ABUNDANCES OF SMALL MAMMALS IN FIR FORESTS IN NORTHEASTERN CALIFORNIA
JEFFREY R. WATERS AND CYNTHIA J. ZABEL
United States Forest Service, Pacific Southwest Research Station, 1700 Bayview Drive, Arcata, CA 95521
We compared abundances of seven species of forest rodents among three types of fir (Abies concolor and A. magnifica) forest: unlogged old-growth, unlogged mature, and shelter- wood-logged old-growth. Small mammals were livetrapped during summers 1991 and 1992 in four grids within each type of forest; grids were located in the Lassen National Forest in northeastern California. Shelterwood-logged forests had been logged 6-7 years previ- ously. Differences in capture rates between unlogged, old-growth and shelterwood-logged, old-growth forests suggest that logging led to significant increases in populations of golden- mantled ground squirrels (Spermophilus lateralis), yellow pine chipmunks (Tamias amoe- nus), and lodgepole chipmunks (Tamias speciosus) but may have led to reduced populations of western red-backed voles (Clethrionomys californicus). Capture rates of Douglas’ squir- rels (Tamiasciurus douglasii), Allen’s chipmunks (Tamias senex), and deer mice (Pero- myscus manicukatus) did not differ significantly between unlogged and shelterwood-logged forests. Capture rate of T. douglasii was significantly greater in mature forests than in old- growth forests in 1992, but we did not detect significant differences between old-growth and mature forests for the other six species. These results illustrate how opening of the canopy and disturbance of the forest floor can lead to significant changes in patterns of abundance of forest rodents.
Key words: Clethrionomys californicus, Peromyscus maniculatus, Spermophilus lateralis, Tamias amoenus, Tamias senex, Tamias speciosus, Tamiasciurus douglasii, logging, small mammals
Four silvicultural systems traditionally changes in abundances of small mammals have been recognized for harvesting and re- following clearcutting. Traditional clearcut- generating forests (Smith, 1986): clearcut, ting, in which all or nearly all live trees are seed-tree, shelterwood, and selection. Nu- cut, is being used less frequently on public merous studies have evaluated effects of forests in the western United States. Silvi- clearcutting on populations of small mam- cultural practices in which live trees and mals (Corn and Bury, 1991; Gashwiler, other forest components such as large snags 1959, 1970; Gunther et al., 1983; Hooven and logs are retained at the time of harvest and Black, 1976; Kirkland, 1977; Raphael, are being used more frequently (Franklin et 1988; Sims and Buckner, 1973; Tevis, al., 1997). 1956a; Van Home, 198 1), but published In a previous paper, we compared den- studies on effects of seed-tree, shelterwood, sities of northern flying squirrels (Glauco- and selection harvests are lacking. Losses mys sabrinus) among three types of fir of food and cover associated with harvested (Abies) forest in northeastern California trees and disturbance of the forest floor, as (Waters and Zabel, 1995): unlogged old- well as increases in food and cover asso- growth, unlogged mature, and shelterwood- ciated with increased production of herba- logged old-growth. In the shelterwood sil- ceous vegetation and shrubs following dis- vicultural system, the mature stand is re- turbance, typically lead to significant moved in at least two cuts, so that regen-
1244 eration of the new stand occurs under a par- was referred to as young by Waters and Zabel, tial forest canopy (Smith, 1986). In this pa- 1995). These forests were dense and had closed per, we compare abundances among the canopies, and virtually no herbaceous plants or three types of forest of seven additional shrubs were present in the understory. Old, dead species of small mammals captured during stems on the forest floor indicated that shrubs were abundant for some period of time after the study of flying squirrels: golden-man- wildfires occurred. The organic soil included tled ground squirrel (Spermophilus lateral- layers of litter and humus. Based on counts of is), lodgepole chipmunk (Tamias specio- growth rings from cored trees, we estimated that sus), yellow pine chipmunk (T. amoenus), most codominant and dominant trees in this for- Allen’s chipmunk (T. senex), Douglas’ est type were 80-100 years old. squirrel (Tamiasciurus douglasii), deer