The Great Basin NaturalistPUBLISHED AT PROVO, UTAH, BY

* BRIGHAM YOUNG UNIVERSITY

ISSN 0017-3614

VOLUME 51 31 DECEMBER 1991 No. 4

Great Basin Naturalist 51(4), 1991, pp. 301-309

IMPACTS OF BLACK-TAILED PRAIRIE DOGRODENTICIDES ON NONTARGET PASSERINES

Anthony D. Apa1,3, Daniel W. Uresk2,4, and Raymond L. Linder1

ABSTRACT. -In 1983 zinc phosphide, strychnine with prebait, and strychnine without prebait were applied toblack-tailed prairie dog (Cynomys Zudovicianus) colonies in west central South Dakota. Short-term (four days later) andlong-term (one year later) impacts of the rodenticides on Horned Larks (Eremophila alpestris) and other granivorousbirds inhabiting prairie dog colonies were evaluated. Horned Larks and 49 other bird species were observed.Immediate impacts reduced Horned Lark relative densities 66% with strychnine only and 55% with prebaitedstrychnine. Zing phosphide caused no measurable reduction. Horned Larks showed no long-term direct impacts.Indirect negative impacts occurred through habitat changes following prairie dog control. The granivorous guildshowed no short- or long-term effects.

Key words: Cynomys ludovicianus, Eremophila alpestris, granivorous guild, zinc phosphide, strychnine.

Poisons applied to oats are the past andpresent primary tool for black-tailed prairiedog (Cynomys ludovicianus) control. Strych-nine was introduced into the United Statesabout 1847, and its success as a rodenticidehas varied (Crabtree 1962). The alkaloid formon grain was recommended by the U. S. De-partment of Agriculture at the beginning ofthe century (Merriam 1902, Crabtree 1962).Inconsistent treatment effects in certain situa-tions and potential hazards to many nontargetspecies (Tietjen 1976) have caused concern.

Zinc phosphide was introduced as a verte-brate pest control agent in 1943 due to strych-nine shortages during World War II (Crab-tree 1962). Following replenished suppliesof strychnine and the development of sodiummonofluoracetate (Compound 1080), use ofzinc phosphide as a field rodenticide wascurtailed until it was developed specificallyfor black-tailed prairie dog control in 1976

(Tietjen 1976). S ince 1976, zinc phosphide hasbeen the only rodenticide available for prairiedog control when there is federal involve-ment.

Several granivorous passerines inhabit black-tailed prairie dog colonies (Agnew et al. 1986).Birds residing on prairie dog colonies thatcould suffer death or illness from poison con-sumption are those of seed-eating guilds (Root1967, Creighton 1974.) This guild could con-sume treated grains. Tietjen (1976) observedHorned Larks (Eremophila alpestris) a n dMourning Doves (Zenaida macroura) on zincphosphide-treated prairie dog colonies, butobservations after treatment did not locateany sick or dead birds. In contrast, Hegdaland Gatz (1977a) found significant mortality ofnontarget seed-eating birds, especially HornedLarks and Mourning Doves, when strych-nine-treated grain was applied to Richardson’sground squirrel (Spermophilus richardsonii)

1South Dakota Cooperative Fish and Wildlife Research Unit, South Dakota State University, Brookings, South Dakota 57006.2United States Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, South Dakota School of Mines, RapidCity, South Dakota 57701.

3Present address: Fish and Wildlife Resources, College of Forestry, Wildlife, and Range Sciences, University of Idaho, Moscow, Idaho 83843.4Address all correspondence to Daniel W. Uresk at the address indicated.

301

302302 GREAT BASIN NATURALIST [Volume 51

colonies. The U.S. Department of Interior(1956) acknowledged that nontarget lossesoccurred following exposure to strychnineand recommended that attempts be made tominimize the effects.

Quantitative estimates that evaluate poisonimpacts on avian residents from prairie dogcontrol have not been fully documented. Theobjectives of this study were to evaluate (1)short- and long-term effects of zinc phosphide(with prebait), strychnine with prebait, andstrychnine without prebait on Horned Larksand the granivorous avian guild and (2) com-pare three rodenticide treatments. Immedi-ate effects on birds by the three rodenticidetreatments have been reported by Uresk et al.(1988).

STUDY AREA

The study area is located on BadlandsNational Park and Buffalo Gap National Grass-lands in west central South Dakota. Climate isconsidered semiarid with an average annualprecipitation of 40 cm at the Cedar Pass Visi-tors Center, Badlands National Park. Approx-imately 80% of the precipitation falls asthundershowers during April to September,and rainfall can be localized or cover largeareas. Temperatures range from -5 C in Jan-uary to 43 C in July, with an average annualtemperature of 10 C.

Raymond and King (1976) described thesoils on the area as sedimentary deposits ofclay, silt, gravel, and volcanic ash. Topo-graphic features consist of rugged pinnacles,vegetated table top buttes, creek gullies, andgrassland basins. Gently rolling grasslandsare located in the northern portion of thestudy area with elevation ranging from 700to 1000 m.

