of
Agriculture Forest Service, Washington, D.C.
Bryant, J. P. 1987. Feltleaf willow-snowshoe hare inter-
actions: plant carbon/nutrient balance and floodplain
succession. Ecology 68: 1319-1327.
HELM AND COLLINS: VEGETATION SUCCESSION AND DISTURBANCE
565
Bryant, J. P., and F. S. Chapin, III. 1986. Browsing-
woody plant interactions during boreal forest plant suc-
cession. Pages 213-225 in Forest ecosystems in the
Alaskan taiga. Edited by K. Van Cleve, F. S. Chapin, III,
P. W. Flanagan, L. A. Viereck, and C. T. Dyrness.
Springer-Verlag, New York.
Clautice, S. F. 1974. Spruce and birch germination on
different seedbeds and aspects after fire in interior
Alaska. M.Sc. thesis, University of Alaska Fairbanks,
Fairbanks, Alaska.
Collins, W. B., and D. J. Helm. 1997. Moose, Alces
alces, habitat relative to riparian succession in the boreal
forest, Susitna River, Alaska. Canadian Field-Naturalist
111: 567-574.
Craighead, F. L., E. L. Young, and R. D. Boertje. 1984.
Stikine River moose study. Wildlife evaluation of
Stikin-Iskut Dams. Alaska Department of Fish and
Game, Juneau.
Deal, R. L., C. D. Oliver, and B. T. Bormann. 1991.
Reconstruction of mixed hemlock-spruce stands in
coastal southeast Alaska. Canadian Journal of Forest
Research 21: 643-654.
Densmore, R., and J. Zasada. 1983. Seed dispersal and
dormancy patterns in northern willows: ecological and
evolutionary significance. Canadian Journal of Botany
61: 3207-3216.
Fenner, P., W. W. Brady, and D. R. Patton. 1985.
Effects of regulated water flows on regeneration of
Fremont cottonwood. Journal of Range Management 38:
135-138.
Gill, D. 1972a. Modification of northern alluvial habitats
by river development. The Canadian Geographer 2:
139-153.
Gill, D. 1972b. The point bar environment in the
Mackenzie River delta. Canadian Journal of Earth
Science 9: 1382-1393.
Helm, D. J., and E. B. Allen. 1995. Vegetation chronose-
quence near Exit Glacier, Kenai Fjords National Park,
Alaska, U.S.A. Arctic and Alpine Research 27:
246-257.
Helm, D. J., E. B. Allen, and J. M. Trappe. 1996.
Mycorrhizal chronosequence near Exit Glacier, Alaska.
Canadian Journal of Botany 74: 1496-1506.
Helm, D. J., and D. E. Carling. 1993. Use of soil transfer
for reforestation on abandoned mined lands in Alaska.
II. Effects of soil transfers from different successional
stages on growth and mycorrhizal formation by Populus
balsamifera and Alnus crispa. Mycorrhiza 3: 107-114.
Helm, D. J., W. B. Collins, and J. C. LaBelle. 1985.
Riparian vegetation succession report. Draft Report.
Alaska Power Authority, Susitna Hydroelectric Project.
Prepared for Harza-Ebasco Susitna Joint Venture. 185
pages.
Johnston, C. A., and R. J. Naiman. 1990. Browse selec-
tion by beaver: effects on riparian forest composition.
Canadian Journal of Forest Research 20: 1036-1043.
Jonsson, B. G., and M. Dynesius. 1993. Uprooting in
boreal spruce forests: long-term variation in disturbance
rate. Canadian Journal of Forest Research 23:
2383-2388.
Moen, R., J. Pastor, and Y. Cohen. 1990. Effects of
beaver and moose on the vegetation of Isle Royale
National Park. Alces 26: 51-63.
Naiman, R. J., G. Pinay, C. A. Johnston, and J. Pastor.
1994. Beaver influences on the long-term biogeochemi-
566
cal characteristics of boreal forest drainage networks.
Ecology 75: 905-921.
Nanson, G. C., and H. F. Beach. 1977. Forest succession
and sedimentation on a meandering river floodplain,
northeast British Columbia, Canada. Journal of
Biogeography 4: 229-251.
Reed, J. C., Jr., and J. C. Harms. 1956. Rates of tree
growth and forest succession in the Anchorage -
Matanuska Valley area, Alaska. Arctic 9: 238-248.
Safford, L. O. 1974. Picea A. Dietr. Spruce. Pages
587-597 in Seeds of woody plants in the United States.
Edited by Schopmeyer, C. S. United States Department
of Agriculture Forest Service, Washington, D.C.
Sigafoos, R. S. 1964. Botanical evidence of floods and
flood-plain deposition. United States Geological Survey,
Professional Paper 485—A.
Teversham, J. M., and O. Slaymaker. 1976. Vegetation
composition in relation to flood frequency in Lillooet
River Valley, British Columbia. Catena 3: 191-201.
Van Cleve, K., F. S. Chapin, III, P. W. Flanagan, L. A.
Viereck, and C. T. Dyrness. Editors. 1986. Forest
Ecosystems in the Alaskan Taiga: A Synthesis of
Structure and Function. Ecological Studies: Analysis and
Synthesis. Springer-Verlag, New York. 230 pages.
Van Cleve, K., C. T. Dyrness, G. M. Marion, and R.
Erickson. 1993. Control of soil development on the
Tanana River floodplain, interior Alaska. Canadian
Journal of Forest Research 23: 941-955.
Van Cleve, K., C. T. Dyrness, and L. Viereck. 1980.
Nutrient cycling in interior Alaska flood plains and its
relationship to regeneration and subsequent forest devel-
opment. Pages 11—18 in Forest regeneration at high lati-
tudes. Edited by M. Murray and R. M. Van Veldhuizen.
United States Department of Agriculture Forest Service,
PNW-107.
Van Cleve, K., L. A. Viereck, and C. T. Dyrness. 1996.
State factor control of soils and forest succession along
the Tanana River in Interior Alaska, U.S.A. Arctic and
Alpine Research 28: 388-400.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Viereck, L. A. 1970. Forest succession and soil develop-
ment adjacent to the Chena River in interior Alaska.
Arctic and Alpine Research 2: 1-26.
Viereck, L. A., C. T. Dyrness, A. R. Batten, and K. J.
Wenzlick. 1992. The Alaska Vegetation Classification.
USDA Forest Service, Pacific Northwest Research
Station, Portland, Oregon, PNW-GTR-286. 278 pages.
Viereck, L. A., C. T. Dyrness, and M. J. Foote. 1993.
An overview of the vegetation and soils of the floodplain
ecosystems of the Tanana River, interior Alaska.
Canadian Journal of Forest Research 23: 889-898.
Walker, L. R., and F. S. Chapin, II. 1986. Physiological
controls over seedling growth in primary succession on
an Alaskan floodplain. Ecology 67: 1508-1523.
Walker, L. R., J. C. Zasada, and F. S. Chapin, III. 1986.
The role of life history processes in primary succession
on an Alaskan floodplain. Ecology 67: 1243-1253.
Wolff, J. O. 1976. Utilization of hardwood browse by
moose on the Tanana floodplain in interior Alaska.
United States Department of Agriculture Forest Service,
Pacific Northwest Forest and Range Experiment Station.
7 pages.
Youngblood, A. P., and J. C. Zasada. 1991. White spruce
artificial regeneration options on river floodplains in inte-
rior Alaska. Canadian Journal of Forest Research 21:
423-433.
Zasada, J. 1986. Natural regeneration of trees and tall
shrubs on forest sites in interior Alaska. Pages 44-73 in
Forest ecosystems in the Alaskan taiga. Edited by K.
Van Cleve, F. S. Chapin, III, P. W. Flanagan, L. A.
Viereck, and C. T. Dyrness. Springer-Verlag, New York.
Zasada, J. C., L. A. Viereck, M. J. Foote, R. H.
Parkenson, J. O. Wolff, and L. A. Lankford, Jr.
1981. Natural regeneration of balsam poplar following
harvesting in the Susitna Valley, Alaska. The Forestry
Chronicle 57: 57-65.
Received 2 November 1995
Accepted 3 March 1997
Moose, Alces alces, Habitat Relative to Riparian Succession in the
Boreal Forest, Susitna River, Alaska
WILLIAM B. Co ..ins!, and D. J. HELM?
‘Alaska Department of Fish and Game, 1800 Glenn Hwy., Palmer, Alaska 99645
Agricultural and Forestry Experiment Station, University of Alaska Fairbanks, 533 E. Fireweed Avenue, Palmer, Alaska
99645
Collins, William B., and D. J. Helm. 1997. Moose, Alces alces, habitat relative to riparian succession in the boreal forest,
Susitna River, Alaska. Canadian Field-Naturalist 111(4): 567-574.
We documented Moose, Alces alces, habitat characteristics relative to boreal forest succession in the Susitna River flood-
plain, Alaska. Early Shrub and Old Poplar (Populus balsamifera) Forest sites were most important to wintering Moose.
Browse availability was the principal factor affecting winter habitat selection by Moose. Feltleaf Willow (Salix alaxensis)
in Early Shrub was the principal browse species, producing approximately 101 kg/ha of browse. In a year of average snow-
fall, 76% of available Feltleaf Willow was utilized. Other important species, High Bushcranberry (Viburnum edule) and
Rose (Rosa acicularis) were abundant in Old Poplar Forest and Birch-Spruce (Betula papyrifera-Picea glauca) Forest, but
were unavailable when snow exceeded approximately 110 cm. Non-vegetated sites, dry sloughs and frozen river channels
had significantly less (p < 0.05) snow accumulation than other sites, making them preferred paths of access during periods
of deep snow. Wind speed did not vary significantly (p < 0.05) between successional stages older than Early Shrub, and
wind did not appear to affect habitat use. Moose exhibited some preference for cover during periods of rest, especially dur-
ing warm sunny days in late winter. While biologically feasible, enhancement of browse production in the Susitna River
floodplain appears logistically impractical and of wrong priority. Habitat enhancement should be focused on upland sites
where fire suppression has altered natural ecosystem functions, not in floodplain where the river continues to maintain a
constant supply and diversity of successional conditions important to Moose and other wildlife.
Key Words: Moose, Alces alces, browse, cover, habitat, snow, vegetation succession, wind, boreal forest, Susitna River,
Alaska.
Erosion and redeposition of land by glacial rivers
are primary factors in maintaining the productivity and
diversity of boreal forest and associated wildlife habi-
tats (Larsen 1980). In regions where fire suppression is
reducing the frequency and extent of forest rejuvena-
tion and diversification, the perpetual influences of
rivers and streams on forest vegetation and wildlife
habitats are of increasing importance. Moose (Alces
alces), Snowshoe Hares (Lepus americanus), Beaver
(Castor canadensis) and other early successional wild-
life are dependent on the availability of early growth
hardwoods established following fire, fluvial events or
other forest disturbances (Kelsal et al. 1977; Peek et al.
1976; Koehler and Brittell 1990). Not only do early
successional wildlife depend on young hardwoods for
food, but they can significantly affect successional
development of boreal forest (Wolff and Zasada 1979;
Bryant 1987; Pastor et al. 1988; Johnson and Naiman
1990; Helm and Collins 1997).
Glacial rivers flow through most major valleys in
southcentral Alaska, occurring within important winter
ranges of Moose. In the Susitna River valley, Moose
prefer floodplain vegetation types in winter (Albert
and Shea 1986). Wintering Moose are also attracted to
low-lying uplands disturbed by recent fires, homestead
or subdivision clearing, and right-of-way construction
(Chatelain 1951; Albert and Shea 1986). Floodplains
are mainstay habitat for Moose during severe winters,
particularly in areas where lack of recent disturbance
in upland forests has led to a decline in browse avail-
ability (Simkin 1975; Bishop and Rausch 1975).
Chatelain (1951) ranked the Susitna Valley as the
most productive Moose habitat in Alaska. Chatelain
observed that high Moose carrying capacity in flood-
plains of the Susitna River and its tributaries was
further augmented by abundant upland browse
resulting from wildfires and homestead clearing in
the early and mid 1900s. By the early 1970s strict
fire suppression and natural succession in old forest
burns and homesteads had reduced browse availabil-
ity causing Moose populations to decline (Bishop
and Rausch 1974).
The objectives of this study were to document
Moose habitat characteristics related to forest suc-
cession in the Susitna River floodplain and the bore-
al forest generally, and to identify how this riparian
system may be best managed for Moose. We consid-
ered Moose to be a key indicator species (Hanley
1993), indicating availability of early successional
habitat and welfare of associated wildlife.
Methods
We measured characteristics of Moose habitat for
a range of successional conditions common to the
lower Susitna River floodplain (62° N, 150° W). We
made most measurements in a winter of average
567
568
snow accumulation, 1992-1993, and in summer
1993. We noted general habitat conditions and dif-
ferences in Moose distribution from 1981 to 1995.
Vegetation/habitat was represented by Early
Shrub stages of early succession; Alder (Alnus spp.)
and Young Poplar (Populus balsamifera) Forest
stages of intermediate succession; and Old Poplar
Forest and Birch (Betula papyrifera)-Spruce (Picea
glauca) Forest stages of late succession. For more
complete descriptions of these successional stages
see Helm and Collins (1997).
We used a twig-count method (Shafer 1965) to
estimate availability and utilization of browse. We
identified browse as twigs occurring at least 0.5 m
above the ground, on stems less than 4 cm dbh.
We point sampled horizontal cover in late winter
and again in mid summer (Collins in press), by using
an 8X monocular to sight the intersection of two
lines on a target 1.5 m above the ground, 15 m from
the observer. Overlap of the line intersection by veg-
etation indicated a point of cover. We also point
sampled vertical cover in winter and summer, but by
using an 8X rifle scope (mounted to a staff and
viewed vertically through a 90° mirror reflection).
Vegetation overlap of the cross-hair intersection
indicated a point of cover.
180
Snow Depth
150
E>/70cm
120
m>110cm
Total Days
=)
ce)
(o)
Ww
(o>)
1979/80
1980/81
1981/82
1982/83
1983/84
1984/85
1985/86
THE CANADIAN FIELD-NATURALIST
Voli iia
Snow depth and hardness were measured with a
Rammsonde penetrometer (Benson 1962; Coady
1974) at 24 - 30 locations within each vegetation
stage, during winter 1992 - 1993 when snow depth
fluctuated around the most recent 15-year average
(Figure 1). We recorded wind speeds during a rela-
tively strong wind. Successive measurements were
made 1.5 m above the ground at 8 m intervals at 90
points arranged in a grid in each site. Each set of 90
recordings was completed simultaneously during a
15-minute period.
Using pellet-group counts (Neff 1968), we mea-
sured Moose use of different successional stages in
spring, immediately following snow melt. We ran-
domly located fifty belt transects (2 X 60 m) in
three representative stands of each successional stage
and searched for pellet groups deposited following
leaf fall. When transect length exceeded the extent of
a vegetation patch, we randomly relocated and com-
pleted the remaining portion elsewhere in the same
vegetation patch.
Results
Early Shrub vegetation produced approximately
110 kg Moose browse per hectare, making it the most
productive successional stage (Table 1). Feltleaf
1988/89
1989/90
1990/91
1991/92
1992/93
1993/94
1994/95
FiGureE 1. Snow depths at Willow, Alaska (White’s Crossing), 1979-1995. Figure is based on data recorded by the
National Oceanic and Atmospheric Administration. Deepest snow in vicinity of study area occurs at this
location.
COLLINS AND HELM: MoosE HABITAT RELATIVE TO RIPARIAN SUCCESSION 569
1997
0°9¢
g°Sc
HANG,
VST
9°V9
Gl,
O'S
LCL
88
9°96
0°08
SAIS)
L'e9
6°9¢
L°96
619
a!
SIL
C V6
OSL
Cl
peziqun
asMOIG
%
08 V1
L78
OLS
€8°0
67 ES
COE
ES fol
els
LLO
S00
88°C
10'¢
66 IT
v10
ITO
CL It
c0'0
vy 8L
col
C9'SL
L380
(-ey 3)
pournsuos
IsMOIg
L69S
6€ CE
eae
[GAS
€S°C8
y9'9V
68°SE
88°L
£60
S00
09'¢
8C£
e881
cS 0
IT 0
S081
s10
€L'601
LOC
88 001
8L9
(;-ey’3y)
aSMOIG
aqqeyreay
(O19) 990LI
(€981) €€691
(O12) OSL
(L687) E€817Z
(L617) OSLOI
(097) €9¥
(CIS) THs
(L87) TSE
(LS6) 61ST
(LvZ) L9v
(99) €8
(OOLZ) F897
(O01) LIT
(LE6) OOrI
(O@PT) EE98
(089) 99IT
1-o18J09Y
“SuIaIS
(9L'T) $80
(€O'1) 8S°0
(trp) TST
(6€'C) 8S°7
(S77) OL'T
(Or'l) 9€°7
(O77) STE
(88°€) ZO'S
(LS'L) pr'l
(Or'T) €L'°0
(6S'1) 767
(76'€) 80'S
(760) 670
(ELV) 69°7
(6S°8) 19’
(S71) 6r'0
,-UI9}S
"SSIM}
posmolg
(ZT) EEE
(ST'T) LV
(€9°8) 06'6
(C9°7) 9G°€
(pr'Z) STE
(PLT) 08°7
(TET) OTE
(t's) L79
(Ce WSs C
(C61) LOZ
(Iv'7) COE
(Z1°8) T8°L
(07 T) 90°
(81'T) S8°7
(OL'9) ST'9
(08°€) 08"€
j-W9}S'S3IM J,
(Sr'0) LS‘0
(TE0) 85°0
(770) vr'0
(€r'0) 09'0
(6r'0) 90'T
(1S'0) TL'0
(S70) Lv'0
(TS'0) €9°1
(Lv'0) 760
(670) Zr'0
(81°0) rr'0
(LEO) 98'0
(ve'0) 790
(81°0) 7S'0
(09°0) 06'T
(TE) €S'T
(3)
WYSIOM
SIM |
(71) OV
(60) 8’€
(SO) EE
(6:0) 6'€
(0) I't
(3'0) I'r
(60) T'€
(91) 6
(60) 6'€
(80) 9'€
(60) T'€
(TOTP
(THG6E
(60) T’€
(L'1) y's
(TDL
SUISMOIG
jo yutod ye
(UI) IoJOWIeIG
[OL
ajnpa WNUANg1A
SIAD]INDIID DSOY
psafiskdod vinjag
ysoroj sonids-yong
[BIOL
ajnpa WNUANG1A
SIUDINNIVD DSOY
yso1oj sejdod pio
[BIOL
ajnpa WNnuUANg1A
a1] 8UD-avAoU X1]DS
sisuaxD]D XYDS
SIADINIIID DSOY
jso1oj rejdod Suno Xx
TIOL
ajnpa WNUsNg1A
IDI] SUD-aDAOU X1]DG
SISUAXD]D XID
SIADINNIID DSOY
JSPLV
[BIOL
aD1]8UD-aDAOU XI]DS
SISUaKXD]D XIV
viafiuvsjog snjndog
qniys Ayreq
[BIOL
soroeds
9301S
‘sisoujuored Ul Sioa prepurys Aq poMoy][oj ose suvoTy “UTe[dpOOT] JOATY vUJISNG SY} JO SaseE}s [PUOTSSODONS JUSIJOJJIP Ul UONeZI[NN pue AjITIGe[IeAe BSMOIG BSOOJ “| FG"
570
TABLE 2. Horizontal! and vertical? cover (%) in different stages of riparian forest succession, Susitna River, Alaska.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Measurements were made in March and July 1993. Means are followed by coefficients of variation in parentheses.
Stage Winter Summer Winter Summer
Early Shrub
8-yr-old 13.8 (37.7) 39:2'(29-3) 0.3 (10.8) 18.2 (49.2)
14-yr-old 28.2 (11.9) 62.0 (10.2) 8.6 (33.7) 68.0 (10.3)
Alder 24.9 (4.2) S555 ©:2) 11.2 4(G:8) 65.0 (8.7)
Young Poplar Forest 26.0 (5.3) 46.9 (4.8) 14.8 (5.2) 62.4 (6.8)
Old Poplar Forest 23.0 (6.8) dhZ (6:1) 23:4 (ES) 32.1 (Qo)
Birch-Spruce Forest 25:1, (62) 48.7 (5.3) ZL 22) 61.3 (8.3)
‘Horizontal cover = % visual obstruction within 15 m horizontal distance, 1.5 m above ground.
Vertical cover = % visual obstruction of the sky as observed from 1.5 m above ground.
Willow (Salix alaxensis) represented 92% of avail-
able browse and 96% of browse consumed in Early
Shrub. Our estimates of total browse and Feltleaf
Willow availability for Early Shrub were almost
identical to the most productive of similar sites mea-
sured in interior Alaska (Wolff and Cowling 1981).
Development of an Alder overstory was associat-
ed with reduced density and availability of Feltleaf
Willow (Table 1). By this point in succession most
Feltleaf Willow and Balsam Poplar surviving the
combined effects of shading and browsing had
grown beyond the reach of Moose, resulting in an
83% decrease in browse availability.
When sites were approximately 50 years age,
Balsam Poplar dominated young forest overstories,
and the density and availability of Alder and tall
browse species had declined (Helm and Collins
1997). Although increases in Rose (Rosa acicularis)
and High Bushcranberry (Viburnum edule) began to
offset losses of tall browse species (Table 1), Young
Poplar Forest was least productive of Moose browse.
Rose and High Bushcranberry increased when
Balsam Poplar overstories opened as a result of tree
mortality. This made Old Poplar Forests second only
to Early Shrub in terms of browse production and
consumption (Table 1).
By the time Birch-Spruce overstories had devel-
oped in late succession, density and availability of
High Bushcranberry had declined by 78 and 70%,
respectively. Rose increased in density by 58% but
decreased in availability by 41% (Table 1). Shrub-
sized Paper Birch produced limited browse in late
succession.
Horizontal cover in winter, viewed 1.5 m above
ground from a distance of 15 m, increased from 0 to
28% within the first 14 years of vegetation succes-
sion. It then remained relatively constant through all
later successional stages (Table 2). Summer hori-
zontal cover reached 62% in 14-year-old Early
Shrub, but declined to approximately 50% in later
succession.
Early Shrub provided little vertical cover in winter
until approximately 14 years age, when canopies
above 1.5 m height began to close (Table 2).
Increase in size and density of White Spruce caused
Old Poplar Forest and Birch-Spruce Forest to have
the greatest winter vertical cover. Summer vertical
cover was greatest in 14-year-old Early Shrub and
Alder, declining slightly later in succession.
Mean wind speeds measured during winter 1.5 m
above ground were greatest for river channels and
gravel bars lacking above-snow vegetation (Table 3).
Wind speed was significantly reduced in vegetated
areas, with wind in Early Shrub being approximately
55% of that in barren areas. Wind speeds in Alder,
Young Poplar Forest, Old Poplar Forests and Birch-
Spruce Forest were approximately 22% of that in
barren areas.
In late February 1993, snow depth in barren loca-
tions was significantly (p < 0.05) less than at all
TABLE 3. Mean wind speeds (kmh!) during a relatively strong wind, 14 March 1993, within different stages of riparian for-
est succession, Susitna River, Alaska.
Stage x
Non-vegetated 8.8 a!
Early Shrub (8-yr-old) 4.8b
Alder ONG
Young Poplar Forest ee
Old Poplar Forest 2.0c¢
Birch-Spruce forest 2.0c¢
' Means followed by a common letter are not significantly different (p < 0.05).
