Management of Amphibians, Reptiles, and Small Mammals in North America

Home , Cow Knob salamander

Abstract.-The status and ecology of Plethodon Forestry Operations and punctatus was investigated in George Washington Terrestrial Salamanders: National Forest, Virginia to determine potential effects of logging. Pitfall traps and mark-recapture Techniques in a Study of the supplemented searching by hand. Elevation, aspect, soil characteristics, and number of cover Cow Knob Salamander, objects (rocks) are the most important features that identify P. puncfatus habitat, Intensive logging Plethodon puncfafusl operations appear to be detrimental to this species.

Kurt A. B~hlrnann,~Christopher A. Pag~e,~ Joseph C. Mit~hell,~and Robert 8. Glasgow5

Increasing emphasis is being placed on Shenandoah and North Mountain ton National Forest, Virginia. Before on conservation and preservation of of western Virginia and eastern West its purchase, between 1911 and 1940, biological diversity worldwide Virginia (Highton, 1972, Tobey, by the U.S. government, this area (Norse et al., 1986; Wilson, 1988). 1985). Most of the known range of was repeatedly logged and burned U.S. federal and state agencies have this recently described species (Leichter, 1987; original land deed become concerned about the bio- (Highton, 1972) is in the George documents). Few virgin stands of diversity of their managed lands and Washington National Forest. Fraser forest remain, and regrowth and log- are directing efforts towards preserv- (1976) compared some aspects of the ging operations has resulted in a mo- ing natural biota. From a manage- ecology of this species with a sympa- saic of mixed hardwoods of various ment perspective, research on am- tric congener, Plethodon hoffmani. ages. phibians and reptiles lags behind that Little else is known of the ecology of We selected five sites of different devoted to game animals, such as this salamander. Because of its rela- aged forest to determine the relative some mammals, birds, and fish (Bury tively small range and unknown abundance of Plethodon punctatus (fig, et al., 1980). This is partly due to a status, P. punctatus was added to the 1) to see if its presence was affected previous lack of interest in these U.S. Fish and Wildlife Service's Cate- by logging. All sites selected had groups, but also because some spe- gory 2 list (U.S. Fish and Wildlife similar aspects (S-SE) and elevation cies can be more difficult to observe Service, 1985). Potential timber har- (914- 1127 m) (table 1). We used or investigate. vesting within the range of this spe- USDA Forest Service compartment The Cow Knob salamander, Ple- cies (USDA Forest Service, 1986) descriptions and maps to aid in se- thodon punctatus, is a dark, moder- prompted us to examine its status in lection of sites and to obtain informa- ately large (to 74 mm snout-vent forest stands of various ages. In this tion on the history, physical and bio- length), woodland, fossorial amphib- paper we report the following as- logical descriptions, and future man- ian (Martof et al., 1982) found only pects of this study: techniques of cap- agement goals for each site. A com- ture and data collection, salamander partment is divided into a series of Paper presented at symposium, Man- agement of Amphibians, Reptiles, and habitat characteristics, and potential stands, each of which defines a for- Small Mammals in North America. [Flag- effects of logging operations. Our ob- ested area of similar tree species by staff, AZ, July 79-2 7, 1988). jective in this paper is to make other composition, age, and stand condi- 2KurtA. Buhlmann is a consulting biolo- researchers aware of the techniques tion. Stand age is defined by the age gist for the US. Forest Service. Buhlmann 's we used and the problems we en- of dominant canopy trees. Final current address is 200 7 N. Main St., Blacksburg, VA 24060. countered in developing useful man- choices of sites were made only after each was checked in the field and 3ChristopherA. Pague is a doctoral can- agement recommendations for the didate in the Ecological Sciences Program, protection of an apparently rare ter- tree age was verified by tree ring Department of Biological Sciences, Old restrial salamander. analysis. Dominion University, Norfolk, VA 23508. In each site we erected drift fence 4Joseph C. Mitchell is a Research Biolo- arrays (Campbell