Bat Use of Remnant Old-Growth Redwood Stands

WLLLARi1 j. ZIELINSa* AND STIE\;EN T. GELL U.S. Forest Service, Pacific Southwest Research Station, Redwood science;, hrcata, CA 95521, U.S.A.

Abstract: ~Ziostof the old-growth redzuood (Sequoia sempen~irens)in Cal@t*tzia has been cut; reg-enerati~zg forests willpt-obably never resemble those that were haniested, and what old growth remains on priuate land occurs in small, isolated remnant patches. The landscapes in which these stands occur dzyfer so markedly from their original condition that their value as habitat to mauy species of wildlzye, iincluding bats, is un- knoufn.Previozks research in zmnfragmented redwoodforests demonstrated that bats use basal hollows itz old- growth redufoodsas roosts. We sought to determine whether bats use similar old-growth trees as roosts when they occur in small, remnant patches of isolated old growth on commercial forest land. We compared bat oc- currence in remnant and contiguoz~sstands by collecting guano in traps suspended in hollows and by moni- toring flight activity with ultrasonic bat detectors. Hollouls in trees zuithin the remnant stands had signzyi- cantly more guano deposited per tree than the trees within the contiguous forest. me mean rzumbers of bat passesper night were statistically indisti~zguishablebetween the two treatments, aldhozk@ mean Fight uctivity in the remnant stands was greater than in the contiguous forest. Bats frequently used basal hollou~sin small ((5 ha) stands of remnant old growth, u~hichmay be due to the closer proximity of remnants to stream courses, to their greater interface with edge where foraging success may be greater, or to the fact that the lower density of hollow-bearirzg trees in remnants than in contiguous forest favored greater use per tree. Sig- nzyicant use of small, residual old-growth redwood provides reason to maintain these remnants in managed landscapes as potentially important habitat for forest bats.

Uso de Reinanentes de Bosques Maduros de Secoyas por Murcielagos Resurnen: La mayorfa de 10s bosques maduros de secoyas (Sequoia sempervirens) de Calzyornia ban sido tal- ados, 10s bosques en regeneration quiza rzunca sean lo gue alguna vez fueron 10s bosques talados y lo gue gueda de bosgue maduro en propiedudes privadas ocurre en parches renzanentes pequefios y aislados. Los paisajes en los cuales estos remanentes ocurren difieren tan marcadamente de las cotzdiciorzes originales que su valor como hdbitat para muchas especies de uida siluestre, incluyendo a los nzurci6lagos es desconocido. Inuestigaciones preuias en bosques no fragmentados de secoyns demostrnron que 10s murcie'lagos utilizan huecos en 10s &boles como perchas. Intentamos determinar si 10s murci6lagos usan drboles similat-es a 10s utilizados en bosques maduros cuando estos ocurren en parches de bosque maduro pequeGos y aislados etz tiemsforestales comerciales. Comparatmos la presencia de mut-cie'lagos en sitios renzanentes y etz sitios con- tiguos mediante la coleccion de guano en tmtnpas suspendidas en agujeros y por motzitoreo de la nctividad de vuelo ~nediantedetectores ultmscjnicos de mu~"ci6lagos.Los agujeros en 10s urboles dentro 6le 10s sitios re- tnanetztes tuuieron significatiuamerzte mas guano depositado por arb01 yue los drboles dentro de bosgues contiguos. El nzimero promedio de murci6lngos que pasaron por rzoche fzce estadisticamente irzdistizzguible entre los dos tratamientos, aungue la actiuidadpromedib de vuelo en 10s bosgues remanentesfae nzayor yue en 10s bosques copztigzcos. Los murcie'lagosfrecue~ztefnenteutilizan agujeros busales epz porciofzes remanentes peqzceAas de bosque madzaro (<5 ha), lo czcalpuede deherse a fa cerctzna proximidad de los I-emanentesn los cauces de an-oyos, su gt-an interface con el borde donde el e'xito deforrajeo &he ser mayor, o a1 hecho de yzme la densidad baja de drboles con agzajeros en las bosques remanetztesfacorece un mqyor uso por drbol. El uso signzyicatiuo de pequeAos bosques residuales maduros de secoyas provee una razon para manterzer estos re- matzentes en paisajes manejados como un hdbitat potencia1 para murcielagos de bosque.

* email u?zieZinski/[email protected]. us. Paper submitted October 17, 1997; revised manz~scriptaccepted Ma!y 20, 1998.