The shelterwood-logged, old-growth forest mouse (Peromyscus maniculatus), and type (hereafter, shelterwood-logged forests) was western red-backed vole (Clethrionomys located in Swain Mountain Experimental Forest. californicus). Our objectives were to eval- The experimental forest was dominated by old- uate effects of these shelterwood harvests growth forests until 1984-1985 when large areas on abundances of forest rodents and deter- were logged to study natural regeneration rates mine differences in abundances of forest ro- using the shelterwood silvicultural system. dents between old-growth and mature fir These timber harvests left a park-like stand structure of widely spaced, large-diameter trees. forests. The ground was intentionally disturbed to ex- pose mineral soil for natural regeneration. Trac- MATERIALS AND METHODS tors with brush blades were used to remove logs Study area.-Thethree types of forest in and disturb soils, and slash was piled and burned which we sampled (unlogged old-growth, un- or broadcast burned. At the time of our study, logged mature,and shelterwood-logged old- grasses, forbs, and low shrubs (primarily Cea- growth) were located within and around Swain nothus cordulatus and Ribes roezlii) had become Mountain Experimental Forest, Plumas Co., in established, and the little organic soil present on the Lassen National Forest in northeastern Cal- the forest floor primarily was characterized by ifornia. This area lies at the southern extent of undecomposed litter. the Cascade Range, and soils are of volcanic or- Sampling-We selected four areas within igin. The unlogged, old-growth forest type each of the three types of forest that were similar (hereafter, old-growth forests) was characterized in elevation and composition of trees, 2 150 m by forests with multi-layered canopies and nu- apart, relatively homogeneous, and sufficiently merous large, decayed logs, stumps, and snags. large. These restrictions precluded random se- Dense patches of small firs occurred in the un- lection from a larger number of potential areas, derstory, but herbaceous plants (e.g., Pyrola pic- so we do not attempt to make inferences beyond ta, Viola purpurea, and Corallorhiza maculata) the areas studied. Elevations of the 12 grids and shrubs (primarily Chrysolepis sempervirens) ranged from ca. 1,800 to 2,000 m. Each area was were uncommon. The organic soil included lay- dominated by white fir (Abies concolor) or a ers of litter and humus and large pieces of bur- mixture of white and red fir (A. magnifica). Jef- ied, decayed wood. Based on counts of growth frey pine (Pinus jeffreyi), ponderosa pine (P. rings on cut stumps in shelterwood-logged areas ponderosa), sugar pine (P. lambertiana), and within Swain Mountain Experimental Forest, we lodgepole pine (P. contorta) occurred at low estimated that the majority of codominant and densities in some stands. The four areas in shel- dominant trees in the old-growth and shelter- terwood-logged forests were located within wood-logged forest types were 200-400 years Swain Mountain Experimental Forest. Three of old. the four areas in the old-growth forest type were The unlogged, mature forest type (hereafter, located in residual stands of old-growth within mature forests) was characterized by even-aged the experimental forest, and the fourth area was forests that grew back after stand-replacement located 6 km southwest of the experimental for- wildfires. Few residual logs, snags, or trees from est. One of the areas in the mature forest type the previous forest were present (this forest type was located in Swain Mountain Experimental 1246 JOURNAL OF MAMMALOGY Vd. 79, No. 4
Forest, and three were located 8 km southwest cm) on the ground at each grid point in 1991. of the experimental forest. Sherman live traps were not used in 1992 be- Within each of the 12 areas we established an cause of damage to traps in 1991 by bears. Bait 11.5-13.4-ha, rectangular-to-square grid (7 by was a mixture of rolled oats and peanut butter 13, 8 by 13, 8 by 12, 9 by 11, or 10 by 10) with (a small amount of molasses was used in 1991 40-m spacing between grid points (size and but not in 1992). After marking and weighing shape of grids varied depending on size and individuals, we determined sex and reproductive shape of stands). Each grid was surrounded by status. S. lateralis and T. douglasii were ear- a strip of similar forest 240 m wide. During tagged, and all other species were toeclipped. surnrner 1991, we sampled vegetation at every Data analyses.