UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influence of conifer release treatments on habitat structure and small mammal populations in south-central… Runciman, James Bruce 1994

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

Item Metadata

Download

Media
831-ubc_1994-0606.pdf [ 1.15MB ]
Metadata
JSON: 831-1.0075186.json
JSON-LD: 831-1.0075186-ld.json
RDF/XML (Pretty): 831-1.0075186-rdf.xml
RDF/JSON: 831-1.0075186-rdf.json
Turtle: 831-1.0075186-turtle.txt
N-Triples: 831-1.0075186-rdf-ntriples.txt
Original Record: 831-1.0075186-source.json
Full Text
831-1.0075186-fulltext.txt
Citation
831-1.0075186.ris

Full Text

INFLUENCE OF CONIFER RELEASE TREATMENTS ON HABITAT STRUCTURE AND SMALL MAMMAL POPULATIONS IN SOUTH-CENTRAL BRITISH COLUMBIA by James Bruce Runciman B.Sc., The University of British Columbia, 1991  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES  FACULTY OF FORESTRY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA August 1994 ©  James Bruce Runciman, 1994  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of  1 —cJ  c_—,cm  The University of British Columbia Vancouver, Canada Date  DE.6 (2/88)  4  /  cj  ABSTRACT  I examined the effects of manual cutting and cut-stump applications of glyphosate herbicide on vegetation, woody debris, and small mammal populations in young mixed-conifer plantations of south-central British Columbia, Canada. The experimental design consisted of nine separate and independent plantations: 3 controls, 3 manual treatments, and 3 cut-stump treatments.  Treatments were conducted between 21 September and 17 October 1992.  Vegetation and woody debris were sampled once within each plantation during the last pre treatment year (1992) and again during the first post-treatment year (1993). Small mammal populations were sampled at three-week intervals within each plantation from September 1991 to October 1993 during snow-free periods.  Total volumes of space occupied by herbs,  coniferous trees, and woody debris were not affected by manual and cut-stump treatments for conifer release. However, both treatments reduced total volumes of shrubs and deciduous trees. The number of pieces of small diameter woody debris increased following the cutting of competing vegetation and increased the complexity of low ground cover on treated plantations. There were no discernable effects of manual or cut-stump treatments on the population size of deer mice (Peromyscus maniculatus), yellow-pine chipmunks (Tamias amoenus), southern redbacked voles (Clethrionomys gapperi), long-tailed voles (Microtus longicaudus), or meadow voles (M. pennsylvanicus). Sex ratios, body weights, reproduction, recruitment, and survival of deer mice were also similar on treatment and control plantations. Changes in habitat structure during the first post-treatment year did not appear to exceed the tolerance of small mammal populations for early successional change. Continued sampling on these plantations will establish whether there are any long-term treatment effects.  ii  TABLE OF CONTENTS ABSTRACT  ii  TABLE OF CONTENTS  .  LIST OF TABLES  .  LIST OF FIGURES  iii iv vi  ACKNOWLEDGEMENTS  .  INTRODUCTION  viii 1  MATERIALS AND METHODS Description of Study Areas Experimental Design Vegetation Woody Debris Small Mammal Populations Demographic Parameters of Small Mammals Treatment Quality Statistical Analyses  4 4 6 9 10 11 13 15 16  RESULTS Vegetation Woody Debris Trappability Population Size Sex Ratios Body Weights Reproduction Recruitment Survival Treatment Quality  19 19 23 27 30 35 35 39 39 44 44  DISCUSSION Experimental Design Vegetation and Woody Debris Small Mammal Populations  46 46 47 48  CONCLUSIONS  52  REFERENCES  54  APPENDIX I  61  .  .  .  .  .  .  .  .  .  iii  LIST OF TABLES Table 1.  Site descriptions for each of the nine plantations included in this study. Site data were provided by the B. C. Ministry of Forests, Salmon Arm Forest District. 7  Table 2.  Minimum unweighted trappability estimates for deer mice and yellow-pine chipmunks on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are mean percent trappability, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate 29  Table 3.  Total number of individuals captured for red-backed voles, long-tailed voles, and meadow voles on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean total number of individuals, the 95 % confidence interval, and the sample size (number of replicates). Flash marks (--) indicate that there were insufficient data to calculate an estimate 34  Table 4.  Sex ratios (proportion of males in population) for deer mice on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean sex ratio, the 95% confidence interval, and the sample size (number of replicates) 36  Table 5.  Length of breeding season (number of weeks between first and last capture of a breeding male or female) for deer mice on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean total number of weeks, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate 40  Table 6.  Proportion of adult male deer mice in breeding condition on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and posttreatment (Summer 1993) sampling. Values are the mean proportion breeding, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate. 41  Table 7.  Proportion of adult female deer mice in breeding condition on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean proportion breeding, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate 42  iv  LIST OF TABLES (continued) Table 8.  Minimum survival estimates for deer mice on control, manual, and cut-stump treatments during pre-treatment (Summer 1991, Winter 1991-1992, and Summer 1992) and post-treatment (Winter 1992-1993 and Summer 1993) sampling. Values are mean minimum survival, the 95 % confidence interval, and the sample size (number of replicates) 45  V  LIST OF FIGURES Figure 1.  Location of the Eagle Bay and Sicamous study areas in south-central British Columbia 5  Figure 2.  Mean absolute volume (m 3 per 0.01 ha) of prominent herbaceous species and total herbs (TOT) on control, manual, and cut-stump treatments in the last pre treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were Queen’s cup (CU), fireweed (EA), wild strawberry (FV), white-flowered hawkweed (HA), white sweet-clover (MA), and bracken (PA). Each value is the mean of three replicates ± the 95 % confidence interval. 20  Figure 3.  Mean absolute volume (m 3 per 0.01 ha) of prominent shrub species and total shrubs (TOT) on control, manual, and cut-stump treatments in the last pre treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were Douglas maple (AG), speckled alder (AT), falsebox (PM), bitter cherry (PE), thimbleberry (RP), and willow (SS). Each value is the mean of three replicates ± the 95% confidence interval 21  Figure 4.  Mean absolute volume (m 3 per 0.01 ha) of prominent deciduous tree species and total deciduous trees (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were paper birch (BP), trembling aspen (PT), and black cottonwood (PC). Each value is the mean of three replicates ± the 95% confidence interval 22  Figure 5.  Mean absolute volume (m 3 per 0.01 ha) of prominent coniferous tree species and total coniferous trees (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were western larch (LO), hybrid white spruce (PE), lodgepole pine (PC), Douglas fir (PM), western red cedar (TP), and western hemlock (TH). Each value is the mean of three replicates ± the 95 % confidence interval 24  Figure 6.  Mean volume (m 3 per 0.01 ha) of woody debris on control, manual, and cutstump treatments in the last pre-treatment year, 1992 (clear bar), and first posttreatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95% confidence interval 25  Figure 7.  Mean number of pieces of woody debris counted (number of pieces per 90 m transect) in each of three diameter classes on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95% confidence interval 26  vi  LIST OF FIGURES (continued) Figure 8.  Mean number of pieces of woody debris counted (number of pieces per 90 m transect) in each of five decay classes on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95% confidence interval 28  Figure 9.  Minimum number alive (MNA) estimates of population size for deer mice on replicate control, manual, and cut-stump treatments. Hatched bar on horizontal axis indicates time of treatments 31  Figure 10.  Minimum number alive (MNA) estimates of population size for yellow-pine chipmunks on replicate control, manual, and cut-stump treatments. Hatched bar on horizontal axis indicates time of treatments 32  Figure 11.  Median weight at sexual maturity (g) for male (0) and female (•) deer mice on control (C), manual (M), and cut-stump (CS) treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95 % confidence interval 37  Figure 12.  Mean adult male body weights (g) for deer mice on control (C), manual (M), and cut-stump (CS) treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95% confidence interval except the manual treatment mean in 1993 (*) which was calculated over two replicates 38  Figure 13.  Total number of recruits for deer mice in control (C), manual (M), and cut-stump (CS) treatment populations during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95% confidence interval except the cut-stump treatment mean in 1991 (*) which was calculated over two replicates 43  vii  ACKNOWLEDGEMENTS  I thank my supervisor, Dr. Tom Sullivan, for his encouragement and support as well as for providing me with the opportunity to conduct applied research. I also thank my committee members, Drs. Alton Harestad and Peter Marshall, for their advice throughout my studies and for reviewing earlier drafts of this manuscript. Drs. Valerie LeMay and Wes Hochachka offered keen insights and constructive criticism on statistical matters as did Markus Merkens on general content and presentation.  I am most grateful to the British Columbia Ministry of Forests,  Research Branch, and the Salmon Arm Forest District Office for financial and logistical support. Additional funding was provided by the Asa Johal Graduate Fellowship in Forestry and the Canadian Forest Products Ltd. Fellowship in Forest Wildlife Management. My field assistants were indispensable and I offer my most sincere thanks and praise to all of them: Michael Burwash, Blair Hammond, Claudia Houwers, Steve Milne, Glen Ozawa, Marion Porter, Erin Roberts, and Gordon Ryznar. Students in the Forestry-Wildlife program provided stimulating discussions and good times to smooth my progress through graduate school.  Finally, I am  grateful and humbled by the love and support that Kiran Kennedy gave to me throughout my studies at U.B.C. This thesis is as much hers as it is mine.  viii  INTRODUCTION  Vegetation management is an important silvicultural practice in North American forests. On intensively managed forest lands, a variety of manual, mechanical, and chemical treatments are applied primarily to young conifer plantations to release naturally regenerating and planted crop trees from competition with angiosperms (Newton & Comeau 1990; McDonald & Radosevich 1992). Although these conifer release treatments alter the structure and composition of the vegetative community to favour commercially valuable tree species (Morrison 1982; Santillo 1987; Lautenschlager 1990), individual competing species are seldom eliminated (May, May & McCormack 1982; Morrison & Meslow 1983; Lautenschlager & McCormack 1989; Newton et a!. 1989; McMillan et a!. 1990; Freedman, Morash & MacKinnon 1993). However, the changes in habitat structure that follow vegetation management (Freedman 1989), and the potential for direct chemical toxicity from herbicides (Newton et al. 1984), suggest that wildlife communities may be affected. Small mammals may be particularly vulnerable to the effects of either direct chemical toxicity or habitat alteration due to their role as primary consumers and the sensitivity of some species to changes in microhabitat (M’Closkey & Fieldwick 1975; Miller & Getz 1977; Dueser & Shugart 1978; Kitchings & Levy 1981; Yahner 1986). Aerial sprays of the herbicide glyphosate are the most commonly prescribed treatment for vegetation management in Canadian forests (Campbell 1990).  