Conservation Biology, Pages 160- 167 Volume 13, No. 1. February 1999 Corsi et nl

editors. Large-scale ecology and conservation biology. Blackwell 210 in D. R. McCullough, editor. Metapopulations and wildlife con- Scientific Prtblications, Oxford, Iinited Kingdom. sesvation. Island Press, Washington. D.G. McKelvey, K., B. R. Noon, and R. H. Larnberson. 1992. Consenration Promberger, C., and W. Schroder, editors. 1993. Wolves in Europe: sta- planning for species occupying fragmented landscapes: the case of tus and perspectives. Munich Wildlife Sociery, Ettal, Gemany. the Northern Spotted Owl. Pages 424-450 in P. &I. Kareiva. J. G. SAS Institute, Inc. 1985. SAS user's guide: statistics. Version 6.04. Gary. Kingsolver, and R. B. Huey, editors. Biotic interactions and global North Carolina. change. Sinauer Associates, Sunderland. Massachusetts. Saunders. D. A., R. J. I-lobbs, and C. Margules. 1991. Biological conse- Xech, L. D. 19'0. The wolf: the ecology and behavior of an endan- quences of ecosystem fragmentation: a review. Consemation Biol- gered species. Natural Histoll; Press. Doubleday, New York. om 518-32. Mech. L, D. 1995. The challenge and opportunirj, of recovering =oLf Sokal. R. R.. and F. J. Rohlf. 1995, Biometq-. W. 11. Freem, New York. populations. Consen~ationBiolog) 9:270-278. Stoms. D. Sf.. F. W. Davis, and C. B. Cogan. 1992. Sensitivity of wildlife habitat mocfels to uncertainties in GIs data. Photogrammetric Engi- Mech, L. D , S. H. Fsitts, G. L. Radde, and W. J. Pad. 1988. Wolf distri- bution and road density in S4innesota. Wildlife Society Bulletin 16: neering & Remote Sensing 58:843-850. 85-87. Stornler, F. A.. and D. H. Johnson. 1986. Introciuction: biornetric ap- proaches to modeling. Pages 159-160 in J, Verner, M. L. Morrison, Meriggi, A., and S. Lovari. 1996. A review of wolf predation in south- and C.J. Ralph, editors. Wildlife 2000: modeling babitat relationships ern Europe: does the wolf prefer wild prey to livestock?Journal of of terrestrial vertebrates. University of Wisconsin Press, Madison, Applied Ecology 33: 1561- 1571. Thiel, R. P. 1985. The relationship between road densities and wolf Mladenoff, D. J., and T. A. Sickely. 1998. Assessing potential gray wolf habitat suitability in Wisconsin. American Midland Naturalist 113: restoration in the northeastern United States: a spatial prediction of 404-107. favorable habitat and potential population levels. Journal of Wild- Tomlin, C. D. 1990. Geographic information systems and cartographic life Management 62:1 - 10. modelling. Prentice-Hall, London. Mladenoff, D. J., T. A. Sickle)?,R. G. Haight, and A. P. Wydeven. 1995. Turner, M. G., G. J. Arthaud, R. T. Engstrom, S. J. Hejl. J. Liu, S. Loeb, A regional landscape analysis and prediction of favorable gray wolf and K. McKelvey. 1995. Usefulness of spatially explicit population habitat in the northern Great Lakes region. Conservation Biology 9: models in land management. Ecological Applications 5: 12- 16. 279-294. Verbyla D. L., and J. A. Litvaitis. 1989. Resampling methods for evaluat- Noss, R. S. 1992. Issues of scale in conservation biology. Pages 240- ing classification accuracy of wildlife habitat models. Environmen- 250 in P. I,. Fiedler and S. K. Jain, editors. Conservation biology the tal Managenlent 13:783-787. theosy and practice of nature conservation, preservation and man- Weaver, J. L., P. C. Paquet, and L. F. Ruggiero. 1996. Resilience and agemmt. Chapman and Hall, New York. conservation of large carnivores in the Rocky Mountains. Conser- Noss, R. F., H. B. Quigley, M. G. Hornocker, T. Mesril, and P. C. Paquet. vation Biology 10:964-976. 1996. Conservation biology and carnivore conservation in the Wiens, J. A. 1996. Wildlife in patchy environments: metapopulations, Rocky Mountains. Conservation Biology 10:949-963. mosaics and management. Pages 53-84 in D. R. McCullough, edi- Pinchera F., L. Boitani, and I;. Corsi. 1997. Application to terrestrial tor. Metapopulations and wildlife conservation. Island Press, Wash- vertebrates of Italy of a system proposed by IUCN for a new classi- ington, D.C. fication of national Red List categories. Biodiversity and Consesva- Zimen, E. 1978. Der Wolf, Mythos und Verhalten. Meyster Verlag, Wien, tion 6:959-978. Austria. Price, M. V , and M. Gilpin. 1996. Modelers, mammologists. and meta- Zimen, E.. and L. Boitani. 1975. Number and distribution of wolves in populations: designing Stephens' kangaroo rat reserves. Pages 217- Italy. Zeitschsift fiir Saugetierkunde 40:102- 112.