-Numbers of captures and re- third point within each grid. Within circular captures varied widely among species and grids. plots, we measured diameter at breast height For several species, there were few or no recap- (dbh) of trees >5-cm dbh and length and mid- tures in at least some of the grids, so we cal- point diameter of all logs with a diameter 2 12 culated capture rate as a measure of relative cm at the midpoint. Trees <5-cm dbh were tal- abundance for each species. Capture rate (CR) lied, not measured. Size of vegetation plot dif- was expressed as number of individuals/l,000 fered among the three types of forest because of trap nights and was determined using the for- large differences in the densities of trees. In old- mula of Nelson and Clark (1973): growth forests, small trees (< 18-cm dbh) were CR = I X l,OOO/(T - S/2), measured within a 6-m radius and larger trees and logs within a 16-m radius. In mature forests, where I = number of individuals captured, T = all trees and logs were measured within a 6-m number of traps multiplied by number of nights radius. In shelterwood-logged forests, all trees traps were open, and S = number of traps and logs were measured within an 18-m radius. sprung by all causes. To evaluate how well cap- We used a spherical densiometer to obtain a rel- ture rates reflected abundance, we computed ative measure of canopy cover at each vegeta- within-species correlations between capture rate tion plot (we used the average of three densiom- and estimated population size for those species eter readings taken 8 m from the center of the and grids in which one or more individuals was plot). Ground cover was estimated using a point- recaptured. Population size was estimated using intercept method; type of ground cover was re- the first-order jackknife estimate (Burnham and corded at 52 points/vegetation plot (13 points/ Over-ton, 1979), which has been shown to be a transect along four transects following cardinal reliable estimator (low bias and precise), even directions). We also dug a small soil pit at each when capture probabilities are low (Rosenberg point in the grid and measured the depth of or- et al., 1995). Capture rate was correlated highly ganic soil (litter, humus, and buried decayed with estimates of population size for the eight wood). species-by-year combinations evaluated; the Each year (1991 and 1992), we livetrapped eight Pearson correlation coefficients ranged small mammals during two trapping sessions; from 0.933 to 0.998, and the eight Spearman six of the grids were trapped during the first ses- ranked correlation coefficients ranged from sion (two grids within each of the three types of 0.952 to 1.000. forest were randomly selected) and the other six Sufficient data were available from both 1991 were trapped during the second session. Trap- and 1992 for only T. speciosus and T. senex. For ping sessions were in August and September in S. lateralis, T. amoenus, P. maniculatus, and C. each year and 15-16 nights long (traps were califomicus, capture data presented are from closed for 2 days in the middle of each session); 1991. Few T. amoenus, P. maniculatus, and C. trapping sessions were long because our primary califomicus were captured in 1992 because we objective was to obtain estimates of population did not use Sherman live traps that year. We did size of flying squirrels, whose capture probabil- not eartag S. lateralis in 1992 because there ities were low (Waters and Zabel, 1995). At each were large differences in abundance among the grid point, we placed two Tomahawk live traps three types of forest in 1991, and we needed to ( 13 by 13 by 41 cm): one on the ground and one reduce total processing time of small mammals. attached to a tree ca. 1.5 m from the ground. We Data for T. douglasii were from 1992 because also placed a Sherman live trap (8 by 9 by 23 only two individuals were caught in 1991.1. November I998 WATERS AND ZABEL-ABUNDANCE OF RODENTS IN FIR FORESTS 1247
TABLE 1.---Meansand standard errors for measures of vegetation in three types of fir forest (n = 4 grids in each type of forest) in northeastern California in 1991. P-values are from ANOVA; means with the same letter did not differ at ar XW.5.