and Christman, gist, Department of Biology, University of 1982) consisting of four 60 cm x m Richmond, Richmond, VA 23 773. Materials and Methods 7.5 sections of aluminum flashing ar- 5Roberf B. Glasgow is the Wildlife Biolo- gist for the George Washington National We conducted this study on Shenan- ranged in a cross pattern. Opposite Forest, Harrison Plaza, P.O. Box 233, Harri- doah Mountain, Augusta and Rock- arms of the cross were separated by sonburg, VA 2280 7. ingham Counties, George Washing- 15 m and all sections were sunk in the ground approximately 10 cm. A ies on reproductive cycles and ecol- punctafus and to obtain information 5-gal plastic bucket was placed in the ogy. Pitfalls were checked and all on range and habitat characteristics. center of each arm and #10 cans were captures (including other vertebrates Results from timed collecting periods placed in the ground on either side of and all invertebrates) were collected allowed comparison among sites and the ends of each arm so that the tops weekly May 5 - June 18,1987, bi- dates of collection. Between April 20 of the pitfalls were flush with the weekly July 7 - November 22,1987 and June 2,1987, we collected data ground surface. Sites A and C con- and monthly December 1987 and on eleven microhabitat variables at tained two drift fence arrays, and the January 1988. Samples were sorted in 67 sites to evaluate those most im- remaining three sites had one array the laboratory and the vertebrates portant in predicting the presence of each. In each pitfall we put 4-10 cm stored in 10% neutral buffered for- this salamander. These variables of 10% formal-dehyde solution to in- maldehyde. Invertebrates were were elevation, aspect, slope, soil sure adequate preservation of the stored in 70% isopropanol. temperature under cover object, soil salamanders. We selected this Hand-collecting supplemented moisture, soil pH, soil description, method to obtain samples of all the drift fence collection and was used to canopy cover, number of cover ob- terrestrial fauna for a range of stud- determine the elevational range of P. jects available within a 2 m circle, type of cover object (e.g., rock, log), and forest type. One site >1 km away from any of the collection sites was selected for estimation of population size and data on individual movements. We searched for salamanders in daylight by turning and replacing all surface objects and at night while they were active on the surface (i.e., during conditions of near 100% relative hu- midity [sensu Hea twole, 1960; Jaeger 19781). Each individual was measured (snout-vent length, tail length to nearest mm), weighed (nearest 0.1 g), the sex determined, assigned to adult or juvenile age-classes, marked by toe-clipping, and released at its cap-

ture site. We marked each cavture1 Figure 1.-Plethodon puncfatus, the Cow Knob salamander, from Shenandoah Mountain, site with survey flags on which the Augusta County, Virginia. Photograph by Kurt A. Buhlmann. salamander's number and capture date were written. We noted all recaptures and measured movements in linear fashion (0.1 m) between capture points.

Results and Discussion

Capture Techniques

Nineteen P. punctatus were caught in the pitfall traps, 2.0% of the total number of salamanders. Of the 17 recorded, 12 were caught in 5-gal. buckets and 5 in #10 cans. Eleven P. punctatus were caught in Site E, six in Site B, and two Site D. None were caught in Site A or the Site C. In con- trast, by hand collecting in areas ad- jacent to Site E, we found 38 P. punctatus in 7.7 man hours of searching. The drift fence method appears only moderately effective in sampling this A salamander. It is feasible that P. punc- MAY JUN JUL AUG SEP OCT NOV DEC tatus is less likely to fall into the pitfalls than other salamander species. Figure 2.