CV
minimum
oooorf
maximum
a
RA HRROD
1997 COLLINS AND HELM: MoosE HABITAT RELATIVE TO RIPARIAN SUCCESSION 574
TABLE 4. Mean depth (cm) and integrated hardness (kg-f cm) of snow within different stages of riparian forest succession,
Susitna River, Alaska. Means are followed by coefficients of variation in parentheses. Measurements were made in late
March 1993 following a winter of typical snow accumulation.
Hardness Depth
Stage (kg-f cm) (cm)
Non-vegetated 4.76 (61.3) 48.1 a! (58.8)
Early Shrub (8-year-old) 1.96 (15.3) 83.0b (4.5)
Alder 1.96 (14.3) 79.5b (10.3)
Young Poplar Forest 1: 95iG28) ITO) O85)
Old Poplar Forest 2.02 (13.4) 77.4b (10.2)
Birch-Spruce Forest ZAZ GS st) I> bwGhi8)
' Means followed by a common letter are not significantly different (p < 0.05).
other sites (Table 4). Snow depths in other succes-
sional stages did not vary significantly, although
snow in Early Shrub tended to be deepest. We
observed little variation in integrated snow hardness
between vegetated stands, but barren sites were more
than twice as hard.
Following winter 1992-1993, Early Shrub and Old
Balsam Poplar Forest had the highest densities of
pellet groups, 203 and 209 per hectare, respectively
(Table 5). Alder had the lowest density of pellet
groups, 108 per hectare. Assuming Moose consume
an average of 5 kg dry weight per day during winter
(Gasaway and Coady 1974) and that they use all
areas for the same purposes, browse utilization indi-
cated that Early Shrub was used 48% more than Old
Poplar Forest. Pellet group densities were propor-
tionately higher than percentages of browse utiliza-
TABLE 5. Mean number of pellet groups per hectare (p.g.
ha!) within different stages of riparian succession, Susitna
River, Alaska. Pellets were counted in early May 1993.
Estimates are based on 50 120 m* plots in each succession-
al stage. Moose days per hectare (m-d‘ha‘') were estimated
by: 1) dividing total p.g. by and assumed defecation rate of
20 p.g. moose! day", and 2) by dividing browse utilization
(see Table 1) by an assumed forage intake of 5 kg moose"!
day".
m-d ha’! m-d ha"!
(based on 20 p.g. (based on
Stage p.g. ha! moose!day!) 5 kg intake)
Non-vegetated n.d.! n.d. n.d.
Early Shrub 203 (126) 10.2 15.7,
Alder 108 (148) 5.4 2.4
Young poplar
forest 161 (137) 8.1 det
Old poplar
forest 209 (133) 10.5 10.7
Birch-spruce
forest 160 (165) 8.0 3.0
'No data were obtained for non-vegetated sites because they
included extensive snow and ice covered river channels and
sloughs which were washed clean of pellet groups before
snow had melted and pellet groups could be counted.
tion in Alder, Young Poplar Forest and Birch-Spruce
Forest.
Discussion
Habitat
In all years, Feltleaf Willow was the principal
browse resource for Moose wintering in the Susitna
River floodplain. Not only was it highly preferred, as
indicated by rates of utilization, but its availability
was less limited by deep snow than were other
species. Balsam Poplar saplings were available in all
years, but were discriminated against by Moose
except during periods of deep snow when other
browse resources were less available.
Rose and High Bushcranberry were important
browse species, but deep snow prevented their uti-
lization in some winters. Moose utilized these
species less in Birch-Spruce than in Old Poplar
Forest, because those in the Birch-Spruce stands
were shorter and smaller diameter and were bent
over and covered by snow earlier than in Old Poplar
Forest. Limited use of Rose and High Bushcranberry
in Birch-Spruce Forest may also have resulted
because those stands occurred primarily in the most
stable portions of the floodplain, farthest from
recently disturbed Early Shrub where Moose pre-
ferred to feed.
Highest percentages of browse utilization occurred
in Young Poplar Forests, even though those stands
were less productive of browse than any other suc-
cessional stage. High pellet group density associated
with numerous bed depressions indicated Moose pre-
ferred Young Poplar Forests and Alder stands for
resting (Collins and Urness 1979 and 1981).
Alder and Young Poplar Forest may have been
preferred resting sites for Moose because they pro-
vided better cover than many Early Shrub stands.
Young Poplar Forest and Alder were successionally
and spatially adjacent to Early Shrub where Moose
preferentially browsed. However, requirement for
cover did not keep Moose from utilizing preferred
Willows within sparsely vegetated stands. Early
Shrub isolated on small islands 100’s of meters from
other vegetation received utilization equivalent to
572
Early Shrub within a few meters of dense cover. The
only other apparent preference of Moose for cover
occurred in late winter on warm, sunny days.
In March and April, Moose frequently lay in the
shade of mature White Spruce during sunny days,
suggesting warmer temperatures and more direct
sunlight caused them to seek shade to reduce heat
stress (Schwab and Pitt 1991). Renecker and Hudson
(1986) observed that Moose were more likely to
show signs of heat stress than cold stress in winter.
They observed increased metabolic rates of Moose
when ambient temperatures exceeded -5.1°C, and
observed Moose panting at 2.2°C. Demarchi and
Bunnell (1995) similarly observed that cow Moose
used denser forest during periods of heat stress in
summer. Forage was not available beneath spruce
and, therefore, we do not believe food attracted
Moose to spruce as it sometimes does to upland
spruce (LeResche and Davis 1973).
Moen (1973) concluded that reduction of wind
velocity is one of the most significant benefits of
cover to animals in winter, usually even more impor-
tant than its effect on radiant energy flux.
Nevertheless, we and Modafferi (personal communi-
cation) have not observed Moose increasing their use
of denser cover in the Susitna floodplain during peri-
ods of wind. One-minute-wind speeds (National
Oceanic and Atmospheric Administration,
Climatological Data, Alaska, 1979-1995) exceeding
28.8 km‘h", the limit to thermoneutrality for Moose
calves at -20° C (Renecker et al. 1978), did not occur
in March and April when Moose used Spruce cover.
Snow on barren gravel bars and ice-covered river
channels was wind compacted and twice as hard as
in vegetated sites, but did not support Moose. Except
for winters 1989-1990, 1990-1991, 1991-1992 and
1994-1995 snow in these sites did not exceed the 70
cm depth Coady (1974) concluded represents a slight
impediment to Moose. Consequently, braided river
channels represented a network of corridors, provid-
ing Moose good access to all parts of the floodplain.
Besides elevating energetic costs to Moose, snow
deeper than 70 cm (Figure 1) reduced browse avail-
ability. Rose and High Bushcranberry represented
22% and 29%, respectively, of all browse available
in early winter, but they were buried when snow
depth reached 70-110 cm. Snow greater than 110 cm
caused Moose to stop browsing in forests by mid
November in winter 1994-1995 and by late
December in winters 1989-1990, 1990-1991, and
1991-1992 (NOAA).
Our estimates of Moose densities based on pellet
group densities were subjective, since we did not
determine defecation rates. We computed Moose
densities, using a rate of 20 pellet groups per day
(Andersen et al. 1992), because we assumed the
quality and availability of forage in fall and early
winter would have elevated food consumption and
THE CANADIAN FIELD-NATURALIST
Vol. 111
defecation rates above those determined in late win-
ter by Franzmann and Arneson (1975), producing a
season average closer to 20.
We became concerned about loss of pellet groups
prior to counting, because Moose regularly used ice
covered river channels and dry sloughs for access to
feeding sites. We observed that spring flooding
washed pellet groups away before snowmelt allowed
pellet groups to be counted. Collins and Urness (1979)
reported that Elk (Cervus elaphus nelsoni) defecated
11 to 18 times more frequently when traveling from
one location to another, causing approximately 40%
of all defecations to occur within 3.5 to 5.6% of the
day as the animals were walking between feeding or
resting sites. We concluded pellet groups could not be
used to estimate overall Moose density in the flood-
plain but that they were useful in conjunction with
other data for describing Moose distribution.
Browse utilization, tracks, beds and fecal deposi-
tion indicated only rare, localized summer use of the
lower Susitna River floodplain by Moose. There
were no indications of Moose in most areas in sum-
mer, an exception being use of several islands north
of Talkeetna in late May and June. This agrees with
Modafferi (1988) who observed a tendency for
radio-collared cow Moose north of Talkeetna to
leave the floodplain as snow receded but then to
return for the period late May through June.
We believe that Moose are currently utilizing the
lower Susitna River floodplain near winter capacity.
During a year of average snowfall, Moose utilized
preferred browse species at approximately 75%, a
level Wolff and Zasada (1979) suggested represents
carrying capacity for similar vegetation. Browse not
utilized was apparently of poorer quality and/or less
efficiently ingested. Although abundant reserves of
less preferred browse species existed, snow depths
exceeding 70-110 cm triggered several events that
combined to decrease food availability and reduce
Moose productivity and survival.
Deep snow eliminated availability of Rose, High
Bushcranberry and short individuals of other browse
species, confined Moose to the most accessible sites,
and accentuated negative energy balance for Moose
by causing them to expend greater energy for move-
ment. Griese (in press) observed a 35% decline in
the Susitna Valley Moose population following the
deep-snow winter 1989-1990. Since that time, a con-
tinuing series of deep-snow winters has resulted in
the population remaining at or below the 1990 level.
Griese (in press) reported that Susitna Valley Moose
have experienced significant winter die-off at least
once each decade starting in the 1950s.
Management implications
Erosion and redeposition of substrates within a
braided river system like the lower Susitna River is a
dynamic process controlled almost entirely by dis-
1997
charge variations and sediment loads (Leopold
1964). Conditions necessary for vegetation succes-
sion to proceed are established as sedimentation ele-
vates sites, reducing flooding frequency. Shifts of
river channels quickly reverse succession, rejuvenat-
ing or eliminating browse stands (Helm and Collins
1997).
Unless flow of the Susitna River is interrupted by
hydroelectric development, there is no need or rea-
sonable opportunity to enhance browse productivity.
While total area of floodplain covered by each suc-
cessional stage may remain relatively constant over
time, and while successional timeframes are general-
ly predictable, life expectancies of specific sites are
unpredictable. Eleven of 20 early successional sites
we monitored were either temporarily denuded or
completely eroded and redeposited down stream dur-
ing the period 1981-1995, effectively rejuvenating
them without human intervention. By contrast,
browse production in upland sites can be efficiently
and predictably enhanced by crushing, cutting, or
prescribed burning (Oldemeyer and Regelin 1987;
Collins 1996).
Browse production within Balsam Poplar Forest
and Birch-Spruce Forest can be stimulated by a com-
bination of overstory removal and scarification
(Zasada et al. 1981; Collins 1996), costs being subsi-
dizea by timber sales. However, costs and accessibil-
ity make it difficult to justify overstory removal in
floodplain solely to enhance browse production.
Felling of Balsam Poplar and Birch must be fol-
lowed by timely scarification to obtain hardwood
density meeting minimal reforestation standards
(Collins 1996). This requires use of heavy equipment
during that time of year when it is not possible to
construct winter roads or ice bridges to cleared sites.
Browse regeneration failed in floodplain stands win-
ter logged as much as 30 years ago, because the sites
were not scarified, and preexistent Bluejoint
Reedgrass (Calamagrostis canadensis) and Alder
(Alnus spp.) increased to exclude other species
(Mitchell and Evans 1966; Collins 1996). Tree har-
vest eliminated opportunity for natural scarification
through uprooting by wind and gravity (Jonsson and
Dynesius 1993; Helm and Collins 1997). Similarly,
we observed regeneration failure in Young Poplar
Forest where most Balsam Poplars were felled by
Beavers.
The rapidity with which early seral vegetation in
floodplains grows out of reach of Moose and/or the
frequency with which it is rejuvenated by fluvial
forces make it impractical to assess the welfare of
Moose and their habitat by traditional methods of
monitoring condition and trend (Stoddart et al.
1975). However, general assessment of browse with-
in the floodplain indicates Moose are near ecological
carrying capacity, being periodically limited by
snow-induced shortages of food (Caughley and
Sinclair 1994). Ecological carrying capacity of the
COLLINS AND HELM: MoosE HABITAT RELATIVE TO RIPARIAN SUCCESSION
573
Susitna Valley as a whole is primarily limited by fire
suppression in lands adjacent to the floodplain.
Moose are limited by frequency of natural distur-
bances supporting establishment of early succession-
al vegetation. Therefore, effective management must
recognize the collective importance of all naturally
occurring disturbances (erosion, forest diseases,
windfall and fire) in maintenance of primary produc-
tivity within the boreal forest (Attiwill 1994).
Managers should first attempt to enhance Moose
habitat by eliminating or altering management prac-
tices which disrupt or prevent natural forces from
maintaining diverse and productive habitat. Direct
efforts to enhance habitat are not only more costly,
but often, as on the Susitna River floodplain, mis-
placed.
Acknowledgments
This project is a contribution of Federal Aid
Wildlife Restoration Projects W-24-3 and W-24-4.
Literature Cited
Albert, S. W., and L. C. Shea. 1986. Moose winter habi-
tat in the lower Susitna Valley, Alaska: Pilot Project on
Habitat Suitability Assessment. Alaska Deptartment of
Fish and Game Technical Report, Number 86-6. 105
pages.
Andersen, R., O. Hjeljord, and B. Saether. 1992. Moose
defecation rates in relation to habitat quality. Alces 28:
95-100.
Attiwill, P. M. 1994. The disturbance of forest ecosys-
tems: the ecological basis for conservative management.
Forest Ecology and Management 63: 247-300.
Benson, C. A. 1962. Stratigraphic studies in the snow and
firn of the Greenland ice sheet. U.S. Army Snow, Ice,
and Permafrost Research Establishment/CRREL Report
70. 93 pages.
Bishop, R. H., and R. A. Rausch. 1974. Moose popula-
tion fluctuations in Alaska, 1950-1952. Le Naturaliste
canadien 101: 559-593.
Bryant, J. P. 1987. Feltleaf willow-snowshoe hare inter-
actions: plant carbon/nutrient balance and floodplain
succession. Ecology 68: 1319-1327.
Caughley, G., and A. R. E. Sinclair. 1994. Wildlife
Ecology and Management. Blackwell Scientific
Publications, Boston. 334 pages.
Chatelain, E. F. 1951. Winter range problems of moose
in the Susitna Valley. Proceedings of Alaska Science
Conference 2: 343-347.
Coady, J. W. 1974. Influence of snow on behaviour of
moose. Le Naturaliste canadien 101: 417-436.
Collins, W. B. 1996. Wildlife habitat enhancement in the
spruce-hardwood forest of the Matanuska and Susitna
River Valleys. Alaska Department of Fish and Game
Federal Aid in Wildlife Restoration Program Report,
Project W-23-5, Study 1.44. 52 pages.
Collins, W. B. in press. Measurement of horizontal cover.
Journal of Range Management.
Collins, W. B., and P. J. Urness. 1979. Elk pellet group
distributions and rates of deposition in aspen and lodge-
pole pine habitats. Pages 140-144 in: Elk Ecology
Symposium. Edited by M. S. Boyce and L. D. Hayden-
Wing. University of Wyoming, Laramie.
574
Collins, W. B., and P. J. Urness. 1981. Habitat prefer-
ences of mule deer as rated by pellet-group distributions.
Journal of Wildlife Management 45: 969-972.
Demarchi, M. W., and F. L. Bunnell. 1995. Forest cover
selection and activity of cow moose in summer. Acta
Theriologica 40: 23-36.
DesMueles, P. 1968. Determination of the number of pel-
let groups voided and the number of beds established by
moose in winter. Le Nauraliste canadien 95: 1153-1157.
Gasaway, W. C., and J. W. Coady. 1974. Review of
energy requirements and rumen fermentation in moose
and other ruminants. Le Naturaliste canadien 101:
227-262.
Griese, H. J. in press. Survey-Inventory Management
Report. Moose. Alaska Department of Fish and Game,
Federal Aid in Wildlife Restoration, Project W-24-4,
Study 1.0. Juneau, Alaska.
Hanley, T. A. 1993. Balancing economic development,
biological conservation, and human culture: the Sitka
black-tailed deer Odocoileus hemionus sitkensis as an
ecological indicator. Biological Conservation 66:
61-67.
Helm, D. J., and W. B. Collins. 1997. Vegetation suc-
cession and disturbance on boreal forest floodplain,
Susitna River, Alaska. Canadian Field-Naturalist 111:
553-566.
Johnson, C. A., and R. J. Naiman. 1990. Browse selec-
tion by beaver: effects on riparian forest composition.
Journal of Forest Research 20: 1036-1043.
Jonsson, B. B., and M. Dynesius. 1993. Uprooting in
boreal forests: long-term variation in disturbance rate.
Canadian Journal of Forest Research 23: 2383-2388.
Kelsall, J. P., E. S. Telfer, and T. D. Wright. 1977. The
effects of fire on the ecology of the boreal forest, with
particular reference to the Canadian north: a review and
selected bibliography. Canadian Wildlife Service
Occasional Paper Number 32. Ottawa, Ontario. 58 pages.
Koehler, G. M., and J. D. Brittell. 1990. Managing
spruce-fir habitat for lynx and snowshoe hares. Journal
of Forestry 88: 10-14.
Larsen, J. A. 1980. The Boreal Ecosystem. Academic
Press, New York. 439 pages.
Leopold, L. B., M. G. Wolman, and J. P. Miller. 1964.
Fluvial processes in geomorphology. W. H. Freemen
and Company, San Francisco. 522 pages.
LeResche, R. E., and J. L. Davis. 1973. Importance of
nonbrowse foods to moose on the Kenai Peninsula,
Alaska. Journal of Wildlife Management 37: 279-287.
Mitchell, W., and J. Evans. 1966. Composition of two
disclimax bluejoint stands in southcentral Alaska.
Journal of Range Management 44: 65-68.
Modaferri, R. D. 1988. Susitna Hydroelectric Project,
Final Report Big Game Studies, Volume I - Moose -
THE CANADIAN FIELD-NATURALIST
Vol tM
Downstream. Alaska Department of Fish and Game.
211 pages.
Moen, A. N. 1973. Wildlife Ecology. W. H. Freeman and
Co. San Francisco. 458 pages.
Neff, D. J. 1968. The pellet-group count method for big
game trend, census, and distribution: a review. Journal
of Wildlife Management 32: 597-614.
Oldemeyer, J. L., and W. L. Regelin. 1987. Forest suc-
cession, habitat management, and moose on the Kenai
National Wildlife Refuge. Swedish Wildlife Research
Supplement 1: 163-179.
Pastor, J., R. J. Naiman, B. Dewey, and P. McInnes.
1988. Moose, microbes, and the boreal forest.
BioScience 38: 770-777.
Peek, J. M., K. L. Uruch, and R. J. Mackie. 1976.
Moose habitat selection and relationships to forest man-
agement in northeastern Minnesota. Wildlife Monograph
48. 65 pages.
Renecker, L. A., and R. J. Hudson. 1986. Seasonal ener-
gy expenditures and thermoregulatory responses of
moose. Canadian Journal of Zoology 64: 322-327.
Renecker, L. A., R. J. Hudson, M. K. Christophersen,
and C. Arelis. 1978. Effect of posture, feeding, low
temperature, and wind on energy expenditures of moose
calves. Proceedings of the 14th North American Moose
Conference and Workshop. Halifax, Nova Scotia.
Schwab, F. E., and M. D. Pitt. 1991. Moose selection of
canopy cover types related to operative temperature, for-
age, and snow depth. Canadian Journal of Zoology 69:
3071-3077.
Shafer, E. L. 1965. The twig-count method for measuring
hardwood deer browse. Journal of Wildlife Management
27: 428-437.
Simkin, D. W. 1975. Reproduction and productivity of
moose. Le Naturaliste canadien 101: 517-526.
Stoddart, L. A., A. D. Smith, and T. W. Box. 1975.
Range Management. McGraw-Hill, Inc. 532 pages.
Wolff, J. O., and J. Cowling. 1981. Moose browse uti-
lization in Mount McKinley National Park, Alaska.
Canadian Field-Naturalist 95: 85-88.
Wolff, J. O., and J. C. Zasada. 1979. Moose habitat and
forest succession on the Tanana River flood plain and
Yukon-Tanana upland. Jn Proceedings of the North
American Moose Conference and Workshop. Edited by
H. G. Cumming. School of Forestry, Lakehead Univer-
sity, Thunder Bay, Ontario 15: 213-245.
Zasada, J. C., L. A. Viereck, M. J. Foote, R. H. Parken-
son, J. O. Wolff, and L. A. Lankford Jr. 1981. Natural
regeneration of balsam poplar in the Susitna Valley,
Alaska. Forestry Chronicle 57: 57-65.
Received 2 November 1995
Accepted 3 March 1997
Detection and Distribution of the Apple Leaf Midge, Dasineura mali,
in Nova Scotia
BRIAN R. EATON! ? and ERIKA BENT?
'Nova Scotia Tree Fruit Research Foundation, Agricultural Centre, Kentville, Nova Scotia B4N 1J5
Agricultural Pest Monitoring, P. O. Box 550, Wolfville, Nova Scotia BOP 1X0
3Present address: Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9 [Author to whom
correspondence should be addressed. ]
Eaton, Brian R., and Erika Bent. 1997. Detection and distribution of the Apple Leaf Midge, Dasineura mali, in Nova
Scotia. Canadian Field-Naturalist 111(4): 575-579.
The Apple Leaf Midge (Dasineura mali (Kieffer)) (Diptera: Cecidomyiidae), a foliar pest of apple trees, originated in
Europe and was first found in Canada in 1964, when the midge was detected in New Brunswick. Adult midges oviposit on
developing apple leaves; the feeding activity of hatched larvae prevents these leaves from unrolling properly, resulting in
reduced photosynthesis. Midge infestations on mature trees are not usually problematic. On nursery stock or young trees,
however, midge infestations can cause stunting and disrupt tree training. Using data from monitoring records for the period
1990 to 1994, we trace the establishment and spread of the Apple Leaf Midge in Nova Scotia. Results from a 1993 survey
of the severity of midge infestations in apple orchards are also presented. Both surveys demonstrate the speed with which
the midge spread from the area of initial detection throughout the apple-growing region of Nova Scotia.
Key Words: Apple Leaf Midge, Dasineura mali, apple orchards, Nova Scotia.
The Apple Leaf Midge, Dasineura mali (Kieffer)
(Diptera: Cecidomyiidae), a foliar pest of apple
trees, originated in Europe and accompanied the
spread of apple from this region to other parts of the
globe. In the USA, D. mali was first detected in
1928 in Massachusetts; by 1933, the midge had
infested an area of approximately 1000 square
miles, which included parts of Massachusetts and
New Hampshire (Whitcomb 1934). By 1936,
orchards in Maine and New York state were
attacked, possibly because of wind dispersal of adult
midges (Whitcomb 1937). In New Zealand, the
midge was initially found in 1950 at Auckland,
Palmerston North, and Tauranga; in Auckland it had
spread from the original infestation site to neigh-
bouring orchard areas by 1952 (Morrison 1953). In
Canada, D. mali was first detected in 1964 in
Charlotte County, New Brunswick. From there it
spread to infect orchards throughout the Saint John
River Valley from Woodstock to the Gagetown-
Queenstown area by 1978 (MacPhee and Finnamore
1978). The midge is presently found in all apple
growing regions of New Brunswick (M. Mazerolle,
personal communication). In 1993, larvae found in
apple leaf samples collected near Clearbrook,
British Columbia were tentatively identified as D.
mali; formal identification awaits the collection of
adult specimens in good condition (H. Philip, per-
sonal communication). Here we report on the estab-
lishment and spread of D. mali in Nova Scotia.