Old-growth Mature Shelterwood Measure of vegetation x SE x SE x SE P Basal area of trees YGcrn dbh (m2/ha) 72.6A 1.7 92.lB 1.5 22.6C 1.6 <0.00 1 Percent canopy cover 66.OA 0.4 78.3B 1.8 24.3C 1.7
We compared measures of vegetation and significantly greater ground cover of forbs, capture rates of small mammals among types of significantly less ground cover of small-di- forest using analysis of variance (ANOVA). ameter and large-diameter logs, and a sig- Equality of variances was first tested using the nificantly thinner layer of organic soil (Ta- Brown-Forsythe homogeneity of variance test ble 1). Compared to grids in old-growth (SAS Institute, Inc.,1997). If variances did not forests, grids in mature forests had signifi- differ significantly among the three types of for- est, we used traditional ANOVA, but if vari- cantly greater basal area and canopy cover, ances differed significantly, we used Welch’s significantly more small-diameter snags and variance-weighted ANOVA (SAS Institute, Inc., significantly fewer large-diameter snags, 1997). For T. speciosus and T. senex, we tested and significantly less ground cover of large- for a year effect using repeated-measures ANO- diameter logs (Table 1). Each of the three VA; if the year effect was not significant, we types of forest had its own characteristic performed ANOVA on the average capture rate distribution of diameter size classes (Fig. 1). between 1991 and 1992 to simplify presentation Mean capture rates were greatest in old- of results. We performed multiple-comparison growth forests for two of the seven species: tests using Tukey’s studentized range test (SAS T. senex and C. californicus. For T. senex, Institute, Inc., 1989) if variances did not differ significantly among types of forest or a vari- repeated-measures ANOVA indicated that ance-weighted, adjusted, Tukey’s multiple-com- the effect of type of forest was not signifi- parison test (SAS Institute, Inc., 1997) if vari- cant, but capture rate was significantly ances differed significantly. For all tests, we greater in 1992 than 1991, and differences used an alpha of 0.05. in capture rate among types of forest also were significantly greater in 1992 (the in- RESULTS teraction term between type of forest and Measures of vegetative structure differed year was significant; Fig. 2). Capture rates greatly among the three types of forest. for C. californicus were low and variable, Compared to grids in old-growth forests, but the difference among types of forest grids in shelterwood-logged forests had sig- was nearly significant (Fig. 3a). Mean cap- nificantly less basal area and canopy cover, ture rate of C. californicus was 5.7 times
1250 JOURNAL OF MAMMALOGY Vol. 79, No. 4 forests suggest that populations of S. later- greater abundances in logged forests (Ahl- alis, T. amoenus, and T. speciosus increased gren, 1966; Gashwiler, 1959, 1970; Hooven significantly following logging and site and Black, 1976; Martell and Radvanyi, preparation, and populations of C. califor- 1977; Raphael, 1988; Sims and Buckner, nicus probably decreased. In a previous pa- 1973; Tevis, 1956a, 1956b; Van Horne, per, we concluded that populations of G. 1981). sabrinus decreased significantly following Other studies have shown that popula- silvicultural treatments within the experi- tions of C. californicus were low in recently mental forest (Waters and Zabel, 1995). logged or young stands (Gashwiler, 1959, Consistent with our results, McKeever 1970; Hooven and Black, 1976; Raphael, (196 1) captured few S. lateralis, T. amoen- 1988; Rosenberg et al., 1994). Truffles are us, and T. speciosus in a stand of old- a primary food of C. californicus (Maser growth fir within Swain Mountain Experi- and Maser, 