UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relationship of small mammal populations to uniform even-aged shelterwood systems Trebra, Charlotte Dorothea von 1994

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

Item Metadata


831-ubc_1994-0647.pdf [ 1.69MB ]
JSON: 831-1.0075258.json
JSON-LD: 831-1.0075258-ld.json
RDF/XML (Pretty): 831-1.0075258-rdf.xml
RDF/JSON: 831-1.0075258-rdf.json
Turtle: 831-1.0075258-turtle.txt
N-Triples: 831-1.0075258-rdf-ntriples.txt
Original Record: 831-1.0075258-source.json
Full Text

Full Text

RELATIONSHIP OF SMALL MAMMAL POPULATIONSTO UNIFORM EVEN-AGED SHELTERWOOD SYSTEMSbyCHARLOrI’E DOROTHEA VON TREBRAB.Sc.(Agr.), The University of British Columbia, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Forestry)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAAugust 1994Charlotte Dorothea von Trebra, 1994In presenting this thesis in partial fulfillment of therequirements for an advanced degree at the University of BritishColumbia, I agree that the Library shall make it freely availablefor reference and study. I further agree that permission forextensive copying of this thesis for scholarly purposes may begranted by the head of my department or by his or herrepresentatives. It is understood that copying or publication ofthis thesis for financial gain shall not be allowed without mywritten permission.(SignatureDepartment of____________________The University of British ColumbiaVancouver, CanadaDate__________ABSTRACTThis study was designed to test the hypothesis that a shelterwood silvicultural systemwould reduce small mammal population levels in 30% and 50% basal area removal stands. Thebenefit of an overstory canopy, or shelterwood, may help reduce the frequent frost problemsobserved in interior Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. glauca) plantations,and improve the success of natural regeneration, if seed predation does not threaten the survivalof the seed supply. Small mammal population levels, seed fall, seed germination, and seedpredation were monitored in order to determine the small mammal population dynamicsassociated with seed predation in the central interior of British Columbia, Canada. Deer mice(Peromyscus maniculatus) and red-backed voles (Clethrionomys gapperi) were the most commonsmall mammal species sampled prior to harvesting, and both species continued to be dominantin the first and second year post-harvest. There was no negative effect on small mammalpopulations due to the different intensities of basal area removal. Seed fall crops were recordedthe first and second year post-harvest; the seed crop in the second year post-harvest wassubstantially larger than the first year post-harvest on treatments and controls. Seed predationvaried between l.4x105 and 3.0x105 seeds/ha/month, and seemed to fluctuate with the seasonalsmall mammal population levels (lower in spring and higher in fall) but not with the amount ofseed crop available. Germination occurred in 62% to 63 % of the seeds that survivedoverwinter, on the seedbeds created during harvesting. Thus, the regeneration success of thestands, with regard to seed survival and establishment, in 30% and 50% basal area removalstands did not seem to be inhibited by the small mammal communities observed on the sites.11TABLE OF CONTENTSABSTRACTLIST OF TABLESLIST OF FIGURES11VviiLIST OF APPENDICES viiiACKNOWLEDGEMENTS ix5Study areasExperimental designSmall mammal population samplingDemographic parametersDensitySurvivalBody mass and sexual maturityReproductionSmall mammal diversitySeed parametersStatistical analysis56899111212131416RESULTS 18TrappabilityPopulation densityAnimal survival .RecruitmentReproduction . .Body massDiversitySeed fallSeed germinationSeed predation. .20222729333838454851INTRODUCTIONObjectivesMETHODS15111TABLE OF CONTENTS (cont.)DISCUSSION 54Experimental design 54Demographic responses of small mammals 56Diversity 61Seed Parameters 62MANAGEMENT IMPLICATIONS 64REFERENCES 66ivLIST OF TABLESTABLE 1. Breakdown of time periods for analysis 10TABLE 2. Total captures of individuals of less common microtines, chipmunks andshrews in the 1991, 1992, and 1993 trapping sessions 19TABLE 3. Jolly trappability estimates for deer mice, red-backed voles, and less commonmicrotine populations. Values shown are the average percent trappability(number caught/Jolly-Seber population estimate) ± 95% confidence limits 21TABLE 4. Average Jolly-Seber population density estimates of deer mice and red-backedvoles. Values shown are mean J-S estimates for the time periods indicated.Overall time period is the entire post-harvest period of study (September 1991 -October 1993). The number of sampling sessions (n) varied between blocks preand post-harvest 1991 23TABLE 5. Mean monthly Jolly survival of deer mice and red-backed voles. Valuesrepresent mean estimates ± 95% confidence limits of different samplingperiods 28TABLE 6. Cumulative number of deer mouse and red-backed vole recruits per ha duringpre-treatment and post-treatment periods of study (199 1-1993) 32TABLE 7. Mean body mass (in grams) at sexual maturity for male and female deer miceand red-backed voles. Averages were taken from each of the nine grids and arebased on the entire period of study 34TABLE 8. Length of breeding season in weeks. The start of the breeding season is thetime of capture of the first scrotal male or lactating female. The end of thebreeding is the week in which the last lactating female was recorded 34TABLE 9. Numbers and proportions of adult deer mice in breeding condition on threereplicates throughout the study period (n=number of individuals sampled) 35TABLE 10. Numbers and proportions of adult red-backed voles in breeding condition onthree replicates throughout the study period (n=number of individuals in thesample) 36vLIST OF TABLES (cont.)TABLE 11. Treatment effects on the percentage of breeding adults in the samplepopulations. Anova F and P values are shown for treatment effects withintrapping years. Treatments were pooled across the replicate blocks. Degrees offreedom=2. P=0.05 37TABLE 12. Mean body mass (in grams) of male adult deer mice in treatment and controlstands throughout the study period (n = number of individuals recorded during theseparate time periods) 39TABLE 13. Mean body mass (in grams) of male adult red-backed voles in treatment andcontrol stands throughout the period of study (n=number of individuals recordedduring the separate time periods) 40TABLE 14. Mean Simpson’s and Shannon-Wiener’s indices of diversity and evenness forvarious time periods over three consecutive trapping years 43TABLE 15. Species richness, and mean Simpson’s and Shannon-Wiener diversity andevenness values pooled for three replicates for the overall study period 44TABLE 16. Mean number of Douglas-fir seeds/ha and number of sound Douglas-firseeds/ha. Seed fall samples were collected in the first and second year postharvest (1992 and 1993, respectively) 47TABLE 17. Survival and germination results of Douglas-fir germination trials. Shownare the mean values of three replicated unlogged controls, 30% and 50% basalarea (BA) removal treatments 48viLIST OF FIGURESFIGuRE 1. Jolly-Seber estimate of deer mouse population size on unlogged controls,30% and 50% basal area removal stands. Vertical solid line indicates the timeof harvesting 24FIGuRE 2. Jolly-Seber estimate of red-backed vole population size on unlogged controls,30% and 50% basal area removal stands. Vertical solid line indicates the timeof shelterwood harvesting 25FIGuRE 3. Cumulative post-harvest (199 1-1993) number of deer mouse and red-backedvole recruits, as observed over three replicates: unlogged controls, 30% basalarea removal shelterwood treatments, and 50% basal area removal shelterwoodtreatments 30FIGuRE 4. Cumulative number of deer mouse and red-backed vole recruits throughoutvarious periods of study: pre-treatment 1991, post-treatment 1991, summer 1992,summer 1993. Values represent means of pooled treatments and controls 31FIGuRE 5. Simpson’s and Shannon-Wiener indices of small mammal species diversityfor different time periods throughout the study. Values represent means ofpooled treatments and controls 42FIGuRE 6. Total number of Douglas-fir and white spruce seeds collected per ha.Distribution of seed rain shown for unlogged controls, 30% and 50% removaltreatments in the first and second year post-treatment. Samples collected in fallof 1992 and 1993 46FIGuRE 7. Fates of the Douglas-fir seeds which were recovered in the germinationtrials. Seeds were sown in fall 1992 and results were recorded in June 1993.Values shown are the mean values of three pooled replicates 50FIGuRE 8. Douglas-fir seed predation per ha per month on the controls and treatments.Values shown are the number of seeds consumed per month in October, 1992,overwinter 1992/93 (01w) and May to October, 1993. Treatment means arepooled across the three replicates 52viiLIST OF APPENDICESAPPENDIx 1. Comparison of Jolly-Seber estimates and MNA direct enumeration of deermouse population levels in the control and treated stands of the UBC replicate. . . 71APPENDIx 2. Comparison of Jolly-Seber estimates and MNA direct enumeration of deermouse population levels in the control and treated stands of the Gavinreplicate 72APPENDIx 3. Comparison of Jolly-Seber estimates and MNA direct enumeration of red-backed vole population levels in the control and treated stands of the UBCreplicate 73viiiACKNOWLEDGEMENTSI thank my supervisor, Dr. Tom Sullivan, for his support, guidance, constantencouragement and optimism. I am most grateful for funding and logistical support from theSilviculture Systems Working Group, Silviculture Branch, Ministry of Forests, Victoria, theForest Sciences Section, Cariboo Region, Weidwood of Canada (Williams Lake), and the UBCAlex Fraser Research Forest. Many thanks to Mike Burwash, Tom Friese, Jim Kurta, PontusLindgren, Markus Merkens, Gord Ryznar, Sandra Sulyma, and Mike and Jenny Tudor for theirassistance with the field work and the good times we had getting the job done. I thank ChrisKohier, Doug Ransome, and especially Bruce Runciman for their informative discussions andFrecirik von Euler for his statistical advice. I am grateful for the patience and support of myfamily during my studies. Special thanks go to Jim Kurta for more than I can mention; all hishelp and support was greatly appreciated during this long haul.ix1INTRODUCTIONDifficulties exist in establishing regeneration of interior Douglas-fir (Pseudotsugamenziesii (Mirb.) Franco var. glauca) in the central interior region of British Columbia, Canada.Poor survival of planted stock in open clearcuts and increasing public pressure for the integrationof non-timber resources in forest management practices has illustrated the need for an alternativesilvicultural system on these sites. Stands containing both Douglas-fir and lodgepole pine (Pinuscontorta Dougl. var. latfolia) are especially suited to shelterwood cutting systems (B.C.Ministry of Forests 1991). In a shelterwood system the mature stand is removed with aminimum of two cuts in order to cull the diseased and damaged trees from the stand, andencourage regeneration of the desired species under the protection of a partial canopy. Themature crop is removed in one or more harvests as the regeneration becomes established andrequires more growing space. The benefit of an overstory canopy, or shelterwood, may helpreduce the frequent frost problems observed in interior Douglas-fir plantations, and improve thesuccess of natural regeneration while retaining the advanced fir regeneration that is usuallypresent (Sutherland 1990).Within the last 20 - 40 years, the sole reliance on the use of natural regeneration inconifer stands has all but disappeared as planting has generally resulted in more rapid rates ofregeneration (Moore 1940; Owston et al. 1992; West 1992). The potential delays or failuresof the crop trees to establish were too great a risk. However, due to the rising costs of artificialregeneration, the problem of establishing Douglas-fir plantations in clearcuts in the interiorregion, and public pressure to reduce the amount of clearcut harvesting, there has been a2renewed interest in the natural regeneration which can occur in a shelterwood system. Althoughone could not expect adequate regeneration from natural seed sources for most conifer speciesin a given year (West 1992), this does not necessarily hold true for a shelterwood system.Williamson (1973) observed coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var.menziesii) seedlings of all possible ages four years after an initial shelterwood cut, although noappreciable cone crops were predicted during that time. Apparently some regeneration hadoccurred during years of predicted low seed fall, demonstrating that it is possible for cone andseed production of the