The vegetation consists of a mosaic of na-tive grasses, forbs, shrubs, and isolatedtrees. Dominant grasses include bluegrama(Bouteloua gracilis), buffalograss (Buchloedactyloides), needleleaf sedge (Carex eleo-char-is), and western wheatgrass (Agropyronsmithii). Common forbs include scarlet mal-low (Sphaeralcea coccinea), American vetch(Vicia americana), dogweed (Dyssodia pap-posa), sage (Salvia reflexa), and prairie sun-flower (Helianthus petiolaris). The dominantshrub species is pasture sagebrush (Artemisiafrigida). Nonnative grasses include cheat-

grass (Bromus tectorum) and Japanese chess(B. japonicus). (Scientific names of plantsfollow Nickerson et al. 1976 and Van Bruggen1976.) Native herbivores are black-tailedprairie dog, mule deer (Odocoileus hemi-onus), Rocky Mountain bighorn sheep (Oviscanadensis), American bison (Bison bison),pronghorn (Antilocapra americana), black-tailed jackrabbit (Lepus californicus),, andwhite-tailed jackrabbit (L. townsendii). Smallrodents include deer mouse (Peromyscusmaniculatus) and grasshopper mouse (Ony-chomys leucogaster). Livestock are not pres-ent in the Park, but bison graze all year. Cattleare allowed to graze the National Grasslandssix months during the growing season eachyear.

METHODS

Eighteen sites on 15 15 prairie dog colonieswere sampled in 1983 and 1984 with 9 treat-ment sites and 9 controls (Uresk et al. 1986).Sites were clustered into three major areas,and each rodenticide treatment had 3 treatedand 3 control sites. Zinc phosphide with pre-bait was applied to the area within BadlandsNational Park because administrative re-straints did not allow the use of strychnine.Four of the zinc phosphide (2 treatment and2 control) sites were clustered and paired ona prairie dog colony of approximately 600 ha.The remaining treatment prairie dog colonywas located northwest of the aforementionedlarger colony, and the control colony wasnortheast of the larger colony in the BuffaloGap National Grasslands. Strychnine withand without prebait treatment was randomlyassigned to the two areas on the NationalGrasslands. The area with prebaited strych-nine was located east and south of Scenic,South Dakota. All strychnine treatment andcontrol sites were on isolated towns rangingfrom 12 to 283 ha. Within each treatmentregime, treatment or control designationwas determined randomly except when theNational Park Service imposed administra-tive restrictions.

Avian populations were sampled on 18 18 per-manent 805 x 62-m (4.9 ha) belt transects,one on each site. Relative densities of birdspecies were estimated using a modification oftechniques developed by Emlen (1971, 1977)and Rotenberry (1982). The observer walked

1991] Rodenticide IMPACTS ON PASSERINES 303

the transect line and counted birds up to 31 mon either side of the line. Sampling beganone-half hour after sunrise and continued forapproximately 4-5 hours; average walkingtime was 15-25 minutes per transect. Sam-pling was conducted on four consecutive daysfor each sampling session. Birds hoveringover and/or flying over the transect were tal-lied. In 1983 four pretreatment sampling ses-sions, prior to poisoning prairie dogs, wereconducted in June, July, August, and earlySeptember. The sampling session in earlySeptember occurred one week prior to poi-soning. The first posttreatment session inlate September commenced four days aftertreatment to evaluate short-term treatmenteffects. Four posttreatment sampling sessionswere conducted in June, July, August, andSeptember of 1984 to evaluate long-termtreatment effects.

Bait Formulation and Application

Poison application was in accordance withfederal label instructions. The untreated oats(prebait) and the poisoned oats were appliedfrom 3-wheel drive ATVs fitted with baitdispensers (Schenbeck 1982). Smaller acre-ages were treated by hand using teaspoons(H. P. Tietjen, USFWS, Denver, Colorado,personal communication).

Prebait consisted of four grams of high-quality, untreated, steam-rolled oats appliedat each prairie dog burrow. Three sites wereprebaited on 20 September 1983 and threeon 21 September 1983. A minimum of 95%of the burrows were prebaited. Prebait wasapplied (<0. 01 m2 area) at edges of prairie dogmounds. Three days after prebait applicationthree sites were treated with 4 g of 2.0% activezinc phosphide steam-rolled oats. On 23 Sep-tember 1983 three additional sites weretreated with 8 g of 0.5% strychnine alkaloidsteam-rolled oats per burrow. The last threesites were treated with strychnine oats on24 September 1983 but were not prebaited.