1988; Maser et al., 1978; Ure mental Forest in 1958 before any of the and Maser, 1982), and we found that truf- shelterwood harvests; these three species, fles were significantly less abundant in shel- however, were captured frequently in open terwood-logged forests than in old-growth stands of ponderosa pine outside of the ex- and mature forests in 199 1 (Waters and Za- perimental forest. Sharples (1983) trapped bel, 1995). C. californicus has been shown small mammals at similar elevations to our to be associated closely with large logs stands in the northern Sierra Nevada and (Hayes and Cross, 1987; Tallmon and Mills, found that T. amoenus and T. speciosus 1994), and there were significantly more were rare in closed-canopy stands but com- large logs in old-growth forests than in ma- mon in more open and disturbed stands. ture and shelterwood-logged forests. Or- Grasses, forbs, and shrubs were more abun- ganic soil also may be an important habitat dant in shelterwood-logged forests than in component for C. californicus (Rosenberg old-growth and mature forests, and these et al., 1994), and lack of a well-developed plants provide a variety of foods such as layer of organic soil may have reduced leaves, flowers, seeds, and fruit that are eat- quality of shelterwood-logged forests as en commonly by S. lateralis and Tamias habitat for C. californicus. (Bartels and Thompson, 1993; McKeever, Unlike T. amoenus and T. speciosus, T. 1964a; Nowak, 1991; Sutton, 1992; Tevis, senex was abundant in old-growth forests. 1952, 1953). Tevis (1953) found that seeds This result is consistent with Sharples of Ceanothus cordulatus and Ribes roezlii (1983), who found that of these three spe- were abundant in the stomachs of T. cies of chipmunks, only T. senex was com- amoenus, T. speciosus, and T. senex. These mon in closed-canopy forest. Raphael shrubs were present in shelterwood-logged (1988) and Tevis (1956a), however, found forests in our study area but not in old- that the abundance of T. senex was greater growth and mature forests. The disturbed in recent clearcuts than in older stands in soil profile and exposed mineral soil also northwestern California. Anthony et al. may have provided favorable burrowing (1987) found that the closely related Town- substrate for these ground-dwelling sciur- send’s chipmunk (T. townsendii) was more ids. abundant in younger forests than in older Seeds and arthropods are primary foods forests in the Oregon Cascade Range. Ro- of P. maniculatus (Gunther et al., 1983; senberg and Anthony (1993), however, Tevis, 1956a; Williams, 1959), whose mean found that density of T. townsendii was sig- capture rate was greatest in shelterwood nificantly greater in old-growth stands than forests. Most studies that have compared in young stands in western Oregon, and populations of P. maniculatus between Hayes et al. (1995) found no significant as- logged and unlogged forests have found sociation between density of T. townsendii November I998 WATERS AND ZABEL-ABUNDANCE OF RODENTS IN FIR FORESTS 1251
and age of stand in the Oregon Coast significant effects of year and the interac- Range. tion between type of forest and year for T. Unlike comparisons between old-growth senex and the large difference in captures and shelterwood-logged forests, we did not of T. douglasii between 1991 and 1992 sug- observe large differences in relative abun- gest that caution should be taken in gener- dances of small mammals between old- alizing about patterns of abundance from 1 growth and mature forests. C. californicus or 2 years of data. may have been less abundant in mature for- Rodents are prey to a variety of preda- ests than in old-growth forests, but we were tors, and changes in populations of rodents