Conservation Biology Volume 13, No. 1, Febn~ary1999 Introduction and exists as young, second- or tbird-growth forest. Old- growth forest that occurs outside the bomdaries of fed- With the exception of extensive stands that may occur in eral and state parks does so in the form of small, remnant reserves such as national parks, remnant patches are stands that punctuate the landscape of young forest. about aU that remain of the original forest cover that pre- These small stands are increasingy valued by ddlife ceded European settlement in the United States. Approxi- managers for the unique and rare elements they pro- mately 85%of the primary forests in the United States and vide-large and sometwes hollow live trees, large snags, 95-9896 of those in the conteminous 48 states have been large logs-and by environmentalists who wish to protect harvested (Postel & Ryan 199I). The result has been an in- original forests that are not cunently in reserves. crease in the number of forest fragments, amount of edge, With the knowledge that bats use hollows in old- and isolation of remnant forest patches (Groom Sr Schu- growth redwood forests, we became interested in deter- nlaker 1993). In many cases, small patches of original for- mining whether this use was iduenced by the type of est are all that remain as examples of particular vegetation stand in which the hollow-bearing tree occurred. For in- types (Groom & Schumaker 1993: Saunders et al. 1993). stance, were trees with basal hollows in contiguous old- These scattered forest remnants cannot function as eco- growth forest used more often than similar trees in small, logical entities. Thus, it is the modified landscapes that old-growth "islands7'surrounded by forests regenerating must he mawaged to retain the full complement of species from harvest? It is commonly assumed that these small adapted to the older and rnore contiguous forests of the stands are so rare as to contribute little habitat value in past. Renmants are the obvious nuclei on which to base the region and, consequently, there have been few ef- tlie restoration of mature forests. Because rnore effort is forts to protect them outside of federal reserves. Re- being directed toward the management and preservation cently though, small "set-asides7'for Northern Spotted of forest remnants, it is necessary to determine how rem- Owls (Strix occidentalis caurina) and Marbled Murrelets nants function to provide habitat to the wildlife species (Brachyrampbus marmoratus) have begun to be con- they are intended to benefit. sidered in habitat conservation plans developed by pri- Previous work has focused on the relationship between vate companies in the redwood region (Simpson Timber stand age and indices of bat abundance, and the result is the Company 1992). %ether these set-asides should include nearly unamous conclusion that older forests favor abm- remnants of original old-growth forest is still a matter of dant and diverse communities of forest bats momas 1988; debate (A. Brickey, personal communication). Fenton et al. 1992; Huff et al. 1993; Kn~sicet al. 1996; We conducted a study to compare the use by bats of Parker et al. 1996). Our interest is different: sought to hollow, old-growth redwood trees in contiguous forest understand how bats use old trees in small remnant patches and in remnant stands to determine the importance of of old growth versus old trees in contiguous, unfragmented these increasingly rare landscape features to the commu- forest. This information may help managers assign value to nity of forest bats in the northern coastal region of Cali- the increasingly rare patches of old (>500 years) redwood fornia. We compared indices of roosting behavior and forest within the extensive matrix of younger stands (5-80 flight activity to evaluate how bats use the remnants of years old) in the north coast of California. the forests that preceded European settlement. Basal hollows in redwood trees are important roost sites for bats in coastal northern California. Hollows form as the result of periodic fires and subseq~ientwood decay O;ritz 1932; Fimey 1996) and can become very large and persist for centuries before the tree falls. Forest-dwelling Study Area bats use the fire-scar cavities in redwood as maternity. day, and night roosts and occupy hollows during every The work was conducted from tlie spring of 1992 through month of the year (Rainey et al. 1992; Gellman & Zielin- the autumn of 1994 in the California Coast Range, approx- ski 1996). Trees with the largest hollow volumes and imately 16 h south of Crescent City, Del Norte County, those nearest to available surface water appear to receive Califolrlia (Fig. 1). mecontiguous forest study area was an the greatest use by roosting bats (Gellman & Zielinski ately 1300-ha portion of Del Norte Coast Red- 1996). Our previous work in the redwood forest of the wood State Park, which adjoins 'md is managed coopera- north coast of California was conducted exclusively in a tively with Redwood National Park. Most of this part of the reserve (Redwood National Park and associated state study area was within 2 km of the ocean and was pdarFl-y parks), and sample trees were selected from among many old-growth redwood forest, with most mature trees ex- in a large, contiguous block of original forest. Tl~esefor- ceeding 500 years of age. Sitka spruce (Picea stichensis), ests are rare across the range of redwood; less than 10% western hendock (Tsuga hete+-ophylla),grand fir (Abies of the range of redwood is occupied by large patches of gt-andis), and Douglas fir (Psuedotsuga menziesiz3 were dense, old-gro~hredwood forest (Fox 1996). Most of less common species in redmroodclominated stands. Red the redwood region is privately owned, has been logged, alder (Alnus rubrum) occurred along most of the streams

Conservation Biology Volume 13, No. I, February 1999 Wilson Creek. Nets were opened about 0.5 hours prior to sunset and kept open for 3 hours or until at least I hour transpired without a bat capture at any net. Ml cap- tured bats were banded and less than 2% were recap- tured. Sampling occurred at each of tl-rese locations from one to five times fronl 12 Xifay 1993 to 9 March 1994.