In Nova Scotia, D. mali was first discovered in an
orchard in Aylesford in 1990, where its presence
was detected by the distinctive damage caused by
larval feeding. Adult leaf midges deposit eggs on
the margins of growing leaves; when the eggs hatch,
the feeding activity of the larvae prevents the leaf
from unrolling properly, resulting in the formation
of a tight roll (a gall), with the edges of the leaves
rolled toward, and parallel to, the midvein. Leaves
attacked in this manner do not function properly,
and eventually become brittle and discoloured;
affected leaves may drop prematurely. Because D.
mali only attacks growing shoots, it is not of major
concern in mature trees, which can usually sustain
this type of damage; in nursery stock or young trees,
however, this pest can cause stunting and disrupt
tree training (Whitcomb 1934). There are usually
two to three generations of the Apple Leaf Midge in
Nova Scotia.
Methods
Distribution of the Apple Leaf Midge in Nova
Scotia’s Annapolis Valley for the period 1990-1994
was determined using presence / absence data. The
majority of this data was taken from the records of
Agricultural Pest Monitoring (APM), a private com-
pany which offers custom pest monitoring programs
to fruit growers in the Annapolis Valley.
Agricultural Pest Monitoring surveys approximately
150 orchard blocks per year; the presence of Apple
Leaf Midge was recorded for each orchard in which
it was encountered. Distribution data was also
obtained during the Intensive Integrated Pest
Management Project (IPM), a technology-transfer
project funded by the Agri-Food Development
Agreement; during this project, eleven orchards
were monitored weekly throughout the growing sea-
son for a period of five years.
575
576
LEGEND 2
Presence/ Absence
ZED &
E
ANNAPOLIS COUNTY
THE CANADIAN FIELD-NATURALIST
Vol. 111
BAY OF FUNDY
ips ces
ae \ IN SOF
4 YQ :
= "OMY
o8*
(lly,
WS
CY)
S
FiGurE 1. Distribution of Apple Leaf Midge infestations in the Annapolis Valley of Nova Scotia for the period 1990-1994.
Each circle represents one orchard. In cases where the midge was found in the same orchard in multiple years, the
year of first infestation is shown. Inset shows the location of study area within the province.
Presence / absence data was also recorded during
a survey of the severity of Apple Leaf Midge infesta-
tions in the Annapolis Valley. This survey was done
between 21 September and 3 November 1993, and
included 57 orchards, some of which were the same
orchards monitored by APM or during the IPM pro-
ject. Orchards were surveyed in autumn because
galls formed by the last generation of the midge are
usually more numerous than those from earlier gen-
erations, making detection of the pest in an orchard
easier. The object of the severity survey was to
determine where in the Annapolis Valley the Apple
Leaf Midge was most prevalent. Orchards from the
eastern (Windsor area) to the western end
(Annapolis area) of the Valley were surveyed, as
well as orchards on the valley floor, and on the
slopes of the North and South Mountain. Severity of
midge infestation in each orchard was rated accord-
ing to the following scale: 0 - no midge galls found,
1 - at least one gall found, 2 - a few galls throughout
the orchard, 3 - a few galls on many trees, 4 - several
galls on all trees, or many galls on a few trees, and 5
- galls on 50% or more of the new growth.
Locations of orchards surveyed for Apple Leaf
Midge presence or infestation severity were plotted
on 1:50 000 topographic maps. For each orchard, the
universal transverse mercator grid (UTM) coordinate
corresponding approximately to the centre of the
orchard was recorded. These coordinates were used
to generate maps of midge distribution and severity
ratings in Nova Scotia.
The distribution map (Figure 1) was constructed
using a process known as buffering, in which a
boundary of a fixed size is drawn around a geo-
graphic point, forming a series of large dots (Roberts
et al. 1993). In mapping the distribution of the leaf
midge, a buffer of 1000 m was drawn around the
point coordinate used to designate each orchard.
Where infested orchards occur in close proximity,
the dots for each orchard blend together, forming
belts where leaf midge were found. In those cases
where the midge was found in the same orchard in
multiple years, only the year in which it was first
detected is mapped. In Nova Scotia, little success has
been achieved in eradicating the Apple Leaf Midge
from infested orchards. Therefore, it is reasonable to
assume that when an orchard becomes infested by
this pest, it generally remains infested.
The map of severity ratings (Figure 2) was con-
structed by assigning a symbol reflecting the severity
of the infestation (the larger the symbol, the worse
the infestation) to the point representing each
orchard. A different symbol was used to designate
orchards where no midge population was found.
1997
LEGEND o
Level of Severity
ANNAPOLIS COUNTY
EATON AND BENT: APPLE LEAF MIDGE IN NOVA SCOTIA
BAY OF FUNDY
577
FIGURE 2. Severity of midge infestations in individual orchards in the Annapolis Valley of Nova Scotia in 1993. The size
of the symbol for each orchard reflects the severity of the infestation, not the size of the orchard. Severity levels
were rated according to a six number scale, with 0 being the lowest severity, and 5 the highest. Inset shows the
location of study area within the province.
Each symbol is centred about the point denoting the
orchard’s coordinate; the size of the symbol does not
reflect the geographic extent of the infestation in that
area, but merely reflects the severity rating assigned
to that orchard.
Results and Discussion
Apple orchards in the Annapolis Valley are con-
centrated in Kings County, which contains approxi-
mately 90% of the Valley’s production by area
(Embree et al. 1984). The Apple Leaf Midge, initial-
ly discovered near Aylesford in 1990, spread quickly
throughout Kings County, infesting an increasingly
large number of orchards (Figure 3). The midge
reached orchards near Kentville (to the east) and
Kingston (just over the Kings/Annapolis County
border to the west) by 1992 (Figure 1). By 1993, the
midge was detected at the far western end of the
Valley, around Annapolis Royal, while it first
appeared at the far eastern end, near Windsor, in
1994. Both these regions are somewhat isolated from
the main apple producing areas.
Results from the 1993 severity survey in general
reflect the history of leaf midge distribution in Nova
Scotia (Figure 2). The most severe infestations were
concentrated around the Aylesford area, with less
acute cases at the ends of the Valley, which the
midge had invaded more recently. Even isolated
orchards on the slopes of the North and South
Mountain had moderate infestations.
Severity ratings of leaf midge infestations should
be interpreted with some caution, as the severity
depends on several factors, including apple tree
variety (Whitcomb 1942), and the amount of new
tissue growth available (Todd 1959). D. mali adults
only oviposit on new growth; when little new
growth is available, fewer leaves are infested. Those
which are attacked, however, have large numbers of
eggs (Whitcomb 1941). The amount of new growth
is influenced, in turn, by climatic conditions, time of
the season, rootstock, tree variety, and orchard man-
agement practices such as pruning and fertilization
regimes. We surveyed midge infestations during the
fall, when no tree growth was occurring, but the
impact of the midge during the past season was evi-
dent. We did not estimate the amount of new growth
which had been produced by trees in individual
orchards.
Both the presence / absence data and the severity
survey suggest that the Aylesford area served as the
initial invasion point of the Apple Leaf Midge into
the Annapolis Valley. Once in the Valley, this pest
Nn
~~
(oe)
Soe
Number of infested orchards
Oo
Oo
20
10
0
1990 1991 1992 1993 1994
Year
FiGuRE 3. Number of orchards sampled by Agriculture Pest
Monitoring company which were infested by Apple
Leaf Midge. Approximately 150 orchards were
sampled annually.
spread quickly throughout the apple-growing region.
This rapid dispersal through a newly invaded area
parallels experiences in the USA and New Zealand
(Whitcomb 1934; Morrison 1953). Infested plant
material has been the suggested route of invasion in
both the northeastern United States, where material
from France was implicated (Felt 1932), and New
Zealand, where the appearance of the midge was
linked to importation of material from Holland
(Morrison 1953). No specific case of human intro-
duction has been suggested for the appearance of the
leaf midge in New Brunswick. There has been some
suggestion that the infestation in British Columbia
may have originated from root stock imported from
Holland, though this has not been confirmed
(H. Philip, personal communication). Movement of
the midge into Nova Scotia (possibly from New
Brunswick) may have resulted from wind dispersal
or the movement of material or equipment (such as
apple bins) into Nova Scotia. The midge would have
had to travel approximately 135 km (straight-line
distance) over the open water of the Bay of Fundy
from infested areas in New Brunswick to reach the
site of first detection in Nova Scotia. Moreover,
because Nova Scotia is an isthmus, and virtually no
apples are grown within 90 km or more of the Nova
Scotia - New Brunswick border, it seems unlikely
that the dispersal of the midge to Nova Scotia is a
natural event. The most probable scenario is that the
midge was introduced to the province via equipment
that was moved from New Brunswick to Nova
Scotia, or infested plant material from other areas
where the midge is found.
In the USA, Whitcomb (1937) found that the
Apple Leaf Midge had spread from its original infes-
tation area in northeastern Massachusetts and south-
eastern New Hampshire in 1933, to southern Maine
THE CANADIAN FIELD-NATURALIST
Vol. 111
(a distance of approximately 36 km) and Rochester,
New York (a distance of approximately 540 km) by
1936. Whitcomb (1937) suggested that the spread of
the midge into Maine may have resulted from wind
dispersal of the adults; he did not make a similar
suggestion for the appearance of the midge in New
York. In British Columbia, the midge spread from
the original 1993 infestation site near Clearbrook, to
orchards in the Yarrow area by 1995, a distance of
about 25 km (H. Philip, personal communication).
The leaf midge, while not usually damaging to
mature trees, does pose a threat to nursery stock and
young plantings. Research in midge control in
Canada has centred on finding pesticides which
would offer good control, and the use of biocontrol
agents. Between 1981 and 1985, 730 adults of
Inostemma contariniae Szelényi (Hymenoptera:
Platygasteridae) and 7472 adults of Platygaster mar-
chali Kieffer (Hymenoptera: Platygasteridae) (from
Austria and Italy) were released in New Brunswick
(Agriculture and Agri-Food Canada 1995). Adult
females of most species in this family of wasps
deposit their eggs within the midge embryo encased
in its own egg; when the parasitoid eggs hatch, the
wasp larvae feed on the larva of the midge, eventual-
ly killing it. The present status of these species in
New Brunswick is unknown.
In 1993, several hundred adults of Platygaster
demades (Walker) (Hymenoptera: Platygasteridae)
from Holland were released in an orchard near
Berwick, Nova Scotia to combat the leaf midge
(Hardman 1993). The parasite survived the Canadian
winter, and P. demades larvae were found infesting
leaf midge larvae in the fall of 1994 and 1995
(Hardman, personal communication). The spread of
the parasite is expected to be slow, but hopefully it
will offer some degree of natural control over the
leaf midge in the future.
Acknowledgments
Thanks are extended to Danny Swim for making
the maps, and to Myrna Blenkhorn, Bill Craig, and
Dick Rogers for assisting in the midge severity sur-
vey. Thanks to Alana Eaton, Mike Hardman, and
two anonymous reviewers for comments on this
manuscript. This work was supported by the
Canada/Nova Scotia Agri-Food Development
Agreement (Project # TIP 264), the Nova Scotia
Tree Fruit Research Foundation, and the Nova Scotia
Department of Agriculture and Marketing, Plant
Industry Branch.
Literature Cited
Agriculture and Agri-Food Canada. 1995. PRIS -
Parasitic and predatory insect releases; choose Index:
agent, search using “Dasineura” and “mali”. [CD-ROM].
Available: CHEM Source, CCInfo Disc, Canadian
Centre for Occupational Health and Safety. 95(3):
records number 1207, 1222, 1280, 1335, 1340, 1382,
1389, 1456, 1465, 1466, 1467, 1468, and 1469.
19977
Embree, C. G., M. J. Blenkhorn, and B. J. McLaughlin.
1984. Changes in the fruit tree inventory of the Annapolis
Valley of Nova Scotia: 1939 - 1982. Nova Scotia Fruit
Growers’ Association, Kentville, Nova Scotia. 30 pages.
Felt, E. P. 1932. Apple leaf curling midge. Journal of
Economic Entomology 25: 932.
Hardman, J. M. 1993. Biological control of mites and
apple leaf curling midge in Holland and in Nova Scotia.
Nova Scotia Fruit Growers’ Association Annual Report
130: 39-47.
MacPhee, A. W., and D. B. Finnamore. 1978. The apple
leaf midge, Dasineura mali (Diptera: Cecidomyiidae), in
New Brunswick apple orchards. Canadian Entomologist
111: 1363-1364.
Morrison, L. G. 1953. Apple leaf-curling midge in New
Zealand. New Zealand Journal of Agriculture 86:
311-315.
Roberts, E. A., F. W. Ravlin, and S. J. Fleischer. 1993.
Spatial data representation for integrated pest manage-
ment programs. American Entomologist 39(2): 92-107.
EATON AND BENT: APPLE LEAF MIDGE IN NoVA SCOTIA
579
Todd, D. H. 1959. The apple leaf-curling midge. Das-
yneura mali Kieffer, seasonal history, varietal suscepti-
bility and parasitism 1955-58. New Zealand Journal of
Agricultural Research 2: 859-869.
Whitcomb, W. D. 1934. The apple leaf-curling midge, a
new pest of apples. Journal of Economic Entomology
27: 355-361.
Whitcomb, W. D. 1937. Apple leaf curling midge. Bulletin
of the Massachusetts State Agricultural Experiment
Station 339: 57-58.
Whitcomb, W. D. 1941. Biology and control of the apple
leaf curling midge. Bulletin of the Massachusetts State
Agricultural Experiment Station 378: 68-69.
Whitcomb, W. D. 1942. Biology and control of the apple
leaf curling midge. Bulletin of the Massachusetts State
Agricultural Experiment Station 388: 63-64.
Received 15 February 1996
Accepted 14 February 1997
Songbird Nest Placement in Vermont Christmas Tree Plantations
JEFFREY W. HUGHES! and FRANKLYN K. HUDSON2
‘School of Natural Resources and Field Naturalist Graduate Program, Department of Botany, University of Vermont,
Burlington, Vermont 05405
2SUNY College of Environmental Studies and Forestry, Syracuse, New York 13210
Hughes, Jeffrey W., and Franklyn K. Hudson. 1997. Songbird nest placement in Vermont Christmas tree plantations.
Canadian Field-Naturalist 111(4): 580-585.
We studied arboreal nest site selection in three Vermont Christmas tree plantations as a way to simplify analysis of vari-
ables that affect nest placement. Nests of House Finch (Carpodacus mexicanus) and Chipping Sparrow (Spizella passerina)
were most common (16 and 8 nests, respectively, of a total of forty), followed by American Robin (Turdus migratorius) (6
nests). Accounting for differences in availability of spruce and fir trees for nesting, five of the six bird species that nested
in the plantations preferred White Spruce over Balsam-Fraser Fir. Overhead shading of nests in spruce, our measure of nest
concealment, was significantly higher than in fir; shading from the side was similar. Almost all bird species exhibited tree-
height and nest-height preferences but height of nest placement tended to increase significantly with tree height.
Concealment of nests from overhead predators, or protection from inclement weather, appeared to be the primary factors in
nest-site selection for most species that nested in these plantations.
Key Words: House Finch, Carpodacus mexicanus, Chipping Sparrow, Spizella passerina, American Robin, Turdus migra-
torius, nest site selection, nesting substrate, edge-species, interior-species, Vermont.
Determining which cues are used by birds to
select nest sites is difficult because many factors
influence suitability of a potential site. Variables
which may influence nest placement (reviewed in
Martin and Roper 1988; Holway 1991; Martin 1993;
Filliater et al. 1994) include availability of food,
competition, substrate, natal imprinting, bird age and
experience, vegetation structure, cover and conceal-
ment, surrounding land use, nest microclimate, rela-
tionship to edge, and vulnerability to nest predation.
These potentially interacting variables impart a
uniqueness to each nest location that greatly compli-
cates efforts to determine specific characteristics of
desired nest locations, or how these characteristics
vary among bird species. We designed a study that
greatly reduced the number of covariates in nest-site
selection (e.g., structural and compositional com-
plexity of the forest) by studying nest placement in
Christmas tree plantations. The plantations that we
selected for study enabled us to compare nesting
preferences directly by substrate height, substrate
location (edge vs interior), and substrate type (spruce
or fir). We also evaluated the extent to which nest
placement was related to shading of nests by foliage.
Methods
Study Sites
Three Christmas tree plantations in central
Vermont were selected for our study of nest-site
selection. One plantation each was located in three
of the region’s main physiographic provinces
(Champlain Lowlands, Green Mountains, and
Vermont Piedmont; Johnson 1980). The climate
varies somewhat by elevation and physiographic
province but the region is generally characterized as
humid continental, with warm summers and long,
cold winters. Weather early in the nesting season
(April and May) is often cold and unpredictable, and
diurnal temperatures may vary by 20° C.
Plantations ranged in area from 29 to 59 ha, with
totals of 877 to 1764 planted trees. All three planta-
tions were established in old fields, and the meadow
buffer which immediately surrounded each planta-
tion (6 to 18 m wide) was mowed two to three times
each year. Two plantations were surrounded by
mature hardwood forest (>50 years since the last
major harvest), composed largely of Sugar Maple
(Acer saccharum), White Ash (Fraxinus americana),
and other northern hardwoods (Braun 1950). The
third plantation was surrounded by hayfields, corn-
fields, and apple orchards.
All three plantations were planted in blocks of
White Spruce (Picea glauca) and Balsam/Fraser Fir
(Abies balsamea/fraseri; Robinson and Thor 1969),
in spacings of 1.8 m between trees. Vegetation
between trees was mowed two or three times each
summer, reaching a maximum height of 15-25 cm.
Trees were trimmed annually in July or August to
maintain desired shape and foliage density. Heights
of trees ranged from 1.0 m to 3.0 m (Table 1).
Nest Placement
All trees in each plantation were carefully sur-
veyed in early April 1994 for arboreal open cup
nests built during the 1993 breeding season. This
enabled us to make a number of measurements with-
580
1997
HUGHES AND HUDSON: SONGBIRD NEST PLACEMENT
581
TABLE 1. Summary of potential nest substrates by species, height, and location relative to edge in three
Vermont Christmas tree plantations, 1994.
fir
# %
Location of trees
Interior 1710 68.1
Edge 803 31.9
Total 2513 —
Height of trees (m)
e015 837 Ba 8
ag) 25-2.0 1124 44.7
(HC,) 2.0-2.5 538 21.4
Gig) 2-5=3.0 14 0.6
out compromising nest success (Gotmark 1992). We
are extremely confident that all nests constructed
during the 1993 breeding season were located and
included in our survey. Nests built before 1993 were
very easily recognized and were not included in any
analysis.
Located nests were identified to bird species
(Nickell 1965; Harrison 1975; Harrison 1978) and
nests with problematic identifications were returned
to the lab for more careful inspection and identifica-
tion. We were unable to identify finch nests to
species so we revisited the plantations in May 1995
and 1996 on non-windy, non-rainy mornings to
determine which finch species were nesting in the
plantations. We found only House Finches
(Carpodacus mexicanus) during those surveys and
therefore assumed that all located finch nests were
built by House Finches. During these visits we also
noted which bird species flushed from which trees.
Several characteristics related to nest location
were recorded for each nest. Each nest tree was iden-
tified to species and tree height class (Table 1), and
nest height and position on the tree (relative to trunk
and lateral branches) were also recorded. Trees
growing in perimeter rows of each plantation were
classified as “edge trees” as were trees growing
along the edge of clearings within the plantation
(i.e., surrounded by fewer than six adjacent trees).
All other plantation trees were classified as “interior
trees.” Both tree species, and all four height classes,
were represented equally at the edge.
Nest Shading
To evaluate the extent to which shading by foliage
might contribute to nest site selection, we measured
shading at each nest (April 1994) using a Sunfleck
Ceptometer. This instrument is a hand-held wand
with multiple light sensors that measures irradiance
in wavelengths corresponding to photosynthetically
active radiation. We limited measurements with the
Ceptometer to the last 10 cm of the wand, the approx-
fir & spruce
spruce combined
# % # %
WS 63.7 2467 66.6
432 36.3 1235 33.4
1189 — 3702 =
368 30.9 1205 S255
DD) 21.2 1376 37-2
550 46.1 1088 29.4
21 1.8 35 0.9
imate diameter of most nests that we encountered.
Overhead light was quantified at each nest location
by placing the wand 3cm above the nest cup with the
light sensors oriented skyward. Side-lighting was
quantified in similar fashion after rotating the wand
90° and 270°. Corresponding measurements were
taken in full light so that percent of relative light
transmission could be calculated (e.g., overhead nest
lighting/overhead open lighting xX 100). The light
transmission values were then converted to percent
shading for ease of discussion (100% - light transmis-
sion % = % of light shaded by foliage).
Statistical Analyses
Differences in nest placement (by species of bird,
species of tree, height of tree, height of nest, shad-
ing, location of nest relative to forest edge) were
evaluated using two-way analysis of variance with
interactions (Sokal and Rohlf 1981) and hierarchical
log-linear chi-square analysis (Bishop et al. 1975;
Haberman 1978; Norusis 1993), a statistically more
conservative approach than repeated chi-square anal-
yses. Comparisons of bird species were limited to
the three species with the most nests: House Finch,
American Robin (Turdus migratorius), and Chipping
Sparrow (Spizella passerina).
Results
Despite a careful and thorough search, only 40
nests, representing at least six bird species, were
located in the 3702 trees sampled (1.08% of all trees
sampled had a nest). No fallen nests were found on
the ground and no tree had more than one nest. All
nests of all species were built against tree trunks
rather than on mid-sections of branches. Thirty-four
nests were identified to species, two were identified
only as sparrow nests, and four were not identified.
Overall, nests of House Finch and Chipping
Sparrow were most common (16 and 8 nests, respec-
tively), followed by American Robin and Song
Sparrow (Melospiza melodia) (6 and 2, respectively).
582
Gray Catbird (Dumetella carolinensis) and White-
throated Sparrow (Zonotrichia albicollis) had one
nest each. The plantations differed significantly
(p< 0.05) in density of nests of different bird
species. Nests of Robin were disproportionately
common in the plantation surrounded by agriculture
(0.23% of the trees had a Robin nest compared to
0.10% of the trees in the plantations surrounded by
forest). Nests of House Finch, in contrast, were more
common in plantations surrounded by forest (0.67%
of the trees had a House Finch nest, compared with
0.23% in the plantation surrounded by agriculture).
These apparent trends, as well as those below, were
also evident when we revisited the plantations during
the 1995 and 1996 breeding seasons and noted which
birds flushed from which trees in which plantations.
Notwithstanding the small number of nests found
in plantation trees, a number of highly significant
trends were revealed. A significant (p <0.001) and
disproportionately greater number of nests were
found in spruce than in fir (2.11% of spruce trees
had a nest but only 0.54% of the firs) and this result-
ed in different bird assemblages in spruce and fir
trees (p < 0.05). Excepting White-throated Sparrow
(only one nest sampled), all bird species placed more
nests in spruce than in fir (p < 0.001; Figure 1).