remaining trees to be stimulated by a shelterwood cut.A shelterwood system produces a structurally diverse stand which allows numerous floraand fauna species to coexist. The opening of the canopy combined with the partial cover of theretained mature crop enhances the vertical structure as well as the species composition withina stand. By not prescribing burns after harvest, the coarse woody debris can accumulate andprovide cover for small mammal species. In addition to the importance of these small mammalspecies as a prey base for raptors, furbearers and larger carnivores, these small mammals maysignificantly help as well as hinder establishment of natural regeneration (Owston et a!. 1992).Most higher plants have evolved with an obligatory symbiotic relationship with root-inhabiting(mycorrhizal) fungi (Maser et al. 1978a). Whereas epigeous mycorrhizal spores are prolific innumber and are dispersed primarily by air currents, hypogeous mycorrhizal fungi are dependentupon small mammals such as deer mice (Peromyscus maniculatus), voles (Microtus spp.), shrews(Sorex spp.) and chipmunks (Eutamias spp.), all avid mycophagists, as primary vectors of sporedissemination (Maser et al. 1978b). The behaviour patterns and habitats used by the differentmycophagists determine the extent of spore dispersal: those small mammals which are more3ubiquitous or edge-dwelling may provide mycorrhizal inoculum from forested areas to adjacentnonforested areas. However, harvesting and silvicultural practices may profoundly change smallmammal population levels and thereby contribute to their destructive potential in youngplantations.A community of these small mammals exist in established forests and the various specieshave diverse feeding and habitat preferences. While some species may benefit from, or adaptto harvesting disturbances, the potential impact on more specialized species such as the southernred-backed vole (Clethrionomys gapperi) is unknown. Red-backed voles are consideredinhabitants of mature forest, and their density may drop precipitously within 1 or 2 yearsfollowing logging or forest fire disturbance (West et al. 1980). Red-backed voles disappearedfrom a small mammal community following a diameter-limit logging trial (Medin 1986), whereasa study on selective cutting (Medin and Booth 1989) found that a 29% volume removal had nosignificant effect on red-backed voles. Corn et al. (1988) compared coastal Douglas-fir forestsof varying age classes (less than 10 years to 450 years of age), and found a positive correlationof red-backed voles with density and basal area of live trees. Red-backed voles were virtuallyabsent from clearcuts whereas deer mice were abundant (Corn et al. 1988). Deer mice areknown to easily survive harvest operations, and hence are able to rapidly repopulate harvestedareas (West 1992). Some studies show that deer mice abound during the early stages of forestsuccession, particularly when ground cover is disturbed during harvest (Lawrence et al. 1961;West 1992), while other findings conclude that numbers of deer mice decrease on logged plotsversus unlogged controls (Medin and Booth 1989). In either case, it is necessary to determinehow management actions will affect the species capable of damaging regenerating conifers in4order to minimize the risk of seed predation.Natural regeneration under the partial canopy of seed trees depends upon seed production,survival of the fallen seed until the following growing season, and germination. Both before andafter seeds fall from the trees they are likely to be eaten by insects, birds and rodents (Smith1986). Deer mice, in particular, are primarily granivorous and highly efficient at finding andconsuming large quantities of conifer seed (Moore 1940; Smith and Shaler 1947; Sullivan 1979a;Hawthorne 1980; Sullivan and Sullivan 1982); some voles and insectivorous mammals such asshrews (Sorex spp.) also consume seeds but are only secondary granivores (Sullivan and Sullivan1982; West 1992). Seed predation by deer mice and other rodents has contributed to the failurein regenerating cutover forest lands as well as overgrazed range lands (Sullivan 1979b). Seedpredation is most severe in years of low seed production; however, animals are not completelyefficient at locating seed and are simply incapable of destroying all seed in a good year (West1992). Emphasis must be put on harvesting strategies that reduce animal damage or seedpredation to tolerable levels by allowing regeneration of trees, but discourage the buildup oflarge populations of potentially damaging species of wildlife (Emmingham et al. 1992). To date,there has been little research on the long-term effects of harvesting methods, other thanclearcutting, on small mammal populations.In order to assess the effects of shelterwood harvesting on small mammal species, it isnecessary to monitor small mammal population dynamics along with natural seed fall, seedpredation, and seed germination in different harvesting treatments to provide some rationale forregeneration, or lack of it, in the years after initial harvest. This study focuses on whether analternative silvicultural system such as a shelterwood harvest can alleviate frost problems5associated with Douglas-fir regeneration as well as minimize small mammal predation on thenatural seed supply.ObjectivesThe objectives of this study were to: (1) test the hypothesis that a shelterwoodsilvicultural system would reduce small mammal population levels in the treated stands; (2)measure natural seed production, seed predation by small mammals and seed germination underuniform, even-aged shelterwood harvesting systems; (3) measure small mammal populationdynamics associated with seed predation; (4) use this information to help understand theregeneration success or failure of the various harvesting treatments.METHODSStudy areasThree study aieas were established within the Cariboo Forest Region, northeast ofWilliams Lake, British Columbia. The Beaver Valley (Gavin site, 52°29’N; 122°37’W) and theBeedy Creek (Skelton site, 52°33’N; 122°04’W) site were located in the Williams Lake ForestDistrict, and the University of British Columbia’s Alex Fraser Research Forest (UBC site,52°29’N; 122°39’W) was located in the Horsefly Forest District. All sites are in the mesicfalsebox (Paxistima myrsinites)-sarsaparilla (Aralia nudicaulis) ecosystem unit of the dry and6warm Sub-boreal Spruce biogeoclimatic subzone (SBSdw1) (Meidinger and Pojar 1991). Theforest cover was primarily mature (107-126 year old) Douglas-fir/lodgepole pine stands withsome white spruce (Picea glauca (Moench) Voss) interspersed. More spruce was present on theUBC site than on the other two sites, indicating a slightly wetter moisture regime (permesic) inthese treatment units (BCMOF 1993a). Topography in the area is flat to gently rolling withelevation ranging from 800 m (Skelton site) to 1050 m (UBC and Gavin sites). The study sitesvaried in size between 20 and 30 ha. This size accommodated treatment blocks as well astreated buffers and surrounds, and reflected operational cutbiock size. In addition, this scale ofshelterwood treatments may have been more representative of the potential impact on the mobilesmall mammal species present, and reflected the effect of the disturbances on these populationlevels more realistically. Generalist small mammal species such as the deer mouse would notbe able to modify their behaviour or home ranges to adjust to the treatments as readily as theymay be able to when study sizes are less than their home range size.Experimental designThe experimental design was a randomized block design with three replicate blocks.Three treatments were interspersed randomly within each block to compare two levels of basalarea (BA) removal (30% and 50%) with an unlogged control. The treatment replicates werespatially segregated, thereby ensuring statistical independence (Huribert 1984). Treatments were1.4 ha in size with 20 m buffers. The buffers were treated with 30% BA removal between eachtreatment in order to minimize any treatment interaction. The location of each of the three7blocks was determined from aerial photographs of the sites.Block locations were selected according to:i) uniformity of ecosystems, topography, and tree species distributionii) low windthrow hazardiii) available access and locationThe target tree species distribution was between 60/40% to 40/60% Douglas-fir/lodgepolepine. Lodgepole pine was selected over Douglas-fir for removal, as Douglas-fir was thepreferred species to regenerate. The southwest corners of each block were permanently markedon a Douglas-fir tree. This acted as a tie point for all corners and block centres. Mechanicalfalling with a feller-buncher and large skidders was used to create wide skid trails which wouldmimic current industrial operations. Trees were not marked prior to cutting, but the loggingcontractors were shown how to use prisms to select stems to meet the 30% and 50% basal arearemoval criteria. The shelterwood cuts were completed between July and early September,1991. The 30% BA removal treatments were considered to be preparatory cuts (Smith 1986),removing most of the trees from the lower crown classes, whereas the 50% removal treatmentswere considered to be seed cuts and removed trees in the overtopped and intermediate crownclasses along with all or part of the codominants (BCMOF 1993a). Pre-harvest BA/ha in thestands ranged from 58.2 to 64.1 m2/ha (490 to 653 m/ha). Residual BA’s in the 30% removaltreatments ranged from 37.9 to 46.2 m2/ha (369 to 495 m3/ha) and 31.8 to 33.1 m2/ha (293 to356 m3/ha) in the 50% removal stands.8Sniall mammal population samplingSmall mammals were captured in Longworth live-traps baited with oats and a slice ofcarrot. Raw cotton was placed in each trap for bedding. A 1-ha (6 x 8 trap stations)checkerboard sampling grid was established on each treatment area on each block, with a live-trap at each of the 48 trap-stations. Trap-stations were set at 14.3-rn intervals and pre-baitedwith oats at least one week prior to initial sampling each year. Pre-baiting encouragedfamiliarization with the traps and helped maximize trappability of the species present. Pretreatment small mammal populations were monitored from June to August, 1991, to ensure thatinitial base-line populations were similar between treatments within replicate blocks. Post-harvest sampling continued at 3-week intervals from September to October, 1991, May toOctober, 1992, and May to October, 1993, on each of the treatments.During trapping periods, traps were set in the afternoon of day 1, checked in the morningand afternoon of day 2, and checked and locked open in the morning of day 3. They remainedlocked open and accessible during the weeks between trapping periods. Animals captured weremarked with serially-numbered, metal ear-tags, with each series being unique to a treatmentarea. Upon capture, species identification, body mass (to ± 0.5 g on a Pesola spring balance),sex, breeding condition, and location of capture were recorded for each animal. Thereproductive condition for males was noted according to palpation of testes, if not obviouslyscrotal or abdominal. Females were considered to be in breeding condition if they wereobviously pregnant (high body mass and a distended lower abdomen), lactating (verified throughpalpation), or had developed nipples and mammae showing signs of nursing, such as matted fur9(Krebs et al. 1969). All animals were released at point of capture immediately following datacollection. These live-traps sampled deer mice, voles, yellow-pine chipmunks (Eutamiasamoenus), shrews and other less common small mammal species.Demographic parametersTo assess the effects of the different intensities of overstory removal, various populationparameters were monitored. Population density (number of animals/ha), body mass,recruitment, survival, reproduction, and diversity were estimated for the two most common(numerically dominant) small mammal species, deer mice and red-backed voles, and for the lesscommon species such as meadow voles (Microtus pennsylvanicus) and heather voles(Phenacomys intermedius) when sufficient data were available. Due to the difficulty indifferentiating between the less common species in the field, the values obtained for the meadowand heather voles were combined for analysis.DensityThe accuracy of the population parameters of mark-recapture techniques is determinedby the probability of an individual animal’s capture, or trappability (Krebs and Boonstra 1984).Trappability was calculated as the number of marked individuals trapped/Jolly-Seber populationestimate (Seber 1982). Population density was estimated by the Jolly-Seber (J-S) model forreasons indicated by Jolly and Dickson (1983). J-S estimates may be unreliable or impossible10to calculate when sample sizes of the marked