Guano Smpfing in Hollow Trees Guano was sanlpled by collecting it on water-perme- able, bkack plastic screen suspended within large trees with basal hollows (Gellrnan & Zielinski 1996). Only trees with basal hollows were sampled. Beyond the rec- ognition of three morphotypes of bat guano (Gellman CZr Zielinski 1996), we were unable to identify the species of bats using the hollows. Tlierefore, our study w-as un- able to address the effects of landscape composition on the diversity of bats or on iildividual species. Sample trees were selected from the contiguous forest in the park by dividing its southern portion into 7-ha blocks (the approximate modal size of the remnant stands) and randomly selecting 18 of these blocks. Each block was traversed along transects spaced 50 m apart to locate trees with basal hollows. Three of these were randomly 7-ha sample un~ts selected, with the constraint that trees m-ithin a block had to be at least 50 m apart and 50 m from the perinle- ," remnant&ands 1 A ter of the block. Fifty-four trees from 18 blocks com- Figure 1. Study area, Del Norte County, Califowzia, de- posed the sample from the contiguous forest. picting the spatial relationship between contiguous From two to five hollow-bearing trees were selected forest in the park and remnant stands (irregular poly- from each remyant stand, depending on its size, resulting goas) in commercial timberlarzd. in a total of 45 sample trees in remnant stands. Three trees were selected from stands 7 ha or smaller, except for two small stands that contained only two suitable trees. Five that drained the study area. Wiilson Creek was the largest trees were selectecl from stands greater than 7 ha (range stream in the area. 7.4-23.5 ha). When field reconnaissance determined that The 12 remnant stands in our study represented most there were more hollow-bearing trees in a stand than were of those stands greater than 1 ha tlzat remained within a necessary to sample, the requl-ed number for each stand large area of private land east and 2-5 km inland of the was selected at mdom. Guano traps were installed in the sample units in the park (Fig. 1). Remnants ranged in first week of June 1993 and checked once a nlonth size from 1.9 to 56.3 lia (mean = 18.8 ha). All occurred though January 1994. Guano was oven-dried (69" C for witlzin a matrix of second- and third-growth forest (20- 2.5 hours) and weighed (to the nearest 1 mg) within 24 to 80-year-old redwood and Douglas fir) on commercial hours of collection. timber-producing land. The remnant stands ranged from Because previous work demonstrated a relationship 0.2 to 4.8 lm (mean = 1.9 kn~)from the boundary of the between both the diameter at breast height (dbh) and contiguous old-grow11 forest in the park (Fig. 1). The the internal volume of the hollow and the use of the tree combined area tbat included all the remnant stands ancl by bats (Gellman & Zielinski 1996), we collected data on the contiguous forest in the park totaled about 6000 ha. these variables from each tree we sampled. Hollow vol- ume was the product of its height, depth. and width. Trapping Transects walked during the process of selecting trees to sample also produced estimates of the availabili~of Because little research had prerriously been done on bats trees with basal hollows deemed suitable for bat use. A in this region of California, we conducted mist-net sur- team of assistants walked parallel transects such that all veys at nine locations along Wilson Creek to establish a trees of sufficient size to contain a hollow were located list of species that may roost in the vicinity. From one to and examined. This complete census was conducted in nine nets (12 m and 6 m) were established along a 50-m 11 of the sample units in the contiguous forest and all 12 reach of stream at each of nine locations along 5 km of of the remnant stands.

Conservation Biology Volunle 13, No. 1. Febnlary 1999 Zielinski E GeEEnun Bat Cse ofold-Grou!tbRefnnants 163

Remote Ultrasound Smpling of Flight Activity analyzed by a nested analysis for unbalanced data @ROC GLM In SAS, Inc.). Individual hollow trees were treated as We used remote ultrasonic bat detectors (habat 11, Tit- replicates within stands, which were nested within treat- ley Electronics, Australia) to monitor the activity of bats ment (contiguous or remnant stand type). The seven in the vicinity of 47 of the same hollows where guano monthly guano weights were treated as repbcates within was sampled (23 in 12 remnant stands and 24 in 15 of a tree. For the ultrasound analysis, the mean number of the contiguous forest sample units) (Hayes & Hounihan bat passes over consecutive nights at a site was used to 1994). Detectors were placed 1 m off the ground and represent bat activity at the sample point. about 5 m from, and oriented toward, the entrance to * the hollow. hlthough we lioped bats emerging from or returning to hollows would be detected by the record- ers, we viewed the method as a means for sampling gen- eral flight activity near the forest floor. Each unit in- cluded a photosensitive switch to turn the unit on at &lark, a tape delay, a cassegte tape recorder, and a 1261 One hundred and forty-two individual bats, identified to gel cell battery. Sufficient tape and battery power were seven species, were captured during 16 nights, for a total available to record throughout the night, although most of almost 260 net hours (Table I). iM~)otisyu~~zanensis, a activity occurred in the hours shortly after sunset. An at- crevice- and cavity-roosting bat (Dalquest 1947; Rainey et tempt was made to record bat activity for three consecu- al. 1992), was the most frequently captured species and tive nights at each sampling location, although at a mi- was captured most commonly during the summer. La- nority of sites only two nights of data were collected sio-tzycteris noctiuagans, which also roosts in tree cavities due to detector malfbnctions. On each sample night, ul- and crevices (Parsons et al. 1986; Barclay et al. 1988; trasound was sampled at an equal number of trees in Campbell et al. 1996), was infrequently captured during remant stands as in contiguous forest, so as to mini- the sutnrner, but males in particular comprised a substan- mize confounding site effects with temporal effects on tial portion of the fd captures. Few bats were captured bat activity. during the winter, although previous work has demon- One observer listened to all tapes and tallied the num- strated a considerable amount of activity by bats during ber of bat "passes" (Thomas 1988; Krusic et al. 1996; the vvinter months in this coastal environment (GeUman 93 Hayes I997), which included all bat vocalizations Zielinski 1996). (search phase, approaches. and feeding buzzes). The ob- server was trained to distinguish the vocalizations of Characteristics of Sampled Trees bats from other ultrasonic sources by recording and studying the vocalizations made by various species of The mean (SD) dbh of trees in tlie sample units within the bats released by hand after capture. No attempt was contiguous forest and the remnant stands was 295.9 (82.0) made to distinguish the species of bats or the types of cm and 271.2 (86.6), respectively (not statistically different, vocalizations (e.g., search phase from feeding buzzes). 1; = 1.00,p = 0.32). Trees in the contiguous forest had larger interior hollow (mean [SD] = 25.5 [4.9] and Analysis 10.73 [4.8] m3, respectively-), but the difference was also statistically indistinguishable ('= 2.78, p = 0.102). Prior to analysis, the data were assessed for departure Wilson Creek was the largest stream course in the study from normality. Both the guano and ultrasound data were area, and the sample units and remnant stands were not

Table 1. Seasonal distribution of the number of bats fma1e:femde) captured in the mist nets on Wilson Creek. Del Norte County, California, 12 May 1993-9 March 1994.