House Finch exhibited the strongest preference for
nesting substrate (nine-fold more nests in spruce on a
per tree basis than in fir) followed by Robin (four
fold more nests in spruce than in fir). Chipping
Sparrow was the least specific in preference for nest-
ing substrate.
% OF TREES AVAILABLE
House Finch
Chipping Sparrow
SPECIES
FIGURE 1. Nest-site selection by songbirds in Christmas
tree plantations of spruce and fir trees in central
Vermont, 1994. “Edge” nest locations were along
the plantation perimeter; “Interior” nest trees were
surrounded by other plantation trees. Column values
= number of trees with nests for a nest location
type/total number of trees sampled for the same nest
location type X 100.
THE CANADIAN FIELD-NATURALIST
Vol Wl
Bird assemblages in interior and edge trees dif-
fered significantly (p < 0.05), with interior trees dis-
proportionately selected (p < 0.05) by all birds for
which nests could be identified (1.32 % of all interi-
or trees had a nest, 0.63 % of all edge trees had a
nest). In relative terms, House Finch exhibited the
weakest preference for interior trees (1.3 X more
nests in interior trees than in edge trees).
The preferred tree height for nesting varied some-
what by species of bird and also by substrate
(p < 0.05), but trees 2.0 — 2.5 m in height (HC3) had
significantly more nests than expected and trees
1.0 — 1.5 m in height had fewer than expected
(p < 0.001; Figure 2a, 2b). Trees in HC3 were pre-
2.5
2.0
a TREE HEIGHT (m)
Lu
cc WM 1.0-1.5
date rare YA 1.5-2.0
= Fa 2.0-2.5
< EF] 2.5-3.0
<
=
<
WL
e)
32
House Finch Robin Chipping Sparrow
SPECIES
25
TREE HEIGHT (m)
2.0
WW MW 1.0-1.5
oO 1.5-2.0
= Ea 2.0-2.5
< 1. 2.5-3.0
= 5
<
op)
LU
aaa
=
LL
©}
3
0.5
0.0
Robin
House Finch
Chipping Sparrow
SPECIES
FIGURE 2. Nest-site selection by tree height in Christmas
tree plantings of fir (a) and spruce (b) in central
Vermont, 1994. Column values = number of nest
trees in a height class/total number of trees sampled
for the same height class X 100. Values over
columns indicate average above-ground height of
nests (+/- standard errors of the mean). Values lack-
ing standard errors had but a single observation.
197]
ferred, sometimes exclusively, by four of the five
identified species which nested in spruces (Figure
2b) and by two of the five species (House Finch and
Robin) which nested in firs (Figure 2a). Chipping
Sparrows exclusively used 1.5 — 2.0 m tall firs (HC,)
and HC, and HC, spruce trees (Figure 2a, 2b). Of the
four bird species that nested in both fir and spruce,
three species (House Finch, Robin, Chipping
Sparrow) exhibited the same height class preference
in both tree species.
Nest height differed by bird species (p < 0.05),
and taller trees had significantly higher nests
(p < 0.01). Height of nests was independent of sub-
strate, however. For example, House Finches that
nested in HC3 trees placed their nests 1.1 m above-
ground, regardless of tree species (Figure 2a, 2b).
Similarly, Chipping Sparrows that nested in HC2
trees placed their nests 0.8m above-ground, irrespec-
tive of tree species. Robins built the highest nests
(average height = 1.2 m).
There was more overhead shading than side shad-
ing of nests, and bird species differed somewhat in
the extent to which their nests were shaded (Figure
3). The average overhead shading of nests ranged
from 92.1% (Chipping Sparrow) to 95.2% (House
Finch) (Figure 3); average shading from the side
ranged from 82.8% (Robin) to 93.1% (House Finch).
The highest and lowest overhead shading for a single
nest was 98.4% (House Finch) and 86.8% (Chipping
Sparrow). The highest and lowest side-shading for a
100
DIRECTION OF SHADING
Hi Overhead
Side
95
90
85
% SHADING BY FOLIAGE
80
House Finch
Robin Chipping Sparrow
SPECIES
Figure 3. Relationship between foliage shading and nest
site selection in Christmas tree plantations of spruce
and fir trees in central Vermont, 1994. Column val-
ues represent average shading of the nest [1 -(trans-
mission to nest/transmission in open) X 100] with
standard errors of the mean. “Overhead” refers to
foliage shading directly above the nest, “Side”
refers to lateral shading.
HUGHES AND HUDSON: SONGBIRD NEST PLACEMENT
583
single nest was 98.3% (House Finch) and 68.8%
(Robin). Overhead shading was significantly higher
in spruces than in firs (p < 0.005), but side-shading
in the two tree nesting substrates was not significant-
ly different. Nests of House Finch were most shaded
overall, but differences among birds were not signifi-
cant (p > 0.05).
Discussion
Location and predation of nests may explain
>50% of the variation in clutch size and annual
fecundity of common North American Passeri-
formes, including the species which nested on our
sites (Martin 1995). Almost certainly nest location,
nest concealment, and nest predation are interrelated
for some species (Martin and Roper 1988). Shading
of nests, our evaluation of nest concealment, varied
somewhat by species of nesting bird (Figure 3), a
pattern that has been noted elsewhere using other
measures of concealment (Holway 1991). Direc-
tionality of nest concealment (overhead vs. from the
side) also varied by species of bird and we suspect
that this reflects differential responses to predator
avoidance.
That all located nests in our plantations were built
next to tree trunks rather than on open branches may
suggest that support of nests or protection from wind
were important factors in nest placement (Collias
and Collias 1984). This seems especially plausible
since many plantation trees had thin branches with
little forking for nest support. We believe that con-
cealment from predators was a more important fac-
tor, however. First, nests varied greatly in mass, size,
shape, and construction (Harrison 1975; Harrison
1978), and equal structural support of nests of the
different species was clearly not needed. Second,
high winds rarely reached interior trees where most
nests were placed because of the sheltering effect of
surrounding trees. Third, almost 30% of all trees sur-
veyed for nests (height classes HC, and HC,) had a
number of stout, forked branches that could have
supported nests much larger than any of those we
found in our study, yet not one nest of any type was
placed on any of these thousands of potential mid-
branch sites.
Density of nests in our plantations was three-fold
higher than that in a Minnesota Christmas tree planta-
tion having Scotch pines (Pinus sylvestris) of similar
height and density (Buech 1982). Nesting birds on
our sites selected certain tree heights and nest heights
preferentially but placement was not related to avail-
ability of sites. Height of nesting substrate has been
shown to affect nesting success (Yahner 1983), per-
haps through its influence on height of nest place-
ment (Kern et al. 1993). In our study, nest height gen-
erally increased with tree height (see also Chapman
1932; Preston and Norris 1947), and most nests were
placed in taller trees. We suspect that this was related
584
to avoidance of ground predators. Nest predation is
the primary source of egg and nestling mortality for
most birds (Martin 1992a, 1992b) and predation can
alter the height at which subsequent nests are placed
(Marzluff 1988). Our nest shading data revealed that
higher nest placement in taller trees did not compro-
mise overhead nest concealment, and may have
improved nest concealment from below.
The degree of preference for substrate and nest
height varied among bird species (Figure 2). Nest
heights of Chipping Sparrows in HC, firs, for exam-
ple, ranged from 0.3 to 1.1 m above-ground. These
variable placements, which tended to be somewhat
lower than nest heights for this species reported else-
where (Chapman 1932; Walkinshaw 1944;
Messersmith 1963), may be explained by some com-
bination of differential natal nest imprinting
(Marzluff 1988), experience (Marzluff 1988), time
of nest building (Walkinshaw 1944), or different
availabilities of tree heights. Nest placement of older
Chipping Sparrows also may be influenced by loss
of previous nests to predation (Reynolds and
Knapton 1988), which may be related to incomplete
concealment of nests (Buech 1982).
Robins on our sites placed their nests lower, and
in shorter substrates, than has generally been report-
ed elsewhere (Howell 1942; Preston 1946; Yahner
1983), where availabilities of nesting substrates were
more varied. This also was the case with House
Finches (Evenden 1957). Robins and House Finches
did exhibit a preference for taller trees in the planta-
tions (Figure 2), but nests were placed in shorter
trees as well. This lack of specificity in nest place-
ment, which has been shown elsewhere to increase
rates of nest predation (Best and Stauffer 1980), may
be beneficial in habitats such as plantations that have
very low structural and compositional diversity.
Consistent and predictable nest placement in homo-
geneous habitats would likely promote increased
nest predation because predators can more easily
develop a search image that reduces search effort
and increases search efficiency.
Many bird species have been shown to exhibit
preferences for nesting substrate (Martin 1993) and
this was the case in our study. Spruce was strongly
preferred over fir as a nesting substrate and this pref-
erence probably was more related to nest conceal-
ment or protection from inclement weather than to
availability of food. Food was equally available to
birds nesting in spruce and fir (both tree species
were similarly dispersed in the plantations), but
shading of nests in the two tree species was very dif-
ferent. Nests in spruces had almost twice as much
overhead shading as those in firs but about equal
shading from the side. Placement of nests in spruce
rather than in fir confers advantages of better protec-
tion from rainfall and snow, and better concealment
from overhead predators such as hawks.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Almost all of the nests on our sites were placed
within the interior of plantations rather than along
the edge, suggesting that birds differentiate between
these habitat locations. Spring weather in Vermont
during nesting is often cold, and this would be
expected to favor nesting on south-facing edges of
trees and plantations where sunlight is more direct.
Such nest placement did not occur on our sites, how-
ever. Disproportionate placement of nests in interior
trees may have been a response to nest parasitism by
Brown-headed Cowbirds (Molothrus ater) (Buech
1982), but cowbirds were sighted (during our May
surveys) only along the edge of the plantation sur-
rounded by agriculture.
Our study of nest placement in Vermont Christmas
tree plantations indicates that several factors strongly
influence nest placement by House Finches,
Chipping Sparrows, and American Robins. Whereas
Chipping Sparrows placed similar numbers of nests
in spruce and fir, House Finches and American
Robins were significantly more inclined to nest in
spruce than in fir. Shading of nests was much higher
in spruce than in fir, suggesting that nest concealment
may be especially important for House Finches and
American Robins. Preferred tree height for nesting
varied by species, with House Finch and American
Robin selecting taller (2.0 — 2.5 m) trees to place
their nests. Chipping Sparrows, in contrast, favored
shorter trees (1.5 — 2.0 m). All nests of all species
were built against tree trunks rather than on mid-sec-
tions of branches, and nests were disproportionately
placed in interior rather than edge trees. We strongly
suspect that preferences for tree species, and location
of nests within the tree and within the plantation,
were related to concealment of nests from nest preda-
tors, and protection from inclement weather.
Acknowledgments
We thank Doug Bechtel and Mike Tetreault for
assistance with data collection and management,
Cathy Borer and Alan Howard for help with statisti-
cal analysis, and Nick Rodenhouse and Dan Ardia
for thoughtful and helpful comments on earlier drafts
of the manuscript. We also thank Dave Capen for
help with nest identifications; Tom Sherry for refer-
ences; and Stephen Bishop, Phillip Kivlin, and
Russell Ray for permitting us to study nesting in
their plantations.
Literature Cited
Best, L. B., and D. F. Stauffer. 1980. Factors affecting
nesting success in riparian bird communities. Condor 82:
149-158.
Bishop, Y. M. M., S. E. Feinberg, and P. W. Holland.
1975. Discrete multivariate analysis: theory and prac-
tice. MIT Press, Cambridge, Massachusetts.
Braun, E. L. 1950. The deciduous forests of eastern
North America. The Blakiston Co., Philadelphia,
Pennsylvania.
1997
Buech, R. R. 1982. Nesting ecology and cowbird para-
sitism of Clay-colored, Chipping, and Field sparrows in
a Christmas tree plantation. Journal of Field Ornithology
53: 363-369.
Chapman, F. M. 1932. Handbook of birds of eastern
North America. D. Appleton and Co., New York, New
York.
Collias, N. E., and E. C. Collias. 1984. Nest building bird
behavior. Princeton University Press, Princeton, New
Jersey.
Evenden, F. G. 1957. Observations on nesting behavior
of the House Finch. Condor 59: 112-117.
Filliater, T. S., R. Breitwisch, and P. M. Nealen. 1994.
Predation on Northern Cardinal nests: does choice of
nest site matter? Condor 96: 761-768.
Gotmark, F. 1992. The effects of investigator disturbance
on nesting birds. Current Ornithology 9: 63-104.
Haberman, S. J. 1978. Analysis of qualitative data.
Volume 1. Academic Press, New York, New York.
Harrison, H. H. 1975. A field guide to Birds’ nests of
285 species found breeding in the United States east of
the Mississippi River. Houghton Mifflin Co., Boston,
Massachusetts.
Harrison, C. J. O. 1978. A field guide to the nests, eggs,
and nestlings of North American Birds. Collins
Publishing, Cleveland, Ohio.
Holway, D. A. 1991. Nest-site selection and the impor-
tance of nest concealment in the Black-throated Blue
Warbler. Condor 93: 575-581.
Howell, J. C. 1942. Notes on the nesting habits of the
American Robin (Turdus migratorius L.). American
Midland Naturalist 28: 529-603.
Johnson, C. W. 1980. The nature of Vermont. University
Press of New England, Hanover, New Hampshire.
Kern, M. D., M. K. Sogge, R. B. Kern, and C. Van Riper
III. 1993. Nests and nest sites of the San Miguel Island
Song Sparrow. Journal of Field Ornithology 64:
367-381.
Martin, T. E. 1992a. Breeding season productivity: what
are the appropriate habitat features for management?
Pages 455-473 in Ecology and Conservation of
Neotropical Migrant Land Birds. Edited by J. M. Hagan
HUGHES AND HUDSON: SONGBIRD NEST PLACEMENT
585
and D. W. Johnston. Smithsonian Institution Press,
Washington, D. C.
Martin, T. E. 1992b. Interaction of nest predation and
food limitation in reproductive strategies. Current
Ornithology 9: 163-197.
Martin, T. E. 1993. Nest predation and nest sites.
Bioscience 43: 523-532.
Martin, T. E. 1995. Avian life history evolution in rela-
tion to nest sites, nest predation, and food. Ecological
Monographs 65: 101-127.
Martin, T. E., and J. R. Roper. 1988. Nest predation and
nest-site selection of a western population of the Hermit
Thrush. Condor 90: 51-57.
Marzluff, J. M. 1988. Do Pinyon Jays alter nest place-
ment based on prior experience? Animal Behavior 36:
1-10.
Messersmith, D. H. 1963. Birds in a red pine plantation.
Wilson Bulletin 75: 235-243.
Nickell, W. P. 1965. Habitats, territory, and nesting of the
catbird. American Midland Naturalist 73: 433-478.
Norusis, M. J. 1993. SPSS for Windows Advanced
Statistics, Release 6.0. SPSS Inc., Chicago, Illinois.
Preston, F. W. 1946. Nesting heights of birds building in
shrubs. Ecology 27: 87-91.
Preston, F. W., and R. T. Norris. 1947. Nesting heights
of breeding birds. Ecology 28: 241-273.
Reynolds, J. D., and R. W. Knapton. 1988. Nest-site
selection and breeding biology of the Chipping Sparrow.
Wilson Bulletin 96: 488-493.
Robinson, J. F., and E. Thor. 1969. Natural variation in
Abies of the southern Appalachians. Forest Science 15:
238-245.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H.
Freeman and Co., New York, New York.
Walkinshaw, L. H. 1944. The Eastern Chipping Sparrow
in Michigan. Wilson Bulletin 56: 193-205.
Yahner, R. H. 1983. Site-related nesting success of
Mourning Doves and American Robins in shelterbelts.
Wilson Bulletin 95: 573-580.
Received 29 October 1996
Accepted 28 January 1997
Reproductive Success of the Common Loon, Gavia immer, on a
Small Oligotrophic Lake in Eastern Canada
DOUGLAS CLAY! and HEATHER CLAY2
'Fundy National Park, Department of Canadian Heritage, P.O. Box 40, Alma, New Brunswick EOA 1B0
2New Brunswick Agriculture, P.O. Box 6000, Fredericton, New Brunswick E3B 5H1
Clay, Douglas, and Heather Clay. 1997. Reproductive success of the Common Loon, Gavia immer, on a small oligo-
trophic lake in eastern Canada. Canadian Field-Naturalist 111(4): 586-590.
Wolfe Lake, Fundy National Park, is marginal Common Loon (Gavia immer) habitat owing to its small size (22 ha) and its
low primary productivity. A pair of Common Loons has nested on Wolfe Lake every year since 1989. The production from
this pair has been 1.4 fledged young per year. The growth of the chicks is near linear for the first six weeks.
Key Words: Common Loon, Gavia immer, reproduction, Wolfe Lake, Bay of Fundy, New Brunswick.
Fundy National Park, New Brunswick was estab-
lished in 1948 to protect 207 km? of the Maritime
Acadian Forest. Despite this protection, Common
Loons (Gavia immer) did not utilize the freshwater
lakes of the park for breeding until 1989 when the
first confirmed nesting took place on Wolfe Lake.
This small (22 ha) oligotrophic lake is located 14.5
km (65° 08’ W, 45° 39’ N) from the Bay of Fundy at
an elevation of 305 m.
There has been no active management of the
aquatic resources of the lake since stocking of Brook
Trout (Salvelinus fontinalis) ceased in 1980. A grad-
ual reduction in recreational use of the lake fol-
lowed. By 1989 the decline in camping activity was
nearly 60% (Clay and Clay 1994*),.
After the loons became resident, a monitoring pro-
gram was implemented to develop our understanding
of their behavior and general ecology. The objective
was to identify factors that influenced the reproduc-
tive success of the newly resident loons and to com-
pare their reproductive success with that from other
North American studies.
Study Area and Methods
Three lakes in Fundy National Park meet the min-
imum size requirements of Common Loons, about
10 ha (or 20 ac) (Barr 1973; Masse 1991*) but only
Wolfe Lake was used by a pair of Common Loons as
a nesting area. Loons were reported occasionally on
Bennett Lake and Tracey Lake, but none became
resident. Wolfe Lake, created by an earth dam on the
upper reach of the East Branch of the Point Wolfe
River, is at the lower size limit of preferred nesting
areas but it does have two small islands with a com-
bined area of <0.1 ha. The lake has little surface
inflow, is shallow with wide shoals and an average
depth of 3.8 m and maximum depth of 8.5m
(Kerekes et al. 1975*).
Water quality data were collected during 1993 and
1994 to ascertain the productivity of the lake.
*See Documents Cited section after Acknowlegments.
586
Analyses were conducted by the Environment
Canada Water Quality Laboratory, Moncton, New
Brunswick.
Bird sighting data prior to the study period (1989-
1996) were gleaned from the Fundy National Park
Bird Sighting Cards (BSC) which were maintained
from 1948 to 1980 with some intermittent reports
after 1980. Observations recorded during this moni-
toring program were opportunistic except during
those periods considered significant. These key peri-
ods included the arrival and departure, nest estab-
lishment, egg laying and chick hatching.
Nomenclature of the development follows that used
by the Canadian Lakes Loon Survey, Long Point
Observatory.
Results
Common Loons were first reported on Wolfe
Lake in 1965 when three were reported on the BSC.
Intermittent sightings of one to two Common Loons
were made through the 1970s but the first indication
of possible breeding was an unverified report of two
loon chicks in 1983. There were no additional
recorded sightings until 1989, after which there have
been annual sightings.
Low productivity was shown by water clarity
readings with a secchi disk that were greater than the
maximum depth. The pH of the lake averaged 6.4,
nitrogen 0.2 ppm, phosphorus 0.005 ppm, potassium
0.3 ppm and dissolved organic carbon 1.2 ppm.
Calculations of fish yield from empirical models
gave a range of 0.25 to 1.55 kg/ha fish production
for Wolfe Lake.
The loons arrived as the ice broke up, often with
50% of the lake still covered in ice (Table 1). Within
the first week several other loons arrived and
remained for a couple of days. Other piscivorous
birds also arrived at this time and were chased by the
loons (Clay and Clay 1994*).
Courtship display as defined by Crowley and Link
(1987), was observed at different times and the
1997 CLAY AND CLAY: REPRODUCTIVE SUCCESS OF THE COMMON LOON 587
activity took place over about a one week period, the
« commencement of which varied from year to year. A
Seate 5.5 a late cold spring as in 1992 and 1996 delayed the
© So eye oo 6 a) courtship and nesting by about two weeks. Eggs
a 3 a a Pe ae E = & = g were usually laid before the middle of June except
oN aS of for 1992 and 1996 when eggs were laid in the last
week of June.
Clutch survival varied over the eight years.
a Hatching dates for surviving chicks ranged between
wd = 2 July and 1 August with mean date of 12 July. All
ba | 5 = 2-0 Z clutches hatching after the middle of July had some
KR § Beeaels Discussion
3 lla ia ay = dE DSS] Ss ;
eo Ql Ss nea-coS wwe ssi s Wolfe lake offers only marginal habitat for
Ban ieee a 2 ae o5 & 5 || 2 Common Loons, being small and unproductive (olig-
= 2 = 7 & 2 otrophic), with limited seasonal surface inflow. The
‘ ae E g water is low in nutrients and dissolved solids
a Ey E RY Ls = (Kerekes and Schwinghamer 1978*; Clay and Clay
a = “2 & 1994*). Barr (1996) estimated that 423 kg of food is
= & required to support a pair and two chicks for 5.5
Puli ca = months. He calculated that a chick required 53 kg of
eS R 5 oS 8 fish to fledge. Adult loons eat approximately 10% of
5 = ie their body weight (Barr 1996) in fish and aquatic
3 2 organisms per day. The average weight of the adult
5 wee ote 2 loons on Wolfe Lake was 5.1 kg (5.250 kg male and
5 = S 4.925 kg female), thus one adult Common Loon
5 so) would require 510 g of food per day or 76.50 kg for
aa) * E a season May through September. Wolfe Lake
6 Rv 0 a a Common Loons would require 12 kg/ha of fish to
= a Cee 5 dults and two chicks.
& = S328 < support two adu
9 = 2) WE RER Ss Ss Total fish yield! of Wolfe Lake was estimated to
2 S a 4 = z 2 a & be 34.1 kg (range 32.02 to 35.25 kg) or 1.55 kg/ha,
3 iS ae cS é = = E § by Clay and Clay (1994*) using the Ryder et al.
22neaRS65=5 =
al_| <9 8@2R R85 s EO e ihren hci alot hn dab
a) § 2ES2sEEL2s é 'In this model, fish yield refers to only what fish could be
ela alaqugzzade x harvested from a water body.
588
TABLE 2. Common Loon chick survival estimates from
various North American studies compared to those of
Wolfe Lake, Fundy National Park.
Authority Chick Number of Young
Survival Fledged Fledged
from Eggs Young per
Nesting Pair
This study 65 % 1.4 92 %
Benjamin and 0.5
Kerekes 1993
Vogel 1993 0.7
(Nova Scotia)
Vogel 1993 0.6
(New Brunswick)
Robinson et al. 1988 112
Titus and van Druff 1981 0.5
Trivelpiece 1979 0.8
Meyer 1994 0.9 78 %
2This value was calculated from Robinson et al. (1988) by
assuming 75% of the nests were successful and then
estimating chick production from all nests (see text).