not able to detect a significant difference may affect populations of predators. Waters with our limited data for that species. T. and Zabel (1995) noted the possibility of a douglasii was the only species whose cap- causal relationship between limited use of ture rate differed significantly between shelterwood-logged forests within Swain these two types of forest. Although fungi Mountain Experimental Forest by Califor- can be an important part of its diet (Mc- nia spotted owls (Strix occidentalis occi- Keever, 1964b), seeds of conifers are the dentalis) and low densities of G. sabrinus, primary food of T. douglasii (Flyger and the owl’s primary prey in the Lassen Na- Gates, 1982; McKeever, 1964b; Smith, tional Forest. American martens (Martes 1970). One hypothesis explaining why T. americana), however, used shelterwood- douglasii was most abundant in mature for- logged forests greater than expected based ests in 1992 is that production of cones was on availability during summer (L. M. Ellis greatest in that forest type. We cannot test and C. J. Zabel, in litt.). On several occa- this hypothesis because we did not quantify sions, we observed martens killing or chas- production of cones, but basal area of co- ing S. lateralis and Tamias in shelterwood- nifers was significantly greater in mature logged forests, and Zielinski et al. (1983) forests than in old-growth and shelterwood- found that primary prey of martens in sum- logged forests. mer at a site in the northern Sierra Nevada The dramatic increase in captures of T. were ground-dwelling sciurids (Spermoph- douglasii between 1991 and 1992 may have ilus and Tamias). Ground squirrels and been related to a prolific crop of fir cones chipmunks, however, are less available as in autumn 1991. The forest floor was blan- prey in winter because they hibernate. keted with bracts of fir cones that autumn. Our results illustrate how patterns of Populations of T. douglasii and red squir- abundance of forest rodents can change sig- rels (Tamiasciurus hudsonicus) are known nificantly following opening of the canopy to fluctuate greatly with variation in cone and disturbance of the forest floor. Our re- crop (Flyger and Gates, 1982; Kemp and sults also illustrate how varied effects of Keith, 1970; Smith, 1970), and Sullivan logging can be, even among closely related and Sullivan (1982) reported a 5-10-fold and relatively few species. Negative effects increase in density of T. douglasii as a re- on some species most likely mean positive sult of supplemental feeding. The signifi- effects on others and vice versa, so caution cantly greater abundance of T. senex in should be taken to avoid simplified gener- 1992 than in 1991 also may have been re- alizations about potential effects of logging lated to the huge crop of fir cones in autumn on communities of small mammals. 1991. We observed T. senex eating seeds from fir cones along the roads during spring ACKNOWLEDGMENTS 1992. Sullivan et al. (1983) found that We thank J. Blythe, M. Brady, M. Engle, E. abundance of T. townsendii in British Co- Geiger, D. Hamilton, T. Hines, A. Kim, L. lumbia increased significantly following Mackelburg, T. A. Mikkelson, J. Nickerson, C. supplementation of food with seeds. The Voss J. Weeks, and M. Whitt for their hard 1252 JOURNAL OF MAMMALOGY Vol. 79, No. 4 work in the field. B. Ditrnan, B. Wales, G. regulation of red squirrel (Tamiasciurus hudsonicus) Smith, and R. Volke of the Lassen National For- populations. Ecology, 5 1:763-779. KIRKLAND, G. L., JR. 1977. Responses of small mam- est provided greatly appreciated support for this mals to the clearcutting of northern Appalachian for- study. We also thank J. Baldwin, J. Blakesley, J. ests. Journal of Mammalogy, 58:600-609. Hayes, B. Laudenslayer, K. McKelvey, B. Van MARTELL, A. M., AND A. RADVANYI. 