-Seasonality of drift fence captures of Plethodon cinereus and P. pu~ctatusaf Site E We observed several individuals (Tomahawk Mountain), George Washington National Forest. Adults and juveniles are in- cluded, but not hatchlings. simpling period is 5 May 1987 to 24 ~anuar~-1988. climbing rocks and tree trunks during nocturnal surface activity. This suggests that this salamander is able to detect precipices and avoid falling into pitfalls. Also, this species may be active on the horizontal surface only for limited periods of time and under specific environmental conditions. Thus, the drift fence technique, which depends on horizontal activity, may not be an effective sampling method for this salamander (R.D. Semlitsch, pers. comm.). Data from pitfall traps, combined with data from hand collecting, can provide information for management decisions. For instance, seasonal trends in surface activity were simi- larly indicated by both drift fence results (fig. 2) and captures based on hand collection (fig. 3). Comparison of P. punctatus with that of its sympa- tric congener P. cinereus (fig. 3) re- veals concordance in seasonal activity and suggests similar responses to surface environmental conditions. Tomahawk Mountain, George Washington National Forest. Black bars represent P. punctatus This information could be used to and bars with diagonal lines represent P. cinereus. Sampling dates are 22 April to 22 Novem- determine the times logging opera- ber 1987. tions would cause the least impact on ion, 1986; Bury and Corn, 1987). Our four recaptures were made. One 17. salamanders at or near the surface. results indicate this technique can be punctatus captured 28 May was re- The benefits of the drift fence tech- effective in mountainous terrain and captured on 15 October. It had nique outweighed the low numbers can be used to gain information on moved 17.4 m. Three salamanders of captures of P. punctatus. We apparently rare terrestrial salaman- were recaptured within ten days of probably would not have otherwise ders. original capture and had moved 5 2 found this species in Site D because If an endangered or otherwise m. Knowledge of movement capabili- there were few surface rocks to turn protected species is the focus of ties by P. puncfafus is an important over. Although the individuals study and cannot be collected, then part of evaluating the consequences caught may have been transients, this slight modifications of the drift of population fragmentation through species does occasionally occur at fence-pitfall design must be made. logging operations. Are salamanders this site. This rest& would not have Traps would need to be checked on a able to move out of a logged area or been obtained by hand-collecting daily basis, or nearly so, in order to repopulate it when suitable habitat alone. release the animals unharmed (Gib- conditions return? We believe mark- The drift fence method also pro- bcns and Semlitsch, 1981). Water or recapture studies can provide useful vided estimates of the relative abun- wet leaves can be placed in the pit- information on rare terrestrial sala- dance of the salamander fauna and falls for cover and moisture. Poten- manders, but realize that data may other species in the community. The tial problems include killing of the need to be collected over several relative numbers of these species and salamanders in the pitfalls by small years and under standardized condi- species groups can generate addi- mammals, especially shrews, and tions in order to provide direct an- tional information on the structure of desiccation. The loss of animals by swers. the community in which the focal shrew or raccoon predation in pitfall species lives. Drift fence techniques containers affects the samples and have been used for a variety of eco- may prevent quantitative compari- Habitat Characteristics logical studies (e.g., Gibbons, 1970; sons among sites. Data obtained Gill, 1978; Pechmann and Semlitsch, from visitation frequencies of every Preliminary evaluation of microhabi- 1986) but only recently to answer three days (Bury and Corn, 1987) to tat data indicate that four site charac- questions about vertebrate communi- once a week (Enge and Marion, 1986) teristics are most important in deter- ties in relation to forest management probably underestimate actual cap- mining the presence of P. punctatus. kg., Bennett et aE., 1980; Gibbons tures. We found P. punctatus at elevations and Semlitsch, 1981; Enge and Mar- The detection of P. punctafus at a between 732 m and 1317 rn (fig. 4). particular site depends on the time of Most sites (87%)with this species oc- f -', year, substrate type, soil depth, soil curred above 960 m. Plethodon punc- Table 2.