(1974) morphoedaphic index. In comparison,
Bennett Lake, with a significant surface inflow, had
a total estimated fish yield of 72.2 kg (range 71.41 to
72.88 kg) or 2.33 kg/ha. Using three other models,
which incorporated total phosphorus (Hanson and
Leggett 1982), total fish yields for a three year mean
ranged from 0.25 to 1.15 kg/ha for Wolfe Lake.
Whatever the absolute values of fish yield chosen,
the estimated range of 0.25 to 1.55 kg/ha fish pro-
duction for Wolfe Lake is at the low end of the range
of 19 oligotrophic lakes of Kelso and Johnson (1991)
and it was only 10% of the loons’ food requirements.
Loons are known to feed on a wide range of aquatic
organisms (Alvo 1986; Parker 1988) and more food
would be available than these simplified estimates of
fish production indicate. The loons must supplement
their nutrient requirements with amphibian and other
foods and they probably fished in nearby lakes.
Single adult flights to and from the lake were noted,
especially departures at dawn and returning after
dusk (Clay and Clay 1994*).
Chick survival was 65% on average for the past
eight years with 11 young surviving from a potential
of 17 eggs. Under normal circumstances chick mor-
tality is rare after three weeks of age (Parker 1988).
Of 12 loon chicks that hatched over the last eight
years, only one died in the first few days after hatch-
ing. Reproductive success which is defined as the
number of fledged young per breeding pair per year
was 1.4, with 92% of the young fledged (Table 2).
Over a five-year period at Grafton Lake, Kejimkujik
National Park, Nova Scotia the average reproductive
success was ().5 (Benjamin and Kerekes 1993*). The
Canadian Lakes Loon Survey reported the average
for New Brunswick and Nova Scotia as 0.64 and
0.74 respectively (Vogel 1993*).
THE CANADIAN FIELD-NATURALIST
Vol. 111
By comparison loon reproductive success in the
USA showed similar variability. In Michigan 70-80%
of loon nests produced at least one chick (Robinson
et al. 1988) with an average brood size of 1.5 chicks
per successful nest (Table 2). Titus and van Druff
(1981) considered it normal for the average Common
Loon pair to produce a single offspring every other
year (0.5 fledged/pair/year). Common Loons in New
York have been found to fledge 0.8 young / pair/ year
(Trivelpiece et al. 1979). In Wisconsin reproductive
success was 0.94 young/nesting pair with 78% of
chicks fledged (Meyer 1994).
The reasons for nest failure or chick loss varied. In
1990 it is believed that only one egg was laid, as a
limited number of shell fragments, sufficient for one
egg, were found in the nest. In 1992 no chicks sur-
vived, probably owing to parental abandonment
caused by human interference both from fishers on
the shore and canoe traffic. Human disturbance can
vary from interference by individuals (intentional or
unintentional) to heavy metal poisoning or entangle-
ment in fishing gear (Clay and Clay 1996*) One
chick was found dead near the nest, presumably
abandoned and the other egg with its small end bro-
ken still had the membrane inside suggesting preda-
tion. In 1993, visitor boating and fishing activity
near the larger island coincided with loon nesting
activity. The loons abandoned their first choice of
nest site and moved to the smaller island. Both eggs
from this first nest on the small island were washed
away. The loons returned to the larger island and
renested. One of the two eggs disappeared.
Nesting loons are more severely affected by slow-
moving boats, especially canoes and fishing boats
that hug the shoreline, than they are by motor boats
(McIntyre 1975). Although loons have been found
to habituate to human disturbance they still leave
their nests, albeit for shorter periods, leaving the
young exposed to natural predators (Christenson
1981). Fishing activity and canoe traffic on Wolfe
Lake from the opening of the fishing season on
Victoria Day in May until the closure of the season
15 September affected the movement, nesting
behavior and the fishing activity of the loons.
During intrusion near their nest sites, incubating
loons slipped off their nest, swam underwater and
emerged at a distance. This left the nest open to pre-
dation, especially by ravens, similar to the observa-
tions of McIntyre (1986).
Unexpected water level fluctuations have contribut-
ed to nest failure. The water level of Wolfe Lake is
relatively stable, however, in recent years increasingly
large fluctuations occurred in spring and summer for
several reasons. Aquifers and springs have dried up in
periods of drought. This reduced inflow coupled with
a “leaky” dam has caused the water level to drop as
much as 45 cm. Heavy rainfall has on occasions
caused a rapid fill of the lake. In 1993, the wooden
1997
Mass (kg)
mean weight of
eae day old chick
gg
0 10 20
CLAY AND CLAY: REPRODUCTIVE SUCCESS OF THE COMMON LOON
589
eo
1995 chick
1996 chick
30 40 50 60
Age (days)
FIGURE | Growth of juvenile Common Loons (Gavia immer), Wolfe Lake, New Brunswick estimated
from banding of approximately 4- and 6-week-old chicks and from the mean weight of day old chicks
from Barr (1996).
gate structures of the dam developed a small breach
that caused a sudden water level drop of about 20 cm
during the loon nesting season. The resident beaver
repaired the breach in the dam after nesting had taken
place. Water levels rose quickly and, together with a
strong wind, washed the eggs from the nest.
In the past it has been unknown whether the same
pair of loons returned to Wolfe Lake each year. A
study in Isle Royale National Park, Minnesota (Evers
1995) found that the monogamous relationships pre-
viously assumed for loons may be questionable. He
observed that one-fifth of banded Common Loons
switched partners or sneaked onto other territories
(Fair 1994). The loons at Wolfe Lake were captured
and tagged in 1995 and recapturing in 1996 con-
firmed that the same pair returned to the lake.
From the known weights-at-age and assuming a
starting chick weight of 90 to 100 g (Barr 1996), the
growth of chicks on Wolfe Lake indicated a near lin-
ear relationship between weeks 1 and 6. Barr (1996)
found loon chicks became relatively independent and
attained near adult size and shape by 11 weeks.
Weights of our tagged chicks showed that juvenile
growth was rapid (Figure 1). This agrees with that
recorded by Barr for hand-fed loons, but appeared
advanced by about two weeks.
The loons of Wolfe Lake are new arrivals to the
area. If the fish population remains stable there
appears to be sufficient fish and other aquatic organ-
isms in Wolfe lake and other nearby lakes to provide
two adult loons and two young with their nutrient
requirements. The biggest problem is human inter-
ference, with more educational programs and better
communication between the guardians of the park
and the public there is hope that these loons will
continue to return to Wolfe Lake.
Acknowledgments
Many volunteers, visitors, and summer employ-
ees have contributed their observations to this moni-
toring program, we thank them for their interest and
efforts. Sedgewick Sinclair, Park Warden, Fundy
National Park, was responsible for coordinating the
first years of the monitoring. Two anonymous
reviewers provided helpful comments on the initial
manuscript.
Documents Cited (marked * in text)
Benjamin, N., and J. Kerekes. 1993. Grouping dynamics
of Common Loons on Grafton Lake (Cecuncia Gowick)
Kejimkujik National Park (27 July-10 October 1993.
Unpublished report Canadian Wildlife Service, Dart-
mouth, Nova Scotia, Canada. 23 pages.
Clay, H., and D. Clay. 1994. Common Loons (Gavia
immer) on Wolfe Lake, Fundy National Park, Canada.
Unpublished manuscript of Parks Canada, Alma, New
Brunswick, Research Notes of Fundy National Park
Number 94—07. 25 pages.
Clay, D., and H. Clay. 1996. Update of the Status of the
Common Loon (Gavia immer) on Wolfe Lake, a small
oligotrophic lake. Unpublished manuscript of Parks
Canada, Alma, New Brunswick, Research Notes of
Fundy National Park Number 96-07. 26 pages.
590
Kerekes, J., and P. Schwinghamer. 1978. Aquatic
resources inventory, Fundy National Park, New
Brunswick Part 2. Limnological conditions. Canadian
Wildlife Service Atlantic Region, Dartmouth, Nova
Scotia. 143 pages.
Kerekes, J., P. Schwinghamer, and P. Lucas. 1975.
Aquatic resources inventory, Fundy National Park, New
Brunswick Part 1. Lake morphometry and limnological
data. Canadian Wildlife Service, Atlantic Region,
Dartmouth, Nova Scotia, 143 pages.
Masse, D. 1991. Four years of monitoring of Common
Loons (Gavia immer) in La Mauricie National Park,
Quebec. Unpublished manuscript of La Mauricie
National Park, Quebec. 9 pages.
Vogel, H. 1993. Results from 1992 survey. Canadian
Lakes Loon Survey Newsletter 11: 13-15.
Literature Cited
Alvo, R. 1986. Lost loons of northern lakes. Natural
History 95: 58-65.
Barr, J. 1973. Feeding biology of the Common Loon
(Gavia immer) in oligotrophic lakes of the Canadian
shield. Ph.D. thesis, University of Guelph, Guelph,
Ontario.
Barr, J. F. 1996. Aspects of Common Loon (Gavia
immer) feeding biology on its breeding ground.
Hydrobiologia 321: 119-144.
Christenson, B. L. 1981. Reproductive ecology of and
response to disturbance by Common Loons in Maine.
M.Sc. thesis. University of Maine, Orono, Maine.
Crowley, K., and M. Link. 1987. Love of loons. Key
Porter Books Ltd., Toronto, Ontario. 96 pages.
Evers, D.C. 1995. Isle Royale Loons. Park Science 15:
20-21.
Fair, J. 1994. Last call. Equinox 77: 52-59.
Hanson, J. M., and W. C. Leggett. 1982. Empirical pre-
diction of fish biomass and yield. Canadian Journal of
Fisheries and Aquatic Science 39: 257-263.
Kelso, J. R. M., and M. G. Johnson. 1991. Factors relat-
ed to the biomass and production of fish communities in
small oligotrophic lakes vulnerable to acidification.
Canadian Journal Fisheries and Aquatic Science 48:
2523-2532.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Kerekes, J. 1990. Possible correlation of summer
Common Loon (Gavia immer) population with the
trophic state of a water body. International Association
of Theoretical and Applied Limnology 24: 349-353.
McIntyre, J. W. 1975. Biology and behaviour of the Com-
mon Loon (Gavia immer) with reference to its adaptability
in a man altered environment. Ph.D. thesis, University of
Minnesota, Minneapolis, Minnesota. 243 pages.
McIntyre, J. W. 1986. Common Loons. Pages 678-695
in Audubon Wildlife Report. Edited by R. L. Silvestro,
National Audubon Society, New York.
Meyer, M. W. 1994. An investigation into the impact of
fish mercury contamination on Common Loon produc-
tivity in Wisconsin. The Loon Call (Spring 1994). North
American Loon Fund, Gilford, New Hampshire. Pages 1
and 7.
Parker, K. 1988. Common Loon reproduction and chick
feeding on acidified lakes in the Adirondack Park New
York. Canadian Journal Zoology 66: 804-810.
Robinson, W.L., J. H. Hammill, H. R. Hill, and T. A.
deBruyn. 1988. The status of the Common Loon in
Michigan. Pages 132-144 in Papers from the 1987
Conference on Common Loon research and manage-
ment. Edited by P. I. V. Strong. North American Loon
Fund, Meredith, New Hampshire.
Ryder, R. A., S. R. Kerr, K. H. Loftus, and H. A. Regier.
1974. The morphoedaphic index, a fish yield estimator -
review and evaluation. Journal Fisheries Research Board
of Canada 31: 663-688.
Titus, J., and L. van Druff. 1981. Response of the
Common Loon to recreational pressure in the Boundary
Waters Canoe Area, northeastern Minnesota. Wildlife
Monograph Number 79, Wildlife Society. 59 pages.
Trivelpiece, W., S. Brown, A. Hicks, R. Fekete, and W.
Volkman. 1979. An analysis of the distribution and
reproductive success of the Common Loon in Adirondack
Park, N.Y. Pages 45-55 in The Common Loon. Proceed-
ings of the North American Conference on Common Loons
Research and Management. Edited by S. A. Sutcliffe.
National Audubon Society, New York, New York.
Received 20 November 1996
Accepted 22 January 1997
The Effect of Egg Coloration on Predation of Artificial Ground Nests
BENOIT JOBIN! and JAROSLAV PICMAN
Department of Biology, University of Ottawa, 30, Marie Curie, Ottawa, Ontario K1N 6N5
'Present address: 39 Champagnat, Lévis, Québec G6V 2B1 (corresponding author).
Jobin, Benoit, and Jaroslav Picman. 1997. The effect of egg coloration on predation of artificial ground nests. Canadian
Field-Naturalist 111(4): 591-594.
Effects of egg coloration on predation of artificial ground nests in upland habitat adjacent to a cattail marsh were explored
by deploying nests containing either a brown or a white chicken egg. Predation rates were similar on the two nest types
after 13 days of exposure. Nest concealment (vegetation height, overhead concealment, vegetation density) did not differ
between surviving and destroyed nests. Our results suggest that coloration of eggs did not affect the outcome of our preda-
tion experiment using artificial ground nests.
Key Words: artificial nest, egg coloration, nest concealment, predation.
Predation is a major cause of nesting failure in
many bird species (Ricklefs 1969) and generally
accounts for high nest losses in waterfowl (Johnson
et al. 1989). To examine the role of predation in
waterfowl nest success, many authors used artificial
nests. Coloration of eggs used in studies of artificial
ground nests has varied considerably. Either white
(Esler and Grand 1993), brown (Jobin 1991), dyed
(Sugden 1987) or painted (Picozzi 1975) domestic
fowl eggs have been used, as well as pheasant eggs
(Jones and Hungerford 1972).
Most birds are visual predators that hunt in day-
light, whereas mammals hunt generally at night and
rely mostly on scent to find their prey (Storaas
1988). Experiments with eggs of different coloration
that were presented simultaneously to corvids did
not demonstrate a significant advantage of cryptic
coloration (Montevecchi 1976; Salonen and
Penttinen 1988). However, Verbeek (1990) showed
that a Northwestern Crow (Corvus caurinus) egg
with a different coloration than the others in the
clutch was subjected to more intense avian predation
pressure. Examination of avian predation on Black-
headed Gull (Larus ridibundus) nests showed that
egg coloration could play a significant role in nest
survival (Tinbergen et al. 1962). Nest and egg con-
cealment should, therefore, be taken into considera-
tion when predation studies are conducted in areas
where avian predators are numerous (Clark and
Nudds 1991). The objective of our study was to
compare predation rates on artificial ground nests
containing either a brown or a white chicken egg.
We also evaluated the effect of concealment on nest
predation.
Study Area
We conducted this study in upland habitats adja-
cent to the Des Laiches marsh, a cattail (Typha lati-
folia) marsh adjoining the Ottawa river between
Gatineau and Angers, Quebec (45° 30’ N, 75° 30’ W).
52)
Upland habitats consisted of scrubland dominated by
willows (Salix petiolaris), Meadow-sweet (Spiraea
alba), Reed-canary Grass (Phalaris arundinacea)
and sedges (Carex spp.), and hayfields dominated by
Phleum pratense, Lotus corniculatus and Vicia crac-
ca. Several hedgerows consisting of Wild Cherry
(Prunus spp.), hawthorn (Crataegus spp.), ash
(Fraxinus americana) and elm (Ulmus rubra) trees
were present between fields. Waterfowl nesting in
the area were Blue-winged Teal (Anas discors),
Mallard (Anas platyrhynchos), and Wood Duck (Aix
sponsa), but no systematic search for nests of these
species was conducted. The main predators in this
area were American Crow (Corvus brachyrhynchos),
Raccoon (Procyon lotor), Striped Skunk (Mephitis
mephitis), Mink (Mustela vison), and Red Fox
(Vulpes vulpes) (Jobin 1991).
Methods
Artificial ground nests were constructed by press-
ing dead vegetation together to form a shallow cup
(external diameter: 30 cm, internal diameter: 12 cm,
depth: 8 cm). One large, fresh chicken egg, either
brown or white, was placed in each nest. We estab-
lished eight nests, 20 m apart, along each of six tran-
sects beginning at the edge of marsh vegetation and
running perpendicular to the marsh in upland habi-
tats. Transects were > 50 m apart and nests were
either located in scrubland or in hedgerows. Nests
with a brown egg and nests with a white egg were
alternated along each transect and the color of the
egg in the first nest of each transect was also alter-
nated between adjacent transects. Special care was
taken not to disturb vegetation around the nests. We
established 48 nests, 24 containing a brown egg and
24 containing a white egg. Actual manipulation of
the nests occurred only when setting them up. No
nest markers were used; natural landmarks allowed
us to find nests in subsequent visits.
592
Four nest-concealment variables were measured
when the nests were established, always by the same
observer. Height of live and dead vegetation above
each nest was measured to the nearest 5 cm.
Vegetation density around nests (within 5 m) was
categorized as light, medium or heavy (see Bowman
and Harris 1980), relative index 1, 2 or 3, respective-
ly. Overhead concealment was defined as percentage
of the nest visible from 1 m above the nest (see
Dwernychuk and Boag 1972). This variable was
rated on a scale of 0 to 4: 0 = completely visible;
1 = 1-10% of the nest concealed; 2 = 11-50% con-
cealed; 3 = 51-90% concealed; 4 = 91-100% con-
cealed. Nests were deployed 7 June 1990, inspected
the following day and then visited four times at
three-day intervals. We inferred that predation had
occurred when an egg was found destroyed or
moved away but still near its nest. Nest-concealment
variables were compared between nests with brown
eggs versus nests with white eggs using Chi-square
test and the non-parametric Mann-Whitney test.
Because of small sample sizes, Fisher’s exact test
was used to compare predation patterns between the
two types of nest. Logistic regression was also per-
formed to identify factors predicting nest success.
Variables measured at each nest and included in the
analyses were egg color, height of live and dead veg-
etation, vegetation density, overhead concealment,
and distance from marsh edge. Statistical analyses
were performed on JMP 3.0 and Fish 1.001 (categor-
ical statistics) statistical softwares for Macintosh
computers. Statistical significance was set at 0.05.
Results
Height of live and dead vegetation, vegetation
density and overhead concealment at the onset of the
experiment did not differ between nests with brown
eggs and nests with white eggs (live vegetation:
U = 329.0, P = 0.3973; dead vegetation: U = 309.0,
P = 0.5564; overhead concealment: x? = 3.78, df = 3,
P = 0.2865; vegetation density: x” = 0.93, df = 2,
P = 0.6286). Predation rates were similar between
nests with brown eggs and nests with white eggs
THE CANADIAN FIELD-NATURALIST
Vol. 111
throughout the study (Table 1). Two nests with
brown eggs and one nest with a white egg were
depredated during the first 24 hours of exposure.
Predation rates were generally high during the next
three days of exposure when 58% of all eggs were
depredated. By the end of the experiment (day 13),
the two nest types had suffered similar losses
(87.5%).
Logistic regression analysis revealed that neither
egg color, nor any concealment variable were good
predictor of nest success (P > 0.60 for all variables;
nests with brown eggs and nests with white eggs
combined; data from successful and destroyed nests
after four days of exposure). The only variable that
entered the model was distance of the nest from the
marsh edge (P = 0.02). Nests located farther from
the marsh were destroyed more rapidly than those
deployed near the marsh. Note that this relative secu-
rity would, however, be temporary because most
nests were nevertheless destroyed after 13 days of
exposure.
Discussion
We expected that nests containing white eggs
would suffer higher predation rates than nests with
brown eggs because white eggs are more visible
from a distance. However, our results demonstrated
that this was not the case. Salonen and Penttinen
(1988) exposed artificial nests containing white and
dyed brown hen eggs to predators, mainly birds
(corvids), of Great Crested Grebe (Podiceps crista-
tus) nests and found no difference in predation rates
on the two types of nest. Janzen (1978) found no evi-
dence for differential predation rates between ground
nests containing white, brown, and dyed blue chick-
en eggs in a deciduous forest in Costa Rica. He con-
cluded that camouflage based on egg color would be
irrelevant if predation occurred mainly at night.
Although egg coloration should theoretically play a
significant role in reducing chances of nest preda-
tion, those two studies and our results failed to
demonstrate differential predation rates between
nests containing conspicuous versus cryptic eggs.
TABLE 1. Number of depredated nests, initially containing one brown or one white chicken egg, at three day
intervals near Des Laiches marsh, Quebec, June 1990.
Brown (n = 24)
Days of exposure n
1 Z
= 16
7 18
10 20
13 21
Number of depredated nests
White (n = 24)
% n %
8.3 1 4.2
66.7 15 62.5
75.0 17 70.8
83.3 20 83.3
87.5 21 87.5
Note: Results from Fisher’s exact test reveal no significant difference (P = 1.00) between predation rates on
nests with brown eggs and nests with white eggs for all five comparisons.
1997
If mammals dominated the predator community in
our study area, one would not expect nests contain-
ing brown and white eggs to be destroyed at differ-
ent rates. During our study, several nests were
destroyed by mammals, mainly Raccoon and Striped
Skunk (B. Jobin, unpublished data). Crows were also
seen regularly in the study area and were responsible
for several nest predation events. Crows would
refrain from landing in areas with dense vegetation
and poor visibility (Picman 1988; Sullivan and
Dinsmore 1990). Sugden and Beyersbergen (1987)
showed that survival of artificial nests was maxi-
mum at a vegetation height of 70 cm. In our study,
mean vegetation height was greater than 100 cm,
which could have reduced nest visibility to flying
crows. In addition, although Sullivan and Dinsmore
(1990) concluded that increases in cover height
above 50 cm did not substantially reduce predation,
74% of nests under such cover height were not
destroyed. Sugden (1987) showed that vegetative
cover which provided less than 30% concealment
was sufficient to reduce depredation from crows on
his artificial nests. He concluded that “...when the
outline of the eggs was broken by plant cover, crows
experienced more difficulty in detecting them”. This
suggests that eggs clearly visible from above are
more readily found and that even a minimum level
of overhead concealment would reduce chances of
being detected. In our experiment, both brown and
white eggs disappeared rapidly during the first four
days of exposure but predation rates recorded after
day 7 were lower and were mostly attributed to
mammalian predation (B. Jobin, unpublished data).
Eggs placed in the nests destroyed in the first stages
of the experiment must have presented fairly unbro-
ken outlines, hence these eggs disappeared rapidly.
The remaining eggs were, therefore, relatively pro-
tected against predators and most particularly against
avian predators.
Predation experiments using artificial nests have
been conducted for many years, and comparison of
results between these experiments is a common
practice. The possible bias brought about by the use
of visually different eggs has however seldom been
addressed. Our data did not demonstrate differential
predation rates between nests containing brown
eggs and nests with white eggs. Note that artificial
nests were deployed at about 16 nests/ha which
largely overestimated actual breeding bird density in
the study area. Sugden and Beyersbergen (1986)
showed that predation on artificial ground nests
increased markedly when nest density exceeded 2
nests/ha. Density-dependant effect in nest predation
experiments was also demonstrated in several other
studies whereas others failed to demonstrate such
effect (see Clark and Nudds 1991). It is possible that
any advantage of egg coloration or nest conceal-
ment could have been overwhelmed by such high
JOBIN AND PICMAN: EFFECT OF EGG COLORATION ON PREDATION
593
nest density, especially in the first four days of
exposure when two-thirds of all nests were
destroyed. However, low predation rates recorded
after day 4 suggests that any density-dependant
effect, if present, was rapidly invalidated.
Nevertheless, similar nest losses occurring on the
two types of nests before and after day 4 suggest
that, in our study area where avian and mammalian
predators were present, egg coloration did not affect
the outcome of our predation experiment using arti-
ficial ground nests.