Glyphosate is a broad  spectrum, relatively non-selective herbicide available commercially as Vision® for use in forest management (Sutton 1978; Roskamp & Reeves 1980; Campbell 1990). Research indicates that, at recommended rates and under normal conditions of use, glyphosate poses minimal toxicological risk to wildlife (Morrison & Meslow 1983; Newton & Dost 1984; Newton et al. 1984; Atkinson 1985) and presents no threat of bioaccumulation (Newton et al. 1984). 1  Furthermore, both field and laboratory studies have found no direct effect of glyphosate on the survival or reproduction of small mammals (Wahigren 1979; Sullivan & Sullivan 1981; Ritchie, Harestad & Archibald 1987; Sullivan 1990a). Small mammal population responses to aerial applications of glyphosate have varied. Results range from studies that found an overall increase in population density following treatments (Anthony & Morrison 1985; McMillan et al. 1990), to those recording a decrease (dough 1987; Ritchie, Harestad & Archibald 1987; Santillo, Leslie & Brown 1989), to others in which no significant changes in density were observed (Suffivan & Sullivan 1982; D’Anieri, Leslie & McCormack 1987; Milton & Towers 1990; Sullivan 1990b).  Where changes in the population densities of small mammals have been  recorded, individual species responded to broadcast glyphosate treatments in accordance with their habitat preferences and the degree to which treatments altered habitat structure (Sullivan & Sullivan 1982; Anthony & Morrison 1985; Clough 1987; D’Anieri, Leslie & Mcdormack 1987; Ritchie, Harestad & Archibald 1987; Santillo, Leslie & Brown 1989; McMillan et al. 1990; Sullivan 1990b). In addition, studies that continued beyond the first post-treatment year found changes were short-term with populations returning to pre-treatment densities within one or two years (Anthony & Morrison 1985; Sullivan 1990b). The herbicides 2,4-D and 2,4,5-T have also been used for site preparation and conifer release in Canadian forests (Freedman et al. 1988; Campbell 1990). Toxicity studies indicate that neither of these herbicides is directly harmful to wildlife at recommended rates of application (Bovey & Young 1980; Hudson, Tucker & Haegle 1984). However, small mammal community composition was altered due to changes in habitat structure following sprays of 2,4-D (Johnson 1964; Tietjen et al. 1967; Johnson & Hansen 1969; Borrecco, Black and Hooven 1979; Spencer & Barrett 1980) and 2,4,5-T (Kirkland 1978; Freedman et al. 1988) over forest  2  plantations and rangelands. Again, these changes were only short-term as populations recovered when vegetation returned to pre-treatment conditions (Johnson & Hansen 1969; Black & Hooven 1974). The use of broadcast herbicide treatments to control competing vegetation on Canadian forest lands is the subject of increasing public debate (Environics Research Group 1989; Freedman 1989; Campbell 1990). This issue is only part of a more general public concern about chemicals in the environment (Lautenschlager 1986; Freedman 1991). Many forest managers have responded to growing public pressure by prescribing a variety of alternative vegetation management treatments including manual cutting and ground-based herbicide applications (Campbell 1990). However, the effects of these alternative systems on forest wildlife remain largely unstudied (Wagner 1993)’.  It is essential that the potential for these alternative  treatments to affect wildlife habitat is determined and that their effects on wildlife populations  are assessed. This study was designed to determine the effects of manual cutting and cut-stump applications of glyphosate on small mammal communities and habitat structure in young mixedconifer plantations of south-central British Columbia, Canada. There were three objectives: (1) to measure changes in volumes of vegetation 2 and woody debris in young plantations subjected to manual and cut-stump treatments for conifer release; (2) to determine the effects of these treatments on the presence and abundance of individual small mammal species inhabiting treated  ‘Two exceptions are Slagsvold (1977) and Oxenham (1983) who compared the effects of manual cutting and broadcast herbicide treatments for conifer release on populations of songbirds and snowshoe hares (Lepus americanus), respectively. 2 W here the term volume is used to describe vegetation, it refers to the estimated volume of space occupied by herbs, shrubs, deciduous trees, and coniferous trees rather than the total stem volumes of merchantable trees. 3  plantations; and (3) to examine some important demographic characteristics, for selected small mammal species, that may be affected by changes in habitat. Vegetation management treatments for conifer release typically create a more open forest habitat with an overstorey dominated by coniferous trees and vigorous growth of both herbs and low shrubs.  If manual cutting or cut-stump applications of glyphosate herbicide alter small  mammal community composition, it is likely that individual species will respond according to their own habitat preferences. Species associated with open, early successional, herb-dominated habitats will increase in abundance while those sensitive to reduced deciduous overstorey cover will decline. Changes in the demographic characteristics of individual small mammal species may help to explain patterns of abundance in the pre- and post-treatment periods.  MATERIALS AND METHODS  Description of Study Areas  This study was conducted in two areas within the Shuswap Highlands of south-central British Columbia, Canada (Fig. 1). Both the Eagle Bay (50° 55’N; 119° 11 ‘W) and Sicamous (50° 52’N; 118° 59’W) study areas were within the Thompson moist-warm subzone of the Interior Cedar-Hemlock biogeoclimatic zone (Lloyd et al. 1990). Topography in both areas was hilly to steeply sloping and site elevations ranged from 550-1237 m. The climate in the region is characterized by warm dry summers and cool wet winters. Mean annual temperature ranges from 2 to 8.7°C (Ketcheson et al. 1991). Mean annual precipitation ranges from 500-1200 mm with up to 50% occurring as snow (Ketcheson et al. 1991). Western red cedar (Thujaplicata)  4  Sicamous  Mabel Lake  Kamloops  0  10  Fig. 1. Location of the Eagle Bay and Sicamous study areas in south-central British Columbia.  5  20km  and western hemlock (Tsuga heterophylla) are the dominant tree species in climax forests while Douglas fir (Pseudotsuga menziesii), lodgepole pine (Pinus contorta), and paper birch (Betula papyrifera) are common seral species (Lloyd et al. 1990).  Nomenclature for plant species  follows Hitchcock & Cronquist (1973). A more detailed description of the vegetation and soils of this region can be found in Lloyd et al. (1990) and Ketcheson et al. (1991).  Experimental Design  Study plots were established in a total of nine young mixed conifer plantations between September of 1991 and October of 1993. These plantations were selected in conjunction with staff from the British Columbia Ministry of Forests on the basis of operational scale, proximity, and their initial similarity in requiring vegetation management treatments. Three plantations were located near Sicamous and the remaining six were located near Eagle Bay (Table 1). The Sicamous plantations were located on an east to south facing slope. These areas were logged between 1977 and 1988 and planted between 1984 and 1989 largely with Douglas fir. Inspection of these plantations indicated that lodgepole pine, western red cedar, and western hemlock were also present with lesser amounts of hybrid white spruce (Picea engelmannil x P. glauca) and western white pine (Pinus monticola). The most common deciduous tree species on these plantations were black cottonwood (Populus trichocarpa), paper birch, and trembling aspen (Populus tremuloides).  Controlled burning was applied to these plantations as site  preparation prior to planting. Study plots were established in all three Sicamous plantations (A C) in September of 1991.  6  ICHmw3  ICHmw3  100.0 ICHmw3  92.4  ICHmw3  100.0  ICHmw3  99.0  Hmw3  -  Control  Lodgepole pine Hybrid spruce  ICHmw3  -  Control  Lodgepole pine Western larch Hybrid spruce Douglas fir  ICHmw3  mw3  91.6  Cut-stump Cut-stump  95.7  Lodgepole pine Hybrid spruce  1987-1988  1,2  1983  33.0  800  N  Eagle Bay  I  Lodgepole pine Douglas fir Hybrid spruce  1987-1990  2  1986-1988  *Site preparation treatments were 1 rough bunch only. windrow and burn, and 4 broadcast burn, 2 rough bunch and burn, 3 **Spies planted are ranked according to the proportion of each species in the total number of seedlings planted.  Biogeoclimatic Subzone  -  Manual  Manual  Cut-stump  Manual  Control  Brushing Treatment  Treatment Quality (%)  Lodgepole pine  Western larch  Douglas fir Hybrid spruce  Lodgepole pine Douglas fir Hybrid spruce  1987  1985-1987  1986  1985  1984-1986  1987  1989 Douglas fir Lodgepole pine  2  2,4  1,2  1,2  1,2,3  1  1984-1985  2  Species Planted**  Year Planted  Site Preparation*  1978  1979-1982  21.5  1983  1977-1981  1983  1985-1988  Year Logged  36.0  23.0  23.0  22.2  47.0  38.3  25.8  Total Area (ha)  580  610  690  760  785  1143  1030  705  Elevation Cm)  N  N  NE  N  N  E  S  E  Aspect  Eagle Bay  Eagle Bay  Eagle Bay  Eagle Bay  Eagle Bay  Sicamous  Sicamous  Sicamous  Location  H  F G  E  D  C  B  A  Plantation  Site Descriptions  Table 1. Site descriptions for each of the nine plantations included in this study. Site data were provided by the B. C. Ministry of Forests, Salmon Arm Forest District.  The Eagle Bay plantations were located on a north to north-east facing slope. These  areas were logged between 1978 and 1988 and planted predominantly to lodgepole pine between 1985 and 1990. Site inspections indicated that western larch (Larix occidentalis), Douglas fir, western red cedar, western hemlock, and hybrid white spruce were also present. Deciduous tree species again were black cottonwood, paper birch, and trembling aspen. A variety of controlled burning treatments were used to prepare these plantations prior to planting. Study plots were established in five of the six Eagle Bay plantations (D-H) in September of 1991 and in the remaining plantation (I) during April of 1992. Between 21 September and 17 October 1992, six of the nine study plantations received either a manual or cut-stump treatment for conifer release and the remaining three plantations were reserved as controls. All treatments were prescribed and conducted by the B. C. Ministry of Forests. On three manually treated plantations (B, D, and E), all trembling aspen, paper birch, and black cottonwood were cut with power-saws except for a few individual stems that were left in openings in the plantations. All stems of willow (Salix spp.), Douglas maple (Acer  glabrum), bitter cherry (Prunus emarginata), and speckled aider Alnus incana) that were within 1 m of a crop tree and were taller than the crop tree were also cut. All other vegetation was left uncut. On three cut-stump treated plantations (C, H, and I), the manual treatment standards also applied but the cut stumps were hand-sprayed with glyphosate as Vision® diluted 2:1 in water and most maple, cherry, and alder were not cut. A small amount of Basacid Blue® dye was added to the spray mixture to mark all treated stumps. Control, manual, and cut-stump treatments were assigned subjectively among the nine plantations according to two operational constraints.  The first of these constraints was a  minimum acceptable outcome anticipated for the cost of operational vegetation management treatments. This constraint required that a cut-stump herbicide treatment be assigned to the most 8  heavily competed of the Sicamous plantations. The second constraint was related to the close proximity of private homes to the Eagle Bay study area. In this instance, cut-stump herbicide treatments were assigned to the two plantations furthest from the private homes to facilitate the approval of herbicide application permits. For the purpose of making statistical comparisons among treatments, it was assumed that no greater initial bias resulted from the subjective allocation of plantations among treatments in the Sicamous and Eagle Bay study areas than would have resulted from a completely randomized experimental design with three-fold replication. This assumption is supported by the fact that plantations were initially selected on the basis of their similarity in size and successional stage as well as the fact that all nine plantations were within the same biogeoclimatic subzone.  