-Seasonal differences in moisture, soil temperature, and tatus occurred on all slopes but were surface abundance of Plethodon weather conditions (see Habitat Re- more common on north-facing as- puncfatus at Flagpole Knob and Skidmore Tract, Shenandoah quirements). A simple survey of sites pects (87% of 11 sites) than east (38% Mountain, George Washington For- by hand searching and rock turning of 131, south (36% of 81, or west as- est. These sites are e 1 km apart. in daylight hours without attention pects (40% of 7). Most of the captures Flagpole Knob is a rocky, grassy to weather and seasonality will un- (67% of 21) were on slopes of 20-45". ridge habitat containing young derestimate actual abundance and Seven sites were on slopes less than oak (Querws sp.) and mapie fail to detect presence of a species. 20" and between 46" and 60". Sites (Acer sp.) pole timber, and Skid- Table 2 contains comparative data without this salamander were on a more is a virgin hemlock flsuga for two sites searched the same day similar range of slopes (< 20°, 28.6%; canadensis)/yellow birch (Betula at different times of the year and 20-45", 57.1 %; > 45", 14.3%). Wear) forest. Numbers of saleman- demcnstrates a strong seasonal ef- Soil temperatures under cover ob- ders are foGbwed by number of fect. In order to construct effective (x = man hours In parentheses. All data jects at sites with P. punctatus are based on hand-collecting re- management plans, the range and 12.3 C, 9.4-16.1, n = 36) were nearly sults. abundance of a terrestrial salaman- identical to temperatures at sites der must be known. Therefore, re- without this species (x = 12.8 C, 9.4- Data Flagpale Skidmore searchers conducting distributional 15.8, n = 15). Soil pH under cover ob- surveys must take seasonal and die1 jects were also similar (with P. punc- June 2 1 1 (0.5) 10 (1,7) changes in surface activity into con- tatus: x = 6.3,5.4-6.8; without P. punc- June 8 0(0,5) 2(2.0) sideration. tatus: x = 6.4,5.8-6.8). Average soil Sept. 28 0 (1.0) 3 (3.0) Results of our 1987 mark-recap- moisture at sites with P. punctntus

\ H' ture efforts are preliminary; only was 37.1% (12- 70%) and 42.8% (24- 80%) at sites without this species. eliminates the moisture-retaining po- habitats. Plethodon cinereus was sig- Soils in which P. punctatus were tential of the soil and leaf litter, al- nificantly less abundant in a clearcut found are characterized by shallow lows an increase in insolation (with a site compared to an old-growth site black humus intermixed with rocks concomitant increase in soil tempera- in a deciduous forest in New York (72% of 39 sites). One site where tures), and increases soil erosion (Pough et al., 1987). eleven salamanders were captured (Bury, 1983). Populations of Plethodon punctatus consisted of brown humus and ex- The use of heavy machinery com- inhabiting rocky substrates with a tensive log cover, but few rocks. pacts soil and destroys leaf litter. thin soil cover may be able to with- Cover objects under which this sala- Enge and Marion (1986) found that stand some logging operations. Our mander was found were rocks < 645 machine site preparation and Site B was logged in a salvage opera- cm2 (13.6%),rocks 645-1290 cm2 clearcutting had little effect on am- tion after ice storm damage. Not all (40.0%),rocks > 1290 cm2 (34.8%), phibian species richness in a Florida trees were removed and the sub- and logs (10.6%).Over 89% of the slash pine forest. However, of the 15 strate was not as damaged as that in captures were found under rock amphibian species they recorded, Site A, which was clearcut. These fac- cover. Number of cover objects none was a terrestrial salamander. tors, combined with the presence of a within a 2 m circle of the captured On Shenandoah Mountain, where seep near the drift fence array, salamander averaged 15.1 (1-45). most of the terrestrial amphibian probably explain the high numbers of Sites without P. punctatus ranged community is comprised of terres- P. punctatus found at Site B com- from 100% rock cover to 0% rock trial salamanders, logging and pared to other logged sites. cover. Sites with canopy cover equal clearcutting are likely to have detri- We found no P. punctatus on Sites to or greater than 50% accounted for mental effects. Salamander abun- A and C for apparently different rea- 88.2% of the captures (n = 52). dance in a 60-100 yr-old deciduous sons. Site A was clearcut, the sub- We found P. punctatus in the fol- forest in another Virginia site was strate was greatly disturbed, and the lowing forest types: mature oak/ more than four times that in 2 yr-old lack