Gejzus and species Afnjt-Ju .lze Ai~gusf- October muember-Jfarch Total ilfyotis yu??zane~zsis 1l.fyofis caliJf'c1t-FZ icus L1fj~otiseco fis Myotis ~folans Lnsionj~terisnoctivng~tns L~siuruscine$-eus Eptesicus fuscus Number of net hours" Number of nights trapping "Omfemale captured in tree bollout. %ne net hour menfzs one dvrz net kt pluce for 1 hour.

Consenration Biology volume 13, No. 1, February 1999 164 B@tUse of Old-Groz~~thIZemnants Zielinski G Ge&man

equidistant from this source of food and water for bats. been increased with additional samples, given the rela- Remnant stands were siwificantly closer to Wilson Greek tively large vasiances around each mean: 56.07 and 61.40, than the sample htsin the contiguous stand (mean min- respectively. Six: sarnple locations recorded a mean num- imum distance = 304 and 1610 m, respectively; t = 7.0, ber of passes per night that exceeded 20, two of which p < 0.00001). were in remnant stands and four in contiguous stands. Hollow-bearing trees were s cmtly more common within the sample units in the contiguous forest (mean = 1.99fia) than in the remant stands (mean == 0.52fia) Discussion (t = 3.85, p = 0.002). Thus, potential roost trees were more readily available in the unfrawented forest than The guano data demonstrate a significantly greater use of within the patches of residual old growth. old-growth trees in residual stands than within the con- tiguous forest. This suggests that either more bats use Guano Deposition and Flight Activity each of these trees or individual bats return to use these trees more frequently than they do trees within the un- The mean (SD) amount of guano deposited was 0.89 fragmented forest in the park. AltbougEl the ultrasound (1.43) g per month per tree in the remnant stands and data were not statistically different, the isolated stands 0.38 (0.69) in the contiguous stands, with the biggest also had a higher index of bat activity (passes per night). difference between treatment groups occurring in the It is clear that bats are making significant use of old- month of June (Fig. 2). Acco~mtingfor the nesting of sam- growth remnants, which make up a small proportion of ple period within tree, tree within stand number, and the landscape. stand number within treatment, the hollow- trees within Little is known about the dynamics of bat use of red- the remnant stands had significantly more guano depos- wood forests and forest remnants. Our work was not de- ited per tree than the trees within the contiguous, park- signed to determine the origin and movements of the land forest (I; = 4.95,~= 0.035). bats that used the stands we sampled. It is therefore pos- Based on the mean of the multiple-night sample for sible that, given the abilities of some species to travel each tree averaged over all trees per stand, the remnant considerable distances (Morrison 1979; Wilkinson 1985; stands averaged 10.02 passes per night, whereas the con- Brigham 1991; de Jong 1994), the bats we detected in tiguous forest sample units averaged 7.80 passes per night the remnants may at times commute between the rem- (Table 2), a difference that was not statistically significant nants and the larger, intact forest in the park. The areas (I; = 1.00, p = 0.33). Power would undoubtedly have between the remnants and the contiguous forest had few old trees but were probably not entirely hostile to bats becat~sethey were composed of regenerating redwoods and hardwoods that in some areas provided dense cover

Remnant 0 Contiguous Table 2. Mean number of bat passes per night, averaged across the trees within each stand, for sample sites within remnant stands and sample units within contiguous forest.

Stand or sample unit number Remnant Contiguousforest

0.0 Jun Jut Aug Sep Uct Nov Bec

Month Figure 2, Meafz z~~eightsclfgztano collectedfro;znbob lows sampled in remnant stands Cn = 45 tmes) and in the contz'guozcsforesL (n. -- 54 trees)ft-om Jzlll3,1993 to Januav 1994 in Llel 1Vorte County, Cal~yor~ztd.Traps were checked the first week ofeuev g?zontJp7so the monZh& daka represent guano depclsited from the firs$ week of tbat MZOB~~~to .the first week of the follozui~zg rszontlg.