Vegetation  Vegetation was assessed within each plantation in the last pre-treatment year (1992) and in the first post-treatment year (1993) to determine the presence and abundance of individual plant species and to estimate changes in vegetative structure following treatments.  Five  permanent 5 x 25-rn transects, each consisting of five contiguous 5 x 5-rn plots, were established within each plantation. These transects were placed randomly so long as they were at least 50 m from the nearest plantation edge or transect and they did not cross roads, skid-trails, or landings. Within vegetation transects, each 25-rn 2 plot contained three nested subplots: the 5 x 5-rn subplot for sampling trees, a 3 x 3-rn subplot for sampling shrubs, and a 1 x 1-rn subplot for sampling herbaceous species (adapted from Stickney 1980, 1985). In addition, tree, shrub, and herb layers were each sampled within height classes of 0-0.25, 0.25-0.50, 0.50-1.0, 1.0-2.0, 9  2.0-3.0, and 3.0-5.0 m (adapted from Walmsley et al. 1980).  Data recorded were the  percentage cover of the ground for each species within each height class of the appropriate subplot.  Individual plants were measured once only within the height class containing their  topmost growth. Vegetation data were summarized as absolute volume estimates (m /0.0l ha) for each 3 plant species and vegetation layer. Percentage cover values were multiplied by the mid-point of their corresponding height class and summed for individual species within subplots. The average of these subplot values within transects gave the volume of a cylindroid that represented the space occupied by each species on each transect. The sum of all species volumes within subplots was also averaged within transects to estimate the absolute volume of the herb, shrub, and tree layers. Sampling of vegetation was conducted in July and August of both 1992 and 1993.  Woody Debris  Both manual and cut-stump treatments for conifer release involve cutting down competing vegetation and can add a complex new layer to post-harvest accumulations of woody debris. Accumulations of woody debris were sampled within each plantation in the last pre-treatment year (1992) and in the first post-treatment year (1993) according to the method of Van Wagner (1968, 1982) for estimating wood volumes on the ground. Five triangular transect lines, each measuring 30-m per side, were located randomly within each plantation using the same criteria as were used for the placement of vegetation transects (adapted from Trowbridge et al. 1986). Each piece of wood intersecting the transect line was tallied and the diameter recorded if greater than 1 cm. A diameter of 0.5 cm was 10  assigned to all pieces less than 1 cm in diameter.  Diameters were measured using metric  callipers or a standard calibrated diameter measuring tape for larger pieces. Each piece was also classified into one of five decay classes ranging from newly fallen wood (Class 1) to well decomposed and broken wood (Class 5) (Triska & Cromack 1980). Pieces of decay class 5 were not included in calculations of debris volume 3 (m / 0.0l ha) on transect lines because well-decayed pieces did not often approximate the cylindrical shape that is an assumption of the Van Wagner (1968, 1982) method. As the sampling of woody debris on transect lines was semi-destructive, these lines were relocated within plantations between the pre- and post-treatment years. Sampling of woody debris was conducted in July and August of both 1992 and 1993.  Small Mammal Populations  Small mammal communities were sampled intensively to obtain detailed information on population size and other demographic characteristics for individual species. A single 1-ha livetrapping grid was located randomly within each plantation as long as it was at least 100 m from the nearest plantation edge. Each grid consisted of 49 stations distributed in a 7 x 7 matrix at a spacing of approximately 14.3 m. A single Longworth live-trap was placed within 1 m of each grid station on all plantations except D at Eagle Bay where one additional trap was placed at alternate stations during a period of high populations of small mammals (>50 individuals captured per grid) in 1992. A sampling regime involving two-night trapping periods at three-week intervals was adopted for this study.  Traps were operated from September to October 1991 and May to  October of both 1992 and 1993. Trapping was suspended only during the application of manual and cut-stump treatments in 1992 and only for the Sicamous plantations. 11  Within individual  trapping weeks, traps were baited with whole oats for food and a slice of carrot for moisture. Coarse brown cotton was supplied as bedding material. Traps were set on the afternoon of day 1, checked on the morning and afternoon of day 2, and checked again on the morning of day 3 when they were locked open and remained accessible to small mammals until the next trapping period. Traps were also pre-baited for two to three weeks prior to the first trapping session each year to permit familiarization of animals to the traps. All small mammals captured, except shrews and weasels, were identified with individually numbered ear tags. Upon capture, ear tag number, species, capture location, weight (to ±0.5 g on a Pesola® spring balance), sex, and reproductive condition were recorded. Male reproductive condition was determined by palpating the testes and was classified as either abdominal or scrotal. Female reproductive condition was determined by the condition of the mammae and nipple size which were classified as either small (non-lactating) or large (lactating) (Krebs, Keller & Tamarin 1969).  All animals were released at their point of capture  immediately following data collection. The small mammal species captured during this study were common shrews (Sorex cinereus), dusky shrews (S. monticolus) , vagrant shrews (S. vagrans), southern red-backed voles (Clethrionomys gapperi), long-tailed voles (Microtus longicaudus), meadow voles (M. pennsylvanicus), deer mice (Peromyscus maniculatus), yellow-pine chipmunks (Tamias amoenus), and ermine (Mustela erminea). However, the analysis of demographic parameters was restricted to voles, deer mice, and chipmunks because shrews, although frequently captured, were not tagged individually and survived poorly in traps and ermine were rarely encountered. Nomenclature for small mammal species follows Nagorsen (1990).  12  Demographic Parameters of Small Mammals  Minimum unweighted trappability (Krebs & Boonstra 1984) and the minimum number of animals known to be alive (Krebs 1966) were calculated for individual populations of deer mice and chipmunks. Minimum trappability estimates the fraction of a total population that is being caught in traps and indicates whether other population parameters estimated are biologically realistic or potentially biased. However, it is a conservative estimate because all animals are weighted equally irrespective of capture history and calculations do not include individuals captured only once or twice. The minimum number of animals known to be alive is an enumeration-based measure of population size.  Minimum number alive (MNA) was  selected for this study because the generally-preferred Jolly-Seber probabilistic estimator (Seber 1982) became unreliable and impossible to calculate for deer mice and chipmunks with low recaptures of previously marked animals (Krebs et al. 1986). Boonstra (1985) found that the MNA and Jolly-Seber techniques provided similar estimates of small mammal population size under field conditions when trappability exceeded 50 percent. The total number of individuals captured was used to compare populations of southern red-backed voles, long-tailed voles, and meadow voles. Neither minimum trappability nor MNA values provided meaningful estimates of presence or abundance for these species due to low total captures within individual trapping weeks and low recaptures over time. Only deer mice were consistently captured in sufficient numbers to estimate population parameters for sex ratios, body weights, reproduction, recruitment, and survival. Sex ratios, median weight at sexual maturity, mean adult male body weights, length of breeding season, the proportion of adults breeding, the total number of recruits, and minimum survival rates were estimated for deer mice in all nine study plantations. 13  Sex ratios were used to describe the  relative abundance of male and female deer mice in each population. Median weight at sexual maturity was calculated separately for males and females in each population using the technique of Leslie, Perry & Watson (1945). These estimates represent the weights at which 50% of males were scrotal and 50% of females were lactating. Mean adult male body weights were calculated as the average weight of all such individuals captured rather than the average weight of all captures to avoid any bias due to some individuals having long capture histories. Mean adult female body weights were not calculated because female weights were confounded by pregnancies and by the proportion of females breeding in each population. Length of breeding season was calculated as the number of weeks during which a male was scrotal or a female lactating, and the proportion of adults breeding was an estimate of the intensity of reproductive activity for deer mice within each population. Individuals captured within a minimum of two consecutive trapping weeks (i.e. they were present in the area for a minimum of three weeks) were considered residents and classified as recruits (Kienner & Krebs 1991). The number of recruits was tallied for individual trapping weeks and total recruitment was the sum of these values. Finally, mean monthly survival was calculated as the geometric mean of minimum survival rates between trapping weeks (Krebs 1966). Minimum survival rates were estimated from the fraction of deer mice captured and released at time t and known to be alive at time t+ 1. The use of the geometric mean, rather than the arithmetic mean, ensured that mean monthly survival was unweighted for differences in the number of animals captured in individual trapping weeks. MNA estimates of population size were calculated separately for deer mice and chipmunks in individual trapping weeks irrespective of year or treatment period.  All other  demographic parameters were estimated for Summer 1991 (September-October 1991), Summer 1992 (May-September 1992), and Summer 1993 (May-October 1993). In addition, minimum 14  survival was calculated for Winter 1991-1992 (October 1991-May 1992) and Winter 1992-1993 (October 1992-May 1993). In some instances there were insufficient data to estimate individual parameters within a time period. These cases are clearly noted. The separation of Summer and Winter seasons within years was necessary to isolate differences in estimated survival due only to the suspension of live-trapping during the winter months.  Trappability, population size,  recruitment, and survival were calculated using Small Mammal Programs for Mark-Recapture Data Analysis (Dr. C. J. Krebs, Department of Zoology, University of British Columbia).  Treatment Quality  Following the application of manual and cut-stump treatments in September and October 1992, treatment quality and field application rates of glyphosate herbicide were determined by staff from the B. C. Ministry of Forests. One fixed plot, 3.99 m in radius, was established per ha in each treated plantation.  Plots were established no later than one week following  treatments. Treatment quality was calculated within plots as the percentage of total treatable stems present that were acceptably treated and these values were averaged across all plots within each plantation.  Application rates of glyphosate herbicide were estimated for cut-stump  treatments as the product of the application rate per stem [1 ml (0.356 g of active ingredient) per 2.5 cm stem diameter], the average number of stems per hectare, and the average stem diameter within plantations.  15  Statistical Analyses  A single-factor analysis of variance (ANOVA) was used to examine both initial differences and change due to treatments for vegetation and woody debris volumes (Zar 1984). The same analysis was also used to investigate initial differences and treatment effects for all small mammal population parameters other than MNA population size.  