Acknowledgments
This work was supported by NSERC operating
grant to J. Picman and by NSERC post-graduate
scholarship, FCAR-NSERC Joint program post-
graduate scholarship and University of Ottawa
scholarship to B. Jobin. We thank the land owners
who allowed us to work on their properties and the
Ministére du Loisir de la Chasse et de la Péche du
Québec for valuable information about the study
area. We thank Y. Leblanc and J. Maltais for con-
structive comments on the manuscript.
Literature Cited
Bowman, G. B., and L. D. Harris. 1980. Effect of spatial
heterogeneity on ground-nest depredation. Journal of
Wildlife Management 44: 806-813.
Clark, R. G., and T. D. Nudds. 1991. Habitat patch size
and duck nesting success: the crucial experiments have
not been performed. Wildlife Society Bulletin 19:
534-543.
Dwernychuk, L. W., and D. A. Boag. 1972. How vegeta-
tive cover protects duck nests from egg-eating birds.
Journal of Wildlife Management 36: 955-958.
Ksler, D., and J. B. Grand. 1993. Factors affecting depre-
dation of artificial duck nests. Journal of Wildlife Man-
agement 57: 244-248.
Janzen, D. H. 1978. Predation intensity on eggs on the
ground in two Costa Rican forests. American Midland
Naturalist 100: 467-470.
Jobin, B. 1991. The impact of human disturbance on nest
predation patterns in freshwater marshes. M. Sc. thesis.,
University of Ottawa, Ottawa, Ontario, Canada. 92
pages.
Johnson, D. H., A. B. Sargeant, and R. J. Greenwood.
1989. Importance of individual species of predators on
nesting success of ducks in the Canadian Prairie Pothole
Region. Canadian Journal of Zoology 67: 291-297.
Jones, R. E., and K. E. Hungerford. 1972. Evaluation of
nesting cover as protection from Magpie predation.
Journal of Wildlife Management 36: 727-732.
Montevecchi, W. A. 1976. Field experiment on the adap-
tive significance of avian eggshell pigmentation.
Behaviour 58: 26-39.
Picman, J. 1988. Experimental study of predation on eggs
of ground—nesting birds: effects of habitat and nest dis-
tribution. Condor 90: 124-131.
Picozzi, N. 1975. Crow predation on marked nests.
Journal of Wildlife Management 39: 151-155.
Ricklefs, R. E. 1969. An analysis of nesting mortality in
birds. Smithsonian Contribution to Zoology 9: 1-48.
594
Salonen, V., and A. Penttinen. 1988. Factors affecting
nest predation in the Great Crested Grebe: field observa-
tions, experiments and their statistical analysis. Ornis
Fennica 65: 13-20.
Storaas, T. 1988. A comparison of losses in artificial and
naturally occurring capercaillie nests. Journal of Wild-
life Management 52: 123-126.
Sugden, L. G. 1987. Effect of disruptive background on
predation of artificial nests by American Crows. Prairie
Naturalist 19: 149-152.
Sugden, L. G., and G. W. Beyersbergen. 1986. Effect of
density and concealment on American Crow predation
of simulated duck nests. Journal of Wildlife Manage-
ment 50: 9-14.
Sugden, L. G., and G. W. Beyersbergen. 1987. Effect of
nesting cover density on American Crow predation of
THE CANADIAN FIELD-NATURALIST
Volo ttl
simulated duck nests. Journal of Wildlife Management
51: 481-485.
Sullivan, B. D., and J. J. Dinsmore. 1990. Factors affect-
ing egg predation by American Crows. Journal of
Wildlife Management 54: 433-437.
Tinbergen, N., G. J. Broekhuysen, F. Feekes, J. C. W.
Houghton, H. Kruuk, and E. Szule. 1962. Egg shell
removal by the Black-headed Gull, Larus ridibundus L.;
a behaviour component of camouflage. Behaviour 19:
74-117.
Verbeek, N. A. M. 1990. Differential predation on eggs
in clutches of Northwestern Crows: the importance of
egg color. Condor 92: 695-701.
Received 11 October 1996
Accepted 13 January 1997
Effects of Hunting and Loss of Escape Cover on Movements and
Activity of Female White-tailed Deer, Odocoileus virginianus
DAVID E. NAUGLE!, JONATHAN A. JENKS!, BRIAN J. KERNOHAN’, AND REX R. JOHNSON!
‘Department of Wildlife and Fisheries Sciences, South Dakota State University, Brookings, South Dakota 57007
Present address: Boise Cascade Corporation, 400 Third Avenue East, International Falls, Minnesota 56649
Naugle, David E., Jonathan A. Jenks, Brian J. Kernohan, and Rex R. Johnson. 1997. Effects of hunting and loss of escape
cover on movements and activity of female White-tailed Deer, Odocoileus virginianus. Canadian Field-Naturalist
111(4): 595-600.
We monitored radio-collared female White-tailed Deer (Odocoileus virginianus) at Sand Lake National Wildlife Refuge,
South Dakota, before and during hunting seasons in 1992 and 1993 to assess the effect of hunting on deer movements and
activity. Unexpected flooding throughout the refuge in 1993 also allowed us to evaluate the effects of loss of escape cover
(i.e., robust emergent wetland vegetation) on deer movements and activity. Although >70% of diurnal locations occurred
in escape cover in 1992, use of escape cover was <28% in 1993 as a result of unexpected flooding that rendered most
escape cover unavailable to deer. In 1992, hunted deer remained in areas similar to those used before the hunting season
and responded to presence and activity of hunters by moving farther into escape cover. In 1993, deer moved to adjacent
private lands when availability of escape cover decreased on the refuge. Deer located on private lands increased move-
ments during the hunting season. In 1992, deer exhibited diurnal activity patterns before and during hunting seasons; cre-
puscular activity patterns were observed before and during hunting seasons in 1993. Differences in activity patterns
before hunting seasons in 1992 and 1993 indicated that availability of escape cover was the major factor influencing deer
activity patterns.
Key Words: White-tailed Deer, Odocoileus virginianus, activity, escape cover, hunting, movements, South Dakota.
White-tailed Deer populations can be maintained
at high (Larson et al. 1978) or low (Swenson 1982)
densities when intensively managed. Availability,
quality, and juxtaposition of forage and cover, as
well as hunter density and human-deer encounter
rate before harvest are important aspects related to
the vulnerability of deer to hunting. Swenson (1982)
suggested using percent cover as an index to vulner-
ability of deer to hunting; however, deer movements
and use of cover during hunting have been highly
variable (Marshall and Whittington 1968; Root et al.
1988). Studies with refuges (i.e., non-hunted areas)
in proximity to hunted areas have indicated either
deer responded to hunting by moving into refuges
(Zagata and Haugen 1973; Kammermeyer and
Marchinton 1976) or refuges received no additional
use during hunting (Pilcher and Wampler 1982; Root
et al. 1988). Similarly, Marshall and Whittington
(1968) concluded that deer movement increased as
hunting increased, while Kufeld et al. (1988)
observed that deer modified habitat use without
increasing movements.
White-tailed Deer are usually crepuscular
(Montgomery 1963; Kammermeyer and Marchinton
1976; Eberhardt et al. 1984). However, deer may
modify activity patterns in response to weather
(Beier and McCullough 1990) and moonlight
(Kammermeyer 1975). White-tailed Deer also may
modify activity in response to hunting or loss of
escape cover; however, such a response has not pre-
viously been documented.
Our objectives were to evaluate the effect of hunt-
ing and loss of escape cover on movements and activi-
ty of female White-tailed Deer. Escape cover was
defined as the tall, dense and rank emergent wetland
vegetation that was located within the refuge we stud-
ied. Inundation of escape cover that occurred through-
out the refuge in 1993 provided an opportunity to
study the effects of flooding on movements and activ-
ity of female White-tailed Deer. Our null hypothesis
was that deer movements and activity would not differ
before and during hunting seasons and that loss of
escape cover would not influence deer movements
and activity. Our study is unique in the nature and
severity of flooding that occurred.
Study Area and Methods
Sand Lake National Wildlife Refuge (SLNWR) is
8693 ha in size and lies within the James River
Lowland in Brown County, South Dakota. Water
control structures that were constructed in the late
1930s to control water levels have created two large
marsh and open water areas that comprise 45% of
SLNWR. Tall, dense and rank stands of emergent
wetland vegetation (i.e., cattail [Typha spp.] and
Common Reed [Phragmites australis]) that we
defined as escape cover were readily available to
deer in fall 1992. However, escape cover on
SLNWR was flooded in 1993 due to heavy summer
rains. Inundated conditions that persisted into fall
1993 rendered escape cover unavailable to deer
during 1993 SLNWR hunting seasons.
51/5)
596
Vegetation communities other than escape cover
on SLNWR include uplands (34%), which are large-
ly Brome-dominated (Bromus inermus) grasslands
intermixed with native species such as Big Bluestem
(Andropogon gerardii), Little Bluestem (A. scopar-
ius), and Indiangrass (Sorghastrum nutans).
Harvested crops (9%) include corn (Zea spp.),
Alfalfa (Medicago sativa), and small grains (e.g.,
Wheat [Triticum aestivum]). The remaining area
consists of open water (11%) and shelterbelt plant-
ings of trees (1%) such as Cottonwood (Populus del-
toides), Green Ash (Fraxinus pennsylvanica), and
Russian Olive (Elaeagnus angustifolia).
Privately-owned land adjacent to SLNWR is
intensively farmed. Agricultural crops constitute
72% of land use. Crops grown on privately-owned
lands are similar to those planted on SLNWR.
Permanent pasture (17%) and idled areas (12%) (i.e.,
Conservation Reserve Program fields and wetlands)
comprise the remaining land area.
Hunting seasons in 1992 and 1993 were similar in
length and timing on SLNWR. Two-deer permits
were issued to 135 new hunters during the two, 6-
day and two, 7-day rifle seasons that were conducted
consecutively (18 November - 13 December 1992;
17 November - 12 December 1993). Deer were hunt-
ed on private lands in a 213 km? area surrounding
SLNWR. Rifle hunting on private lands was one, 9-
day season in which 900 hunters were issued either
one- or two-deer permits. The hunting season on pri-
vate lands was from 28 November to 6 December
1992 and 27 November to 5 December in 1993.
We digitized ground-truthed upland vegetation
communities from 1:8000 scale black-and-white
aerial photographs into a PC ARC/INFO (Environ-
mental Systems Research Institute, Incorporated,
[ESRI], Redlands, California, USA) geographic
information system (GIS). Maximum acceptable root
mean square error (i.e., mapping error) was 0.03
(ESRI 1991). We obtained wetland vegetation com-
munities and open water areas as a digital coverage
(United States Department of the Interior, Bureau of
Reclamation, Bismarck, North Dakota, USA), which
was combined with upland vegetation communities
in the GIS. Robust emergent wetland vegetation
communities that we defined as escape cover were
coded separately from open water areas and other
vegetation communities that did not contain robust
emergent wetland vegetation.
We captured deer in modified Clover traps
(Clover 1956) and radio-collared them (Telonics,
Incorporated, Mesa, Arizona, USA) during the sum-
mers of 1992 and 1993 and the winter of 1993
(Naugle et al. 1995). We received signals with a
Telonics model TR-2 receiver (Telonics, Incorpor-
ated, Mesa, Arizona, USA) and estimated azimuths
with a vehicle-mounted null antenna system fitted
with an integrated azimuth locating device (Hallberg
THE CANADIAN FIELD-NATURALIST
Vol. 111
et al. 1974; Balkenbush and Hallett 1988). We esti-
mated accuracy of our telemetry system at a distance
of 866 m by placing a transmitter at locations
unknown to the observer. We recorded 10 indepen-
dent azimuths for each of three observers. The stan-
dard deviation of the error angle was 0.7.
We monitored 12 radio-collared females 16
October to 13 December (430 locations) in 1992 and
17 radio-collared females 1 October to 12 December
(817 locations) in 1993. Number of females moni-
tored in 1992 was reduced to 8 (2 yearlings, 6 adults)
because of harvest (n = 1) and movement to non-
hunted areas (n = 2). We excluded from analyses one
female that only used SLNWR nocturnally. In 1993,
our sample was reduced to 13 adult females due to
harvest (n = 4).
We also monitored 5 radio-collared males 16
October to 13 December (224 locations) in 1992 and
10 radio-collared males 1 October to 12 December
(458 locations) in 1993. In 1992, number of males
monitored was reduced to 4 (3 yearlings, 1 adult)
because of harvest (n = 1). In 1993, our sample was
reduced to 3 males (2 yearlings, 1 adult) due to har-
vesi GES):
We estimated deer locations by triangulation from 2
to 5 known receiver locations (e.g., road intersections)
with Program XYLOG (Dodge and Steiner 1986). We
did not use deer locations in analyses that required
>20 minutes to collect. Mean 95% confidence ellipse
was 0.001 ha for 1992 locations (n = 406) and 0.0004
ha for 1993 (n = 764). Mean distance to signal source
was 1034 m in 1992 and 727 m in 1993.
We determined independence between locations
by intensively monitoring three deer for 48 hours in
fall 1992. We tested cumulative distances between
locations for each deer in a 48-hour intensive moni-
toring period for autocorrelation to determine time of
independence (Swihart and Slade 1985). We
obtained independence and representation of loca-
tions over the 24-hour period by locating each deer
every 20 hours (i.e., monitoring period). We located
deer in the same order each monitoring period and
lagged the start time of each monitoring period until
monitoring had begun with each hour of the day.
We selected diurnal locations (0730 - 1700 hr
before hunting, 0800 - 1700 h during hunting) to
evaluate effects of hunting on deer. We combined
diurnal locations with vegetation communities to
compare deer locations before hunting to locations
during hunting with Multiresponse Permutation
Procedures (MRPP) (Mielke et al. 1981) contained
within Program BLOSSOM (Slauson et al. 1991).
We calculated individual 95 percent home ranges
before and during hunting each fall with an adaptive
kernel method (Worton 1989) contained within
Program CALHOME (United States Department of
Agriculture, Forest Service, Pacific Southwest
Research Station and California Department of Fish
597
ER
Cov
OF
Loss
AND
TING
: HUN
ON:
eas
A
SELL
KERN
S,
JENK
GLE,
NAU
oh
199
dis-
used ions ie as
A) We locati ee 63)
c . Ee ewe .
ia. US SsiV CAL = ca-
ifornia, succe Eoae 23 i ie e
li n d d 0. m
Ca ee late use = ho
, tw Icu e E= te - we
Game d be ca nit. VW 0(S Icula 92; w
and Hoe anne ae nd 3 to ca in 19 nd 19
ce) Oe g hu Au2 ing a er ved ing a in
tan durin deer hunti r de mo nae eee
and dex et ting disances mo es per f cover
in n i tin eo T to
an ions | hu to ons un dg ve
locati and des Pe as nated ding by
ions d ine iVi €
on n = 0. ions to n iV W
LL nges (SE r ee tions i and d deer. in
se ra 44 ) lo d di loca er er ces
ZA =< SE = ew me oe ape ca dt tance
S ee Ww ing d e of in esc dete d dis tip
a oe ] umm t edg ions ts to ed, aft ars. cury- If
8 = 9 ee, deat Bai ee 88;
8 > ot ES the ber o ende distan tans w mine h 19 ‘
a S 1 Le num indep IZ, € CO ipp deter ug day
2 sed i nge s f escap See ae for set
Bs a he ted o rs Mice ee a ee ena for
st lla llows anc e ri ta!
| n . ees ich a Beier adjust out ye tive
g = jaa SN a Rad S wh tive : We es sa activ ime ae
D @ itches, Cc . eric t ier
5 EZ En der were a Cae Spe Se a
= mal r e an e € 3 la at
S es ae s WwW ing p 1-0 e th
3) 1€ ng Si tin 0 Orr ime
3 © iS Vv vi 6 u 00 nc tim en
2 < es S ae ‘ote a vals (ex, Sea iss se bie
a - ae 990). 3 been Avion: at fema
foe) : = : S
z a aa ions e Die time th
aa Cu ela es rso O it
Be | — = eva les a lso u een p ring | ten-
v a Yn fema We o betw nd du centra tion
i Pe a ibu
= g E 5 rela ctive Es lative (.e., 75%
> Se a Ss ( ea. S :
20 A eq | ac Ge nce Jin ROMS )in in
Ss 6 oo = M differe f the : epee anges Be
2B a sults cted sion 0 PS.2 ring TUE ee d
aie S mee oe anal eae femal oe a
S me e fe € in 1 e
S 0 f| n € re f a (e}
gal 2 re oeeeees 35% (n = 13) “nce ee
a2 ea O m (n tiv h Ne g
ae 2 n of p f fe 5% ec fore inde urin 05)
9 2 ~ 2 ions 0 nd 8 oe Our i er d <0) er
sae | 5 me 8) a 1993, g tha ie he Salsa )
38 2 | ent retest Pence Set
al| 2 g oo 992 er 5) ( also nd in es Ta 2%
2S as) fe) a 1 gy 0 d a in 2 ih
el iS es |e Oo la 0. e€ 2 é S 9 S
a5 ea “122 were ens moy 199 sci se awe 1993,
o= S in Cc 1 e
z Z at ets of lo ua sents J 11%
oe | aokeel edge ey Arie ie
= es oS fore to .O5) COV ing 1 ales diu d
a5 = Re Sate nue eae ee fe an
2 ahs ess Dis ased ( of Seen ver by _In ; eer by
os see ge Se r incre se co ing A% N ing.
ss) e i Si = p a S x inc nal u 85% cape hunt as 9 f SL ae
3 =~ 5 r a
3 ep a = = = as a ze of ring hn SS in ee ey
2-5 < ase Ua Q be al u 27% fe diurn 5% S ou iurn
25 = 6 ||S ee a aa ane WR 1594, and 3 ene 2 1992
@ 2 Bw ere oe SLN g aA ae. is Se 19 er in cape
3 We | Lis ing. i Vv S
eon ) RS: ) nti 6 a Jo ee I-
2 3 S a 3 sa Si U se a as Det: al men 55 scape outsid dio-co t
Ges 8 ‘) Ss A 0 = ae ee er increased 5 © cated for ra ns tha
an ks exh wag = ber o lente eer lo sizes catio
2S 5 ea mA pisses N ape e of 80% Sawtlin
= 2 o <|% . Il am = -=| esc le us hile Low valu ing.
Ie PEELE: ere aeons i hu
9,0 oo DS ae 5 5 was ere rec durin
23 = ee Pres ee: males preci
=| = tus Se ve) 4 < = = 5 1S) COV ma fore
ae Ie Aa BES Ss red be
a a esaaas la urred
EE are ih
Sj oO Gy Se ce} CAMA)
ae on Ht
SpE = EE > =
5s Sick
S3 aa) s Pe
<8 RS a
ir) tS
are
ian
—) =
fea} 39}
ae
598
100
NOCTURNAL
DIURNAL
80
60
PERCENT ACTIVE
40
0 3 6 Sigma als
TIME. (hours)
THE CANADIAN FIELD-NATURALIST
Vol. 111
FO Ooo Female 1992
ne es Male 1992
——-—- Female 1993
—— Male 1883s
18.21 24
FiGurE 1. Activity (% active) of female and male White-tailed Deer during fall 1992 and 1993 at Sand
Lake National Wildlife Refuge, South Dakota.
We obtained activity for 98% (n = 421) of female
locations in 1992 and 91% (n = 743) in 1993. We
also obtained activity for 86% (n = 193) of male
locations in 1992 and 92% (n = 421) in 1993.
Activity of females and males in 1992 was positively
correlated (77 = 0.471, n = 8, P = 0.060) (Figure 1).
Activity of females and males in 1993 also was posi-
tively correlated "G2 ="0.782, n= 8, P= 01004)
(Figure 1). Activity was not correlated for females
GES2vs 1995 4G= =O 24 ln oe 0 or
males (1992 vs. 1993) (7? = 0.045, n = 8, P= 0.614)
(Figure 1). Activity of females before and during
hunting in 1993 was positively correlated (r? =
0.879, n = 8, P = 0.001) (Figure 2). Activity of
females before hunting in 1992 and during hunting
in 1993 was negatively correlated (77 = 0.801, n = 8,
P = 0.003) (Figure 2). Activity of females before
hunting in 1992 and 1993 also was negatively corre-
lated (77 = 0.685, n = 8, P= 0.011) (Figure 2).
Discussion
Although deer may move to refuges when hunted
despite abundant cover within hunted areas
(Kammermeyer and Marchinton 1976), deer move-
ments during the hunting season on SLNWR were
similar to those reported by Kufeld et al. (1988) and
Root et al. (1988); hunted deer remained within areas
similar to those used before the hunting season. Deer
use of escape cover on SLNWR was high before and
during hunting in 1992. Deer used SLNWR almost
exclusively when water within the James River was at
normal levels. Similar use of escape cover (80-90% of
diurnal locations) has been reported for White-tailed
Deer in Tamarack (Larix laricina) swamps of south-
central Wisconsin (Larson et al. 1978).
Deer home ranges and movements increased dur-
ing hunting in 1992 as deer responded to presence
and activity of hunters by moving farther into escape
cover. Decreased deer use of escape cover on
SLNWR from 1992 to 1993 indicated that availabili-
ty of escape cover was reduced by unusually high
water levels in 1993. Deer movement onto private
lands was reflected in the increased number of diur-
nal deer relocations in areas other than escape cover
from 1992 to 1993. However, deer adapted different-
ly to hunting in 1993 than 1992 even though home
ranges and movements increased similarly during
hunting seasons in both years.
Despite the shorter hunting season on private lands,
most deer movement that caused increases in home
ranges and movements occurred during this period.
Habitats on private lands may provide deer with
immediate concealment similar to escape cover on
SLNWR; however, habitats on private lands that pro-
vide such concealment were generally smaller in size
than larger expanses of escape cover found on
SLNWR. Habitats such as treebelts only constituted 1-
2% of private land use, but deer use indicated that tree-
1997
100
NOCTURNAL
80
60
40
PERCENT ACTIVE
20
Oi 8 6 OP Aza io
TIME (hours)
NAUGLE, JENKS, KERNOHAN, AND JOHNSON: HUNTING AND LOSS OF COVER
599
Gores roaeecc 92-Before Season
enya 92-Durlng Season
——-—- 93-Before Season
93-Durlng Season
18 21 24
FIGURE 2. Activity (% active) of female White-tailed Deer before and during hunting seasons at Sand
Lake National Wildlife Refuge, South Dakota, 1992 - 1993.
belts were an important habitat during hunting (Naugle
1994). Nevertheless, such small and linear habitats do
not constitute high quality escape cover. As a result,
deer located on private lands in 1993 increased move-
ments during hunting. In east-central South Dakota,
Sparrowe and Springer (1970) also reported increased
movements of hunted deer in habitats similar to those
on private lands adjacent to SLNWR.
Activity of females in 1992 differed from that
reported for deer in forested-agricultural
(Montgomery 1963; Kammermeyer and Marchinton
1976) and sagebrush- (Artemesia spp.) steppe habi-
tats (Eberhardt et al. 1984). In 1992, females were
most active on SLNWR at mid-day before and dur-
ing hunting (Figure 2). High diurnal activity also has
been reported for an unhunted Columbian White-
tailed Deer (O. v. leucurus) population in
Washington (Suring and Vohs 1979).