Consenration Biology Volume I3,liio. 1. February 1999 and probably some temporar)r roosting sites. But, we crease in tree mortalie and tree fall that occurs as patch have no data on the use of younger redwood forests as size decreases (Chen & Franlclin 1992). The net result roost sites for bats. Regardless of whether the bats were was that there were fewer basal hollows ak~ailableas resident in the vicinity of the remmt or traveled there roost sites, which may be responsible for the greater per from elsewhere, we must still explain their significantly capita use by bats. This occurred despite the fact that the higher use of hollow-bearing trees in the rerpsnant than in mean volume of hollows in the remaat stands was sub- the unfragmented ibrest. VEre considered (1) the proxim- stantially smaller than that in the contiguous stands, a ity to foraging areas, (2) the proxinlit_V to water, and (3) characteristic that has previously been reported to be the densiq of adtable roost trees. negatiueQSassociated with bat use (Gellman & Zielinski W%en foraging, bats often move along forest edges 1996). It is possible that the increased use by bats of hol- nlore than within the forest interior (Black 197d; Kunz cR- lows in resid~ialstands was due to the concentration, or Martin 1982: de Jong 1994; Crampton & Barclay 1996). itcornb et al. 1981; Lemkuhl et al. 1991), This may facilitate orientation but it may also rnaximize of individuals disturbed by tirnber harvest elsewhere into contact with prey (Lirnpens & Upteyn 1991). The noc- remaining habitat. Most of the harvest occurred 20-30 turnal aerial insects that are typical prey for bats are more years prior to the study, howet~er,before most of the abundant along forest edges and in recently disturbed ar- bats could have been born, so this explanation is less eas than in forest interiors (Cross 1988; Grindal 1996), likely. Despite the fact that old-growih remnants were and bats concentrate where prey are abundant (de Jong & rare in the landscape and contained fewer and smaller Ahlkn 1991). Remnants with suitable roost sites may offer hollows than did the contiguous stands, they retained sig- some bats an opportunity to minimize the distance be- nificant value to forest bats. tween foraging and resting locations. Furthermore, bats Our data demonstrate that small remnants of original or may use trees in remnant stands as night roosts-betpreen old-growth forest continue to function as important habi- foraging bouts-and return to the contiguous forest to tat for forest bats. This conclusion agrees with the work roost during the day. This behavior could also result in of Crampton and Barclay (1996), who found that ,Myotis higher guano deposition sates in. the remnant stands, If activity levels did not change substatltially following for- tlzis w-ere occurring we might expect that vocalizations in est fragmentation, and of Fenton et al. (1992), who found the remnant stands would be distributed more evenly that bat captures generally remained high as long as some throughout the night and that those in the contiguous original forest ren~ained.Erickson and West (1 996) found stands would occur primarily at the onset of activity. The that Llilyotis activity was greater in mature stands. but temporal pattern of bat passes, however, showed no clear there was no dEerence for a number of other species. differences; most activity at both stand types occurred in We do not believe, however, that there is anything inher- the hour or two immediately after sunset. ently attractive about the remnants that resulted in the in- Not only were the basal l-rollows within the remants creased use of basal hollows in trees that occur there. closer to edge habitat that may provide diverse foraging Xeither is there reason to suspect that a landscape domi- opportunities, but they also were five times closer to the nated by young, developing forest with a few remnants largest stream course in the study area than were hollows would provide better habitat for forest bats than an in- in the contiguous forest. After they emerge in the tact, continuous forest; substantial evidence exits to the evening, bats usually seek open water for drinking and contrq (Thomas 1988; Fenton et al. 1992; Huff et al. foraging (Christy & West 1993), and our earlier study 1993; I(rusic et al. 1996; Parker et al. 1996). The lower demonstrated greater use of redmrood trees that were availability of basal hollows in the remants and their for- closer to available water (Gellman & Zielinski 1996). Bats tuutous proximity to water probably explain why individ- drink at open water and are also attracted to insect prey ual hollows in remnants received greater use by bats abundant in riparian areas (Cross 1988; Thomas & West compared to those in hollows in the parkland reserve. 1989), &vo reasons to expect bats to use roosts that min- The Wilson Creek watershed is an example of how an ex- imize the distance among foraging. drinking, and resting tremely modified landscape can continue to provide hab- areas. ,Wyotis yamanensis, in particular, is closely associ- itat for bats when most but not all of the large-cavity ated with stream courses as both feeding and d roosting structure has been eliminated. Orrr data provide sites mrigkam et al. 1992). an indication of the value of semxants to forest bats, The Perhaps the most persuasive reason to expect greater practice of harvesting the remnants to "clean up" all the use of basal hollow-s in the remnants is the fact that there miscellaneous fragments of old growth in a landscape are fewer of them per unit area than in the contiguous and to bring all the stands into rotation for efficient man- forest. And they occurred in a stand type-the rem- agement will probably affect bats and other wildlife. Mar- nant-that w-as itself a rare element in the landscape mo- bled Mumelets and Northern Spotted Owls have also saic. The lower density of hollows within the remnant been reported to either nest or occur in small remnant stands was probably due to isolated incide~ltsof individ- stands of old-gro~hredw-ood (ftfiuer & Ralph 1995; L. ual tree selection hamest and also to the inevitable in- Diller, personal comunication).

Conservation Biology Volume 13, No. I, Febntary 1999 166 But Gse of OH-Grou~tbRemzftants Zieli~ski t? Gell~~zut?