of canopy cover prevented mois- hickory (38.5% of 13), oak/maple/ and 6-7 yr-old clearcuts (Blymer and ture retention. The fact that P. punc- birch (62.5% of 8), oak/pine (33.3% of McCinnes, 1977). Bury (1983) found tatus occurred on the same ridge in a 3),young oak/maple/ hemlock (50% that terrestrial salamanders were nearby hardwood stand suggests this of 81, virgin hemlock/yellow birch more abundant in old growth com- salamander mavJ have occurred on (100% of 2), hemlock/maple/bass- pared to logged redwood forest Site A prior to logging. Site C was wood (62.5% of 81, white pine (0% of 2), and grassy balds (20% of 5). Of the site characteristics we examined, the following appear to be most important in identifying P. punctatus habitat: elevation, aspect, soil characteristics, and number of cover objects (rocks). Habitats of terrestrial salamanders differ among species and, in some cases, among geographic areas within species (e.g., Semlitsch, 1980; Tilley, 1973). Data derived from the literature for management studies and plans must be used with caution. Baseline habitat and life history studies should be conducted on the focal species at the location in question before developing management plans. 457-610 611-762 763-814 915-1067 Q68-1219 1220 + Effects of Logging Elevation ( m) Tree removal effects the terrestrial Figure 4.-Elevational distribution of Plethodon punctatus on Shenandoah Mountain, George salamander community in several Washington National Forest. Solid bars represent sites where P, punctatus was not found and ways. Removal of canopy cover ban with diagonal lines represent sites where this species was found. logged 30 years ago but was re- Seasonal and daily activity vossible1 detrimental effects of planted with white pine (table 1).The patterns of salamander activ- clearcutting on terrestrial amphibi- logging operation and change in ity must be taken into con- ans. Bulletin of the Maryland Her- vegetation type may have affected sideration when surveys are petological Society 13:79-83. the salamander populations previ- conducted to determine Bury, R. Bruce. 1983. Differences in ously present. However, because of range and population abun- amphibian populations in logged the lack of rocky substrate, we can- dance. and old growth redwood forest. not disprove the hypothesis that P. Northwest Science 57:167-178. punctatus may not have occurred Project proposals to federal Bury, R. Bruce, Howard W. there historically. and state agencies should Campbell, and Norman J. Scott, Jr. Plethodon punctatus appears to oc- contain a two step process, a 1980. Role and importance of cur in greatest abundance on rocky field survey phase to obtain nongame wildlife. Transactions of sites that contain virgin hardwoods baseline data on ecology and the 45th North American Wildlife (Site E) and sites that are not heavily life history and an experi- and Natural Resources Confer- disturbed by logging operations (Site mental phase in which log- ence, 45:197- 207. B). Clearcutting and associated dis- ging or other concerns are Bury, R. Bruce, and Paul Stephen turbance does appear to eliminate evaluated. The design of the Corn. 1987. Evaluation of pitfall populations of this salamander. Sala- experimental phase should trapping in northwestern forests: mander mortality can be minimized be based on the results of the trap arrays with drift fences. Jour- if logging operations are conducted field survey. nal of Wildlife Management outside the seasonal activity period. 51:112- 119. If size of the area logged is small, or Campbell, Howard W., and Stephen Acknowledgments if the area is logged in a mosaic, or if P. Christman. 1982. Field tech- corridors are allowed to remain, rein- niques for herpetofaunal commu- We are grateful to the following vasion may eventually be possible nity analysis. p.193-200. In Herpe- people for field assistance: Christian from peripheral populations when tological Communities. N. J. Scott, A. Buhlmann, Kara S. Buhlmann, suitable conditions return. Fragmen- Jr., editor. U.S. Fish and Wildlife Lana C. Buhlmann, Susan J. Fortuna, tation of the limited range of P. punc- Service, Wildlife Research Report Joshua C. Mitchell, and Scott M. tatus by a patchwork of clearcuts 13. Smith. David A. Young deserves spe- could seriously affect its long-term Enge, Kevin M., and Wayne R. Mar- cial thanks for helping install the