Kammermeyer and Marchinton (1976) suggested
that deer fed nocturnally in open agricultural fields
because they were more secure in exposed habitats at
night. Although diurnal deer use of habitats outside of
escape cover 1992 was low, most deer located outside
of escape cover were active. Deer may have experi-
enced a similar “psychological security” diurnally
while in close proximity to escape cover in 1992.
Suring and Vohs (1979) also reported that presence of
dense cover in the vicinity of forage attracted deer.
Although diurnal activity was high throughout fall
in 1992, deer were more active during hunting than
before hunting. Major changes in activity between
periods in 1992 occurred during crepuscular hours.
Deer increased crepuscular activity during hunting
while maintaining high diurnal activity. Increased
activity during hunting may be related to hunting or
other factors such as search strategies of females dur-
ing rut (Holzenbein and Schwede 1989). However,
rutting activities probably were not responsible for
variation in activity patterns before and during hunt-
ing in 1992 because similar responses were not evi-
dent before and during hunting in 1993.
Activity of females during high water levels in
1993 followed crepuscular patterns (Figure 2) simi-
lar to those reported for deer in other habitats
(Montgomery 1963; Kammermeyer and Marchinton
1976; Eberhardt et al. 1984). Crepuscular activity
patterns of females before and during hunting in
1993 (Figure 2), which were positively correlated,
indicated that female activity was not affected by
hunting. Changes in female activity that may have
occurred during hunting would not have been detect-
ed due to the short hunting season on private lands.
Breeding season occurred coincident with hunting
on SLNWR and surrounding private lands. However,
activity patterns of females and males that were posi-
tively correlated within fall 1992 and 1993 and activ-
ity patterns of males and females that were not corre-
lated (i.e.; fall 1992 and 1993; Figure 1) indicated
that the shift from a diurnal activity pattern in 1992 to
a crepuscular activity pattern in 1993 was not
600
attributable to rut. Relyea and Demarais (1994) also
noted similar variation in intersexual activity patterns
during rut for Mule Deer (Odocoileus hemionus croo-
ki) despite differences in activity from prerut to
postrut. Consequently, activity patterns before hunt-
ing in 1992 and 1993 (Figure 2) that were inversely
related indicated that availability of escape cover was
the major factor influencing deer activity patterns.
Acknowledgments
We thank L. D. Flake, K. F. Higgins, and L. A.
Rice for their review of our manuscript. J. J. Kelly
and J. J. Sahli assisted with data collection. This
study (7560) was funded by Federal Aid to Wildlife
Restoration Fund, Project W-75-R through South
Dakota Department of Game, Fish and Parks.
Funding also was provided by the South Dakota
Agricultural Experiment Station and the National
Rifle Association. Support was provided by South
Dakota Cooperative Fish and Wildlife Research
Unit, South Dakota State University, and Sand Lake
National Wildlife Refuge (United States Fish and
Wildlife Service). This is South Dakota Agricultural
Experiment Station Publication Number 3000.
Literature Cited
Balkenbush, J. A., and D. L. Hallett. 1988. An improved
vehicle-mounted telemetry system. Wildlife Society
Bulletin 16: 65-67.
Beier, P., and D. R. McCullough. 1990. Factors influenc-
ing White-tailed Deer activity patterns and habitat use.
Wildlife Monographs 109: 1-51.
Beier, P., and D. R. McCullough. 1988. Motion-sensitive
radio collars for estimating White-tailed Deer activity.
Journal of Wildlife Management 52: 11-13.
Clover, M. R. 1956. Single gate deer trap. California Fish
and Game 40: 199-201.
Dodge, W. E., and A. J. Steiner. 1986. XYLOG: a com-
puter program for field processing locations of radio-
tagged wildlife. Fish and Wildlife Technical Report 4,
Washington, D.C.
Eberhardt, L. E., E. E. Hanson, and L. L. Cadwell.
1984. Movement and activity patterns of mule deer in
the sagebrush-steppe region. Journal of Mammalogy 65:
404-409.
Environmental Systems Research Institute. 1991. PC
ARC/INFO STARTER KIT User’s Guide, Version 3.4D
Plus. Environmental Systems Research Institute,
Incorporated, Redlands, California.
Hallberg, D. F., F. Janza, and G. Trap. 1974. A vehicle-
mounted directional antenna system for biotelemetry
monitoring. California Fish and Game 60: 172-177.
Holzenbein, S., and G. Schwede. 1989. Activity and
movements of female White-tailed Deer during the rut.
Journal of Wildlife Management 53: 219-223.
Kammermeyer, K. E. 1975. Movement-ecology of White-
tailed Deer in relation to a refuge and a hunted area. Ms.
thesis, University of Georgia, Athens. 114 pages.
Kammermeyer, K. E., and R. L. Marchinton. 1976. The
dynamic aspects of deer populations utilizing a refuge.
Proceedings of the Annual Conference of the South-
eastern Association of Game and Fish Commissioners 29:
466-475.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1988.
Influence of hunting on movements of female mule deer.
Journal of Range Management 41: 70-72.
Lariviere, S., J. Huot, and C. Samson. 1994. Daily activi-
ty patterns of female black bears in a northern mixed-for-
est environment. Journal of Mammalogy 75: 613-620.
Larson, T. J., O. J. Rongstad, and F. W. Terbilcox.
1978. Movement and habitat use of White-tailed Deer
in southcentral Wisconsin. Journal of Wildlife
Management 42: 113-117.
Marshall, A. D., and R. W. Whittington. 1968. A tele-
metric study of deer home ranges and behavior of deer
during managed hunts. Proceedings of the Annual
Conference of the Southeastern Association of Game
and Fish Commissioners 22: 30-46.
Mielke, P. W., K. J. Berry, P. J. Brockwell, and J. S.
Williams. 1981. A class of nonparametric tests based
on multiresponse permutation procedures. Biometrica
68: 720-724.
Montgomery, G. G. 1963. Nocturnal movements and
activity rhythms of White-tailed Deer. Journal of
Wildlife Management 27: 422-427.
Naugle, D. E. 1994. Density, movements, and habitat use
of White-tailed Deer at Sand Lake National Wildlife
Refuge, South Dakota. M.Sc. thesis, South Dakota State
University, Brookings. 71 pages.
Naugle, D. E., B. J. Kernohan, and J. A. Jenks. 1995.
Seasonal capture success and bait use of White-tailed
Deer in an agricultural-wetland complex. Wildlife
Society Bulletin 23: 198-200.
Pilcher, B. K., and G. E. Wampler. 1982. Hunting sea-
son movements of White-tailed Deer on Fort Sill
Military Reservation, Oklahoma. Proceedings of the
Annual Conference of the Southeastern Association of
Fish and Wildlife Agencies 35: 142-148.
Relyea, R. A., and S. Demarais. 1994. Activity of desert
mule deer during the breeding season. Journal of
Mammalogy 75: 940-949.
Root, B. G., E. K. Fritzell, and N. F. Giessman. 1988.
Effects of intensive hunting on White-tailed Deer move-
ment. Wildlife Society Bulletin 16: 145—151.
Slauson, W. L., B. S. Cade, and J. D. Richards. 1991.
User manual for BLOSSOM statistical software.
National Ecology Research Center, United States Fish
and Wildlife Service, Fort Collins, Colorado.
Sparrowe, R. D., and P. F. Springer. 1970. Seasonal
activity patterns of White-tailed Deer in eastern South
Dakota. Journal of Wildlife Management 34: 420-431.
Suring, L. H., and P. A. Vohs, Jr. 1979. Habitat use by
Columbian White-tailed Deer. Journal of Wildlife
Management 43: 610-619.
Swenson, J. E. 1982. Effects of hunting on habitat use by
mule deer on mixed-grass prairie in Montana. Wildlife
Society Bulletin 10: 115-120.
Swihart, R. K. and N. A. Slade. 1985. Testing for inde-
pendence of observations in animal movements. Ecology
66: 1176-1184.
Worton, B. J. 1989. Kernel methods for estimating the uti-
lization distribution in home-range studies. Ecology 70:
164-168.
Zagata, M. D., and A. O. Haugen. 1973. Pilot Knob
State Park-a winter deer haven. Iowa State Journal of
Research 47: 199-217.
Received 12 November 1996
Accepted 25 March 1997
A New Surgical Technique for Implanting Radio Transmitters in
Beavers, Castor canadensis
MICHELLE WHEATLEY
Taiga Biological Station, Department of Zoology, University of Manitoba, Winnipeg, Manitoba R3T 2N2
Current address: Sahtu Renewable Resources Board, Box 490, Norman Wells, Northwest Territories XOE OVO
Wheatley, Michelle. 1997. A new surgical technique for implanting radio transmitters in Beavers, Castor canadensis.
Canadian Field-Naturalist 111(4): 601-606.
A surgical method of intraperitoneal implantation of radio transmitters using a ventral-lateral muscle-split technique is
described. Forty-two transmitters were implanted in beavers weighing 6.4 to 20.5 kg (x = 11.4 + 3.9 kg). Recaptured
implanted animals showed good growth with no post-operative complications at 3 to 36 months. Transmitter life span was
46 to 718 days (xk = 213.1 + 155.6 days) excluding animals which left the study area in the first 30 days. This method is
suitable for all ages of Beavers, and is easily performed in the field.
Key Words: Beaver, Castor canadensis, radio transmitters, radio-telemetry, surgical implantation.
Beavers are aquatic, nocturnal mammals and, in
many parts of their range, confined to lodges and
under-ice movement during winter. Radio-telemetry
is an ideal method for the study of their movements.
Guynn et al. (1987) discussed some of the difficul-
ties associated with external transmitters. Externally
mounted transmitters have been used on Beavers, but
with only limited success and they are restricted to
adult individuals. Tail collars (Busher 1975;
Courcelles and Nault 1983; and Wheatley 1989) and
neck collars (Lancia 1979) may cause abrasions or
more serious injury to the animal and may only stay
attached for a short period of time (Busher 1975;
Wheatley 1989). For longer term data gathering,
implanted transmitters are required. Davis (1984)
and Davis et al. (1984) reported on a method of
implanting intraperitoneal transmitters in Beavers
using a dorsal incision. They also attempted two sub-
cutaneous implants with poor success.
For my study of Beaver movements in the taiga of —
southeastern Manitoba (Wheatley 1989, 1994,
1997a, 1997b, 1997c) I required a method of
implanting transmitters that incurred a low risk for
the Beavers and was feasible under field conditions.
This paper describes the method developed, using a
ventro-lateral incision, the post-operative history of
the Beavers and performance of the transmitters.
Methods
Beavers were captured in Hancock live traps and
held in a darkened building at Taiga Biological
Station (TBS) for 1.5 to 12 hours before surgery. All
surgery was carried out at TBS. The surgical
approach was developed in consultation with a sur-
geon (Dr. B. Wheatley, F.R.C.S. Ed.) and a veteri-
narian (Dr. L. Webster, D.V.M., Pembina Veterinary
Hospital, Winnipeg, Manitoba) and was based on the
standard “gridiron” muscle split human appendecto-
my approach (Farquharson 1962). Prior to beginning
the project, initial dissection of Beaver cadavers
showed that for a successful muscle split, the inci-
sion needed to be closer to the costal margin than in
human subjects. All work was approved by the
University of Manitoba Animal Care Committee and
animals were cared for in accordance with guidelines
of the Canadian Council on Animal Care (1980).
The implantable transmitters (Austec Electronics,
Edmonton, Alberta) consisted of the transmitter and
a C-cell lithium battery encased in beeswax and then
in biologically inert Elvax (Mini-mitter Co.,
Sunriver, Oregon). The finished cylindrical package
measured about 6 cm x 4 cm and weighed about
80 g. Each package was sterilised in 70% alcohol for
2 hours and rinsed in sterile water immediately
before insertion.
I anaesthetized Beavers with simultaneous intra-
muscular injections of ketamine hydrochloride and
acepromazine maleate (Lancia et al. 1978). Dosage
was 20 mg ketamine hydrochloride per kg body
weight and a standard dose of 2.5 mg acepromazine
maleate for all Beavers under 18 kg. Beavers larger
than 18 kg were given 4 mg of acepromazine
maleate. Booster doses of 50 to 80 mg ketamine
hydrochloride were given intramuscularly as neces-
sary to maintain anaesthesia. I also injected 40 mg of
lidocaine, a local anaesthetic subcutaneously along
the incision line.
I performed the surgery in a log laboratory build-
ing, log and mosquito screen “gazebo” or in an
2.4 X 3 m wall tent in the field. Following anaesthe-
sia, the Beaver was placed on the operating table and
the front and hind legs were constrained to prevent
involuntary movement. The eyes were protected
with an antibiotic ointment (Rogar-mycine,
Rogar/STB Inc., London, Ontario), and a wet paper
towel was wrapped around the Beaver's tail to pre-
601
602
vent overheating (Lancia et al. 1978). All personnel
wore Sterile surgical gloves and masks during the
operation. Instruments were sterilized between oper-
ations by boiling for 30 minutes, then heating in an
oven at 120°C for at least 1 hour.
I clipped the fur from an area 3 cm wide and 8 cm
long on the ventral surface immediately caudal to the
right costal margin and perpendicular to the midline.
I left about 4 mm of fur to help prevent the Beaver
from catching the stitches during grooming (Dr. L.
Webster, personal communication). The underfur of
the Beaver prevents the Beaver’s skin from getting
wet (Djoshkin and Sofonow 1972) and leaving a
small amount of fur helps keep the incision dry dur-
ing healing. I washed the area with a 30:1 solution of
Savlon and then rinsed with 70% ethanol. Excess
moisture was blotted up. A plastic incise drape
(Steri-drape, 3M Medical-Surgical Division, St.
Paul, Minnesota) was placed over the Beaver, pro-
viding a sterile field. I cut a hole in the plastic over
the site for the incision.
I made the incision using a # 22 scalpel blade after
separating the fur in a line to expose the skin. The
incision started 1.5 to 2 cm lateral to the midline and
3 to 4 cm caudal to the costal margin. It extended
about 8 cm away from the midline, angling slightly
towards the costal margin.
The connective tissue between the skin and the
external oblique muscle was cleared using fine scis-
sors, separating the two layers and exposing the
muscle layer. The external oblique muscle was
opened using a muscle-split technique. A small,
blunt haemostat was placed against the muscle and
was opened with gentle downward pressure, along
the line of the muscle fibres. This was repeated as
necessary until the internal oblique muscle could be
viewed through the split in the external oblique. The
external oblique muscle was separated from the
internal oblique muscle and the internal oblique was
then split by blunt dissection in the same manner as
the external oblique. The internal oblique and
transversus abdominis muscles were separated and
the transversus abdominis muscle was split in the
Same manner as the other muscle layers. In all
Beavers, the peritoneum was attached to the inner
wall of the transversus abdominis muscle. It was
therefore opened at the same time as the transversus
abdominis was split.
Once the opening in the transversus abdominis
muscle and the peritoneum was large enough, I
clipped haemostats on the peritoneum at each end of
the incision to facilitate control and to prevent fur-
ther splitting while inserting the sterilized transmitter
through the incision. The transmitter was gently
pushed towards the midline and a 1 million IU dose
of Penicillin G Sodium (Crystapen, Glaxo) was
poured into the peritoneal cavity.
I sewed up the transversus abdominis muscle and
THE CANADIAN FIELD-NATURALIST
Vol. 111
peritoneum as one layer with a continuous 3 ‘0’
plain or chromic gut suture. The internal oblique and
external oblique muscles were sewn up as separate
layers in turn in a similar manner. I sutured the skin
with 15 to 20 interrupted stitches using 3 ‘0’ silk
sutures. Derapen (Ayerst Laboratories, Montreal,
Quebec), a penicillin antibiotic, was given intramus-
cularly at a dose of 0.22 ml per kg body weight
(66 000 IU per kg). To facilitate visual identification
of individuals, I also attached uniquely numbered
and coloured ear tags (Monel Number 3) to both ears
(Miller 1964).
The Beaver was returned to the live trap. For most
of the 48 hours after surgery, it was kept in a cool
dark building to allow for post-operative and post-
anaesthetic recovery and because in many cases the
location of the home lodge was not known. Beavers
frequently swam long distances (up to 5 km) imme-
diately after release. After 24 hours, the Beaver, in
the trap, was placed partly in the water for 4 to 5
hours and Trembling Aspen (Populus tremuloides)
branches were supplied. The incision site was
inspected before the animal's release at the capture
site.
Results
Transmitters were implanted in 42 Beavers weigh-
ing 6.4 to 20.5 kg with a mean weight of 11.4 + 3.9 kg
(Table 1). Beavers became alert and active between
2 and 5 hours post-operatively, but then became
sluggish at 6 to 8 hours and remained sluggish until
30 to 36 hours post-op. By 48 hours after surgery
they were again alert and active.
With this technique, the range of transmitters was
found to be from 0.1 km (transmitter and receiver in
dense alder bog) to 2 km (receiver on high ridge).
Range was about | km on the lake and about 4 km
from the air.
Ten Beavers disappeared from the study area
within 1 month (Table 1). The 30 remaining Beavers
(two received two transmitters during the study)
were active 3 months to 4 years after surgery. I
recaptured four implanted Beavers a total of seven
times in the 3- to 36-month period after surgery. All
the Beavers had a palpable scar, but no other sign of
the incision site. Beaver 1848 increased in weight
from 13.2 kg to 15.9 kg over the first 3 months after
implantation and to 17.7 kg 35 months after implan-
tation. She bore her first kits 2 years after implanta-
tion. Beaver 1846 increased in weight from 9.5 kg to
11.4 kg in 11 months. This Beaver may have had a
low weight on recapture because the capture was
shortly after breakup, when winter weight loss may
be evident (Aleksiuk and Cowan 1969a, 1969b).
Beaver 294 (later retagged 1510) increased in weight
from 11.4 kg to 20.45 kg over a 36-month period.
Beaver 1611 (later retagged 1516) was an adult at
the time of implantation and recapture 12 months
1997
WHEATLEY: IMPLANTING RADIO TRANSMITTERS IN BEAVERS 603
TABLE 1. Age class, sex and weight of Beavers implanted with radio-transmitters in this study, and lifespan and fate or sta-
tus of transmitters. The term "signal disappeared" refers to animals where there was no visual confirmation of the animal
still being present after loss of the signal. "Transmitter failed" refers to animals which were observed visually after loss of
the signal.
Beaver # Age Class
1846
1834
1848
B280
B286
B434
B294
B481
B420
B292
B109
B181
B290
B491
B433
B403
1631
1603
1646
1680
1686
1649
1641
1617
1682
1611
1684
1623
1697
1516
Yearling
Yearling
2-year-old
Yearling
Yearling
2-year-old
2-year-old
2-year-old
Yearling
2-year-old
Yearling
Adult
2-year-old
Adult
Yearling
2-year-old
Yearling
Yearling
Yearling
2-year-old
2-year-old
Yearling
Yearling
Adult
2-year-old
Adult
Yearling
Yearling
Yearling
Adult
Sex
Male
Female
Female
Female
Female
Male
Male
Male
Male
Female
Male
Male
Female
Male
Male
Female
Male
Female
Female
Female
Female
Male
Male
Male
Female
Male
Male
Female
Male
Male
Weight
(kg)
9.5
10.0
132
eZ
128
ede
OS)
20.5
10.5
10.9
10.0
18.6
Date
Implanted
08-06-87
12-06-87
03-07-87
06-07-87
12-05-88
25-08-88
26-09-88
22-05-89
24-05-89
25-05-89
26-05-89
27-05-89
30-05-89
09-06-89
10-06-89
11-06-89
04-05-90
05-05-90
05-05-90
11-05-90
12-05-90
14-05-90
17-05-90
17-05-90
18-05-90
28-05-90
29-05-90
30-05-90
09-06-90
18-05-91
Last signal
18-11-87
30-05-89
24-10-88
23-07-87
14-10-89
29-08-88
09-03-89
03-12-89
17-06-89
02-06-89
21-08-89
31-08-89
15-07-89
27-10-89
14-08-89
27-10-89
12-05-90
15-05-91
11-05-90
27-05-90
09-09-90
14-04-91
17-09-90
01-11-90
14-08-90
11-05-91
12-05-91
17-07-90
18-06-90
Fate / Status
transmitter failed, 164 days, recaptured
08-05-88, 11.4 kg
signal disappeared, 718 days
transmitter failed, 479 days, recaptured
28-09-87, 15.9 kg, 23-10-88, 17.2 kg,
28-05-89, 17.3 kg lactating, 13-05-90,
17.7 kg lactating
signal disappeared, 17 days*
signal disappeared, 519 days
signal disappeared, 4 days, located
30-03-89 during aerial survey, 36 km
from capture site, 218 days
transmitter failed, 164 days, new
transmitter, 25-09-91, see 1510
signal disappeared, 196 days
signal disappeared, 24 days*, trapped 48
km away on different watershed, 10-91
signal disappeared, 8 days*
signal disappeared, 87 days
transmitter failed, 96 days, observed
visually
signal disappeared, 46 days
transmitter failed, 140 days, animal seen
06-90
signal disappeared, 65 days
signal disappeared, 138 days
signal disappeared, 8 days*, last seen
heading south
signal disappeared, 10 days*
deceased, see text, transmitter
functioning
signal disappeared, 16 days*
signal disappeared, 120 days
transmitter failed, 334 days, trapped at
different lodge, 17-11-91, 16.4 kg,
see text
signal disappeared, 123 days
signal disappeared, 168 days
signal disappeared, 88 days
transmitter failed, 348 days, new
transmitter, 18-05-91, see 1516
signal disappeared, 348 days
transmitter failed, 48 days,
visual observations
signal disappeared, 9 days*
transmitter continuing, 05-92
Continued
604
TABLE 1. (Continued)
Weight Date
Beaver # Age Class Sex (ke) alent
1526 Yearling Female 8.6 23-05-91
1532 Yearling Female ES) 24-05-91
1534 Yearling Male 6.8 25-05-91
1524 Yearling Male 6.8 26-05-91
1502 Yearling Female 9.5 29-05-91
1528 Yearling Male aS 02-06-91
1539 Yearling Female 11.4 17-09-91
1520 Kit Female 6.8 23-09-91
1522 Kit Male 6.8 24-09-91
1518 Kit Male 6.8 24-09-91
1510 Adult Male 20.5 25-09-91
1508 Adult Female Mad 25-09-91
THE CANADIAN FIELD-NATURALIST
Vol. 111
Last signal Fate / Status
08-09-91 signal disappeared, 108 days
01-06-91 signal disappeared, 8 days*,
heading north
08-09-91 signal disappeared, 106 days
03-07-91 signal disappeared, 38 days
18-03-92 beaver killed by Otters, functioning
transmitter found on land, buried in
snow (apparently by Red Fox), 294 days,
temperature -18°C
23-04-92 signal disappeared, 333 days
20-09-91 signal disappeared, 3 days*,
heading west
= transmitter continuing, 05-92
= transmitter continuing, 05-92
= transmitter continuing, 05-92
28-04-92 transmitter failed, 217 days,
observed visually
23-04-92 transmitter continuing, 05-92
* Signals which disappeared within 30 days were assumed to have dispersed.
later. He had lost a small amount of weight, going
from 20.45 kg to 18.6 kg. Again this may be the
result of winter weight loss because the capture was
only shortly after breakup.
Beaver 1649 was trapped in a Conibear trap by the
local trapper 18 months after implantation. This ani-
mal had increased in weight from 8.2 kg to 17.3 kg.