population are very small, despite incorporationof the small sample correction factor of Seber (1982) in the calculations (Krebs 1991).Therefore, direct enumeration of minimum number of animals known to be alive (MNA) (Krebs1966) was calculated in order to distinguish between actual population trends and the creationof artifacts, if any. Those J-S values which seemed biologically unreasonable were substitutedwith MNA counts before seasonal or treatment density averages were calculated. Small mammalpopulation analysis and seed results were compared for various time periods (Table 1). Datacollection was carried out from June 1991 to October 1993, during the snow-free months of theyear. Population density was estimated for pre-and-post-harvest 1991, summer 1992, summer1993, and the overall post-harvest density. Pre-treatment density estimates include only one ortwo sampling sessions. Therefore, the density analysis was restricted to post-harvest treatmentcomparisons rather than pre-and-post-treatment analysis.TABLE 1. Breakdown of time periods for analysisTime Period j DatePre-treatment 1991 June to August 1991Post-treatment 1991 September to October 1991Winter 1991 November 1991 to April 1992Summer 1992 May to September 1992Winter 1992 October 1992 to April 1993Summer 1993 May to October 1993Overall period of study September 1991 to October 1993(post-harvest)11SurvivalFurther demographic comparisons and statistical tests incorporated the J-S densityestimates for reasons indicated by Jolly and Dickson (1983), Nichols and Pollock (1983), andEfford (1992). Survival estimates indicate the disappearance of individuals from the markedpopulation, due to either emigration or mortality. Mean monthly survival was obtained by ageometric mean of the observed survival rates during various time periods (Krebs 1991).Thus, the survival rates were weighted according to the number of individuals recorded inthe designated time periods. Survival estimates may, at times, exceed unity due to thepeculiarity of the J-S density values. As in the comparisons of the density estimates,biologically unreasonable J-S survival estimates (>1.00) were compared with minimumsurvival estimates (based on MNA population enumeration values) for determining the lowerlimit of the survival.Animals which were able to establish themselves as residents of the sampledpopulations were defined as recruits. All animals captured for the first time were categorizedas newcomers; those animals which were captured for a minimum of two consecutivetrapping sessions (i.e. they were present in the area for a minimum of three to six weeks)were considered residents and classified as recruits (Kienner and Krebs 1991).12Body mass and sexual maturityThe mean adult body mass at which more than 50% of the animals captured weresexually mature was used to distinguish juvenile and adult age categories. Body mass atsexual maturity was derived from a breakdown of weight classes for each sex and eachspecies. Mean body masses from each grid were averaged to obtain the mass at sexualmaturity for deer mice and red-backed voles for the entire period of study. Spatial andtemporal comparisons of body mass were based on the mean mass of the resident adult malesrecorded on the grids, and were averaged over the three replicates to analyse treatmentdifferences. Only male body mass was used in the analysis to avoid complications ofundetected pregnancies in female data.ReproductionTwo measures were calculated to evaluate reproductive rate and condition: (i) lengthof the breeding season, and (ii) proportion of adults in breeding condition during the trappingseasons. Length of breeding season (in weeks) was estimated for 1992 and 1993 trappingyears only, as sampling sessions were concurrent across the replicates and sampling wascontinuous for the duration of the snow-free portion of these years only. The start of thebreeding season was defined as the time of capture of the first scrotal male or lactatingfemale and the end of breeding was the week in which the last lactating female was recorded.In the event that no lactating females were captured in the last few weeks of trapping, the13breeding season was assumed to be 3 weeks longer than the last recorded capture of apregnant female. Comparisons of the proportions of adult males and females in breedingcondition were based on 1992 and 1993 trapping sessions and averaged over the threereplicates for analysis. Trappability, population density, recruitment, survival rates, andreproduction estimates were calculated using Small Mammal Programs for Mark-RecaptureData Analysis (Krebs 1991).Small mammal diversitySpatio-temporal alpha (within ecosystem) diversity comparisons of small mammalspecies were analysed between treatments and controls in the SBSdw biogeoclimatic zone ofB.C. over the course of three trapping years. In this study, species diversity was measuredin terms of richness (number of species present) and evenness (the relative abundance of eachspecies in the community). The mathematical indices used, Simpson’s D index and theShannon-Wiener (S-W) H’ statistic, are measures of heterogeneity and incorporate richnessand evenness in their calculations (Magurran 1988). Conventionally, maximum evenness isconsidered to be the most diverse even though maximum evenness is not generally observedin natural systems (Kohler 1993). High evenness could be considered either less diverse, ifgreater differences among species is considered to be most important (Wood 1994), or morediverse because of less dominance by any one species. For the purpose of this study, thegreater the value of evenness, the greater the diversity was assumed to be.14Shannon-Wiener’s H’ statistic and Simpson’s D index are both based on proportionalabundances of species and may be subject to bias caused by fluctuations in abundance ofdominant or rare species. The Shannon-Wiener function is most sensitive to changes in theabundance of rare species in the community sample (Pielou 1966). The larger the value ofH’, the greater the uncertainty of correctly predicting the species of the next individualcollected in the sample and the greater the diversity. Simpson’s index is biased towardsdominance, and thereby shows greater sensitivity to changes in abundance of the mostdominant species in the sample. Simpson’s index (1-D) is related to the probability that twoindividuals picked at random belong to the same species and ranges from 0 (lowest diversity)to almost 1 (Simpson 1949). Comparisons of diversity were analysed using the mean valuesfor the 1992 and 1993 trapping periods and included only the species which were marked.Seed parametersSeed fall traps were placed at 10 randomly selected small mammal trap stationlocations on each of the treatments and controls to measure natural seed fall. The trapsconsisted of a 0.37 m2 frame box covered with 1-cm mesh hardware cloth and windowscreen on the bottom. This design permitted seeds to fall into the traps while keeping seedpredators out. Traps were placed onto the sites in late August 1992 and sampled in October1992 and 1993. Samples were dried and Douglas-fir and white spruce seeds separated fromthe other litter. Cutting tests were performed to identify sound seeds (those containing anembryo and thus potentially able to germinate). The number of sound seeds was represented15as a total value and as a percentage of seeds per hectare to use in further comparisons andcalculations.To determine the suitability of the microclimate for seedling growth under the partialshelterwood canopy, 10 rodent-proof exciosures were placed at the same locations as the seedfall traps on each block in October 1992. The exciosures were 0.37 m2 in area, and werecovered with 0.8-cm mesh hardware cloth which extended 5 cm into the ground on all sides.This exclosure also provided an indication of invertebrate seed predation. Ten stratifiedDouglas-fir seeds were scattered randomly on the seedbed within each exclosure (the litterlayer was not removed) and seed locations were marked with toothpicks. The total numberof seeds placed in the germination exclosures was 270 270 seeds/ha (10 seeds/0.37 m2 x 10000 m2/ha). Germination success was recorded in the following growing season (June 1993)as the percentage of seeds which germinated per ha.Ten 0.37-rn2 seed predation quadrats were installed in the vicinity of the seed fall andseed germination exciosures. Ten Douglas-fir seeds were distributed in each of the quadratsin mid-fall, 1992. Seed locations were marked with wooden toothpicks placed approximately0.5 cm from the seed. Quadrats were sampled every three weeks during the snow-freeportion of 1993 to observe whether or not the seeds had been removed from the plots oreaten. Remains of eaten Douglas-fir seeds were removed at each sampling session andreplaced with whole seeds. The predation rate was calculated as the rate of disappearance ofthe seeds in the quadrats per day. This method has been used successfully by Sullivan(1979a) to quantify predation rate and to identify seed predators. The predation rate perquadrat was extrapolated to the total number of seeds eaten or removed per hectare per16month. The seeds may have been eaten by seed-eating birds. This would have beenapparent by the nature of the seed destruction.Statistical analysisDespite the correlation between population measurements taken on the sameindividuals when animals were captured repeatedly in successive trapping sessions, theanalysis concentrated on the comparison of treatment means during certain time periods ofthe study. Thus, an analysis of variance (ANOVA) was performed on the means of varioustime periods to examine possible treatment effects. Pre-and-post-harvest 1991 samplingsessions were examined for biological significance but were not used in the statistical analysisbecause of the unbalanced sampling sessions between the blocks. The data collected in 1991were not collected in simultaneous sampling sessions, and the number of sampling sessionsvaried between the replicate treatments. The use of these data to represent replicatedtreatments and to test these results for significance would have ensued in temporalreplication. The statistical analysis was restricted to first and second year post-treatment(1992 and 1993) contrasts rather than pre-and-post treatment comparisons. An overall postharvest analysis of some parameters incorporated the final trapping week of 1991, and all ofthe 1992 and 1993 trapping sessions. For the deer mouse population density analysis, MNAvalues were substituted for one biologically infeasible I-S value in the UBC replicate(Appendix A) and for three J-S values in the Gavin replicate (Appendix B) prior to analysis.One MNA value was substituted for an unrealistic J-S density estimate of red-backed voles in17the UBC replicate (Appendix C).The balanced sampling design with three different harvesting intensities was analysedfor treatment differences with an analysis of variance (ANOVA). An ANOVA was carriedout on the mean values obtained from each grid for the first and second year post-treatmentin order to detect significant differences between treatments, blocks, and time periods.Arcsine transformations were performed on proportional data to fit the binomial proportionsto a normal distribution before proceeding with the analysis. When the results between theinteractions were significant, multiple contrasts using Bonferroni t-tests of differencesbetween means (Schlotzhauer and Littell 1987) were used to identify which of the meanswere significantly different from each other. Confidence intervals (95%) compareddifferences in percent trappability and percent survival of the various small mammalpopulations. Mean values and comparisons were based on three replicates for eachtreatment. For comparisons between treatments, the replicates were pooled and thecumulative numbers of the first and second years post-harvest were analysed. Statisticalanalysis was conducted using the SAS statistical analysis package (Schlotzhauer and Littell1987). In all statistical comparisons the level of significance was set at PO.O5.18RESULTSDeer mice and red-backed voles were the most abundant small mammals on the gridsand the majority of the analysis concentrated on these two species. Two less commonmicrotines, heather voles and meadow voles, were recorded on each of the nine grids but inlow numbers (Table 2). Heather voles and meadow voles were combined for demographicanalysis because of the low numbers of individuals captured and the difficulty ofdifferentiating the species in the field. Very low numbers of yellow-pine chipmunks werecaptured on most of the grids (less than 6 individuals per grid), with one exception (50%removal treatment at the Gavin site) where 15 individuals were recorded. Two species ofshrews, Sorex cinereus and Sorex monticolus were