In all analyses of  variance, plantations were random factors nested within treatments which were fixed factors. For pre-treatment tests of difference in vegetation and woody debris volumes, the individual transects sampled were further nested as random factors within plantations. Two pre-treatment null hypotheses were testable using the analysis of variance (ANOVA) for vegetation and woody debris. The first pre-treatment hypothesis stated that there were no initial differences in the absolute volume of herbs, shrubs, deciduous trees, coniferous trees, or in the volume of woody debris when replicate plantations were grouped according to future treatments.  If this first hypothesis was rejected, the second stated that there were no initial  differences in the volume of vegetation or woody debris among plantations irrespective of treatment type. To investigate changes in vegetation and woody debris following treatments, pre-treatment volumes were subtracted from post-treatment volumes on individual plantations and an analysis of variance performed on the differences. As there was only one small mammal sampling grid per plantation, the only pre-treatment hypothesis tested by analysis of variance (ANOVA) was that there were no initial differences in small mammal population parameters when replicate plantations were grouped according to future treatments.  To analyze changes in small mammal population parameters following  treatments, pre-treatment values were again subtracted from post-treatment values on individual plantations and an analysis of variance performed on the differences. 16  A randomization test (RANDMTZE) was used to detect non-random change in MNA population size for deer mice and chipmunks following treatments (Edington 1987; Manly 1991). For this study, the null hypothesis tested was that any changes in population size from pre- to post-treatment did not differ among the three treatment groups.  More specifically, the null  hypothesis stated that the same differences in population size (pre- to post-treatment) would be observed if each week’s MNA estimates were shuffled randomly among study plantations and treatment groups. To test the null hypothesis, a value was calculated to quantify average pre to post-treatment change in population size within treatment groups. A separate test statistic (S) was then calculated for each pair-wise comparison of treatment groups as described in Appendix I. The error distribution of these test statistics was determined separately for each pair-wise comparison by randomly shuffling population size estimates across plantations but within trapping weeks and calculating new test statistics (S’) for each randomization. The probability value (F) associated with each pair-wise comparison was the proportion of 5000 randomizations (Carpenter et al. 1989; Manly 1991) that produced a new test statistic (5’) greater than, or less than, the original test statistic (S). Randomization methods are particularly well-suited for detecting non-random change in studies with both large-scale manipulations and undisturbed control areas, little or no replication of experimental units, and lengthy paired time-series of data from individual treatment and control systems (Carpenter et al. 1989). As this technique determines the error distribution of its test statistics by randomly reordering the data set, it is not bound by the assumptions of random sampling, normally distributed populations, and equality of variances that often restrict the use of parametric statistics (Carpenter et al. 1989; Manly 1991). Also, randomization tests were robust to serial autocorrelation because pre- and post-treatment estimates of population size were each averaged to one number prior to calculating the test statistics. Finally, randomization 17  tests are generally as powerful, and in some cases have more statistical power, than their parametric counterparts (Edington 1987; Manly 1991). Unless otherwise noted, means and comparisons are based on three replicates for each treatment. For each analysis of variance (ANOVA), replicate measurements are the mean value for all transects sampled or, for small mammals, the mean of all individuals monitored on each trapping grid. When no animals were captured on a site, that replicate was omitted from further analysis. Where data from two pre-treatment years were available, each year was analyzed separately and differences due to treatment were calculated between the last pre-treatment year and the first post-treatment year. All proportions or percentage data were arcsine transformed before performing analyses of variance to better approximate a normal distribution. Analyses of variance were conducted using the SAS statistical analysis package (SAS Institute Inc. 1988). Randomization tests (RANDMIZE) were conducted using a program designed, written in ANSI C, and compiled using version 3 of the Borland C/C + + compiler by Dr. W. Hochachka of the Department of Zoology at the University of British Columbia. Alpha (a) was set at the 0.05 level for all analyses of variance.  An adjusted significance level of 0.0083 was used for  randomization tests to control for the simultaneous comparison of three treatments (Neter & Wasserman 1974).  18  RESULTS  Vegetation  The mean absolute volume of herbs was not significantly different among plantations in 1992 which was the only pre-treatment sampling year for vegetation (F =O. 18, P=O.83) (Fig. 26 2).  Prominent herb species on pre-treatment plantations included Queen’s cup (Clintonia  unWora), fireweed (Epilobium angustfolium), wild strawberry (Fragaria virginiana), whiteflowered hawkweed (Hieracium albflorurn), white sweet-clover (Melilotus alba), and bracken (Pteridium aquilinum). The average total volume of herbs appeared lower on manually and cutstump treated plantations than on controls in the first post-treatment year but the change was not statistically significant (F =4.29, P=0.06). Reduced post-treatment herb volumes were largely 26 due to a decline in fireweed and bracken on treated plantations. All six prominent herb species increased on control plantations between the pre- and post-treatment years. The mean absolute volume of shrubs was also not significantly different among plantations in 1992 (F =0.85, P=0.47) (Fig. 3). Prominent pre-treatment shrubs included 26 Douglas maple, speckled alder, falsebox (Pachistirna myrsinites), bitter cherry, thimbleberry (Rubus parvjflorus), and willow. Pre- to post-treatment change in the total volume of shrubs was significantly different among treatments (F, =7.62, P=0.02). 6  Shrub volumes were  between 26% and 40% lower on manually and cut-stump treated plantations than on controls in 1993 due to reduced height and cover of bitter cherry, thimbleberry, and willow. The mean absolute volume of deciduous trees was not significantly different among plantations in 1992 (F =2. 19, P=0. 19) (Fig. 4). The deciduous tree species present in pre 26  19  25 Control  20 15 10 5 0  -I  C  CU  EA  FV  HA  MA  PA  TOT  CU  EA  FV  HA  MA  PA  TOT  CU  EA  FV  HA  MA  PA  TOT  25 20  S S C  C  15 10 5 0  ci)  25 20 15 10 5 0  Species Figure 2.  Mean absolute volume (m 3 per 0.01 ha) of prominent herbaceous species and total herbs (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were Queen’s cup (CU), fireweed (EA), wild strawberry (FV), white-flowered hawkweed (HA), white sweet-clover (MA), and bracken (PA). Each value is the mean of three replicates ± the 95% confidence interval.  20  60 50 40 30 20 10 0  AG  Al  PM  PE  RP  SS  TOT  AG  Al  PM  PE  RP  SS  TOT  RP  SS  TOT  CS  60  60 Cut— Stump  AG  Al  PM  PE  Species Fig. 3. Mean absolute volume (m 3 per 0.01 ha) of prominent shrub species and total shrubs (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were Douglas maple (AG), speckled alder (Al), falsebox (PM), bitter cherry (PE), thimbleberry (RP), and willow (SS). Each value is the mean of three replicates ± the 95% confidence interval.  21  250 200 150 100 50 0  o 0  a)  $ S  BP  PT  PC  TOT  BP  PT  PC  TOT  BP  PT  PC  TOT  250 200 150  -4  o  100  .50  50  •0 ‘cii ‘U C)  250 200 150 100 50 0  Species 3 per 0.01 ha) of prominent deciduous tree species and total deciduous trees Fig. 4. Mean absolute volume (in (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were paper birch (BP), trembling aspen (PT), and black cottonwood (PC). Each value is the mean of three replicates ± the 95% confidence interval.  22  treatment plantations were paper birch, trembling aspen, and black cottonwood. Changes in the total volume of deciduous trees were significantly different among treatment groups in the year following manual and cut-stump treatments (F = 8.74, P=0.0l). 26  Dramatic reductions in  volume for all three deciduous tree species on treated plantations and positive growth on controls each contributed to this difference. The mean absolute volume of coniferous trees was not significantly different among plantations during pre-treatment sampling (F = 1.04, P=0.40) (Fig. 5). 26  Prominent pre  treatment conifer species were western larch, hybrid white spruce, lodgepole pine, Douglas fir, western red cedar, and western hemlock. Average pre- to post-treatment change in the total volume of coniferous trees was similar on control, manual, and cut-stump treatments (F = 1.15, 26 P= 0.37).  Woody Debris  Average debris volumes showed more variability on replicates assigned to cut-stump treatments than on controls and manual treatments in 1992 (Fig. 6).  However, there were no  significant pre-treatment differences among plantations in the volume of woody debris = 1.89, P=0.23). 26 (F  Post-treatment volumes of woody debris appeared lower on control,  manual, and cut-stump treatments in 1993 than in 1992 but the change was not statistically significant (F = 1.65, P=0.26). 26 The average number of pieces of woody debris tallied in each of three diameter classes on control, manual, and cut-stump treatments is shown in Figure 7. The number of pieces counted in the 5-25 cm and >25 cm size classes was similar across years irrespective of  23  250 200 150 100 50 0  LO  PE  c  PM  TP  TH  TOT  LO  PE  PC  PM  TP  TH  TOT  TP  TH  TOT  c13  C  250 200  S S •  C  150 100 50  c’j c)  250  Cut—Stump 200 150  100 50 0  LO  PE  Pc  PM  Species Fig. 5. Mean absolute volume (m 3 per 0.01 ha) of prominent coniferous tree species and total coniferous trees (TOT) on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Prominent species were western larch (LO), hybrid white spruce (PE), lodgepole pine (PC), Douglas fir (PM), western red cedar (TP), and western hemlock (TH). Each value is the mean of three replicates ± the 95% confidence interval.  24  10 9 8 765  0  4 32 1 0  CONTROL  MANUAL  CUTSTUMP  Treatment  Figure 6.  Mean volume (m 3 per 0.01 ha) of woody debris on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95 % confidence interval.  25  10000  Control 1000 100 10  C)  1  S  10000  a)  C  1000  C) a)  Z4  100  a) C)  a)  10  ‘4-  a)  S  : <5  5—25  >25  Manual  a) a)  1  1  :i <5  5—25  >25  z a)  10000  Cut— Stump 1000  100  10  1  <5  5—25  >25  Diameter (cm) Fig. 7. Mean number of pieces of woody debris counted (number of pieces per 90 m transect) in each of three diameter classes on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95% confidence interval.  26  treatment type. However, in the smallest diameter class (<5 cm), the mean number of pieces again changed little between years on controls but increased by 3.5 and 2 times pre-treatment values on manual and cut-stump treatments, respectively. The mean number of pieces counted in each of five decay classes on control, manual, and cut-stump treatments indicates a similar pattern (Fig. 8). The number of pieces in decay classes 2 (mild decay) through 5 (well-decomposed and broken) showed similar magnitude and direction of yearly change irrespective of treatment. In contrast, the average number of pieces in decay class 1 (newly fallen) increased to 21 times the pre-treatment mean on controls while increasing by 1419 and 896 times pre-treatment values on manual and cut-stump treatments, respectively.  Trappability  Pre-treatment trappability for deer mice was not significantly different among treatment groups in either 1991 (F =3.61, P=0.1O) or 1992 2 25 (F 6=0. 