Remnants are the only old growth that occur in many fmgmented and t~nfragmentedaspen mrxed%uod stancis of ct~f-fcr- watersheds. They should be viewed as the nuclei for the enr ages Pages 238-259 in R W R Narcla) and R 91 Brigham, edr- restoration of habitat. or at least as stepping stones in a tors Bats and forest3 s) mposium 'tfml\rtq of Forests, Yictoria, Brit- rsb Cdolttmbsa,Chwada management schenle to lifzk larger units of forest man- C russ, S P 1988 Riparian s) stems and small mammals and bats Pages aged for late-seral stmcture and hnctiun. A similar value 93- 112 in K J Raectcke, editor Streamsicte managenlent riparian has been recognizecl for small, isolated lira~aentsof trop- wildlife ancf forestq interactions <,or,mbution 59 Institute of Tor- ical forest, despite the fact that they may not be able to cst Resources, t nir erc~tjof \% ashington, Seattle Dalqtlest, %/ W 1947 Notes on the natural btstt~r)of the hat. .Z[j~c>tis supporr all species (Turner & Corlett 19961. Remants yamanensis in <,alifc>miawith a description of a new race The that are close to protected parkland. like those consid- ~ericanMidland ru'aruralrst 38:224-7.-~-' ered here, may actually expand the effective size of the de Jong. J 1994 1)iatribution patterns and habitat use b) bats in rels park for species that can izxove easily between areas. It is tion to landscape heterogeneity. and consequences for consemi- aDuarent fkom the number of soecies associated with tion Ph.D. dissertatio~lSwedish University of Xgric~~ltrrralSctences, 3. Vppsala. late-sen1 forest and whose has reduced by de Jong, J., and I. hhlen 1991. Factors affecting the distributlon pattern tinlber l~arvestG7.S. Forest Service & U.S. Bureau of Land of bats in Upplanci, central Sweden. Holarctic Ecolo~).2:92-96. ik2dnagerneraL 1994) that many species in addition to bats Erickson, J. L., and S. I). West. 1996. Managed forests in the western would benefit from protecting and linking the best of the Gascactes: the effects of sera1 stage on bat habitat use patterns. remaining fragments of original forest. Pages 21 5-227 in R. M. R. Barclay and K. h~l.Briglxim, editors. Bats ancl forests symposium. Ministry of Forests, Victoria, British Colum- bia, Canacla. Fenton, M. B., L. Acharya, D. Audet, M. B. 6. Hickey, C. Merriman. Acknowledgments M. K. Obrist, and I>. M. Syme. 1992. Ivllostomid bats (Chiroptera: Phyllostomiclae) as inclicators of habitat disntption in the Neo- We thank T. Lawlor for advice during the inception of tropics. Biotropica 24:440-446. Finney. M. A. 1996. Development of fire scar cavities on old-growth the study and IM. Kampratli, hi. McKenzie, C. Shimizu, T. coast redwood. Pages 96-98 in J. Leblanc, eciitor. Coast Redwood Lesh, R. Noyes, and C. Ogan for assistance in the field. 6. Forest Ecology and Management, University of California, Berkelev. Hodgson and R. Eads provided consultation in wiring and FOX, L. 1996. Current status and tiistrihution of c~astredwood. Pages packaging the bat cietector components. C. Carroll, L. 18-19 in J. LeBkanc. editor. Coast Reclwood Forest Ecology and Diller. M. McKenzie. and 13. Welsh ~rovideclcomments Mai~agemcnt,University of California, Berkeley Fritz. E. 1933. The role of fire in the retiwood region. Journal of For- on early drafts of the manuscript, and R. Schlexer helped estry 29:368-380. prepare the figtlres. Logistical assistance provided Gehan, S. T., and W. 5. Zielinski. 1996. Ilse by bats of old-growth red- L. Diller and J. Momber, who facilitated access to prop- wood hollows on the north coast of California. Journal of Mammal- erty owned by the Sirnpson Timber Company. M. Moore og). 77:255-265. facilitated the funding provided by the california L)~~~~-Grindal. S 1). 1996 Habitat use by bats in fragmented forests. Pages 34. ment of Transportation. Additional funding was provided 260-272 in R. R. Barclay and R. M. Brigham, editors. Bats and forests symposium. Ministry of Forests, 'Cictoria, British Columbia, by the Pacific Southwest Research Sration of the U.S. For- Canada. est Service. Groom. M.J., and N. Schumaker. 1993. Evaluating landscape change: patterns of worldwicie deforestation anct local fragmentation. Pages 24-44 in P. M. Karieva, J. G. Kingsolver, and R. B. Huey, editors. Bi- Literature Cited otic interactions and global cliange. Sina~~erAssociates. Sunder- land, Massachusetts. Barclay, R. M., P. A. Faure, and D. R. Farr. 1988. Roosting behavior and Hayes, J. P. 1997. Temporal variation in activity of bats and the design roost selection by migrating silver-l~airedbats (Lasion3~cteri.snocti- of echolocation-monitoring studies. Journal of Mamnlalogy 78: uag~g~s).Journal of Mammalogy 69:82 1-825. 514-524. Black, H. L. 1974. A north temperate bat community: stmcture ancl Hayes. J. P., and P. Hounihan. 199q. Field use of the habat I1 bat- prey popukations. Journal of Mammalom 55: 138- 157. detector system to monitor bat activity-. Bat Research News 35: 1-3. Brigham, R. M. 1991. Flexibility in foraging and roosting behaviortr by Huff, M. H.,J. F. Lehmkuhl, and J. A. Young. 1993. Linking bat activity the big brown bat (Ef~tesic~rsfusczks). Canadian Journal of Zoology with landxape-level attribtites using GIs. Northwest Science 67: 13 1. 69:117-121. Ksusic, R. A,. M. Yamasaki, C. D. Neef~~s,and P. J. Pekins. 1996. Sat Brigham, R. M.,H. D. J. Pi. Aldsidge, and R. L. Mackey. 1992. Variation habitat use in White Mountain National Forest. Journal of Wildlife in babitat use and prey selection by Yuma bats, i14~0ti~yumanen- Management 60:625-63 1. sis. Journal of Mammalogy 73:640-645. Kunz, T. N.,and R. A. Martin. 1982. Plecotus toujnsendii. Mammalian Campbell, L. A,, J. G. I-Pallett, and M.A. O'Conneil. 196.Consemation species 175. .imerican Society of Mammalogists, New York. of bats in managed forests: use of roosts by Lasicit~cteisnocliva- Lemkuhl, J. F., L. F. Ruggiero. and P. A. Hall. 1991. Landscape-scale pat- gans. Journal of Mammalog-y 77976-984. terns of forest fragmentation and wildlife richness ancl abundance <;hen, J., and J. F. Franklin. 1992. Microclimatic pattern and basic bio- in the southern Washington Cascade range. Pages 425-442 in L. F. logical responses at the clearcut edges of old-groa~thDouglas-fir Ruggiero, K. B. Attbry, A. B. Carey. and M. H. Huff, editors. Wildlife stands. Northwest Environmental Journal 6:424-425. and vegetation of tinmanaged Douglas-f-ir forests. General technical Christy, R. E.. ant! S. D. West. 1993. Biology of bats in Douglas-fir for- report PW-GTR-385. tJ.S. Forest Service, Pacific Northsliest Re- ests. General technical report PW-GTR-308. t!.S. Forest Service, search Station. Portland, Oregon. Pacific Northwest Research Station, Portland, Oregon. Limpens, H. J., and K. Kapteyn. 1991. Bats, their behaviour and linear Crampton, L. H., and R. M. Rarclay. 1996. Habitat selection by bats in Landscape elements. Myotis 29:39-48.