recorded on the majority of the grids.Shrews often did not survive capture, and were not marked if they did. Thus, the shrew dataare presented in Table 2 but further population analysis was not possible.19TABLE 2. Total captures of individuals of less common microtines, chipmunks andshrews in the 1991, 1992, and 1993 trapping sessions.UBC GAVIN SKELTONSPECIES Ctrl 30% 50% CtrI 30% 50% Ctrl 30% 50%Microtines Pre-trtmt .91 3 0 0 2 2 0 1 1 1Post-trtmt ‘91 0 5 0 0 1 1 18 5 31992 6 12 3 1 1 0 2 1 51993 1 10 3 0 7 3 1 1 0Total 10 27 6 3 11 4 22 8 9Chipmunks Pre-trtmt ‘91 0 0 0 1 2 3 0 0 0Post-trtmt ‘91 0 0 0 0 0 0 0 0 11992 0 0 0 2 0 6 1 0 11993 0 1 0 1 3 6 0 1 2Total 0 1 0 4 5 15 1 1 4Shrews Pre-trtmt ‘91 0 0 0 0 0 0 0 0 0Post-trtmt ‘91 0 0 0 1 4 2 10 9 81992 0 4 4 10 26 3 6 3 61993 2 0 1 7 11 2 2 3 17Total 2 4 5 18 41 7 18 15 31Pre-harvest 1991: n=1 to 2Post-harvest 1991: n=2 to 4Summer 1992: n=8Summer 1993: n=8Total: n=1720TrappabilityTrappability (based on the J-S density model) was calculated for deer mice, red-backed voles and the less common microtines for three different time periods: (1) overalltrappability during the entire period of study, (2) trappability during each of summer 1992and summer 1993. Overall trappability of deer mice and red-backed voles was generallyhigh throughout the duration of the study (deer mice: 61.2 ± 19.7% to 90.6± 9.3 %; red-backed voles: 70.1 ± 19.3% to 90.7 ± 7.2%), whereas less common microtines such asheather voles and meadow voles exhibited a far lower trappability overall (10.2 ± 0.1 % to47.2 ± 21.6%). Deer mouse and red-backed vole trappability overall seemed higher in the30% basal area removal stands but this difference was not significantly different (overlapping95% confidence limits). There was no difference in trappability of deer mice, red-backedvoles or other microtines in relation to the intensity of basal area removal between the firstand second year post-treatment (1992 and 1993).21TABLE 3. Jolly trappability estimates for deer mice, red-backed voles, and lesscommon microtine populations. Values shown are the average percent trappability(number caught/Jolly-Seber population estimate) ± 95% confidence limits.Overall * Summer 92 Summer 93Deer miceUBC - Ctrl 61.2 (41.5 - 80.8) 60.2 (25.5 - 95.0) 72.5 (44.9-100.1)30% 90.6 (81.4-100.0) 91.9 (82.0-101.8) 85.9 (63.2-108.6)50% 82.9 (67.6-98.2) 100.0 (100.0- 100.0) 71.9 (44.9 -98.8)Gavin - Ctrl 64.6 (50.3 - 78.9) 79.0 (61.4 - 96.6) 45.6 (23.8 - 67.4)30% 78.2 (64.4-92.1) 88.4 (74.7-102.1) 71.0 (43.3-98.7)50% 69.8 (58.4 - 81.2) 80.3 (65.3 - 95.2) 76.4 (65.2 - 87.5)Skelton - Ctrl 76.5 (66.3 - 86.7) 72.8 (52.9 - 92.6) 86.5 (75.4 - 97.6)30% 89.9 (84.2 - 95.5) 94.0 (89.6 - 98.4) 81.6 (68.4 - 94.8)50% 88.4 (81.7 - 95.1) 92.9 (87.5 - 98.2) 85.8 (75.5 - 96.0)Red-backed volesUBC - Ctrl 84.2 (70.6-97.7) 80.0 (47.8- 112.2) 86.6 (70.0- 103.2)30% 82.8 (68.5-97.2) 81.9 (51.2-112.5) 87.4 (67.7-107.1)50% 70.1 (50.8 - 89.4) 76.3 (53.3 - 99.2) 52.8 (10.1 - 95.4)Gavin - Ctrl 90.7 (83.5-97.8) 91.6 (78.6-104.6) 94.1 (84.8- 103.4)30% 87.1 (79.0 - 95.2) 93.0 (82.0 - 104.0) 86.0 (69.6 - 102.4)50% 76.8 (64.9 - 88.7) 78.6 (50.0 - 107.3) 67.9 (50.3 - 85.4)Skelton - Ctrl 72.2 (54.8 - 89.5) 69.0 (30.8 - 107.2) 94.8 (82.3 - 107.2)30% 86.0 (78.4 - 93.5) 94.8 (82.3 - 107.2) 76.0 (63.0 - 89.3)50% 85.0 (75.8-94.2) 91.9 (78.9- 104.9) 83.0 (65.8- 100.2)Other microtinesUBC - CtrI 31.6 (8.6- 54.6) 50.0 (5.3-94.7) 12.5 (-17.1 -42.1)30% 47.2 (25.6-68.7) 54.1 (12.1 -96.1) 32.9 (3.9-61.8)50% 21.1 (0.9-41.2) 25.0 (-13.7-63.7) 25.0 (-13.7-63.7)Gavin - Ctrl 14.3 (-2.0 - 30.6) 12.5 (-17.1 - 42.1) 0.0 (0.0 - 0.0)30% 33.3 (11.3-55.3) 12.5 (-17.1 -42.1) 37.5 (-5.8-80.8)50% 10.2 (0.0 - 0.2) 0.0 (0.0 - 0.0) 13.1 (-0.2 - 0.4)Skelton - Ctrl 36.9 (14.8 - 58.9) 37.5 (-5.8 - 80.8) 0.0 (0.0 - 0.0)30% 42.9 (19.8- 65.9) 50.0 (5.3 -94.7) 12.5 (-17.1 -42.1)50% 38.1 (15.4 - 60.7) 62.5 (19.2 - 105.8) 0.0 (0.0 - 0.0)* entire post-harvest period (September 1991 - October 1993)22Population densityThe different intensities of basal area removal had no effect on deer mouse populationlevels (Fig. 1). Deer mouse population levels were not different between treatments andcontrols in immediate post-harvest (1991:F2,=0.92, P=0.47), one year post-harvest (1992:F22=1.12, P=0.41), or two years after harvest (1993: F22 =2.54, P=0.19). A slightdecrease in density between pre-treatment and post-treatment 1991 was observed in theunlogged controls as well as in the treated stands (10% to 50% lower density post-harvest) intwo of the three replicates (Table 4). However, there were no significant differences inpopulation trends or density levels between the first or second year post-harvest (F12 =0.00,P= 1.00). Deer mouse populations appeared to increase on two of three replicates in 1992,but the differences were not significant (F22=0.08, P=0.93). The increase in density in the50% treatment (Gavin replicate) in 1992 seemed to be greater than that of the othertreatments or controls but this trend was not repeated on other replicates or in the followingtrapping year. There were no detectable differences between deer mouse population densitiesin any of the controls or treatments by the second year post-harvest. Mean numbers of deermice ranged from 5.6 - 12.8 animals/ha in the overall post-harvest period.Thirty and fifty percent basal area removal intensities did not negatively affect theoverall red-backed vole populations (Fig. 2). A substantial increase in red-backed voles/hawas recorded immediately post-harvest 1991 in the controls and treatments (Table 4). Meannumber of red-backed voles post-harvest 1991 escalated to 40.9 animals/ha in the controls,61.3 animals/ha in the 30% removals, and 62.2 animals/ha in the 50% removals.23TABLE 4. Average Jolly-Seber population density estimates of deer mice and red-backed voles. Values shown are mean J-S estimates for the time periods indicated.Overall time period is the entire post-harvest period of study (September 1991 -October 1993). The number of sampling sessions (n) varied between blocks pre- andpost-harvest 1991.Pre-harvest Post-harvest Summer Summer1991 1991 1992 1993 Overall(n=1 to 2) (n=2 to 4) (n=8) (n=8) (n=17)Deer miceUBC - Ctrl 15.0 2.0 7.1 5.1 5.630% 5.0 2.0 10.6 4.9 7.150% 5.0 2.3 11.6 10.4 10.0Gavin - Ctrl 15.5 7.2 9.6 9.4 9.130% 12.5 3.7 7.9 8.5 7.550% 18.7 8.9 14.0 12.3 12.5Skelton - Ctrl 3.0 11.0 10.0 13.1 11.430% 5.0 12.1 11.5 14.6 12.850% 3.0 7.6 9.4 13.6 10.7Red-backed volesUBC - Ctrl 26.0 36.0 4.6 8.8 9.930% 31.0 48.1 6.6 10.2 12.850% 22.0 48.5 5.7 5.2 10.2Gavin - Ctrl 10.5 47.0 7.0 9.4 14.430% 26.8 70.4 8.0 14.8 20.750% 14.5 76.4 8.5 11.8 20.6Skelton - Ctrl 9.0 39.7 4.3 2.4 10.630% 10.0 65.4 9.0 10.4 20.850% 30.9 61.6 7.9 5.9 17.824111111 11111MAY OCT1993FIGuRE 1. Jolly-Seber estimate of deer mouse population size on unlogged controls,30% and 50% basal area removal stands. Vertical solid line indicates the time ofharvesting.No. of animals/haDEER MICECONTROL I 30 % 50 %30IJBC0ri 111111 11111 I II 111111111 I iiAUG OCT. MAY OCT MAY OCT1991 1992 1993No. of animals/ha30. GAVIN:0 T1Th I I I I I I I I I I I I I I I I I I I I I I I I I IJULY OCT MAY OCT MAY OCT1991 1992 1993No. of animals/ha3020100SKELTONJULY OCT1991MAY OCT199225FIGURE 2. Jolly-Seber estimate of red-backed vole population size on unloggedcontrols, 30% and 50% basal area removal stands. Vertical solid line indicates thetime of shelterwood harvesting.I I I III III II 1 11111 I IAUG OCT MAY OCT MAY OCT1991 1992 1993No. of animals/ha100JULY OCT MAY OCT MAY OCT1991 1992 1993No. animals/ha.L.J1.,80SKELTON.No. of animals/haRED-BACKED VOLESCONTROL -±- 30 % 50 %100806040200UBCJULY OCT MAY OCTI I I I IMAY OCT1991 1992 199326Post-harvest 1991 red-backed vole density values were 7.3-fold to 12.6-fold greaterthan 1992 population levels. Although the increases in density in 1991 (post-harvest) werenotable on the controls as well as the treatments, the mean number of animals/ha was 1.5times greater on the treatment areas than the controls. An analysis of variance betweentreatments for one trapping session (the final trapping week in 1991 was balanced in all threereplicates) indicated that the treatment population densities differed significantly from thecontrol populations (F2,=108.30, P=0.00) and each of the replicate blocks was significantlydifferent from each other (F2,=137.82, P=0.00). The increases observed immediately post-harvest were temporary, and by 1992, red-backed vole populations had declined in thecontrols and treatments and had reached similar densities (4.6 to 9.0 animals/ha). Therewere no significant differences between treatments (1992: F2,=5.35, P=0.07) or betweenthe replicate blocks (1993: F22 =3.72, P=0. 12) within the first year post-harvest.The population trend of red-backed voles exhibited in 1993 was similar to that of1992: treatment populations remained at control levels. There was no difference betweentreatments (F2,=4.74, P=0.09) or between the replicate blocks (F22=5.88, P=0.06) twoyears post-harvest. A comparison of pooled treatments between 1992 and 1993 detected nosignificant differences in red-backed vole density between the two post-harvest treatmentyears (F1,2=1.77, P=0.31). Sampling years reflected seasonal fluctuations in populationlevels (lower after the winter months and higher in the fall at the end of the breedingseason), but no significant treatment differences during the overall period of study.27Animal survivalJolly-Seber survival of deer mice and red-backed voles was estimated for the winterand summer of the first and second year post-treatment as well as for the entire post-harvestperiod of study (Table 5). Comparisons of the means did not result in any survival valuesgreater than 1.00, thus it was not necessary to calculate minimum survival rates in additionto the Jolly survival rates. Overall survival of deer mice post-treatment was not significantlydifferent between unloggeci controls and treatments (controls mean monthly survival rate:0.84; 30% removal: 0.74; 50% removal: 0.80). Mean deer mouse survival appeared todecline between first and second year post-harvest, but no significant treatment effect wasdetected (based on overlapping confidence limits). In both trapping years, there was nosignificant difference in mean deer mouse over-winter survival. Mean survival rates in thesecond year post-harvest (summer 1993) were significantly lower than mean over-wintersurvival (winter 1992-93) estimates on treatments and controls on one of the three replicatesblocks (Skelton). There was no significant difference between deer mouse and red-backedvole overall survival on the treatment and control stands. No treatment effect was detectedbetween mean over-winter survival and summer survival between the two post-harvesttrapping years (Table 5).28TABLE 5. Mean monthly Jolly survival of deer mice and red-backed voles. Valuesrepresent mean estimates ± 95 % confidence limits of different sampling periods.Winter 1991 Summer 1992 Winter 1992 Summer 1993 Overall *Deer miceUBC - Ctrl 1.00 0.72 0.83 0.68 0.82(0.75- 1.25) (0.32- 1.12) (0.63-1.02) (-0.89-2.26) (0.45- 1.19)30% 0.84 0.76 0.85 0.34 0.64(0.58-1.10) (0.6-0.92) (-0.26- 1.58) (-0.81 - 1.48) (0.16- 1.13)50% 1.00 0.75 0.92 0.49 0.81(0.75 - 1.25) (0.68 - 0.83) (0.85 - 1.00) (-1.07 - 2.04) (0.34 - 1.28)Gavin - Ctrl 1.04 0.63 0.91 0.75 0.85(0.84- 1.23) (0.55-0.7) (0.83-0.99) (-1.13-2.63) (0.46- 1.23)30% 1.00 0.57 1.00 0.51 0.73(0.75 - 1.25) (0.4 - 0.74) (0.75 - 1.25) (-0.33 - 1.36) (0.50 - 0.97)50% 1.02 0.62 0.90 0.55 0.79(0.79 - 1.24) (0.5 - 0.74) (0.81 - 0.99) (0.38 - 0.73) (0.71 - 0.87)Skelton - Ctrl 0.89 0.81 0.94 0.67 0.84(0.33- 1.45) (0.69-0.93) (0.89-0.99) (0.53-0.81) (0.67- 1.01)30% 0.87 0.73 0.96 0.78 0.84(0.76 - 0.97) (0.56 - 0.89) (0.91 - 1.00) (0.66 - 0.89) (0.79 - 0.90)50% 0.81 0.80 0.92 0.61 0.80(-0.21 - 1.84) (0.67-0.93) (0.85-0.99) (0.48-0.74) (0.48- 1.12)Red-backed volesUBC - Ctrl 0.61 0.39 0.77 0.62 0.59(0.61 - 0.61) (-0.23 - 1 .00) (-0.68 - 2.21) (0.43 - 0.82) (0.21 - 0.98)30% 0.57 0.68 0.77 0.37 0.60(-9.44- 1.58) (0.36- 1.00) (0.64-0.91) (0.16-0.57) (-2.68-3.88)50% 0.57 0.75 0.73 0.46 0.62(0.57 - 0.57) (0.66 - 0.84) (0.54 - 0.91) (-0.81 - 1.73) (0.31 - 0.94)Gavin - Ctrl 0.63 0.47 0.73 0.51 0.59(-5.25-0.65) (0.27-0.67) (0.54-0.91) (0.31 -0.71) (-1.15-2.34)30% 0.59 0.51 0.79 0.72 0.65(0.59 - 0.59) (0.36 - 0.67) (0.65 - 0.93) (0.56 - 0.89) (0.59 - 0.71)50% 0.55 0.76 0.81 0.57 0.66(0.55- 0.55) (0.42- 1.10) (0.67-0.95) (0.37- 0.77) (0.58- 0.75)Skelton - Ctrl 0.56 0.64 0.81 0.36 0.59(0.56 - 0.56) (0.22 - 1.07) (0.74 - 0.88) (-0.78 - 1.50) (0.24 - 0.94)30% 0.64 0.57 0.85 0.69 0.69(0.64 - 0.64) (0.35 - 0.78) (0.81 - 0.88) (0.56 - 0.82) (0.63 - 0.75)50% 0.59 0.59 0.66 0.62 0.61(0.59 - 0.59) (0.42 - 0.76) (0.41 - 0.91) (0.41 - 0.82) (0.53 - 0.70)* entire post-harvest period (September 1991 - October 1993)29RecruitmentRecruitment of deer mice was not