survival. ion. 1986. Effects of clearcutting drift fence arrays. The rangers and and site preparation on herpe- foresters at the Dry River District tofauna of a north Florida flat- provided logistical support, equip- Conclusions woods. Forest Ecology and Man- ment, and help with site selection. agement 14:177- 192. This study was supported by a Chal- Because of budget and time con- Fraser, Douglas F. 1976. Coexistence lenge Cost Share from the U.S. Forest straints, our study attempted to ob- of salamanders in the genus Ple- Service. Additional support was pro- tain baseline data and evaluate the thodon: A variation on the Santa vided by the Nongame Wildlife and effects of logging simultaneously. We Rosalia theme. Ecology 57:238-251. Endangered Species Program of the offerthe following conclusions to re- Gibbons, J. Whitfield. 1970. Terres- Virginia Department of Game and searchers and managers who must trial activity and the population Inland Fisheries. study a salamander whose ecology is dynamics of aquatic turtles. little known. American Midland Naturalist 83:404-414. Multiple capture techniques Literature Cited Gibbons, J. Whitfield, and Raymond should be used when study- D. Semli tsch. 1981. Terrestrial drift ing an apparently rare terres- Bennett, Stephen H., J. Whitfield Gib- fences with pitfall traps: an effec- trial salamander. bons, and Jill Glanville. 1980. Ter- tive technique for quantitative restrial activity, abundance and sampling of animal populations. The life history and basic diversity of amphibians in differ- Brimleyana 7:l-16. ecology of the study species ently managed forest types. Gill, Douglas E. 1978. The metapopu- needs to be understood be- American Midland Naturalist lation ecology of the red-spotted fore the project's experimen- 103:412-416. newt, Nofophfhalmus viridescens tal design can be erected to Blymer, Michael J., and Burd S. (Rafinesque). Ecological Mono- evaluate logging effects. McGinnes. 1977. Observations on graphs 48:145-166. Heatwole, Harold. 1960. Burrowing Semlitsch, Raymond D. 1980. Geo- ability and behavioral responses to graphic and local variation in desiccation of the salamander, PIe- population parameters of the thodon cinereus. Ecology 41:661- slimy salamander Plethodon gluti- 668. nosus. Herpetologica 36:6-16.

Highton,- Richard. 1972. Distribu- Tilley, Stephen G. 1973. Life histories tional interactions among eastern and natural selection in popula- North American salamanders of tions of the salamander the genus Plethodon. p. 139- 188. In Desrnogna thus ochrophaeus. Ecology The Distributional History of the 54:3-17. Biota of the Southern Appalachian Tobey, Franklin J. 1985. Virginia's Mountains. Part 111: Vertebrates. amphibians and reptiles, a dis- Virginia Polytechnic Institute and tributional survey. Privately pub- State University, Research Divi- lished, Virginia Herpetological So- sion Monographs 4. ciety, 114 p. Jaeger, Robert 6.1978. Plant climb- USDA Forest Service. 1986. Final ing ability by salamanders: peri- land and resource management odic availability of plant- dwelling plan for the George Washington prey. Copeia 1978:686- 6%. National Forest. U.S. Department Leichter, Bill. 1987. The land nobody of Agriculture. wanted. Highlands of the Vir- US. Fish and Wildlife Service. 1985. ginia's Magazine, Autumn 1987/ Endangered and threatened wild- Winter 1988, pp. 3-7. life and plants: review of verte- Martof, Bernard S., William M. brate wildlife. Federal Register Palmer, Joseph R. Bailey, and Jul- 50:37958-37967. ian R. Harrison, 111. 1980. Am- Wilson, Edward 0.1988. Biodiver- phibians and Reptiles of the Caro- sity. National Academy Press. linas and Virginia. University of Washington, D.C. North Carolina Press, Chapel Hill, 264 p. Norse, Elliott A., Kenneth L. Rosen- baum, David S. Wilcove, Bruce A. Wilcox, William H. Romme, David W. Johnston, and Martha L. Stout. 1986. Conserving Biological Diversity in our National Forests. The Wilderness Society, Washing- ton, D.C., 116 pp. Pechrnann, Joseph H. K., and Ray- mond D. Semlitsch. 1986. Die1 activity patterns in the breeding mi- gra tions of win ter-breeding anu- rans. Canadian Journal of Zoology 64:1116-1120. Pough, F. Harvey, Ellen M. Smith, Donald H. Rhodes, and Andres Collazo. 1987. The abundance of salamanders in forest stands with different histories of disturbance. Forest Ecology and Management 20: 1-9.