17, P=0.85) (Table 2). Pre- to post-treatment differences in trappability were statistically significant for this species (F 26 =7.37, P=0.02) but, overall, post-treatment trappability remained high with a minimum average value of 74.1 % on manual treatments. Minimum trappability estimates were generally low (<60%) for yellow-pine chipmunks and varied considerably on treatments and controls between years (Table 2).  However,  trappability was not significantly different among treatment groups in either 1991 (F 23 =0.68, P=0.57) or 1992 (F =0.34, P=0.72). Pre- to post-treatment change in trappability was also 26 not statistically significant for chipmunks (F =0.73, P=0.51). 26  27  10000 Control 1000  II  \ N  100 T  10  C) a)  1  i  N  TI\]  II  N N N N  [N  N TN \1 2  3  I  I  N  N 4  N 5  10000 a)  Manual  s-I  J2  a)  I  1000  N N  C) a)  100  10  N N N  1  N  a) C)  a)  a)  s-I  N  N  z a)  I  N  TT  N N N I\1  2  N N N N  N N  N  N  3  ml  N  N  N N N  \\  4  5  10000 Cut— Stump 1000  100  Figure 8.  N N  I  N N N  IN  T  1  [N  INi  2  N N jN N I N N I N NIN N N  N  T[  10  1  x  N N N  N -  3  4  5  Decay Class Mean number of pieces of woody debris counted (number of pieces per 90 m transect) in each of five decay classes on control, manual, and cut-stump treatments in the last pre-treatment year, 1992 (clear bar), and first post-treatment year, 1993 (lined bar). Each value is the mean of three replicates ± the 95% confidence interval.  28  Summer 1991 Summer 1992 Summer 1993  Tamias amoenus  Summer 1991 Summer 1992 Summer 1993  Peromyscus maniculatus  0 31.5 12.5  67.1 75.5 91.6  mean  0 316.2 198.8  57.9 42.5 36.0  95% C.I.  Control  2 3 3  3 3 3  n  16.6 33.3 50.0  60.5 81.7 74.1  mean  259.3 143.4 124.2  43.4 8.8 10.8  95% C.I.  Manual  3 3 3  3 3 3  n  33.3 66.3 56.0  94.4 79.0 80.7  mean  --  25.0 72.9  70.5 12.1 40.4  95% C.I.  Cut-stump  1 3 3  2 3 3  n  Table 2. Minimum unweighted trappability estimates for deer mice and yellow-pine chipmunks on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are mean percent trappability, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate.  Population Size  Deer mouse populations displayed annual trends in abundance across replicates in two years of pre-treatment sampling (Fig. 9).  In 1991, the first pre-treatment year, minimum  number alive estimates of deer mouse population size increased from an average of 5.1 individuals per ha in early September, to a peak of 8.4 individuals per ha in the third sampling week, before declining to a mean of 1.8 individuals per ha in late October (n = 8 plantations). In 1992, deer mouse population size showed a general pattern of increase across replicates from 4.2 individuals per ha in the first sampling week in May to an average of 6.6 individuals per ha in the last pre-treatment sampling week in September (n=9 plantations). Randomization testing detected no non-random change following treatments in the population size of deer mice in pairwise comparisons of controls and manual treatments (two-tailed P=0.86), controls and cut-stump treatments (two-tailed P=0.55), or manual and cut-stump treatments (two-tailed P=0.41). Average deer mouse population size on controls and manual treatments became more similar in 1993, averaging 5.1 and 5.6 individuals per ha, respectively (n=3 plantations per treatment). Cut-stump treatments became less similar to controls and manual treatments in 1993 with an average deer mouse population size of 3.5 individuals per ha. Minimum number alive estimates of population size for yellow-pine chipmunks were highly variable across replicates in two years of pre-treatment sampling (Fig. 10). Chipmunk population size averaged 1.6 and 2.8 individuals per ha in 1991 (n=6 plantations) and 1992 (n=9 plantations), respectively. Post-treatment population size for chipmunks averaged 2.0, 1.6, and 4.5 individuals per ha in 1993 on controls, manual, and cut-stump treatments, respectively (n = 3 plantations per treatment).  Randomization testing detected no significant  change in the population size of yellow-pine chipmunks in pair-wise comparisons of controls and 30  60 50 40 30 20 10 0  SO  MJJ  SO  MJJSO  SO  MJJ  SO  MJJSO  60 50 40  z S  30 20 10 0  60  Cut— Stump  50  nI  oC  40 30 20 10 0  SO 1991  MJJ  SO  1992  MJJSO 1993  Fig. 9. Minimum number alive (MNA) estimates of population size for deer mice on replicate control, manual, and cut-stump treatments. Hatched bar on horizontal axis indicates time of treatments.  31  20  15  10  5  0  SO  MJJ  20  MJJSO  Ivianual  oB  -4  SO  15 ci)  S z S  SO  MJJ  SO  MJJSO  20 Cut— Stump 15  E Fig. 10.  oC -  SO 1991  MJJ  SO  1992  MJJSO 1993  Minimum number alive (MNA) estimates of population size for yellow-pine chipmunks on replicate control, manual, and cut-stump treatments. Hatched bar on horizontal axis indicates time of treatments.  32  manual treatments (two-tailed P=0.84), controls and cut-stump treatments (two-tailed P=0.28), or manual and cut-stump treatments (two-tailed P=0.41). For red-backed voles, the total number of individuals captured was not significantly different among treatment groups in 1991 (F 11 = 1.53, P=0.43) or 1992 (F 21 =0.29, P=0.79) (Table 3). Pre- to post-treatment change in the total number of individuals captured was also not significant for this species 11 (F =0.49, P=0.61). Total abundance of long-tailed voles was similar among treatment groups in 1991 (F, 1 =4.25, P=0.32), 1992 (F =0.25, P=0.81), and 21 between the last pre-treatment and first post-treatment years (F 21 =0.22, P=0.83). Meadow voles also had similar total numbers of individuals captured on controls, manual, and cut-stump treatments in each of 1991 (F =3.86, P=0.20) and 1992 (F 22 = 1.24, P=0.40). Changes in 23 meadow vole abundance were not statistically significant in the year following treatments =0.91, P=0.44). 12 (F The three vole species captured in this study were found only rarely on more than five plantations in any given year. As a result, statistical degrees of freedom were reduced and the probability that an analysis of variance would detect statistically significant differences among treatments was quite low. Using mean squares values from the preceding tests, and the figures reproduced in Zar (1984) from Pearson & Hartley (1951), estimates of statistical power (1 -3) for tests of difference in vole numbers were consistently less than 0.10.  33  Summer 1991 Summer 1992 Summer 1993  Microtus pennsylvanicus  Summer 1991 Summer 1992 Summer 1993  Microtus longicaudus  Summer 1991 Summer 1992 Summer 1993  Clethrionomys gapperi  3.0 1.3 5.6  3.0 1.0 3.0  2.0 2.0  --  mean  25.4 1.4 17.9  ----  12.7  ---  95% C.I.  Control  2 3 3  1 1 1  0 1 2  n  11.0 6.5 4.5  5.0 7.0 10.0  17.0 30.0 11.5  mean  --  57.1 14.4  14.9  ---  88.9 279.5 95.2  95% C.I.  Manual  1 2 2  1 1 3  2 2 2  n  --  3.5 5.0  2.5 7.5 34.5  --  2.0 14.0  mean  --  --  19.0  6.3 69.8 400.2  ----  95% C.I.  Cut-stump  2 1 0  2 2 2  1 1 0  n  Table 3. Total number of individuals captured for red-backed voles, long-tailed voles, and meadow voles on control, manual, and cut stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean total number of individuals, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate.  Sex Ratios  The proportion of male deer mice in individual populations was not significantly different among treatment groups in either 1991 (F =0.73, P=0.52) or 1992 (F 25 = 1.66, P=0.26) 26 (Table 4). Pre- to post-treatment differences in sex ratios were also not significantly different for deer mice (F =0.08, P=0.92). 26  Body Weights  There were insufficient data to estimate the median weight at sexual maturity, or any other weight-based parameters for deer mice, in the first season of pre-treatment sampling (Summer 1991) (Fig. 11). However, there were no statistically significant differences in median weight at sexual maturity for male or female deer mice among pre-treatment groups in the Summer 1992 sampling period (Males: =1.74, 26 P=0.25; Females: =O.48, F 26 P=0.63). F Weight at sexual maturity for male and female deer mice averaged 20.3 and 19.7 g, respectively, across all replicates in 1992 and did not change significantly due to treatments in 1993 (Males: F =O.33, 2 6 P=0.73; Females: =1.57, 26 P=0.28). Median weight at sexual F maturity was used to delineate adult deer mice from juveniles in other estimates of breeding activity including mean adult male body weights and the proportion of adults breeding. Mean adult body weights for male deer mice did not significantly differ among plantations in 1992 (F =O.69, P=0.53), nor did they change significantly in the year following 26 treatments (F =0. 17, P=0.85) (Fig. 12). As no adult male deer mice were captured on site 25 D in Eagle Bay during 1993, weights were averaged over only two replicates for manual  35  Summer 1991 Summer 1992 Summer 1993  Peromyscus maniculatus  95% C.I. 0.18 0.20 0.11  mean 0.64 0.54 0.56  Control 95% C.I. 0.80 0.09 0.31  mean 0.50 0.59 0.62  n 3 3 3  95% C.I. 0.41 0.26 0.13  mean 0.55 0.47 0.47  3 3 3  Cut-stump  n  Manual  2 3 3  n  Table 4. Sex ratios (proportion of males in population) for deer mice on control, manual, and cut-stump treatments during pre treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean sex ratio, the 95% confidence interval, and the sample size (number of replicates).  Fig. 11.  -  -  M  Summer 1991  C  TooFewData  CS  0  • Females  Summer 1992  M  CS  §TÔ}T  Males  C  Summer 1993  lvi  CS  Median weight at sexual maturity (g) for male (0) and female (•) deer mice on control (C), manual (M), and cut-stump (CS) treatments during pre treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95% confidence interval.  30  35  40  Fig. 12.  30-  35  M  Summer 1991  C  Too Few Data  CS  c  Summer 1992  M  CS  C Summer 1993  Cs  Mean adult male body weights (g) for deer mice on control (C), manual (M), and cut-stump (CS) treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95% confidence interval except the manual treatment mean in 1993 (*) which was calculated over two replicates.  .  40-  45  50  treatments in the first post-treatment year.  Reproduction  Length of breeding season could not be estimated for deer mice in 1991 because sampling was limited to September and October in the first year of study (Table 5).  There were no  significant differences in the length of breeding season for deer mice among plantations in 1992 =3.12, P=0.11). However, breeding activity may have extended beyond the first and last 26 (F trapping weeks on plantations assigned manual treatments in the second pre-treatment year. Length of breeding season did not change significantly for deer mice from 1992 to 1993 =O.33, P=0.73). 26 (F The proportion of adult male deer mice breeding during 1992 ranged from 0.40 to 0.95 among treatment groups but differences were not statistically significant (F =3.41, P=0. 11) 25 (Table 6). Similarly, the proportion of adult female deer mice breeding in each population did not differ significantly during 1992 (F =O.52, P=O.61) (Table 7). 26  Neither sex showed a  significant change in the proportion of adults breeding due to treatments in 1993 (Males: =O.13, 2 F 4 P=O.88; Females: =2.89, 25 P=0.14). F  Recruitment  The average total number of deer mouse recruits appeared higher on controls and manual treatments than on cut-stump treatments in each of the three years of study (Fig. 13). Mean yearly recruitment ranged from 15 to 35 individuals during pre-treatment sampling and numbered  39  Summer 1991 Summer 1992 Summer 1993  Peromyscus maniculatus 20 20  --  mean 11.3 15.5  --  95% C.I.  Control  3 3  --  n 25 25  --  mean 0 0  --  95% C.I.  Manual  3 3  --  n 20 24  mean 4.3 4.3  --  95% C.I.  Cut-stump  3 3  --  n  Table 5. Length of breeding season (number of weeks between first and last capture of a breeding male or female) for deer mice on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean total number of weeks, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate.  Summer 1991 Summer 1992 Summer 1993  Peromyscus maniculatus 0.95 0.90  --  mean 0.20 0.39  --  95% C.I.  Control  3 3  --  n 0.40 0.53  --  mean 3.18 0.46  --  95% C.I.  Manual  2 3  --  n  --  0.91 0.80  mean 0.35 0.43  95% C.I.  