affected by the intensity of the shelterwoodtreatments in the first (1992) or second (1993) year post-disturbance. Cumulative post-harvest recruitment of individual deer mice appeared to be greater in the highest intensity ofbasal area removal and lowest in the unlogged controls (Fig. 3). However, recruitment inthe various intensities of basal area removal was not significantly different in the first andsecond year after disturbance (treatments 1992: F2,=1.44, P=O.34; 1993:F22=3.50,P=O. 13). Recruitment of deer mice was noticeably lower prior to treatment andimmediately post-treatment 1991 (Fig. 4). Differences between the means of the replicateblocks were not significant within trapping years (treatments 1992: F22 =2.66, P=O. 18;1993:F2,=2.49, P=O.20).There were no significant differences in recruitment of red-backed voles between thedifferent intensities of shelterwood treatments in the first or second year after logging (1992:F2,=O.39, P=O.70; 1993:F2,=2.27, P=O.22) (Fig. 4). The number of red-backed volesrecruited into the sample populations exhibited an opposite pattern to deer mouse recruitmentimmediately post-harvest: mean recruitment (cumulative) increased substantially on controlsand treatments in the months immediately post-harvest (1991) and increased with intensity ofdisturbance on all three replicates (Table 6). Cumulative recruitment per sampling yearappeared to return to pre-treatment levels by the first year post-harvest and began to increasemarginally, on controls and treatments, by the second year post-harvest.30FIGuRE 3. Cumulative post-harvest (199 1-1993) number of deer mouse and red-backed vole recruits, as observed over three replicates: unlogged controls, 30% basalarea removal shelterwood treatments, and 50% basal area removal shelterwoodtreatments.DEER MICEControl 30% removal 50% removalNumber of recruits12010080UBCflED-BACKED VOLES30% removal : 50% removalGavin SkeltonControlNumber of recruitsI604020ZZZZL-J0UBC Gavin Skelton31FIGURE 4. Cumulative number of deer mouse and red-backed vole recruitsthroughout various periods of study: pre-treatment 1991, post-treatment 1991, summer1992, summer 1993. Values represent means of pooled treatments and controls.DEER MICEControls 30% removalsNumber of recruits10080604020050% removalsI IPre-trtiut ‘91 Post-trtmt ‘91 Summer 1992 Summer 1993ControlsNumber of recruitsRED-BACKED VOLES30% removals [E] 50% removals100806040200Pre—trtint ‘91 Post-trtmt ‘91 Summer 1992 Summer 199332TABLE 6. Cumulative number of deer mouse and red-backed vole recruits per haduring pre-treatment and post-treatment periods of study (1991-1993).UBC GAVIN SKELTONCtrI 30% 50% Ctrl 30% 50% Ctrl 30% 50%Deer micePre-trtmt 1991 15 5 6 19 13 19 3 6 3Post-trtmt 1991 0 0 1 3 3 3 13 12 11Summer 1992 15 23 25 26 23 34 21 21 20Summer 1993 13 17 37 27 25 34 32 32 35Overall 28 40 62 53 48 68 53 54 55Red-backed volesPre-trtmt 1991 35 35 26 19 35 26 6 12 16Post-trtmt 1991 34 59 59 59 82 100 65 100 109Summer 1992 20 24 17 28 32 31 16 30 28Summer 1993 42 52 23 57 71 51 15 35 35Overall 72 94 52 87 120 94 45 77 81Number of sampling sessions per site:Pre-treatment 1991: UBC=2, Gavin=3, Skelton=2Post-treatment 1991: UBC=2, Gavin=3, Skelton=4Summer 1992: all sites = 8Summer 1993: all sites = 9Overall post-harvest = 1833ReproductionMean adult body mass at which 50% or more of the individuals captured were atsexual maturity was averaged over all grids for each sex and species (Table 7). Males ofboth species were sexually mature at a lower body mass than females (deer mice: 16.8 g formales vs 20.2 g for females; red-backed voles: 20.8 g for males vs 22.2 g for females).Deer mice reached sexual maturity at a lower body mass than red-backed voles.There was little difference in the length of the deer mouse or red-backed volebreeding seasons between treatments within trapping years (Table 8). However, both speciesshowed a considerable increase in breeding season length in the second year post-harvest(1993) versus the first year post-harvest (1992). For the majority of the grids, the red-backed vole breeding season seemed marginally longer than that of the deer mouse: frommid- to late-May to the end of September (1992) or October (1993) for red-backed voleswhereas deer mice had, at times, finished breeding as early as the beginning of July (1992)or on occasion as late as the end of October (1993). In 1992, the length of the deer mousebreeding season ranged from 6.0 to 15.0 weeks (average 11.0); in 1993 the breeding seasonwas 12.0 to 24.0 (average 20.3) weeks long. Red-backed voles were in breeding conditionfrom 9.0 to 21.0 weeks (average 14.3) in 1992 and for 18.0 to 24.0 (average 21.3) weeks in1993. Thirty and 50% shelterwood removals did not have a significant effect on the lengthof the breeding season in either of the trapping years for red-backed voles or deer mice(AN0vA, treatment, deer mice, 1992:F22=0.43, Pt0.68; 1993:F2,=0.54, P=0.62; redbacked voles, 1992:F,=0.34, P=0.73; 1993: F2,=1.00, P=0.44).34TABLE 7. Mean body mass (in grams) at sexual maturity for male and female deermice and red-backed voles. Averages were taken from each of the nine grids and arebased on the entire period of study.Male FemaleDeer mouse 16.8 20.2Red-backed vole 20.8 22.2TABLE 8. Length of breeding season in weeks. The start of the breeding season isthe time of capture of the first scrotal male or lactating female. The end of thebreeding is the week in which the last lactating female was recorded.Deer mice Red-backed voles1992 1993 1992 1993UBCControl 6 12 15 1830% removal 15 18 18 2150% removal 9 21 15 24GavinControl 12 24 15 2430% removal 9 18 9 2450% removal 15 21 12 24SkeltonControl 12 21 9 1830% removal 15 24 21 2150% removal 6 24 15 181992 1993 1992 1993Pooled replicatesControls 10 19 13 2030% removals 13 20 16 2250% removals 10 22 14 2235The overall percentage of male deer mice in breeding condition ranged from 59.0% to92.6% and the percentage of females in breeding condition ranged from 66.7% to 100.0%(Table 9). The percentage of adult male red-backed voles in breeding condition ranged from50.0% to 90.0% (Table 10) and females ranged from 44.4% to 87.3%. The low percentagesof adult red-backed voles breeding may have been due to the high influx of animalsimmediately after logging (Table 10).TABLE 9. Numbers and proportions of adult deer mice in breeding condition onthree replicates throughout the study period (n=number of individuals sampled).CONTROLS 30% REMOVALS 50% REMOVALSM* n F* n M n F n M n F nUBCPre-trtmt ‘91 0.00 2 0.00 3 0.00 0 0.00 0 0.00 0 1.00 1Post-trtmt ‘91 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0Summer 1992 0.75 8 1.00 7 0.60 5 1.00 8 1.00 6 1.00 8Summer 1993 0.89 9 0.71 7 1.00 4 1.00 8 0.90 21 0.91 11Overall 0.82 17 0.86 14 0.78 9 1.00 14 0.93 27 0.95 19GAVINPre-trtmt ‘91 0.43 7 0.71 7 0.00 0 0.67 6 0.67 3 0.57 7Post-trtmt ‘91 0.00 6 0.00 3 0.00 1 0.00 0 0.00 0 0.00 0Summer 1992 0.89 9 1.00 1 1.00 6 0.88 8 1.00 2 0.86 22Summer 1993 1.00 15 1.00 5 0.86 7 1.00 8 0.86 14 0.81 16Overall 0.77 30 0.67 9 0.86 14 0.94 16 0.88 16 0.84 38SKELTONPre-trtmt ‘91 0.25 4 0.00 0 1.00 1 0.50 2 1.00 2 0.00 0Post-trtmt ‘91 0.00 3 0.00 1 0.00 0 0.00 0 0.00 1 0.00 2Summer 1992 0.64 11 1.00 7 1.00 1 0.70 10 0.54 13 1.00 2Summer 1993 0.64 25 1.00 15 0.88 25 0.74 31 0.82 22 0.88 17Overall 0.59 39 0.96 23 0.88 26 0.73 41 0.69 36 0.81 21*M= males F = females36TABLE 10. Numbers and proportions of adult red-backed voles in breedingcondition on three replicates throughout the study period (n=number of individuals inthe sample).CONTROLS 30% REMOVALS 50% REMOVALSM* n F* n M n F n M n F nUBCPre-trtmt’91 1.00 6 0.67 12 0.75 4 0.79 19 1.00 4 0.58 12Post-trtmt 91 0.00 1 0.00 6 0.00 3 0.29 7 0.00 2 0.18 11Summer 1992 1.00 5 0.92 12 1.00 4 1.00 7 1.00 1 0.73 15Summer 1993 0.93 14 0.94 18 0.87 15 0.80 25 1.00 3 0.90 10Overall 0.90 20 0.78 36 0.77 22 0.74 39 0.67 6 0.61 36GAVINPre-trtmt’91 0.63 8 0.75 8 0.46 13 0.53 15 0.56 9 0.63 8Post-trtmt’91 0.00 5 0.15 13 0.67 3 0.18 11 0.50 4 0.17 12Summer 1992 1.00 5 0.85 13 0.91 11 0.71 14 1.00 6 1.00 8Summer 1993 0.94 16 0.78 27 0.86 21 0.82 28 0.85 13 0.75 20Overall 0.77 26 0.64 53 0.86 35 0.66 53 0.83 23 0.63 40SKELTONPre-trtmt ‘91 0.50 4 0.00 0 0.89 9 0.80 5 0.56 9 0.25 4Post-trtmt’91 0.00 7 0.18 17 0.08 12 0.44 9 0.20 10 0.35 20Summer 1992 1.00 3 0.75 4 0.86 14 0.92 12 0.83 12 0.94 16Summer 1993 1.00 4 1.00 6 0.94 17 0.97 34 0.88 8 0.88 17Overall 0.50 14 0.44 27 0.67 43 0.87 55 0.63 30 0.70 53M*= males F* = females37There were no significant treatment effects on the percentage of adult deer mice orred-backed voles breeding within the 1992 or 1993 trapping years (Table 11). The majorityof the individual red-backed voles captured were juvenile and were never recorded inbreeding condition during the course of this study. Eighty-five percent of the individual malered-backed voles and 78% of the females captured were juvenile, thus the percentage ofadults in breeding condition reflected only a small proportion of the red-backed volepopulations on the study sites. The majority of the male deer mice captured were adults(68.9%), but the female deer mice sampled on all treatments were largely juvenile (63.0%).TABLE 11. Treatment effects on the percentage of breeding adults in thesample populations. Anova F and P values are shown for treatment effectswithin trapping years. Treatments were pooled across the replicate blocks.Degrees of freedom=2. P=0.05.Species F statistic P valueP. maniculatus1992 males 0.54 0.62females 1.80 0.281993 males 0.29 0.76females 0.37 0.71C. gapperi1992 males 1.80 0.28females 0.22 0.811993 males 0.67 0.56females 1.49 0.3338Body massThere were no significant treatment differences in the mean body mass of adult, maledeer mice within the first and second years post-harvest (AN0vA treatment 1992: F2,=1.28,P=0.37; 1993:F2,=0.57, P=0.61). The body mass of adult deer mice after the harvestingtreatments were uniform across the treatments and ranged from 19.8 to 20.4 g (Table 12).Adult red-backed voles seemed to be heavier than deer mice: male body mass of red-backedvoles ranged from 22.5 g to 25.0 g (Table 13). There was no significant treatmentdifference observed in the mean adult body mass of red-backed voles within the two post-harvest sampling years (AN0vA treatment 1992:F22=2.06, P=0.24; 1993:F2,=2.30,P=0.22).DiversityValues for the calculation of Shannon-Wiener and Simpson’s measures ofheterogeneity were obtained from absolute numbers of individuals in the various timeperiods. Over the course of the study (1991-1993) shrews were observed on each of the ninegrids (Table 2) but they were not marked for identification. Hence, it was not possible todetermine if shrews were new individuals or repeated captures of the same individuals.Therefore, shrews could not be included in the diversity measurements or comparisons.Suffice it to say they were present in low numbers on each of the grids. The concordance ofS-W and Simpson’s indices in the patterns of diversity throughout the study periods isillustrated in Figure 5.39TABLE 12. Mean body mass (in grams) of male adult deer mice in treatment andcontrol stands throughout the study period (n=number of individuals recorded duringthe separate time periods).overallGAVIN14 20.2 0.6 16 20.1 1.0 28 20.3 0.5overall 20 20.9 0.6 18 20.5 0.7 22 19.3 0.4CONTROLS 30% REMOVALS 50% REMOVALSn Mean S.E. n Mean S.E. n Mean S.E.UBCpre-trtmt ‘91 5 20.4 1.5 2 19.5 0.5 2 18.0 1.0post-trtmt ‘91 0 0.0 0.0 0 0.0 0.0 1 18.0 0.0summer 1992 5 21.8 0.6 9 19.2 0.7 10 17.9 0.5summer 1993 9 19.3 0.7 7 21.3 2.1 18 21.6 0.6pre-trtmt ‘91 9 20.4 0.7 1 17.0 0.0 3 21.0 1.2post-trtmt ‘91 1 23.0 0.0 2 21.5 2.5 4 20.0 0.4summer 1992 9 20.8 1.0 11 20.3 0.8 8 18.8 0.5summer 1993 11 20.9 0.7 7 20.9 1.3 14 19.6 0.6SKELTONpre-trtmt ‘91 3 22.3 0.3 2 21.5 1.5 3 23.3 2.6post-trtmt ‘91 7 19.3 0.6 4 17.5 0.3 5 19.0 0.8summer 1992 9 18.9 0.7 4 19.0 1.2 7 19.9 0.8summer 1993 25 20.1 0.4 16 20.6 0.6 25 20.0 0.6overall 34 20.1 0.4 21 20.1 0.5 32 19.9 0.540CONTROLS 30% REMOVALS 50% REMOVALSn Mean S.E. n Mean S.E. n Mean S.E.UBCpre-trtmt ‘91 5 21.8 0.6post-trtmt ‘91 1 24.0 0.0summer 1992 2 23.0 2.0summer 1993 6 24.7 0.9overall 9 24.2 0.7 12 25.0 0.7 4 21.5 1.9GAVINpre-trtmt ‘91 5 25.4 1.3post-trtmt ‘91 4 21.5 0.5summer 1992 3 24.3 0.3summer 1993 10 24.2 0.96 25.8 1.21 25.0 0.06 24.0 0.910 23.3 0.86 24.2 0.93 22.3 0.72 22.5 1.55 22.2 0.6overallSKELTON13 24.2 0.7 16 23.6 0.6 7 22.3 0.5pre-trtmt ‘91 2 26.0 1.0post-trtmt ‘91 4 22.5 0.7summer 1992 2 22.5 1.5summer 1993 4 29.8 6.85 25.0 1.39 22.7 0.75 24.4 1.06 24.5 1.26 24.5 0.85 21.8 0.44 23.8 0.64 23.8 0.9TABLE 13. Mean body mass (in grams) of male adult red-backed voles in treatmentand control stands throughout the period of study (n=number of individuals recordedduring the separate time periods).3 24.3 1.52 21.5 0.52 24.0 0.010 25.2 0.83 24.0 1.72 19.5 351 22.0 0.02 24.0 0.0overall 7 26.4 4.0 11 24.5 0.8 8 23.8 0.541Species richness varied between 1 and 4 species throughout the study years (Table15). Diversity and evenness was lowest on all grids immediately following the loggingdisturbance in 1991 (post-treatment 1991), presumably due to the high proportion of red-backed voles present at this time (Table 14). There was no significant difference in diversitydue to the logging disturbance detected in the first sampling year post-harvest (AN0vA,Simpson’s index, treatment 1992: F22 =2.06, P=0.24; S-W index, treatment 1992:F2,=0.34, P=0.73). In the second sampling year post-harvest, Simpson’s index indicatedno treatment differences (AN0vA, Simpson’s index, treatment 1993:F22=6.54, P=0.05),whereas the S-W index was significantly greater on the logged treatments (30% and 50%removals had equal diversity in 1993) than on the unlogged controls (AN0vA, S-W index,treatment 1993:F2,=7.92, P=0.04) (Fig. 5). S-W and Simpson’s evenness seemed higheston all grids the first year after disturbance (summer 1992). Despite the initial decline inevenness immediately post-harvest (fall of 1991, when red-backed voles greatly outnumberedthe other species recorded in the traps), no overall differences in species evenness werefound to be attributable to the shelterwood treatments by the first and second year postharvest.42FIGuRE 5. Simpson’s and Shannon-Wiener indices of small mammal speciesdiversity for different time periods throughout the study. Values represent means ofpooled treatments and controls.