The transmitter was encased in many layers of con-
nective tissue and had a large multi-branching vascu-
lar bundle leading to it and wrapped around it. A
small scar was visible on the peritoneum and internal
wall of the transversus abdominis and on the inner
side of the skin.
All 10 Beavers that left the study area during the
first 30 days after surgery were yearling or 2-year-
old animals (Table 1). Two are known to have dis-
persed: B434s signal was located 36 km away from
the capture site during an aerial survey; and B420
was trapped in Ontario 48 km away, two years later.
Three other Beavers (1631, 1532 and 1539) were last
seen heading away from the study area.
Excluding the nine non-relocated dispersers, but
including the relocated disperser, transmitter life var-
ied from 46 to 718 days and averaged 213.1 days
(+155.6 days) (Table 1). These values should be
viewed as minimum lifespan of the transmitters
because some of these Beavers may also have dis-
persed and therefore the transmitters did not neces-
sarily cease to function at the shorter periods. Five
transmitters were still active at the conclusion of the
study, including one in an adult female (1508) who
had successfully produced kits while carrying the
active transmitter.
One Beaver ceased to move six days after release.
The body was located three days later. Weight loss
(including dehydration after death) was 2.3 kg from
the original weight of 6.9 kg (10 days earlier). An
autopsy showed all internal organs to be healthy
except for the first 30 cm of the small intestine
which appeared gangrenous. The transmitter was
found lying between the liver and duodenum.
Beaver 1502 was apparently killed by Otters in
March 1992. The transmitter, completely devoid of
blood or tissue, was found on land, about 20 m from
shore, buried in snow, apparently by a Red Fox.
There were a few small teeth marks in the wax of the
transmitter casing, but despite the -18°C tempera-
tures, the transmitter continued to function 294 days
after implantation.
Discussion
Implanted transmitters can be used in almost any
size animal because they provide no hindrance to
growth and there is no potential for snagging on veg-
etation. The greatest risk to the animal is infection
due to surgery, but this can be reduced by sterile
technique and antibiotics. I found no evidence of
infection in any of the recaptured Beavers. Guynn et
al. (1987) reported one death due to adhesion to the
large intestine with resultant blockage of the lumen
in Davis's (1984) study. Guynn et al. (1987) reported
that after six weeks, intraperitoneally implanted
197
transmitters appeared to be encapsulated in fibrous
tissue, and my finding in the Beaver trapped by the
local trapper would agree with this. Once encapsulat-
ed, there appears to be little effect on the animal
(Guynn et al. 1987), as also evidenced by pregnancy
in implanted animals which I recaptured.
As in this study, Davis et al. (1984) found little
problem with intraperitoneal implants. Their method
used a dorsal incision rather than ventral. This could
be more dangerous to the animal. From my observa-
tions of Beaver over a 6-year period, I have found
that their natural repertoire of movements during
grooming and feeding tends to stretch the dorsal
muscles and skin more than the ventral muscles and
skin. In addition, the kidneys and blood vessels on
the dorsal wall of the abdomen may present a greater
chance for complications with a dorsal incision. In
the ventral-lateral method I describe, no large blood
vessels or organs are in the vicinity and this reduces
the risk of complications, especially when surgery is
performed in the field. With this technique there is
also natural closure of the layers because of the mus-
cle split and because the layers of muscle fibres are
at right angles to each other.
Kenward (1987) reported that a mid-ventral, or
linea alba, incision is the usual means of access to the
peritoneal cavity. This method has been used in physi-
ological studies of Beaver (Dyck 1991). While this
procedure may contribute to a faster operation,
because there are fewer layers, other problems may
arise. Connective tissue (the linea alba) is virtually
avascular and therefore slower to heal than muscle
and provides for a much weaker incision after sutur-
ing than the muscle-split technique (Farquharson
1962). The alternating directions of the split muscle
layers with the ventral-lateral incision provide a much
better seal than the linea alba incision. In addition, the
transmitter may sit directly over a mid-ventral incision
(Kenward 1987) and thereby affect healing, whereas
with the off-centre incision, this is less likely. More
abrasion of the incision site is also likely with a mid-
ventral incision because it is on the lowest part of the
abdomen, and is more likely to catch on the ground
and vegetation than the off-centre incision.
The method I describe is also relatively blood-free.
Since only the skin is cut, there is very little bleeding
and it is possible to see most blood vessels associated
with the muscle layers and avoid them. This is very
important in the field setting where such items as
cauterizing equipment are not readily available.
Recent developments in animal surgery suggest
some possible improvements to the anaesthetic and
analgesic regime used in this study. Greene et al.
(1991) describe a method of anaesthesia using
ketamine initially with subsequent tracheal intuba-
tion and halothane gas in oxygen administration.
However, gas anaesthesia is not practical in the field
setting. Jalanka et al. (1990) describe a combination
drug regimen of Medetomidine and Ketamine, which
WHEATLEY: IMPLANTING RADIO TRANSMITTERS IN BEAVERS
605
may be effective for Beavers although no doses are
given for Beavers. The sluggishness noted in the
post-operative period may result from lack of pain
control (Dr. N. Caulkett, DVM, Department of
Veterinary Anesthesiology, Radiology and Surgery,
University of Saskatchewan, personal communica-
tion). The area of pain control and pain assessment
in animals has presented many problems (Flecknell
1994). Recent studies in this area have suggested
administration of buprenorphine or bupivacaine to
relieve post-operative pain (Liles and Flecknell
1993). The addition of some form of post-operative
anaesthesia to the surgical method described in this
paper would be advised.
In summary, the surgical technique described
herein provides a safe, relatively straightforward
method of implantation of radio-transmitters in
Beavers. Since this technique avoids major organs
and blood vessels and can be performed rapidly, it is
well suited for use in field settings. The addition of
analgesic agents post-operatively is advised.
Acknowledgments
My supervisor W. O. Pruitt, Jr., provided advice
and encouragement throughout this project. L.
Webster and B. Wheatley provided advice on the
surgical technique. D. Bourrier, M. Campbell, J. Fry,
S. Robertson, and F. Weil assisted with surgery and
field work. Bill Conley, on whose Registered
Trapline this research was conducted, provided data
on the retrapped animal and valuable advice.
Financial support was provided by the Taiga
Biological Station Research Trust; Wildlife Branch,
Manitoba Natural Resources; Manitoba Naturalists
Society and Manitoba Chapter, The Wildlife Society.
R. Riewe, C. Jay and M. Shoesmith reviewed earlier
versions of this manuscript and N. Caulkett gave
constructive input and suggestions.
Literature Cited
Aleksiuk, M., and I. Me T. Cowan. 1969a. Aspects of
seasonal energy expenditure in the Beaver (Castor
canadensis Kuhl) at the northern limit of its distribution.
Canadian Journal of Zoology 47: 471-481.
Aleksiuk, M., and I. Mc T. Cowan. 1969b. The winter
metabolic depression in arctic beavers (Castor canaden-
sis Kuhl) with comparisons to California beavers.
Canadian Journal of Zoology 47: 965-979.
Busher, P. E. 1975. Movements and activities of Beavers,
Castor canadensis, on Sagehen Creek, California. M.A.
thesis, San Francisco State University, California. 86
pages.
Canadian Council on Animal Care. 1980. Guide to care
and use of experimental animals. Canadian Council on
Animal Care, Ottawa. 2 volumes.
Courcelles, R., and R. Nault. 1983. Beaver programs in
the James Bay area, Quebec, Canada. Acta Zoologica
Fennica 174: 129-131.
Davis, J. R. 1984. Movement and behavior patterns of
beaver in the Piedmont of South Carolina. M.Sc. thesis,
Clemson University, Clemson, South Carolina.
606
Davis, J. R., A. F. Von Recum, D. D. Smith, and D. C.
Guynn, Jr. 1984. Implantable telemetry in beaver.
Wildlife Society Bulletin 12: 322-324.
Djoshkin, W. W., and W. C. Safonow. 1972. Die Biber
der alten und neuen Welt. A. Ziemsen Verlag,
Wittenberg Lutherstadt, East Germany.
Dyck, A. P. 1991. Bioenergetics and thermoregulation of
beavers (Castor canadensis). M.Sc. thesis, University of
Manitoba, Winnipeg.
Farquharson, E. L. 1962. Textbook of operative surgery.
E and S. Livingstone, Edinburgh. 949 pages.
Flecknell, P. A. 1994. Refinement of animal use - assess-
ment and alleviation of pain and distress. Laboratory
Animals 28: 222-231.
Greene, S. A., R. D. Keegan, L. V. Gallagher, J. E.
Alexander, and J. Harari. 1991. Cardiovascular
effects of halothane anesthesia after diazepam and
ketamine administration in beavers (Castor canadensis)
during spontaneous or controlled ventilation. American
Journal of Veterinary Research 52: 665-668.
Guynn, D. C., Jr., J. R. Davis, and A. F. Von Recum.
1987. Pathological potential of intraperitoneal transmit-
ter implants in beavers. Journal of Wildlife Management
51: 605-606.
Jalanka, H. H., and B. O. Roeken. 1990. The use of
Medetomidine, Medetomidine- Ketamine combinations,
and Atipamezole in nondomestic mammals: a review.
Journal of Zoo and Wildlife Medicine 21: 259-282.
Kenward, R. 1987. Wildlife radio tagging: Equipment,
field techniques and data analysis. Academic Press,
London. 222 pages.
THE CANADIAN FIELD-NATURALIST
Vol. 111
Lancia, R. A. 1979. Year-long activity patterns of radio-
marked beaver (Castor canadensis). Ph.D. dissertation,
University of Massachusetts, Amherst, Massachusetts.
Liles, J. H., and P. A. Flecknell. 1993. The influence of
buprenorphine or bupivacaine on the post-operative
effects of laparotomy and bile-duct ligation in rats.
Laboratory Animals 27: 374-380.
Miller, D. R. 1964. Colored plastic ear markers for
beavers. Journal of Wildlife Management 28: 859-861.
Wheatley, M. 1989. Ecology of beaver (Castor canaden-
sis) in the Taiga of southeastern Manitoba. M.Sc. thesis,
University of Manitoba, Winnipeg, Manitoba. 167 pages.
Wheatley, M. 1994. Boreal beavers (Castor canadensis):
Territoriality, home range, food habits and genetics of a
mid-continent population. Ph.D. thesis, University of
Manitoba, Winnipeg, Manitoba. 349 pages.
Wheatley, M. 1997a. Beaver, Castor canadensis, home
range size and patterns of use in the taiga of southeastern
Manitoba: I. Seasonal variation. Canadian Field-
Naturalist 111: 204-210.
Wheatley, M. 1997b. Beaver, Castor canadensis, home
range size and patterns of use in the taiga of southeastern
Manitoba: II. Sex, age and family status. Canadian
Field-Naturalist 111: 211-216.
Wheatley, M. 1997c. Beaver, Castor canadensis, home
range size and patterns of use in the taiga of southeastern
Manitoba: III. Habitat variation. Canadian Field-
Naturalist 111: 217-222.
Received 25 November 1996
Accepted 10 February 1997
The Condition and Trend of Aspen, Populus tremuloides, in Kootenay
and Yoho National Parks: Implications for Ecological Integrity
CHARLES E. KAY
Department of Political Science, Utah State University, Logan, Utah 84322-0725
Kay, Charles E. 1997. The condition and trend of aspen, Populus tremuloides, in Kootenay and Yoho National Parks:
Implications for ecological integrity. Canadian Field-Naturalist 111(4): 607-616.
Aspen (Populus tremuloides) communities were measured in and near Yoho and Kootenay National Parks to determine
condition and trend. Most aspen stands were heavily invaded by conifers as they had not burned in 60 years or more due to
modern fire suppression and the elimination of aboriginal burning. Aspen is also declining due to repeated ungulate brows-
ing, primarily by Elk (Cervus elaphus). Even where disturbed by logging and burning outside the parks, many aspen stands
failed to produce new stems greater than 2 m tall because all the suckers were repeatedly browsed. Only where ungulate
numbers were low was aspen able to successfully regenerate. Aspen, though, is not “seral,” as that term is commonly used
because the species seldom grows from seed due to its demanding seed bed requirements. This, coupled with high biodi-
versity, makes aspen an excellent indicator of ecological integrity as mandated by Parks Canada statute.
Key Words: Aspen, Populus tremuloides, Kootenay National Park, Yoho National Park, ecological integrity, Elk, Cervus
elaphus.
Aspen (Populus tremuloides) is an excellent indi-
cator of ecological integrity because the species sel-
dom grows from seed due to its demanding seed bed
requirements (Perala 1990). In fact, there are no
known instances of aspen clones having established
from seed anywhere in the southern Canadian
Rockies or in the Intermountain Western U.S. during
the period of recorded history (Kay 1993). It is
thought that environmental conditions have not been
conducive to seedling growth and clonal establish-
ment since shortly after the glaciers retreated 10 000
or more years ago (McDonough 1979, 1985; Perala
1990; Jelinski and Cheliak 1992; Mitton and Grant
1996). This means that the aspen clones found
throughout the southern Canadian Rockies today
have likely maintained their presence on those sites
for thousands of years via vegetative regeneration.
Thus, aspen may be among the oldest living organ-
isms on Earth (Mitton and Grant 1996). In fact,
Peterson et al. (1995*:14—17) classified aspen as
old-growth ancient forest.
Aspen seedlings are more common in the northern
Canadian Rockies (Peterson and Peterson 1992,
1995) and there may be “windows of opportunity”
that allow seedling establishment at infrequent, 200
to 400 year or longer, intervals (Jelinski and Cheliak
1992: 728), but successful sexual reproduction of
aspen is still exceedingly rare (Mitton and Grant
1996). Aspen trees invariably occur as clones in
which all the individual trees (ramets) are genetical-
ly identical, having grown from a common root sys-
*See Documents Cited section following Acknowledgments.
tem by vegetative shoots (Shepperd and Smith
1993). If aspen clones are lost due to forest succes-
sion or other factors, there are no known means of
reestablishment (Kay et al. 1994). As a relatively
short-lived tree (< 150 years), long-lived aspen
clones are often dependent on periodic disturbance
such as fire to stimulate vegetative regeneration via
root suckering, and to reduce conifer competition
(Bartos and Mueggler 1979, 1981; Bartos et al.
1991, 1994; Shepperd 1993; Shepperd and Smith
1993). In the absence of fire or other disturbance,
most aspen clones in the southern Canadian Rockies
will eventually be replaced by more shade-tolerant
species. Thus, the condition and trend of aspen pro-
vides information not only on an area’s fire history,
but also addresses the question of whether past fire
suppression practices have had a significant impact
on park resources (Walker and Irons 1993). While
70 or more years of active fire suppression may not
have had a detectable effect on coniferous species
(Masters 1990), fire exclusion may have had a
greater impact on aspen communities (DeByle et al.
1987; Bunnell 1995).
In addition, aspen provides highly palatable for-
age for Elk (Cervus elaphus) and other ungulates in
the Canadian Rockies (Nelson and Leege 1982; Poll
et al. 1984; Timmermann 1991). Aspen is, however,
sensitive to repeated browsing. High-density Elk
populations commonly strip bark from mature aspen
and severely browse aspen suckers, thus preventing
stand regeneration which may eventually lead to the
607
608
loss of aspen clones (Krebill 1972; Olmsted 1977,
1979; Weinstein 1979; DeByle 1985; Kay 1990;
Shepperd and Fairweather 1994). Unlike herbaceous
plants, the long-term grazing and fire histories of
aspen communities can also be judged from histori-
cal and repeat photographs (Kay and Wagner 1994).
Moreover, aspen communities support an array of
other species and may have the highest biodiversity
of any forest type in the Canadian Rockies (DeByle
and Winokur 1985; Peterson and Peterson 1992,
1995*; Stelfox 1995). Bird communities, for
instance, vary with the size, age, and location of
aspen clones, as well as with grazing intensity and
history (Young 1973, 1977; Flack 1976; Winternitz
1980; Daily et al. 1993; Johns 1993; Westworth and
Telfer 1993; Pojar 1995; Stelfox 1995). If aspen is
lost, many birds and small mammals will decline;
some precipitously (Ehrlich and Daily 1993).
In Yoho National Park (established 1886) and
Kootenay (established 1920), aspen is common on
lower-elevation montane slopes with southerly or
westerly aspects (Kuchar 1978*; Achuff et al. 1984)
— areas that are also rated as prime winter habitat for
Elk and other ungulates (Poll et al. 1984; Van
Egmond 1990). To the east in Banff National Park,
high ungulate populations are believed to be having a
negative effect on that park’s aspen communities
(Cowan 1944*, 1947, 1950; Flook 1964, 1970; Kay
and White 1995). In U.S. national parks, such as
Yellowstone and Rocky Mountain, Elk have had a
major impact on aspen, often eliminating the species
from many areas (Olmstead 1977, 1979; Kay 1985,
1990; Hess 1993; Kay and Wagner 1996). Aspen,
however, has not previously been studied in
Kootenay or Yoho. Moreover, none of the montane
areas in either park have burned in more than 60
years due to active fire suppression programs (Kay
and White 1995), which raises the prospect that Yoho
and Kootenay’s aspen communities may be declining
due to advancing forest succession. To address these
and other questions, I measured and surveyed aspen
in and near Kootenay and Yoho National Parks.
Methods
I conducted a systematic survey of aspen commu-
nities in Kootenay and Yoho National Parks during
September 1994 and September 1995. Both parks are
situated immediately west of the continental divide
in British Columbia between 50° 30’ — 51°40’ N and
116° 10’ — 117° 15’ W. Due to both park’s thick
coniferous forests, steep terrain, and limited budget
it was not possible to undertake a random survey of
aspen stands. Instead, sampling was confined to
established trails, old fire-roads, and along park
highways. Each trail or road was first driven or
walked and all aspen stands plotted on 1:50 000
topographic maps. Then a representative number of
stands was selected for detailed measurement.
THE CANADIAN FIELD-NATURALIST
Vol. 111
At each aspen community that was sampled dur-
ing this study, a 2x30 m belt transect was placed
perpendicular to the slope in the stand’s center. To
facilitate data recording, I subdivided each 30 m
transect into 3 m segments and then recorded the
number of live aspen stems by size classes within
each 3 m segment. The following size classes were
used: (1) stems less than 2 m tall, (2) stems greater
than 2 m tall but less than 5 cm diameter at breast
height (DBH), (3) stems between 6 and 10 cm DBH,
(4) stems between 11 and 20 cm DBH, and (5) stems
greater than 21 cm DBH. Ages of aspen within each
size class were determined by counting annual rings.
The ages of large aspen were obtained with the aid
of an increment bore while smaller stems were cross-
sectioned, usually those less than 5 cm DBH. Stems
less than 2 m tall were not aged.
Within each stand, the following information was
also recorded: (1) elevation as determined from
topographic maps; (2) Universal Transverse
Mercator (UTM) grid coordinates, again estimated
from topographic maps; (3) aspect; (4) estimated
slope in percent; (5) estimated stand size in meters;
(6) bark damage — percent of stems that exhibited
old black-scar, ungulate bark damage and the percent
of stems with new or recent bark damage — wounds
that had not yet healed over with black-colored bark,
usually less than two years old (Krebill 1972); (7) an
estimate of the mean percent of each stem that had
been damaged by ungulate bark stripping; (8) if the
stand had newly regenerated stems greater than 2 m
tall but less than 5 cm DBH, an estimate of the per-
cent that showed evidence of ungulate highlining —
where the ungulates browse off all the lower branch-
es as high as the animals can reach, usually 2 m; and
(9) the percent of stems less than 2 m tall that exhib-
ited ungulate browsing.
Items 6 to 9 provided an estimation of past ungu-
late use. Only Elk or Moose (Alces alces) strip-off
and eat the bark of aspen, and bark damage usually
occurs during winter when other foods are in short
supply (Krebill 1972). Neither Mule (Odocoileus
hemionus) nor White-tailed deer (O. virginianus)
strip aspen bark, but both species of deer, as well as
Elk and Moose browse aspen. Since at least 1940,
however, Elk have dominated the ungulate commu-
nities in Kootenay and Yoho, especially during win-
ter (Poll et al. 1984; Van Egmond 1990). In areas
with high Elk populations, bark stripping can be so
severe that the lower 2 m of aspen trunks are black
instead of their normal white coloration (Kay 1990;
Kay and Wagner 1994).
In addition, at each stand the number and species
of conifers was recorded on the 2 x 30 m belt transect
that was used to count aspen stems. Conifers were
recorded by the same five size classes that were used
for aspen. Total percent conifer canopy cover in each
stand was also estimated according to guidelines
1997
established by Mueggler (1988). Finally, understory
species canopy cover was estimated for each stand
but those data are not reported here (see Kay 1996*).
This was part of a larger project to assess long-
term ecosystem states and processes in the southern
Canadian Rockies (Kay et al. 1994; Kay and White
1995), but here I only report the results of my aspen
research in and around Yoho and Kootenay. Aspen
outside Yoho and Kootenay was included because
other studies have found a marked inside-outside
park difference in aspen community dynamics due to
differences in ungulate use. Inside Yellowstone
National Park, for instance, the area occupied by
aspen has declined approximately 95% since park
establishment in 1872 and even burned aspen stands
have failed to successfully regenerate due to repeat-
ed ungulate browsing (Romme et al. 1995; Kay and
Wagner 1996). Outside that park, however, where
hunting limits Elk numbers, aspen stands have suc-
cessfully regenerated without fire or other distur-
bance and display characteristics of climax commu-
nities (Kay 1985, 1990). The same is true in
Colorado’s Rocky Mountain National Park (Hess
1993). By measuring aspen stands in the same
drainage with similar histories of disturbance, but
with different histories of ungulate use, it is possible
to determine if climatic change, fire suppression, or
grazing is primarily responsible for any observed
differences in community structure (Kay 1990).
Kuchar (1978*), Achuff et al. (1984), Poll et al.
(1984), Van Egmond (1990), and Tymstra (1991)
provided information on vegetation, wildlife, and cli-
matic conditions in Kootenay and Yoho.
Results
A total of 269 aspen stands were measured in or
near Kootenay (n=168) and Yoho (n= 101)
National Parks. Most aspen stands were heavily
invaded by conifers, primarily White Spruce (Picea
glauca), Lodgepole Pine (Pinus contorta), or
Douglas Fir (Pseudotsuga menziesii); mean conifer
canopy cover = 38% (SEM = 2.6%). Ungulates have
also had a significant impact on these aspen commu-
nities. Only where ungulate use was low had aspen
stands been able to successfully regenerate —
defined as producing new stems more than 2 m tall.
In Kootenay National Park, aspen successfully
regenerated in the Columbia Valley where there are
few Elk, but not in the Kootenay Valley where most
of the park’s Elk winter. While in Yoho National
Park, except for a handful of stands, no aspen com-
munities successfully regenerated.
Even clear-cut stands were not able to successful-
ly regenerate in the Cross River drainage south of
Kootenay National Park. Although logging and asso-
ciated soil disturbance increased sucker densities 60
fold (mean = 291 stems/ha unlogged vs. 17 337
stems/ha logged), aspen height growth was limited
KAY: THE CONDITION AND TREND OF ASPEN
609
Y ~
& 20,000 1,250
ne) c
= 16,295 1,042 @
o” 1,000 O
: 15,000
oe S
= 750 £
2 >
% 10,000 7
5 c
a 500 &
8 5,000 =
Say: 3,334 250 ©
es 1,500 ”
a 0 Bone none a