Cut-stump  3 3  --  n  Table 6. Proportion of adult male deer mice in breeding condition on control, manual, and cut-stump treatments during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean proportion breeding, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate.  t’.)  Summer 1991 Summer 1992 Summer 1993  Peromyscus maniculatus 1.00 0.60  --  mean 0 0.80  --  95% C.I.  Control  3 3  --  n --  0.86 0.62  --  95% C.I.  0.80 0.75  mean  Manual  3 3  --  n 0.88 1.00  --  mean  --  0.48 0  95% C.I.  Cut-stump  3 2  --  n  Table 7. Proportion of adult female deer mice in breeding condition on control, manual, and cut-stump treatments during pre treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Values are the mean proportion breeding, the 95% confidence interval, and the sample size (number of replicates). Hash marks (--) indicate that there were insufficient data to calculate an estimate.  Fig. 13.  i)  (I,  0  10  20  30  40  50  60  70  80  90  M  Summer 1991  C Cs M  Summer 1992  C  CS  M  Summer 1993  C  CS  Total number of recruits for deer mice in control (C), manual (M), and cut-stump (CS) treatment populations during pre-treatment (Summers 1991 and 1992) and post-treatment (Summer 1993) sampling. Each value is the mean of three replicates ± the 95% confidence interval except the cut-stump treatment mean in 1991 (*) which was calculated over two replicates.  H  z  ,0  a)  C  a)  h. C)  .4-)  100  110  120  35, 34, and 23.3 individuals on control, manual, and cut-stump treatments, respectively, in 1993. These differences were not statistically significant for either pre-treatment year (1991: =0.67, P=0.55; 1992: F, 5 , 2 F = 1.15, P=0.37), nor for pre- to post-treatment change in 6 recruitment (F =0.01, P=0.98). 26  Survival  There were no consistent seasonal or yearly patterns in the minimum survival of deer mice on control, manual, and cut-stump treatments (Table 8). Overall, pre-treatment summer survival averaged 0.26 and 0.53 in 1991 and 1992, respectively, and was not significantly different among treatment groups in either year (Summer 1991: = 25 1.13, P=0.39; Summer F 1992: F =0.44, 2 6 P=0.66). Pre-treatment winter survival was similar among populations in 1991-1992 (Winter 1991-1992: F 25 =0.63, P=0.56) and averaged 0.21 across all replicates. There were no significant changes in Summer or Winter survival of deer mice following treatments (Summer 1993: =O.92, 26 P=0.44; Winter 1992-1993: F F =0.99, 2 5 P=0.43).  Treatment Quality  Treatment quality averaged 99.7±1.4 percent (mean±95 % confidence interval) on manually treated sites and 93.2±5.4 percent on cut-stump treatments.  Application rates of  glyphosate herbicide were 1.77, 2.87, and 1.65 kg of active ingredient per ha on cut-stump replicates C, H, and I, respectively.  44  mean 0.26 0 0.54 0.72 0.28  Peromyscus maniculatus  Summer 1991 Winter 1991-1992 Summer 1992 Winter 1992-1993 Summer 1993  n 3 3 3 3 3  95% C.I. 0.38 0 0.53 0.45 0.73  Control 95% C.I. 0.24 1.29 0.28 0.17 0.13  mean 0.33 0.30 0.63 0.79 0.64  Manual  3 3 3 3 3  n 0.14 0.39 0.43 0.48 0.55  mean 1.50 5.00 0.79 1.04 0.23  95% C.I.  Cut-stump  2 2 3 3 3  n  Table 8. Minimum survival estimates for deer mice on control, manual, and cut-stump treatments during pre-treatment (Summer 1991, Winter 1991-1992, and Summer 1992) and post-treatment (Winter 1992-1993 and Summer 1993) sampling. Values are mean minimum survival, the 95% confidence interval, and the sample size (number of replicates).  DISCUSSION  Experimental Design  This study is uncommon among investigations of small mammal and wildlife habitat responses to vegetation management in that several separate and independent forest plantations were sampled for at least one year prior to imposing replicated treatments.  However, the  comparison of vegetation, woody debris, and small mammal population responses among treatment groups still rests on the assumption that initial differences among the nine study plantations were very few, or that the allocation of treatments to plantations was sufficiently unbiased that experimental design controlled for initial differences.  Natural variation was  apparent among the nine study plantations. While assigning cut-stump treatments to plantations furthest from private homes in Eagle Bay almost certainly did not create any marked experimental bias, the most heavily competed plantation in Sicamous was subjectively assigned a cut-stump treatment. Pre-treatment variation among plantations and non-random allocation of treatments increased experimental error and reduced the power of statistical tests to detect differences, especially for the population size of voles. Even so, treatments were interspersed and there was little indication that initial differences within treatment groups were any greater, or lesser, than initial differences among groups. The only certain option to increase statistical power in this study would have been to increase the number of replicates for each treatment and that was not possible. The familiar constraints of limited time and funding, as well as a lack of suitable sites in reasonably close proximity, prevented additional replication of treatments. In any event, replication and interspersion of treatments, the independence of spatially segregated plantations, and random sampling within plantations, all permitted parametric statistics to be 46  applied without committing pseudoreplication of treatments (Huribert 1984).  Vegetation, and Woody Debris  Both manual and cut-stump treatments effectively released regenerating conifers from deciduous competition.  Changes in vegetative structure and composition that followed these  treatments clearly favoured naturally regenerating and planted coniferous tree species. Ninetynine to 100 percent of competing stems were cut during manual treatments.  On cut-stump  treated plantations, between 91 and 96 percent of competing stems were cut and sprayed with glyphosate herbicide. The greatest reductions in absolute volume of vegetation were recorded for designated target species. These species included bitter cherry and willow in the shrub layer and all three deciduous tree species: paper birch, black cottonwood, and trembling aspen. Thimbleberry, a non-target shrub species, was reduced in volume on manual treatments because many plants were crushed under the weight of cut deciduous tree stems.  Cut stems also  hindered the growth of fireweed and bracken, both tall herbs, and lowered total herb volumes on all treated plantations. Treatment quality was generally lower on cut-stump treatments than on manual treatments because, in accordance with treatment guidelines, most maple, cherry, and alder were not cut. However, as re-sprouting of deciduous trees was only observed on manually treated plantations, both prescriptions had similar overall effects on habitat structure in the first post-treatment year. Volumes of woody debris were similar among plantations in the last pre-treatment and first post-treatment years. However, dramatic additions of small diameter woody debris were recorded on treated plantations following the cutting of deciduous trees and shrubs and this increased the complexity of low ground cover. Large increases in the number of newly fallen 47  pieces of woody debris did not increase estimated debris volumes on treated plantations because the Van Wagner (1968, 1982) estimator is more sensitive to greater piece size than to greater numbers of pieces (Caza 1993).  Small Mammal Populations  Changes in vegetative structure and increased complexity of woody ground cover may affect the distribution and abundance of small mammal populations (M’Closkey & Fieldwick 1975; Miller & Getz 1977; Dueser & Shugart 1978; Kitchings & Levy 1981; Yahner 1986). In this study, deer mice were the most abundant species on all nine plantations over three years of sampling. Population size of deer mice was unaffected by manual or cut-stump treatments for conifer release, at least in the first post-treatment year. This result is consistent with most studies of deer mice in forest plantations aerially sprayed with the herbicide glyphosate (Sullivan & Sullivan 1982; Anthony & Morrison 1985; dough 1987; D’Anieri, Leslie & McCormack 1987; Santillo, Leslie & Brown 1989; Sullivan l990b). Similarly, deer mouse population size was unaltered by treatments of 2,4-D (Tietjen et al. 1967; Johnson & Hansen 1969) and 2,4,5-T (Kirkland 1978; Freedman et al. 1988) for range improvement and conifer release, respectively. In contrast, other studies indicate that deer mouse population size may increase (McMillan et al. 1990; Milton & Towers 1990) or decrease (Ritchie, Harestad & Archibald 1987) following glyphosate treatments in forest plantations. The total abundance of deer mice was also found to be greater in plantations treated with a combination of 2,4-D and atrazine in the first post treatment year (Black & Hooven 1974; Borrecco, Black & Hooven 1989); however, yearly change in population size was similar on treatments and controls. The overall resilience of deer mouse population size to conifer release treatments may be due to the generalist habits of this 48  species.  Although research indicates that deer mice are sensitive to changes in vegetative  structure (Miller & Getz 1977; Vickery 1981) and protective cover (Anderson 1986), they are a common species, and are often abundant, in a wide variety of habitats including swamps, deserts, scrubby woodlands, and mature forests (Baker 1968).  Despite the clear effects of  manual and cut-stump treatments on vegetation and woody debris, sufficient cover and forage may have persisted on treated plantations that habitat quality was undiminished for deer mice. Population size of yellow-pine chipmunks was similar among plantations but highly variable over two years of pre-treatment sampling. Chipmunk population size again varied, but remained similar among treatment groups, in the first post-treatment year. In the only previous study of yellow-pine chipmunk responses to conifer release treatments, population size was unaltered following applications of 2,4-D and atrazine (Black & Flooven 1974; Borrecco, Black & Hooven 1979). Related species, including Townsend chipmunks (Tamias townsendii) and least chipmunks (Tamias minimus), have also been found in similar abundance on control sites and areas treated with glyphosate (Sullivan & Sullivan 1982; Anthony & Morrison 1985; Sullivan 1990b) and 2,4-D (Tietjen et al. 1967). Although Johnson & Hansen (1969) caught fewer least chipmunks on two range areas treated with 2,4-D, they did not collect pre-treatment data at either location so the cause of these differences could not be determined. Since yellowpine chipmunks typically inhabit open coniferous forests and shrub habitats (Sutton 1992), the effects of clearcut harvesting are probably a more important alteration of habitat for this species than are vegetation management treatments in young forest plantations. Manual and cut-stump treatments for conifer release did not appear to alter plantation habitats beyond the tolerance of chipmunks for early successional change. I found no significant differences among treatment groups in any year for the total number of captures of southern red-backed voles, long-tailed voles, or meadow voles. Previous 49  studies indicate that vole populations may increase, decrease, or show no long-term change following applications of glyphosate (Sullivan & Sullivan 1982; Anthony & Morrison 1985; dough 1987; D’Anieri, Leslie & Mcdormack 1987; Santillo, Leslie & Brown 1989; Sullivan 1990b), 2,4-D (Johnson 1964; Tietjen et al. 1967; Johnson & Hansen 1969; Spencer & Barrett 1980), 2,4,5-T (Kirkland 1978; Freedman et al. 1988) or 2,4-D and atrazine in combination (Black & Hooven 1974; Borrecco, Black & Hooven 1979). Where changes have occurred in the abundance of voles following these treatments, they have usually been attributed to shifts in the structure and composition of vegetation that matched or contrasted with the habitat preferences of individual vole species. For example, meadow voles were less abundant in forest plantations and experimental enclosures treated with 2,4,5-T (Kirkland 1978) and 2,4-D (Spencer & Barrett 1980), respectively, where there was a reduction in vegetative cover. In a similar fashion, red-backed voles were reduced in abundance on glyphosate-treated areas where there was, again, a reduction in vegetative cover and exposure of the soil to evaporative drying (dough 1987; D’Anieri, Leslie & Mcdormack 1987; Santillo, Leslie & Brown 1989; McMillan et al. 1990; Sullivan 1990b). Although some studies indicate that voles are sensitive to reduced protective cover (Eadie 1952; Birney, Grant & Baird 1976; Anderson 1986; Merkens, Harestad & Sullivan 1991), descriptions of vole population responses to conifer release and range improvement have often been drawn from comparisons of unreplicated treatments (Spencer & Barrett 1980; Sullivan & Sullivan 1982; Anthony & Morrison 1985; Clough 1987; D’Anieri, Leslie & Mcdormack 1987; Sullivan 1990b), small sample sizes (<10 individuals) (Kirkland 1978; D’Anieri, Leslie & Mcdormack 1987; Freedman et al. 1988; Santillo, Leslie & Brown 1989; Sullivan 1990b), or data collected in a single season of sampling (Spencer & Barrett 1980; Clough 1987; D’Anieri, Leslie & Mcdormack 1987). While these problems are difficult to overcome in field studies, it is important to recognize that they may compromise any conclusions 50  drawn from comparisons of treatment effects on small mammals (Lautenschlager 1993). In my study, differences in the abundance of red-backed, long-tailed, and meadow voles were more likely due to the general scarcity of these species and chance differences between populations than to distinct differences in habitat structure following manual and cut-stump treatments. Population size alone may be a misleading indicator of habitat quality for small mammals (Van Home 1983). However, the majority of individuals in a population must be captured for other estimates of population condition to be accurate (Krebs & Boonstra 1984). In my study, deer mice were the only species with sufficiently and consistently high trappability to justify the estimation of sex ratios, median weight at sexual maturity, mean adult male body weights, length of breeding season, the proportion of adults breeding, the total number of recruits, and minimum survival rates. All of these parameters were similar among treatment groups during two pre treatment sampling years and in the first post-treatment year. Similar results have been obtained in other studies.  Sullivan & Sullivan (1981, 1982) and Sullivan (1990a, 1990b) found little  difference in body weights, reproduction, recruitment, or survival of deer mice on control and treatment areas following aerial sprays of glyphosate in coastal British Columbia. In addition, the mean number of foeti and placental scars did not differ significantly for deer mice on control areas and areas treated with glyphosate (Ritchie, Harestad & Archibald 1987) or 2,4-D (Johnson & Hansen 1969). Newton et al. (1984) found that glyphosate residues in the visceral and body tissues of deer mice had declined to minimum detectable levels within 28 days of aerial sprays over hardwood forests.  Consequently, any effects due to the ingestion of glyphosate-treated  vegetation by deer mice would almost certainly be observed in the first post-treatment year. Since estimated application rates of glyphosate on cut-stump treatments were similar between this study [2.1±1.7 kg of active ingredient per ha (mean±95 % confidence interval)] and those of Sullivan & Sullivan (1981, 1982), Ritchie, Harestad & Archibald (1987), and Sullivan (1990a, 51  1990b) (range  1.2-3.0 kg of active ingredient per ha), the lack of change in deer mouse  demographic characteristics was not unexpected.  CONCLUSIONS  Both manual and cut-stump treatments were effective silvicultural prescriptions for conifer release. Changes in vegetative structure and composition that followed these treatments clearly favoured naturally regenerating and planted coniferous tree species. Overall, there were no discernable effects of manual or cut-stump treatments on small mammal communities in the first post-treatment year.  Continued sampling on these plantations will determine whether  manual or cut-stump treatments have any long-term effects on small mammals. Small mammals are an important source of prey for small carnivores and raptorial birds (Fitzgerald 1977; Craighead & Craighead 1956). In particular, voles of the genus Microtus are primary prey  species for ermine (Simms 1979a), long-tailed weasels (Mustelafrenata) (Quick 1951; Simms, 1979b), and marten (Martes americana) (Zielinski, Spencer & Burnett 1983).  Even in the  absence of change in the size of small mammal populations, altered habitat structure may affect their predators. Future studies should examine the effects of conifer release at these higher trophic levels. This is the first replicated study to describe the effects of manual or cut-stump conifer release treatments on small mammal populations and habitat structure in young plantation forests. The results are encouraging in that small mammal population parameters were maintained at pre treatment levels for at least one post-treatment year.  However, forest managers should be  cautious in prescribing operational-scale silvicultural treatments across intensively managed forest landscapes. Treatments should be separated in both time and space to avoid unnecessarily severe 52  short-term alterations to wildlife habitat.  53  REFERENCES Anthony, R.G. & Morrison, M.L. (1985). Influence of glyphosate herbicide on small mammal populations in western Oregon. Northwest Science, 59, 159-168. Atkinson, D. (1985). Toxicological properties of glyphosate-a summary. The Herbicide Glyphosate (Edited by E. Grossbard & D. Atkinson), pp. 127-133. Butterworths, London. Anderson, P.K. (1986). Foraging range in mice and voles: the role of risk. Canadian Journal of Zoology, 64, 2645-2653. Baker, R.H. (1968). Habitats and distribution. Biology of Peromyscus (Rodentia) (Edited by J.A. King), pp. 125-167. American Society of Mammalogists, Special Publication Number 2. Birney, E.C., Grant, W.E. & Baird, D.D. (1976). Importance of vegetative cover to cycles of Microtus populations. Ecology, 57, 1043-105 1. Black, H.C. & Hooven, E.H. (1974). Response of small mammal communities to habitat changes in western Oregon. Wildlife and Forest Management in the Pacific Northwest (Edited by H.C. Black), pp. 177-186. School of Forestry, Oregon State University, Corvallis. Boonstra, R. (1985). Demography of Microtuspennsyivanicus in southern Ontario: enumeration versus Jolly-Seber compared. Canadian Journal of Zoology, 63, 1174-1180. Borrecco, J.E., Black, H.C. & Hooven, E.F. (1979). Response of small mammals to herbicide-induced habitat changes. Northwest Science, 53, 97-106. Bovey, R.W. & Young, A.L. (1980). The Science of 2,4,5-T anti Associated Phenoxy Herbicides. John Wiley and Sons Inc., New York. Campbell, R.A. (1990). Herbicide use for forest management in Canada: where we are and where we are going. Forestiy Chronicle, 66, 355-360. Carpenter, S.R., Frost, T.M., Heisey, D. & Kratz, T.K. (1989). Randomized intervention analysis and the interpretation of whole-ecosystem experiments. Ecology, 70, 1142-1152. Caza, C.L. (1993). Woody Debris in the Forests of British Columbia: a review of the literature and current research. British Columbia Ministry of Forests, Land Management Report Number 78. Clough, G.C. (1987). Relations of small mammals to forest management in northern Maine. Canadian Field-Naturalist, 101, 40-48.  54  Craighead, J.J. & Craighead, F.C. (1956). Hawks, Owls, and Wildlfe. Stackpole Company and Wildlife Management Institute. D’Anieri, P., Leslie, D.M. Jr. & McCormack, M.L. Jr. (1987). Small mammals in glyphosate-treated clearcuts in northern Maine. Canadian Field-Naturalist, 101, 547-550. Dueser, R.D. & Shugart, H.H. Jr. (1978). Microhabitats in a forest-floor small mammal fauna. Ecology, 59, 89-98. Eadie, W.R. (1952). Response of Microtus to vegetative cover. Journal of Mammalogy, 34, 263-264. Edington, E.S. (1987). Randomization Tests. 2nd edition. M. Dekker, New York. Environics Research Group (1989). 1989 National survey of ‘anadian public opinion on forestry issues. Unpublished Report. Environics Research Group Ltd., Toronto. Fitzgerald, B.M. (1977). Weasel predation on a cyclic population of the montane vole (Microtus montanus) in California. Journal of Animal Ecology, 46, 367-397. Freedman, B. (1989). Environmental Ecology. Academic Press, San Diego. Freedman, B. (1991). Controversy over the use of herbicides in forestry, with particular reference to glyphosate usage. Journal of Environmental Science and Health, C8, 277286. Freedman, B., Poirier, A.M., Morash, R. & Scott, F. (1988). Effects of the herbicide 2,4,5T on the habitat and abundance of breeding birds and small mammals of a conifer clearcut in Nova Scotia. Canadian Field-Naturalist, 102, 6-11. Freedman, B., Morash, R. & MacKinnon, D. (1993). Short-term changes in vegetation after the silvicultural spraying of glyphosate herbicide onto regenerating clearcuts in Nova Scotia, Canada. Canadian Journal of Forest Research, 23, 2300-23 11. Hitchcock, C.L. & Cronquist, A. (1973). Flora of the Pacific Northwest. University of Washington Press, Seattle. Hudson, R.H., Tucker, R.K. & Haegle, M.A. (1984). Handbook of Toxicity of Pesticides to Wildflfe. 2nd edition. United States Department of the Interior, Washington, D.C., Resource Publication 153. Hurlbert, SH. (1984). Pseudoreplication and the design of ecological field experiments. Ecological Monographs, 54, 187-211. Johnson, D.R. (1964). Effects of range treatment with 2,4-D on food habits of rodents. Ecology, 45, 24 1-249. 55  Johnson, D.R. & Hansen, R.M. (1969). Effects of range treatments with 2,4-D on rodent populations. Journal of Wildlife Management, 33, 125-132. Ketcheson, M.V., Braumandi, T.F., Meidinger, D., Utzig, G., Demarchi, D.A. & Wikeem, B.M. (1991). Interior cedar-hemlock zone. Ecosystems of British Columbia (Edited by D. Meidinger & 3. Pojar), pp. 167-18 1. British Columbia Ministry of Forests, Victoria. Kirkland, G.L. (1978). Population and community responses of small mammals to 2,4,5-T. United States Department of Agriculture, Pacific Northwest Forest and Range Experiment Station, Research Note PNW-314. Kitchings, J.T. & Levy, D.J. (1981). Habitat patterns in a small mammal community. Journal of Mammalogy, 62, 814-820. Kienner, W. & Krebs, C.J. (1991). Red squirrel population dynamics. I. The effect of supplemental food on demography. Journal of Animal Ecology, 60, 961-978. Krebs, C.J. (1966). Demographic changes in fluctuating populations of Microtus caltfornicus. Ecological Monographs, 36, 239-273. Krebs, C.J., Keller, B.L. & Tamarin, R.H. (1969). Microtus population biology: demographic changes in fluctuating populations of Microtus ochrogaster and M. pennsylvanicus in southern Indiana. Ecology, 50, 587-607. Krebs, C.J. & Boonstra, R. (1984). Trappability estimates for mark-recapture data. Canadian Journal of Zoology, 62, 2440-2444. Krebs, C.J., Gilbert, B.S., Boutin, S., Sinclair, A.R.E. & Smith, J.N.M. (1986). Population biology of snowshoe hares. I. Demography of food-supplemented populations in the southern Yukon, 1976-84. Journal of Animal Ecology, 55, 963-982. Lautenschlager, R.A. (1986). Forestry, herbicides, and wildlife. Is Good Forestry Good Wildlife Management? (Edited by J.A. Bissonette), pp. 299-308. Maine Agriculture Experimental Station, Miscellaneous Publication 689. Lautenschlager, R.A. (1990). Red raspberry (Rubus idaeus L.) ecology: germination, growth, and interactions with white spruce (Picea glauca (Moench) Voss) growth. Ph.D. dissertation, University of Maine, Orono. Lautenschlager, R.A. (1993). Response of wildlife to forest herbicide applications in northern coniferous ecosystems. ‘anadian Journal of Forest Research, 23, 2286-2299. Lautenschlager, R.A. & McCormack, M.L. Jr. (1989). Herbicide release may increase plant species diversity. Forest and Wildkfe Management in New England-what can we afford? (Edited by R.D. Briggs, W.B. Krohn, 3.0. Trial, W.D. Ostrofsky & D.B. Field), p. 242. Maine Agriculture Experimental Station, Miscellaneous Report 336. 56  Leslie, P.R., Perry, J.S. & Watson, J.S. (1945). The determination of the median weight at which female rats reach maturity. Proceedings ofthe Zoological Society ofLondon, 473488. Lloyd, D.A., Angove, K., Hope, G. & Thompson, C. (1990). A Guide for Site IdentWcation and Interpretation ofthe Kamloops Forest Region. 