— ControlsSIMPSON’S INDEX30% removals 50% removals- j.Post-trtmt ‘911.510.50— ControlsSHANNON-WIENER INDEX30% removal 50% removal10. ‘91 Summer 1992 Summer 1993Pre-trtint ‘91 Post-trtmt ‘91 Summer 1992 Summer 1993UBCGAVINSKELTONPERIODSControl30%50%Control30%50%Control30%50%Simpson’sindexPre-treatment0.500.220.320.580.490.560.600.530.35Post-treatment0.’sEvennessPre-treatment0.730.440.610.760.640.820.820.750.50Post-treatment0.460.—IC,,I;o.4•0.—ccl,rID>00rM1)C’,0a)44TABLE 15. Species richness, and mean Simpson’s and Shannon-Wiener diversityand evenness values pooled for three replicates for the overall study period.Pre-treatment Post-treatment Summer Summer Overall1991 1991 1992 1993Species richnessControls 3 to 4 1 to 3 3 to 4 3 3 to 430% removal 2 to 4 2 to 3 3 4 450% removal 2 to 3 2 to 4 3 to 4 3 to 4 3 to 4Simpson’s indexControls 0.56 0.29 0.58 0.44 0.5430% removal 0.41 0.17 0.56 0.51 0.5350% removal 0.41 0.11 0.58 0.55 0.56Simpson’s evennessControls 0.77 0.46 0.82 0.64 0.7430% removal 0.61 0.27 0.83 0.68 0.7150% removal 0.64 0.16 0.82 0.78 0.77Shannon-Wiener indexControls 1.28 0.75 1.33 0.97 1.2830% removal 0.99 0.48 1.24 1.26 1.3050% removal 0.95 0.35 1.35 1.26 1.34S-W evennessControls 0.75 0.53 1.01 0.61 0.6930% removal 0.63 0.35 0.78 0.63 0.6550% removal 0.68 0.21 0.79 0.73 0.7245Seed fallTotal seed fall for Douglas-fir and white spruce was recorded in the fall of the firstand second year post-treatment (Fig. 6). The 1992 seed crop ranged from 5.4x 10 to6.2x105Douglas-fir seeds per hectare, and a negligible amount of white spruce seed wasrecorded on 6 of the 9 treatment units. There were no detectable treatment differences intotal Douglas-fir seed fall in the first year post-harvest (AN0vA, treatment 1992: F22=1.37,P=O.78). The seed crop in the second year post-harvest was considerably larger than that ofthe first year (Douglas-fir seed crops ranged from 1.6x10 to 4.6x10 seeds per hectare, andwhite spruce seed was recorded on all 9 treatments at densities of up to 1 .8x105 seeds perhectare), but again there was no identifiable pattern of seed fall intensities observed betweenthe treatment units. The number of Douglas-fir seeds recorded in the fall of 1993 appearedto be greatest on the unlogged controls and lowest on the 50% basal area removal stands(3.6x106 seeds/ha vs 3.0x106 seeds/ha), but there was no significant treatment effect detectedin the amount of seed fall (AN0vA, treatments 1993: F22 =1.17, P = 0.40).46FIGuRE 6. Total number of Douglas-fir and white spruce seeds collected per ha.Distribution of seed rain shown for unlogged controls, 30% and 50% removaltreatments in the first and second year post-treatment. Samples collected in fall of1992 and 1993.1992Total seedfall Douglas-firseeds/ha (thousands)[,.::::] White spruceUBC GAVIN SKEL UBC GAVIN SKEL UBC GAVIN SKELControls 30% Removal 50% Removalseeds/ha (millions)6543210VBC GAVIN SKEL UBC GAVIN SKEL IiBC GAVIN SKELControls 30% Removal 50% Removal47The percentage of sound Douglas-fir seeds varied between the two sampling years:36.5% to 42.0% of the seeds collected in the first year post-harvest were sound, and 71.0%to 78.0% of the seeds collected in the second year post-harvest were sound (Table 16). Theviability of the seeds seemed to increase with the larger seed crop. The percentage of soundseeds did not vary significantly between treatment units within either of the two post-harvestsampling years (AN0VA, treatments 1992: F22 =0.08, P=0.92; treatments 1993:F2,=2.61,P=0. 19).T.BLE 16. Mean number of Douglas-fir seeds/ha and number of sound Douglas-firseeds/ha. Seed fall samples were collected in the first and second year post-harvest(1992 and 1993, respectively).Total seed fall No. sound seeds Percentage sound1992controls 160 360 58 558 36.530% removal 235 135 72 973 31.050% removal 262 162 109 910 41.91993controls 2 838 738 2 213 513 78.030% removal 2 443 243 1 756 757 71.950% removal 2 504 504 1 802 702 72.048Seed germinationSeed germination exciosures were placed on the grids in the fall of 1992 and sampledin June 1993. Not all of the seeds survived in the germination exciosures until the samplingperiod, thus two values are presented: 1) the percent germination of the total number ofseeds placed in the germination exciosures, and 2) the percent germination of the number ofseeds which survived until the following spring (Table 17). The percentage of seeds whichsurvived in the exciosures and germinated was presumed to be the actual germinationpotential of the sites.TABLE 17. Survival and germination results of Douglas-fir germination trials.Shown are the mean values of three replicated unlogged controls, 30% and 50% basalarea (BA) removal treatments.SEEDS CONTROLS 30% REMOVALS 50% REMOVALSNo. sown 100 100 100No. recovered 47.3 51.0 59.0No. germinated 29.3 32.3 37.3No. consumed 3.7 5.7 4.0by insectsOnly 47% to 59% of the seeds sown in the germination trials were actually located inthe spring sampling session. Germination occurred on all treatments and on all sampledseedbeds. The percentage of seeds that germinated was highly uniform between the49treatments: 62.0% of the seeds that were found on the controls and 63.4% of the seeds foundon the 30% and 50% basal area removal treatments germinated in the following growingseason (Fig. 7). There was no significant treatment effect detected in the germinationresults, but the three replicate blocks were significantly different (AN0vA, treatments:F22=O. 8, P=0.77; blocks: F22=11.29, P=0.02). Insect predation was observed on 6.8 to11.1 % of the seeds which survived in the germination sampling plots, but it was not possibleto distinguish between pre-dispersal insect predation from the nursery seed and on-site insectpredation. Survival of the seeds in the germination exciosures was jeopardized by variousunforeseen hazards: some of the protective covers were trampled by moose or deer, andothers were punctured by falling branches or windfall, thereby exposing the seeds to smallmammal or bird predation. Still other seeds may have been lost in the litter layers, as theseeds were carried down through the forest profile during the winter months.50FIGURE 7. Fates of the Douglas-fir seeds which were recovered in the germinationtrials. Seeds were sown in faIl 1992 and results were recorded in June 1993. Valuesshown are the mean values of three pooled replicates.30.28%CONTROLS61.97%25.49%30% REMOVAL STANDS63.40%29.94%50% REMOVAL STANDS63.28%11.11%iIiii1 Gerrninants I_,__i Ungerminated seed Insect predation51Seed predationMean monthly seed predation rates in the second year post-disturbance averaged1.4x105 seeds/ha/month on the controls, and 1.9x105 seeds/ha/month on the 30% and 50%removal areas. Mean monthly seed predation rates throughout the snow-free months werenot significantly different between treatments (AN0vA, treatments, 1993: F2,=2.98, P=0.06)(Fig. 8). The highest seed predation rates were recorded in the fall of 1992 (controls:1.6x105 seeds/ha/month; 30% removals: 2.7x105 seeds/ha/month; 50% removals: 2.1x105seeds/ha/month) but these values were based on a single sampling session and may have ahigh human error factor as it was the first time the seed predation plots were sampled.Overwinter mean monthly seed predation was noticeably lower than predation during thesnow-free months, but again no treatment effect was discernable in the winter predation rates(controls: 2.6x104 seeds/ha/month; 30% removals: 3.4x10 seeds/ha/month; 50% removals:2.8x104 seeds/ha/month). Some seeds remained in the sampling plots throughout the wintermonths and were identified in the following spring.An increase in predation rates was observed in the early fall of 1993 on all grids, butthis increase subsided during the last two sampling sessions of the year. The increase inpredation during the fall was larger on the 30% and 50% removal treatments than on theunlogged controls, but there was no significant treatment effect on the overall seed predation.52FIGuRE 8. Douglas-fir seed predation per ha per month on the controls andtreatments. Values shown are the number of seeds consumed per month in October,1992, overwinter 1992/93 (o/w) and May to October, 1993. Treatment means arepooled across the three replicates.Control I 30% removal )I( 50% removal400300200100Predation/ha (thousands of seeds/month)00 Nov. - April S 0 Nov.I I I I I I I I I IMJ J A1992 o/w 199353The number of seeds placed in the seed predation quadrats was less than the totalnumber of seeds collected in the seed fall traps per hectare on all grids. Thus, it wasassumed that the seed predation quadrats did not encourage any additional predation or attractadditional animals onto the sites. On average, 32 to 35% of the total seeds consumed withinthe sampling quadrats were eaten by small mammals, but 65 to 67% of the predation couldnot be attributed to a specific group of seed eaters, as the seeds were removed from thesample plots. It is probable that a combination of seed-eaters was responsible for the totalpredation, as seed-eating birds such as the dark-eyed juncos (Junco hyemalis) were alsoobserved on the sites.54DISCUSSIONExperimental designThe random interspersion of the three treatment replicates within each of the threeblocks (study sites) ensured statistical independence of the replicates, appropriate testing ofthe hypotheses in question, and avoidance of pseudoreplication (Hurlbert 1984). Theincreased sample size provided by additional replicates may have increased the precision ofthese results, and possibly improved the power of the statistical tests, but additional replicatesmay not have been economically justifiable. The control treatments may not have beenentirely effective in providing an indication of undisturbed population trends or densities.The movement of animals between the treatment units within the replicate blocks and thedisturbance by researchers during the intense sampling and data collection on the sites mayhave jeopardized the validity of the controls. The sites were sampled during the snow-freeportions of the year (May to October) only, but there was considerable activity on the studysites during these months. In addition, the removal of trees in order to minimize further pestproblems (some trees were removed from buffer zones in an attempt to contain a Douglas-firbeetle (Denroctonus pseudotsugae) attack) may have sacrificed the soundness of some of thebuffer zones between the treatments.Burton (1994) conducted a survey of surface materials in the fall of 1993 to estimatethe relative abundance of seedbed materials in the control and treatment stands. Disturbancewas recorded on the harvesting treatments as well as on the unlogged controls, but the55disturbance was understandably greater on the harvested areas (2% on the unlogged controls,10% on the 30% BA removals, and 12% on the 50% BA removals) (Burton 1994). Therewas reasonable uniformity in site characteristics and no significant block differences inpopulation levels of small mammals between the treatments in the first or second year posttreatment, and little indication that the soundness of the treatments was compromised. Thelogistics of the logging operation required that the harvesting be staggered over a number ofweeks across the sites, and thus the sampling sessions during the treatment year wereirregular between the replicates. In