1977. Changes in Home, and B. Zielinski for their helpful reviews. small mammal populations after clearcutting of northern Ontario black spruce forest. The Canadian Field-Naturalist, 91:41- -46. ITERATURE ITED L C MASER, C., AND Z. MASER. 1988. Mycophagy of red- AHLGREN, C. E. 1966. Small mammals and reforesta- backed voles, Clethrionomys californicus and C. tion following prescribed burning. Journal of For- gapperi. The Great Basin Naturalist, 48:269-273. estry, 64:614-618. MASER, C., J. M. TRAPPE, AND R. A. NUSSBAUM. 1978. ANTHONY, R. G., E. D. FORSMAN, G. A. GREEN, G. Fungal-small mammal interrelationships with em- phasis on Oregon coniferous forests. Ecology, 59: WITMER, AND S. K. NELSON. 1987. Small mammal populations in riparian zones of different-aged co- 799-809. niferous forests. The Murrelet, 68:94-102. MCKEEVER, S. 1961. Relative populations of small mammals in three forest types of northeastern Cal- BARTELS, M. A., AND D. P. THOMPSON. 1993. Sper- mophilus lateralis. Mammalian Species, 440: 1-8. ifornia. Ecology, 42:399-402. BURNHAM, K. P. AND W. S. OVERTON. 1979. Robust -. 1964a. The biology of the golden-mantled estimation of population size when capture proba- ground squirrel, Citellus lateralis. Ecological Mono- bilities vary among animals. Ecology, 60:927-936. graphs, 34:383-401. CORN, P. S., AND R. B. BURY. 1991. Small mammal - . 1964h. Food habits of the pine squirrel in communities in the Oregon Coast Range. Pp. 241- northeastern California. The Journal of Wildlife 254, in Wildlife and vegetation of unmanaged Doug- Management, 28:402-404. las-fir forests (L. F. Ruggiero, K. B. Aubry, A. B. NELSON, L., JR., AND F. W. CLARK. 1973. Correction Carey, and M. H. Huff, technical coordinators). for sprung traps in catch/effort calculations of trap- United States Forest Service General Technical Re- ping results. Journal of Mammalogy, 54:295-298. port, PNW-285: l-533. NOWAK, R. M. 199 1. Walker’s mammals of the world. FLYGER, V., AND J. E. GATES. 1982. Pine squirrels Fifth ed. The Johns Hopkins University Press, Bal- (Tamiasciurus hudsonicus and T. douglasii). Pp. timore, Maryland, 1:l-642. 230-238, in Wild mammals of North America: bi- RAPHAEL, M. G. 1988. Long-term trends in abundance ology, management, and economics (J. A. Chapman of amphibians, reptiles, and mammals in Douglas- and G. A. Feldhamer, eds.). The Johns Hopkins Uni- fir forests of northwestern California. Pp. 23-31, in versity Press, Baltimore, Maryland, 1,147 pp. Management of amphibians, reptiles, and small FRANKLIN, J. F., D. R. BERG, D. A. THORNBURGH, AND mammals in North America: proceedings of the J. C. TAPPEINER. 1997. Alternative silvicultural ap- symposium (R. C. Szaro, K. E. Severson, and D. R. proaches to timber harvesting: variable retention Patton, technical coordinators). United States Forest harvest systems. Pp. 11l-l39, in Creating a forestry Service General Technical Report, RM- 166: l-458. for the 2 1st century: the science of ecosystem man- ROSENBERG, D. K., AND R. G. ANTHONY. 1993. Differ- agement (K. A. Kohm and J. F. Franklin, eds.). Is- ences in Townsend’s chipmunk populations between land Press, Washington, D.C., 475 pp. second- and old-growth forests in western Oregon. GASHWILER, J. S. 1959. Small mammal study in west- The Journal of Wildlife Management, 57:365-373. central Oregon. Journal of Mammalogy, 40: 128- ROSENBERG, D. K., W. S. OVERTON, AND R. G. AN- 139. THONY. 1995. Estimation of animal abundance when - 1970. Plant and mammal changes on a clear- capture probabilities are low and heterogeneous. The cut in west-central Oregon. Ecology, 5 1: 1018-1026. Journal of Wildlife Management, 59:252-261. GUNTHER, I? M., B. S. HORN, AND G. D. BABB. 1983. ROSENBERG, D. K., K. A. SWINDLE, AND R. G. AN- Small mammal populations and food selection in re- THONY. 