2 volumes. British Columbia Ministry of Forests, Land Management Handbook Number 23. Manly, B.F.J. (1990). Randomization and Monte Carlo Methods in Biology. Chapman & Hall, New York. May, T.A., May, E.E. & McCormack, M.L. Jr. (1982). Plant response to herbicides in clearcut conifer strips. Proceedings of the Northeastern Weed Science Society, 36, 205208. McDonald, P.M. & Radosevich, S.R. (1992). General principles of forest vegetation management. Silvicultural Approaches to Animal Damage Management in Pacific Northwest Forests (H.C. Black, Technical Editor), pp. 67-91. United States Department of Agriculture, Pacific Northwest Research Station, General Technical Report PNW GTR-287. M’Closkey, R.T. & Fieldwick, B. (1975). Ecological separation of sympatric rodents (Peromyscus and Microtus). Journal of Mammalogy, 56, 119-129. McMillan, L.M., Connor, J.F., Timmermamn, H.R., McNichol, J.G. & Krishka, C.S. (1990). Small mammal and lesser vegetation response to glyphosate tending in north central Ontario. Ontario Ministry of Natural Resources, Northwestern Ontario Forest Technology Development Unit, Thunder Bay, Technical Report 57. Merkens, M., A.S. Harestad & Sullivan, T.P. (1991). Cover and efficacy of predator-based repellents for Townsend’s vole, Microtus townsendii. Journal of Chemical Ecology, 17, 401-4 12. Miller, D.H. & Getz, L.L. (1977). Factors influencing local distribution and species diversity of forest small mammals in New England. Canadian Journal of Zoology, 55, 806-814. Milton, G.R. & Towers, J. (1990). Relationships of songbirds and small mammals to habitat features on plantation and natural regeneration sites. Canadian Institute of Forestry, Nova Scotia Department of Natural Resources, Antigonish. Morrison, M.L. (1982). The response of avian communities to herbicide-induced vegetation changes, western Oregon. Ph.D. dissertation, Oregon State University, Corvallis. Morrison, M.L. & Meslow, E.C. (1983). Impacts of forest herbicides on wildlife: toxicity and habitat alteration. Transactions of the North American Wildlife and Natural Resources Conference, 48, 175-185. 57  Nagorsen, D. (1990). Mammals. The Vertebrates ofBritish Columbia (Edited by R.A. Cannings & A.P. Harcombe), pp. 39-43. Royal British Columbia Museum, Heritage Record Number 20, British Columbia Ministry of Environment, Wildlife Branch, Wildlife Report Number R24. Neter, J. & Wasserman, W. (1974). Applied Linear Statistical Models: regression, analysis of variance, and experimental designs. R.D. Irwin Inc., Homewood, Illinois. Newton, M. & Dost, F.N. (1984). Biological and Physical Effects of Forest Vegetation Management. Washington Department of Natural Resources, Olympia. Newton, M., Howard, K.M., Keipsas, B.R., Danhaus, R., Lottman, C.M. & Dubelman, S. (1984). Fate of glyphosate in an Oregon forest ecosystem. Journal ofAgricultural and Food Chemistry, 32, 1144-1151. Newton, M., Cole, E.C., Lautenschlager, R.A., White, D.E. & McCormack, M.L. Jr. (1989). Browse availability after conifer release in Maine’s spruce-fir forests. Journal of Wildlife Management, 52, 643-649. Newton, M. & Comeau, P.G. (1990). Control of competing vegetation. Regenerating British Columbia’s Forests (Edited by D.P. Lavender, R. Parish, C.M. Johnson, G. Montgomery, A. Vyse, R.A. Willis & D. Winston), pp. 256-265. University of British Columbia Press, Vancouver. Oxenham, R. (1983). A study of the effect of competing vegetation and browsing on early plantation performance. Nova Scotia Department of Lands and Forests, Forest Technical Note 5. Pearson, E.S. & Hartley, H.O. (1951). Charts for the power function for analysis of variance tests, derived from the non-central F-distribution. Biometrika, 38, 112-130. Quick, HF. (1951). Notes on the ecology of weasels in Gunnison County, Colorado. Journal of Mammalogy, 32, 281-290. Ritchie, D.C., Harestad, A.S. & Archibald, R. (1987). Glyphosate treatment and deer mice in clearcut and forest. Northwest Science, 61, 199-202. Roskamp, G. & Reeves, J.W. (1980). Brush control in the 80’s. Journal ofArboriculture, 6, 32 7-331. Santillo, D.J. (1987). Response of small mammals and songbirds to herbicide-induced habitat changes on clearcuts in Maine. M. S. thesis, University of Maine, Orono. Santfflo, D.J., Leslie, D.M. Jr. & Brown, P.W. (1989). Response of small mammals and habitat to glyphosate application on clearcuts. Journal of Wildlfe Management, 53, 164172. 58  SAS Institute Inc. (1988). SAS/STAT User’s Guide. Release 6.03 Edition. (Edited by K.P. Ingraham, R.C. Luginbuhi, S.D. Schlotzhauer & H. Watts). SAS Institute Inc., Cary, North Carolina. Seber, G.A.F. (1982). The Estimation of Animal abundance and Related Parameters. 2nd edition. Charles Griffin & Company, London.  Simms, D.A. (1979a). Studies of an ermine population in southern Ontario. Canadian Journal of Zoology, 57, 824-832. Sinims, D.A. (1979b). North American weasels: resource utilization and distribution. Canadian Journal of Zoology, 57, 504-520. Slagsvold, T. (1977). Bird population changes after clearance of deciduous scrub. Biological Conservation, 12, 229-244. Spencer, S.R. & Barrett, G.W. (1980). Meadow vole population response to vegetational changes resulting from 2,4-D application. American Midland Naturalist, 103, 32-46. Stickney, P.F. (1980). Data base for post-fire succession, first 6-11 years. United States Department of Agriculture, Intermountain Research Station, General Technical Report INT-GTR-62. Stickney, P.F. (1985). Data base forpost-fire succession on the Sundance burn, northern Idaho. United States Department of Agriculture, Intermountain Research Station, General Technical Report INT-GTR-189. Suffivan, T.P. (1990a). Influence of forest herbicide on deer mouse and Oregon vole population dynamics. Journal of Wildlife Management, 54, 566-576. Suffivan, T.P. (1990b). Demographic responses of small mammal populations to a herbicide application in coastal coniferous forest: population density and resiliency. Canadian Journal of Zoology, 68, 874-883. Suffivan, T.P. & Sullivan, D.S. (1981). Responses of a deer mouse population to a forest herbicide application: reproduction, growth, and survival. Canadian Journal ofZoology, 59, 1148-1154. Suffivan, T.P. & Sullivan, D.S. (1982). Responses of small mammal populations to a forest herbicide application in a 20-year-old conifer plantation. Journal ofApplied Ecology, 19, 95-106. Sutton, D.A. (1992). Tamias amoenus. Mammalian Species, 390, 1-8. Sutton, R.F. (1978). Glyphosate herbicide: an assessment of forestry potential. Forestry Chronicle, 54, 24-28. 59  Tietjen, H.P., Haivorsen, C.H., Hegdal, P.L. & Johnson, A.M. (1967). 2,4-D herbicide, vegetation, and pocket gopher relationships. Black Mesa, Colorado. Ecology, 48, 634643. Triska, F.J. & Cromack, K. Jr. (1980). The role of wood debris in forest and streams. Forests: fresh perspectives from ecosystem analysis, Proceedings of the 40th Annual Biology Colloquium, pp. 17 1-190. Oregon State University Press, Corvallis. Trowbridge, R., Hawkes, B., Macadam, A. & Parminter, J. (1986). Field Handbook for Prescribed Fire Assessments in British Columbia: logging slash fuels. British Columbia Ministry of Forests, Land Management Handbook Number 11. Van Home, B. (1983). Density as a misleading indicator of habitat quality. Journal of Wildflfe Management, 47, 893-901. Van Wagner, C.E. (1968). The line intersect method in forest fuel sampling. Forest Science, 14, 20-26. Van Wagner, C.E. (1982). Practical aspects of the line intersect method. Canadian Forest Service, Petawawa National Forest Institute, Forest Information Report pl-X-12. Vickery, W.L. (1981). Habitat use by northeastern forest rodents. American Midland Naturalist, 106, 111-118. Wagner, R.G. (1993). Research directions to advance forest vegetation management in North America. Canadian Journal of Forest Research, 23, 23 17-2327. Wahigren, J.R. (1979). The effects of the herbicide Roundup® on reproduction in mice. B.S. Agriculture thesis, University of British Columbia, Vancouver. Walinsley, M.E., Utzig, G., Void, T. & Van Barneveld, J. (1980). Describing Ecosystems in the Field. British Columbia Ministry of Environment and Ministry of Forests, Land Management Report Number 7. Yahner, R.H. (1986). Microhabitat use by small mammals in even-aged forest stands. American Midland Naturalist, 115, 174-180. Zar, J.H. (1984). Biostatistical Analysis. 2nd edition. Prentice-Hall, New Jersey. Zieliuski, W.J., Spencer, W.D. & Burnett, R.H. (1983). Relationship between food habits and activity patterns of pine martens. Journal of Mammalogy, 64, 387-396.  60  APPENDIX I Randomization testing for changes in small mammal population size following manual and cutstump treatments: test statistics and P values. 3  Prior to calculating test statistics, data were standardized across replicate plantations to eliminate potential differences among trapping weeks other than treatment effects. First, within each trapping week and irrespective of treatment type, MNA estimates of population size were turned into values relative to the median value for that trapping week. Next, differences from the median were standardized to a common “variance” within each trapping week, by dividing each value by the absolute value of the median difference between the overall median and all observed values. This was equivalent to creating a standard normal deviate with a mean equal to 0 and standard deviation equal to 1 except that medians were substituted for means in all appropriate calculations. Medians were chosen over means because medians are less sensitive to outliers. After data were standardized, test statistics were calculated as follows. First, a data matrix was created with each row containing the records for one replicate plantation and each column containing standardized MNA population size estimates for a single trapping week. Second, pre and post-treatment data were averaged separately within rows. Third, again within rows, the average post-treatment value was subtracted from the average pre-treatment value to indicate change in population size for individual plantations. Fourth, these estimates of pre- to post-treatment change in population size were averaged within control, manual, and cut-stump  This description of data manipulation and calculations for randomization testing borrows 3 extensively from an informal manual produced by Dr. W. Hochachka (Department of Zoology, University of British Columbia) and distributed with his computer program for randomization analysis. 61  treatments to combine data from replicates.  Finally, pair-wise differences were calculated  between control and manual treatments, control and cut-stump treatments, and manual and cut stump treatments; these pair-wise differences were the test statistics (S) for the observed data. Once test statistics were calculated from the observed data, the data matrix was independently and randomly shuffled within each trapping week (i.e. within columns of the data matrix). After the data were randomized, new versions of the test statistics (S’) were calculated  and compared to the test statistics from the observed data. A record was kept of whether the observed test statistics (S) were larger or smaller than the test statistics created from the randomization. This procedure continued until the data set had been randomized 5000 times. Five thousand iterations should be sufficiently large that the test statistics from the observed data were compared with all possible permutations of the data set in proportion to the relative likelihood of each possible permutation. The final probability value associated with each pairwise comparison between treatments (F) was equal to the proportion of times that the test statistic from the observed data was larger, or smaller, than the test statistic generated from the randomized data sets. Throughout this analysis, no assumptions were made as to the value of missing data. As a result, the occurrence of missing data values was shuffled among replicates within trapping weeks in exactly the same manner as were other data values. Also, for chipmunks, there was so little variability in population size within some trapping weeks that the median-based “variance” used in standardizing data became zero and subsequent calculations failed. As a result, the absolute value of the mean deviation of observed values from the median population size within trapping weeks was used instead of the median deviation to standardize data for this species.  62  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075186/manifest

Comment

Related Items