addition, the pre-treatment sampling sessions were notbalanced in number or timing across the replicates. The results of the pre-and post-treatmentdata in 1991 were notable biologically but could not analysed statistically. Greater emphasison synchronizing the number and timing of the sampling sessions across the replicates wouldhave enabled the pre-treatment and immediate post-treatment results to be interpreted withmore confidence. Extending the duration of this study to two years post-treatment providedcomparatively long-term results of the effects of the shelterwood harvests on small mammalpopulations and seed parameters, but follow-up sampling is recommended to observe smallmammal population and seed fall responses throughout the second and final shelterwoodpasses.56Demographic responses of small mammalsThe results of the population analysis of the small mammal species present on thestudy sites indicated that, other than the initial influx of red-backed voles immediately post-harvest (which may have been attributable to a population cycle), the first pass of theshelterwood system did not significantly affect small mammal population levels or dynamicsin the treated stands. It was presumed that chipmunks, shrews, and the less commonmicrotines did not constitute a large component of the small mammal community in thepartially logged stands during the first two years post-harvest, and played a minor role inseed predation at that time. The less common microtines were only occasionally recordedduring the trapping sessions and, as a result, the trappability of these species was very low(10.2 to 47.2%). The trappability of the predominant species (deer mice and red-backedvoles) was consistently high (ranging from 61.2 to 90.7%) throughout the time of study andacross the treatments. Thus, the accuracy and reliability of the additional demographicparameters calculated for these common species was presumed to be high.The habitat requirements of the deer mice and red-backed voles appeared to be met,at least initially, with the habitat created within a shelterwood system. It is plausible that thedemographic responses of the less common species to these treatments were similar to thoseof the deer mice and red-backed voles, and also may not have been negatively affected by theremoval treatments. Throughout the post-harvest period of study, the mean number of deermice ranged from 6 to 13 animals/ha. As deer mice tend to be more common on recently57cut areas or early successional stages than in mature forest (Tevis 1956; Gashwiler 1970;Krefting and Ahlgren 1974; Sullivan and Krebs 1981; Gunther et al. 1983; Scrivner andSmith 1984; Medin 1986; Walters 1991), they have imposed a serious threat to naturalregeneration on clearcut areas. On the contrary, the habitat alterations resulting from thepartial cuts did not appear to attract additional deer mice to the sites or cause a great increasein density of seed-eater populations.Red-backed voles are known to collect, cache, and presumably consume conifer seed(Askham 1992) and seemingly fare well in a partially logged environment. Results from thisstudy showed dramatic increases in red-backed vole population levels immediately post-harvest 1991, on the treatments as well as on unlogged controls: 3-fold increase from preharvest levels on the controls and 50% removal treatments, and 4-fold increases from preharvest levels on the 30% removal treatments. This increase in density was temporary, andby the first year post-harvest, the average population levels had subsided from a range of 41to 62 voles/ha to a range of 6 to 10 voles/ha.Discrepancies exist with regard to the periodicity and duration of small mammalcycles, but fluctuations in small mammal populations are common and are presumed to occurevery 3-4 years (Krebs 1966). Thus, it is possible that the 1991 “super-populations” of redbacked voles may have reflected a vole peak. Whereas red-backed vole densities of 5 to 15individuals per hectare are common (Medin, 1986; Medin and Booth 1989), our studyshowed as many as 70 individuals per hectare on some sites in October 1991.58As the high population levels in the fall of 1991 were positively correlated with theintensity of harvesting disturbance (density levels on the 30% and 50% removal sites were1.5 times greater on the treated areas than the controls), the initial shelterwood removal mayhave affected the red-backed vole habitat in various ways:a) by providing a temporary increase in local forage and/or cover with the falling ofselected trees. This may have provided increased insect and/or seed availability bydislodging them from the canopy or exposing them in the duff;b) by providing a more readily available food source for rodents. The soildisturbance associated with the removal of felled trees may have exposedhypogeous fungi to a greater extent than was naturally present;c) by increasing thermal and security cover for small mammal species byleaving coarse woody debris on the sites after logging. In addition, theundisturbed ground cover remaining after harvesting (as indicated by the lowpost-harvest soil disturbance levels) may have contributed to suitableoverwintering sites for the red-backed voles (West et al. 1980).These habitat alterations may explain the magnitude of post-treatment populationdensities encountered in 1991. It is important to note that this increase in vole densityoccurred to a varying extent on the unloggeci controls as well as on the treatment sites, as did59the overwinter (199 1-1992) population decline. Therefore, the oscillating population trendsand the poor over-winter survival of red-backed voles cannot be wholly attributable to theeffects of harvesting and instead must be reflecting some sort of cycle. Red-backed volescontinued to exhibit seasonal population fluctuations (lower after winter, higher in the fall)despite the surge in density in the fall of 1991. Thus, the populations were merely present inlower numbers in the first and second year post-harvest, as would be expected following acyclical peak, but were not displaced by the harvesting treatments.The seasonal variation in deer mouse survival may have been a result of the habitatvariety created by the partially-cut stands. The habitat alterations created with shelterwoodharvesting were apparently favourable for a variety of predator species. Pine marten (Martesamericana), red fox (Vulpesfulva), wolves (Canis lupus), and raptors were observed in thecontrols and treatment stands throughout the study. As deer mice are active year round, thewinter snow cover may have provided greater protection from predators and enabled highersurvival in the winter than could be expected during the snow-free months, but no data werecollected for such an analysis as it was beyond the scope of this study. The two majorspecies showed no significant difference in survival on the controls or basal area removaltreatments for the entire post-harvest period of the study. There was no significant seasonaldifference observed in the survival of red-backed voles.Recruitment of deer mice showed no difference between the first and second yearpost-harvest, and there was no detectable treatment effect between the stands. Recruitmentof deer mice was lowest immediately following the disturbance (post-harvest 1991) butincreased in subsequent years, possibly indicating that the deer mice can rapidly repopulate60disturbed areas. Recruitment of red-backed voles was greatest immediately after logging(post-harvest 1991), and exhibited a positive correlation of recruitment with residual basalarea in the stands. This correlation paralleled the growth in red-backed vole populationlevels at this time. The post-harvest 1991 increase could not be analysed statistically becauseof the unbalanced sampling sessions during this time period, and the observed increase inrecruitment at this time had subsided to pre-treatment levels by the first year post-treatment(summer 1992).Deer mouse and red-backed vole males were sexually mature at a lower mass (age)than females on all sites. However, adult body mass did not exhibit any discernabledifferences due to treatment effects for the deer mice or the red-backed voles. The length ofthe breeding seasons were not significantly different across treatments but were considerablylonger, for both species, in the second year post-harvest than the first (mean number ofweeks breeding in 1992 versus 1993, deer mice: 11.0 versus 20.3 weeks, respectively; red-backed voles: 14.3 versus 21.3 weeks, respectively). Breeding seasons may be shortenedduring times of stress in a population, and hence any increase in breeding season length maybe favourable in terms of the health of the species.There was a wide range in the percentage of adults in breeding condition throughoutthe course of the study, but there were no significant differences attributable to treatmenteffects. The lowest percentages of red-backed voles in breeding condition was recordedimmediately post-harvest, at which time the population densities were escalating to thehighest levels recorded in the course of the study. However, a high proportion of the redbacked voles captured during this study were juvenile and were never recorded as sexually61mature. Only 15% of the male red-backed voles and 22% of the females captured during thestudy were recorded as adults. Thus, the analysis of the adult individuals reflected only asmall proportion of the animals present throughout the study. The deer mice populationssampled consisted of predominantly adult males (68.9%), but the females sampled werelargely juvenile (63.0%). Grids with lower quality habitat have been known to exhibit highanimal densities, but van Home (1983) noted that this may be due to an irruption of juvenilesconsisting largely of immigrants which have been forced into lower-quality, or “sink”habitats. The information collected in this study was not sufficient for determining whetherthe fates of the individuals were attributable to death or dispersal, and high proportions ofjuveniles were recorded on the controls as well as the treatment units.DiversityDespite the debate of whether diversity may or may not be a reasonable objective innatural resources management (Magurran 1988; Hunter 1990; Burton et al. 1992), speciesdiversity and evenness was measured to assess treatment effects. Diversity measures andspecies evenness measures were lowest on all grids immediately following the harvestingoperations, presumably due to the great influx of red-backed voles at this time. Thereappeared to be greater species diversity on the unlogged controls pre-and-post-treatment 1991and throughout 1992, but the differences between the controls and treatments were notsignificant. The second year after disturbance, species diversity was greater on the loggedtreatments than on the unlogged controls.62Seed ParametersSeed fall measurements indicated that a small Douglas-fir crop was produced in 1992and a larger crop of Douglas-fir and white spruce was produced in 1993 on all grids. Totalseed fall in 1992 was marginally greater in the treatments than the unlogged controls butthese differences were not significant. The larger seed crop in 1993 showed only minordifferences between treatments but there was a notable increase in the proportion of viableseeds in the seed crop. Surface materials in each of the three study site locations weredominated by organic materials (80 to 90% cover), and this did not change markedly withthe harvesting treatments (Burton 1994). There was no detectable treatment effect on thegermination potential of the post-disturbance seed beds in the stands.Seed predation increased dramatically on all sites in the early fall of 1993, at whichtime deer mouse and red-backed vole populations were at their seasonal peaks. However,there did not seem to be a large influx of new animals at this time, and by late fall seedpredation had subsided considerably. This decline in predation in the sampling plotscoincided with the timing of the natural seed crop dispersal in the fall. Some Douglas-firseed survived in the exposed sampling plots throughout the winter months. Snow coverpromoted seed survival by concealment, regardless of the small mammal density, and as thevulnerability of the seeds decreased, the rate of seed predation was also reduced. Seedpredation due to insects was negligible on all sites. Seeds missing from the plots may havebeen eaten by various seed-eating birds, but accurate identification of the various birds63observed in the stands was not obtained.Seed predation varied with the seasonal fluctuations of the seed-eater populations, butthe predation rates did not seem to be influenced by the amount of seed available on thesites. There was no observed change in small mammal population parameters or communitystructure in relation to the amounts of