1994. Habitat associations of California red- lation to timber harvest practices in the western Cas- backed voles in young and old-growth forests in cade Mountains. Northwest Science, 57:32-44. western Oregon. Northwest Science, 68:266-272. HAYES, J. F!, AND S. F! CROSS. 1987. Characteristics of SAS INSTITUTE, INC. 1989. SAS/STAT user’s guide, logs used by western red-backed voles, Clethriono- version 6. Fourth ed. SAS Institute, Inc., Cary, North mys californicus, and deer mice, Peromyscus mani- Carolina, 2:891- 1686. culatus. The Canadian Field-Naturalist, 101:543- - . 1997. SAS/STAT software: changes and en- 546. hancements through release 6.12. SAS Institute, Inc., HAYES, J. P., E. G. HORVATH, AND P. HOUNIHAN. 1995. Cary, North Carolina, 1,162 pp. Townsend’s chipmunk populations in Douglas-fir SHARPLES, F. E. 1983. Habitat use by sympatric species plantations and mature forests in the Oregon Coast of Eutamias. Journal of Mammalogy, 64:572-579. Range. Canadian Journal of Zoology, 73:67-73. SIMS, H. P. AND C. H. BUCKNER. 1973. The effect of HOOVEN, E. F., AND H. C. BLACK. 1976. Effects of clear-cutting and burning of Pinus banksiana forests some clearcutting practices on small-mammal pop- on the populations of small mammals in southeast- ulations in western Oregon. Northwest Science, 50: ern Manitoba. The American Midland Naturalist, 90: 189-208. 228-231. KEMP, G. A., AND L. B. KJZITH. 1970. Dynamics and SMITH, C. C. 1970. The coevolution of pint: squirrels November I998 WATERS AND ZABEL-ABUNDANCE OF RODENTS IN FIR FORESTS 1253
(Tamiasciurus) and conifers. Ecological Mono- tions to logging of Douglas-fir. Journal of Mammal- graphs, 40:349-371. ogy, 37: 189-196. SMITH, D. M. 1984. The practice of silviculture. John - 19566. Effect of a slash burn on forest mice. Wiley & Sons, New York, 527 pp. The Journal of Wildlife Management, 20:405-409. SULLIVAN, T. P., AND D. S. SULLIVAN. 1982. Population URE, D. C., AND C. MASER. 1982. Mycophagy of red- dynamics and regulation of the Douglas squirrel backed voles in Oregon and Washington. Canadian (Tamiasciurus douglasii) with supplemental food. Journal of Zoology, 60:3307-3315. Oecologia, 53:264-270. VAN HORNE, B. 198 1. Demography of Peromyscus SULLIVAN, T. P., D. S. SULLIVAN, AND C. J. KREBS. maniculatus populations in seral stages of coastal 1983. Demographic responses of a chipmunk (Eu- coniferous forest in Southeast Alaska. Canadian tamias townsendii) population with supplemental Journal of Zoology, 59:1045-1061. food. The Journal of Animal Ecology, 52:743-755. WATERS, J. R., AND C. J. ZABEL. 1995. Northern flying SUTTON, D. A. 1992. Tamias amoenus. Mammalian squirrel densities in fir forests of northeastern Cali- Species, 390: l-8. fornia. The Journal of Wildlife Management, 59: 858-866. TALLMON, D., AND L. S. MILLS. 1994. Use of logs with- in home ranges of California red-backed voles on a WILLIAMS, 0. 1959. Food habits of the deer mouse. remnant of forest. Journal of Mammalogy, 75:97- Journal of Mammalogy, 40:415-419. 101. ZIELINSKI, W. J., W. D. SPENCER, AND R. H. BARRETT. 1983. Relationship between food habits and activity TEVIS, L., JR. 1952. Autumn foods of chipmunks and golden-mantled ground squirrels in the northern Si- patterns of pine martens. Journal of Mammalogy, 64:387-396. erra Nevada. Journal of Mammalogy, 33: 198-205. - . 1953. Stomach contents of chipmunks and Submitted 20 October 1997. Accepted 30 January mantled squirrels in northeastern California. Journal I998 of Mammalogy, 34: 3 16-324. -. 1956a. Responses of small mammal popula- Associate Editor was Troy L. Best.