Conseniltion Riolrtgy Volume 13, No. I, February 1999 Miller C L . and C J Ralph 1995 Rel,~tionshipof marbled murrelets nng project on fragmented landscapes a rettew of current mfor- u~ithhabttdt characteristrcs at lnlanci sttei, in C dlifornra Page%205- marlon Biological C:on.lenration 64: 185- 192 215 m C J R;lph, G L Htmt, tl C, Raphnel, andJ F Rart edrtors Simpson Timber C:omprmy 1992 Ilabttat consenation phn for the Ecology nnd consen7atlon of the %Parbled Wurrelet C,enerdt technl- Nonhem Spotted O%l on the <,alifi,rnra tirnberlanils of the Sinlpsctn cal report PSW-GTR-152 1! 5 Fore% Sen Ice Pacific \o~rtbn7e(.tRe- Timber C,ompan? S~mpsonTimber <,ompan). Korbrl, California search 'station, klbdny tal~fomta Ttiomar. D M 1988 The distrib~ttionot hats in different ages of I>~LI- lilomstsrt, D W 19'9 Apparent male defense ctf tree hullov~srn the g1a.i-fir forests Jo~~nlalof tX rlcllife Managrnent 42:619-626 " fmtt bat, I-I~;TZ~CU~~~JC-C;PPZL~!ZL~IZSI.~Joi~rnal of X4ammalctgj 60: 11-1 5 Thomas 11 % , and S D West 1989 Sampllng methods for bats Gen- Parker D 1 , J A C oak, and S W Leuis I996 Wfects of timber har- eral tethnrcat report, PNW-245 I S Forest Senice, Pacrfic torth- vest on bat acttt ltt m 5c)tltheastern 41,aska s trnlperate ratnfc>rests west Research Station, Portland, Clregon Pages 27'-292 m K M R Barcb) and R ?/I Bnghdnl, editor4 Bat4 Turner, I M , and R T C.orlett 1996 The consenatlot1 lal~~eof small and farests syrnp0~1um M~ntctqof Forests, Vlctorid, Bnr~sh( olurn- isolated fragments of lot.iland troplcal rarn forest Trcnds m Ect~l- hia C,nnada om and Evolution lll:330-333 Pnr4oris H J , D A Smrth, and R F %%rttam 1986 Maternit! colotlres t 5 Forest Sen~iceand 1 S Bureau of Land tlanagernent 194 41naI of srl?er-haired bats, laszonj~cte?"zsrzoctzrs~xg~~rzs in Ontario and supplemental entlronmetltal impact statement on management of Saskatchewan Journal of %ammaion 67:598-600 habitat far 1;tte-success~onaland oldgrottth forest related species Postel S and J C R! 'in 1991 Reforming fore5tq Pages 74-92 in L wrthin the range of the Northern Spotted Chi Portland, Oregon Stdrke, edstor State ctf the world 1991 a %orldwatch lnstrtute re- \!%ttcomb, R F J F Ljncfi, dncl M K iCi~mh~ewci1981 Iftecm of lor- port on progress toward a s~~stalnablesociety W W horton, New est f~~gmentationon avifauna of the eastern deciduous forest Pages York 124-205 in R L Burgess md D M Sharpe, editors Forest islanil clj- Ralney, W E , E D Pierson, M Colberg, 'xnd J H Bdrcla) 1992 Bats nanzics m man-dominettedlandscapes, Springer-Verlag, New York sn hollow redwoocis seasonal use and role In nrltrsent transfer into Wilkinson. G S 1985 The social organization of the common vampire old growth commultirties Rat Rewarch hews 3371 bat I Pattern and cause of association Beha~ioralEcology and So- Saunders. 11 A R J Hobbs, and G W Arnold 1993 Tlie Kellerber- ciobiology 121 1 l - 122

Consemtion Biology Volume 13, No. 1, February 1999