conifer seed available on the sites during the two post-harvest sampling years. The population levels of the most common small mammal speciesdid not vary with the increased seed crop. Seed predation rates reflected seasonal populationfluctuations rather than seed fall intensities, i.e. the rate of seed predation per ha per monthdid not increase with the larger seed crop in the second year after harvesting. Seed predationrates seemed less severe at times when seed-eater populations were at annual low points,such as late winter or early spring. Douglas-fir seeds survived the winter on all grids andhad the potential to germinate on the seedbeds available and in the microclimate created withshelterwood harvesting. Thus, the regeneration success of the stands, with regard to seedsurvival and establishment, in the basal area removal intensities tested during this study didnot seem to be inhibited by the small mammal communities recorded during this study.64MANAGEMENT IMPLICATIONSBy examining the habitat requirements of the seed predators and observing theirseasonal population fluctuations, this study showed that it may be possible to manage forseed-eaters and still allow natural regeneration to succeed. Changes in habitat due to theshelterwood harvesting did not appear to be severe enough to negatively affect the generalist,rapidly reproducing small mammal species monitored during this study. It is important torecognize that complete elimination of animal damage is neither practical nor necessary(Owston et al. 1992). Prevention of severe seed predation can be mitigated by monitoringseed crops and scheduling harvests in years preceding high seed production. Over longertime frames, harvest operations could be scheduled during good seed years when seedproduction would be sufficient to overcome losses due to seed-eaters (Janzen 1971, Shearerand Schmidt 1971). Comparison of these seed and small mammal results with regenerationdata collected by the B.C. Ministry of Forests (Williams Lake) would determine if theregeneration observed during the first two years post-harvest was considered adequate forrestocking the harvested areas, or if it was necessary to supplement the natural regenerationwith planted stock.Alternative silvicultural systems such as the shelterwood system are a step towardsproviding diversity within the managed forest landscape. By combining the maintenance ofecological systems with the extraction of timber, shelterwood systems present an option tothe preservation versus timber production stalemate that dominates forest landscapes in B.C.65(Hopwood 1991). By removing the overstory canopy in a series of cuts, forest cover isprovided during the establishment of regeneration and the advance growth on the sites isincorporated into the next rotation. Shelterwood systems may increase the component ofDouglas-fir in future stands and have the potential to integrate the management of timbervalues with non-timber values such as wildlife, water supply, range, recreation, andesthetics, which are demanded from these interior sites.66REFERENCESAskham, L.R. 1992. Voles in Silvicultural Approaches to Animal Damage Management inPacific Northwest Forests. Edited by H.C. Black. PNW-GTR-287. USDA ForestService, Pacific Northwest Research Station, Portland.B.C. Ministry of Forests. 1991. Silvicultural systems: their role in British Columbia’s forestmanagement. Victoria, B.C.B.C. Ministry of Forests. 1993. Uniform shelterwood system for even-aged Douglasfir/lodgepole pine stands in the SBSdw1 subzone. Establishment report. WilliamsLake, B.C.Burton, P.J. 1994. Seeclbed surveys and installation of seedbed germination trials in SBSuniform shelterwood plots. Draft Report. Dept. of Forest Sciences, University ofBritish Columbia, Vancouver.Burton, P.J., A.C. Balisky, L.P. Coward, S.G. Cumming and D.D. Kneeshaw. 1992. Thevalue of managing for biodiversity. The Forestry Chronicle 68(2):225-237.Corn, S.P., Bury, R.B., and T.A. Spies. 1988. Douglas-fir forests in the CascadeMountains of Oregon and Washington. in The abundance of small mammals relatedto stand age and moisture. Proc. Mngmt. of amphibians, reptiles and small mammalsin North America. USDA Forest Service Report RM No. 166:340-352.Efford, M. 1992. Comment. Revised estimates of the bias in the MNA estimator.Canadian Journal of Zoology 70:628-631.Emmingham, W.H., R. Hoithausen, M. Vomocil. 1992. Silvicultural systems and standmanagement in Silvicultural Approaches to Animal Damage Management in PacificNorthwest Forests. Edited by H.C. Black. General Technical Report PNW-GTR287. USDA Forest Service, Pacific Northwest Research Station, Portland.Gashwiler, J.S. 1970. Plant and mammal changes on a clearcut in west-central Oregon.Ecology 51(6):1018-1026.Gunther, P.M., B.S. Horn, and G.D. Babb. 1983. Small mammal populations and foodselection in relation to timber harvest practices in the western Cascade Mountains.Northwest Science 57:32-44.Hawthorne, D.W. 1980. Wildlife damage and control technique. Wildlife ManagementTechnique Manual:41 1-439.67Hopwood, D. 1991. New forestry practices: a report for Management Practices Team of theOld Growth Strategy Working Group.Hunter, M.L. 1990. Wildlife, forests, and forestry - principles of managing forests forbiological diversity. Prentice-Hall, Inc. Englewood Cliffs, New Jersey. 370 pp.Huribert, S.H. 1984. Pseudoreplication and the design of ecological field experiments.Ecological Monographs 54(2): 187-211.Janzen, D.H. 1971. Seed predation by animals. Annual Review of Ecology and Systematics2:465-492.Jolly, G.M. and J.M. Dickson. 1983. The problem of unequal catchability in mark-recapture estimation of small mammal populations. Canadian Journal of Zoology61:922-927.Kohier, C.M. 1993. Influence of large-scale food supplementation on diversity of rodentcommunities. M.Sc. Thesis. Dept. of Forest Sciences, University of BritishColumbia, Vancouver.Klenner, W. and C.J. Krebs. 1991. Red squirrel population dynamics. I. The effect ofsupplemental food on demography. Journal of Animal Ecology 60:961-978.Krebs, C.J. 1966. Demographic changes in fluctuating populations of Microtus calfornicus.Ecological Monographs 36:239-273.Krebs, C.J. 1991. Small Mammal Programs for Mark-Recapture Data Analysis. Universityof British Columbia, Dept. of Zoology, Vancouver.Krebs, C.J., B.L. Keller, and R.H. Tamarin. 1969. Microtus population biology:demographic changes in fluctuating populations of M. ochrogaster and M.pennsylvanicus in southern Indiana. Ecology 50:587-607Krebs, C.J. and R. Boonstra 1984. Trappability estimates for mark-recapture data.Canadian Journal of Zoology 62:2440-2444.Krefting, L.W., and C.E. Ahlgren. 1974. Small mammal and vegetation changes after firein a mixed conifer-hardwood forest. Ecology 55:1391-1398.Lawrence, W.H., Kverno, N.B., Hartwell, H.D. 1961. Guide to wildlife feeding injurieson conifers in the Pacific Northwest. Western Forestry Conservation Assoc.Portland, OR.68Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton UniversityPress. Princeton, New Jersey. 179 pp.Maser, C., J.M. Trappe, and Nussbaum. 1978a. Fungal-small mammal interrelationshipswith emphasis on Oregon coniferous forests. Ecology 59(4):799-809.Maser, C., J.M. Trappe, D.C. Ure. 1978b. Implications of small mammal mycophagy tothe management of western coniferous forests. Trans. North Am. Wildl. and Nat.Resour. Conf. 43:78-88.Meidinger, D. and 3. Pojar. 1991. Ecosystems of British Columbia. Research Branch,Ministry of Forests, Victoria, British Columbia. Special Report Series Number 6.Medin, D.E. 1986. Small mammal responses to diameter-cut logging in an Idaho Doulgasfir forest. Res. Note INT-362, USDA Forest Service, Intermountain ResearchStation.Medin, D.E. and G.D. Booth. 1989. Responses of birds and small mammals to single-treeselection logging in Idaho. Res. Note INT-408, USDA Forest Service, IntermountainResearch Station.Moore A.W. 1940. Wild animal damage to seed and seedlings on cut-over Douglas-fir landsof Oregon and Washington. Technical Bulletin No. 706, USDA Forest ServiceWashington, D.C. 28 pp.Nichols, J.D. and K.H. Pollock. 1983. Estimation methodology in contemporary smallmammal capture-recapture studies. Journal of Mammalogy 64(2):253-260.Pielou, E.C. 1966. The measurement of diversity in different types of biological collections.Journal of Theoretical Biology 13:131-144.Owston, P.W., S.P. Smith, and W.I. Stein. 1992. Stand establishment in SilviculturalApproaches to Animal Damage Management in Pacific Northwest Forests. Edited byH.C. Black. General Technical Report PNW-GTR-287. USDA Forest Service,Pacific Northwest Research Station, Portland.Schlotzhauer, S.D. and R.C. Littell. 1987. SAS system for elementary statistical analysis.SAS Institute Inc., Cary, NC.Scrivner, J.H. and H.D. Smith. 1984. Relative abundance of small mammals in foursuccessional stages of spruce-fir forest in Idaho. Northwest Science 58:171-176.69Seber, G.A.F. 1982. The Estimation of Animal Abundance and Related Parameters. 2idEd., Charles Griffin and Co., London.Shearer, R.C. and W.C. Schmidt 1971. Ponderosa pine cone and seed losses. Journal ofForestry 69:370-372.Simpson, E.H. 1949. Measurement of diversity. Nature 163:688.Smith D.M. 1986. The Practice of Silviculture. 8th Edition. John Wiley and Sons, Inc.,New York 527 pp.Smith, C.F., and Shaler. 1947. The influence of mammals and birds in retarding artificialand natural re-seeding of coniferous forests in the U.S. Journal of Forestry 45:36 1Sullivan, T.P. 1979a. The use of alternative foods to reduce conifer seed predation by thedeer mouse (Peromyscus maniculatus). Journal of Applied Ecology 16:475-495.Sullivan, T.P. 1979b. Repopulation of clearcut habitat and conifer seed predation by deermice. Journal of Wildlife Management 43:861-871.Sullivan, T.P. and C.J. Krebs. 1981. An irruption of deer mice after logging of coastalconiferous forest. Canadian Journal of Forestry Research 1 1(3):586-592.Sullivan, D.S., and T.P. Sullivan. 1982. Effects of logging practices and Douglas-firseeding on shrew populations in coastal coniferous forest in B.C. Canadian FieldNaturalist 96(4):455-461.Sutherland, C., C. Burger, and K. Day. 1990. Uniform shelterwood systems for even-agedDouglas-fir/lodgepole pine stands in the SBSk subzone. Working Plan.Tevis, L., Jr. 1956. Responses of small mammal populations to logging of Douglas-fir.Journal of Mammalogy 37(2): 189-196.van Home, B. 1983. Density as a misleading indicator of habitat quality. Journal of WildifeManagement 47(4): 893-901.Walters, B.B. 1991. Small mammals in a sub-alpine old-growth forest and clearcuts.Northwest Science 65:27-31.West, S.D. 1992. Seed-eating mammals and birds in Silvicultural Approaches to AnimalDamage Management in Pacific Northwest Forests. Edited by H.C. Black. GeneralTechnical Report PNW-GTR-287. USDA Forest Service, Pacific Northwest Researchtation, Portland.70West, S.D., R.G. Ford and J.C. Zasada. 1980. Population response of the northern red-backed vole (Clethrionomys rutilis) to differentially-cut white spruce forest. ResearchNote PNW-362, USDA Forest Service, PNW Forest and Range Experiment Station.Williamson, R.L. 1973. Results of shelterwood harvesting of Douglas-fir in the Cascades ofwestern Oregon. USDA Forest Service Pacific Northwest Forest and RangeExperiment Station, Portland, OR. 13 pp.Wood, P.M. 1994. The priority of biological diversity conservation in forest land usedecision making. Chapter 4, Ph.D. Dissert. University of British Columbia,Vancouver, B.C.71APPENIIx 1. Comparison of Jolly-Seber estimates and MNA direct enumeration ofdeer mouse population levels in the control and treated stands of the UBC replicate.DEER MICE‘CONTROL -• 30 % 50 %No. of animals/ha50UBC4030100 —r•••—i—•••— N+__;._.7— - , ,AUG OCT MAY OCT MAY OCT1991 1992 1993— CONTROL I 30 % * 50 %No. of animals/ba40MNA DENSITY ESTIMATES30..,.,.... ., ..0AUG OCT MAY OCT MAY OCT1991 1992 199372APPENDIx 2. Comparison of Jolly-Seber estimates and MNA direct enumeration ofdeer mouse population levels in the control and treated stands of the Gavin replicate.DEER MICE— CONTROL I 30 % --*- 50 %No. of animals/ha50GAVIN0——i——i—— 111111111111111111 1111111111111111JULY OCT MAY OCT MAY OCT1991 1992 1993— CONTROL -+- 30 % —*--- sO %No. of animals/ha40MNA DENSITY ESTIMATES300— i-Ti I I I I i • I I I I I I I I I I I I I I I I I I I I I I I I I I IJULY OCT MAY ocr MAY OCT1991 1992 199373APPENDIX 3. Comparison of Jolly-Seber estimates and MNA direct enumeration ofred-backed vole population levels in the control and treated stands of the UBCreplicate.RED-BACKED VOLESCONTROL -+- 30 % - 50 %No. of animals/ha12010080 UBC60402011111111111 iiII;iii I 1111.1 IIAUG OCT MAY OCT MAY OCT1991 1992 1993CONTROL + 30 % — 50 %No. of animals/ha120100MNA DENSITY ESTIMATESAUG OCT MAY OCT MAY OCT1991 1992 1993


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items