Statistical Analyses

Analysis of covariance was used to compareeach treated group (cluster) of sites with itsrespective control group. Applications of re-peated measures were examined but requiredconstant response through time; i.e., no inter-action between time and treatment. Thesedata did not show a constant response through

time and had interactions; therefore we usedcovariance adjustments for individual post-treatment sampling sessions. Pretreatmentobservations were used as covariates. Effect ofrodenticide treatment for each time point wasestimated as the covariance-adjusted differ-ence between treated and control sites foreach rodenticide. After obtaining an overallrejection of the hypothesis of no treatmenteffect, contrasts for each rodenticide treat-ment compared to its control were evaluatedfor significance based on a variance estimatedonly from the sites in each cluster (variancewas heterogeneous among clusters). If thecorrelation between pretreatment and post-treatment observations was not significant(cx < 0.20), then estimates of treatment effectswere based on the difference between post-treatment and pretreatment observations(repeated measures). This analysis uses theinteraction between time and treatment asthe indicator of the significant change due totreatment (Green 1979). Rodenticides werecompared by forming pairwise contrasts of thecontrasts obtained for the individual rodenti-cide treatments. Randomization procedures(Edington 1980, Romesburg 1981) based on10,000 random permutations of the data pairsamong treatment groups were used to esti-mate statistical significance of the various con-trasts. Variance of a contrast was calculated asthe sum of the variances of the means in thecontrast, with calculated individual variancesbased on the covariance and homogeneousvariance assumptions appropriate for the par-ticular variable.

Because omission of any effect due to poi-soning was considered more serious than thepotential incorrect declaration of a significanttreatment effect, Type II error protection wasproduced by testing each contrast individu-ally. However, some Type I error protectionwas afforded by testing individual contrastsonly after first observing a significant (P = .10)overall test of treatment differences usinganalysis of covariance (Carmer and Swanson1973). Individual contrasts were consideredbiologically significant at P = .20. Althoughadmittedly unconventional, for the number ofsites available for study, this significancecriterion produces a power (probability ofdetecting a true difference) of approximately0.80 for a contrast twice as large as its stan-dard error. This was considered a reasonable

308 GREAT BASIN NATURALIST [Volume 51

(Euphagus cyanocephalus) Blackbirds; Ves-per Sparrows; Western Meadowlarks; andMourning Doves. Poisoned grains treatedwith zinc phosphide also posed a lesser threatto other seed-eating birds (Tietjen 1976, Heg-dal and Gatz 1977b).

ACKNOWLEDGMENTS

Thanks are extended to Nebraska NationalForest and Badlands National Park for provid-ing study areas. Partial funding of this studywas provided by National Agricultural Pesti-cide Impact Assessment Program (NAPIAP)and Nebraska National Forest (InteragencyAgreement IAG -56).

LITERATURE CITED

AGNEW, W., D. W. URESK, AND R. M. HANSEN. 1986. Floraand fauna associated with prairie dog colonies andadjacent ungrazed mixed grass prairie in westernSouth Dakota. Journal of Range Management 39:135-139.

A p a , A. D., D. W. URESK, AND R. L. LINDER. 1990. Black-tailed prairie dog populations one year after treat-ment with rodenticides. Great Basin Naturalist 50:107-113.

BENT, A. C. 1968. Life histories of North American fly-catchers, larks, swallows, and their allies. DoverPublications Inc., New York.

CARMER, S. G. 1976. Optimal significance levels for appli-cation of significant difference in crop perfor-mance trials. Crop Science 16: 95-99.

CARMER, S. G., AND M. R. SWANSON. 1973. An evaluation often pairwise multiple comparison procedures byMonte Carlo methods. Journal of American Statis-tics Association 68: 66-74.

COURTSAL , F. R. 1983. Pigeons (Rock Doves). PagesE35-E41 in R. M. Timm, ed., Prevention andcontrol of wildlife damage. Great Plains Agricul-ture Council Wildlife Resources Committee, Ne-braska Cooperative Extension Service, Lincoln.

CRABTREE, D. G. 1962. Review of current vertebratepesticides. In: Vertebrate Pest Control Confer-ence Proceedings, California Vertebrate PestControl Tech, Sacramento. 390 pp.

CREIGHTON, P. D. 1974. Habitat exploration by an avianground-foraging guild. Unpublished dissertation,Colorado State University, Fort Collins.

DuBois , A. D. 1935. Nests of Horned Larks and Long-spurs on a Montana prairie. Condor 37: 56-72.

EDINGTON , E. S. 1980. Randomization tests. MarcelDekker, Inc., New York. 287 pp.

E m l e n , J. T. 1971. Population densities of birds derivedfrom transect counts. Auk 88: 323-342.

_ 1977. Estimating breeding season bird densitiesfrom transect counts. Auk 94: 455-468.

G i e z e n t a n n e r , J. B. 1970. Avian distribution and popula-tion fluctuations on the shortgrass prairie of northcentral Colorado. Unpublished thesis, ColoradoState University, Fort Collins.

GREEN, R. H. 1979. Sampling design and statistical meth-ods for environmental biologists. John Wiley andSons, New York. 257 pp.

Hegdal, P. L., AND T. A. Gatz. 1977a. Hazards to seed-

_

eating birds and other wildlife associated with sur-face strychnine baiting for Richardson’s groundsquirrels. EPA Final Report under InteragencyAgreement EPA-IAG-D4-0449.1977b. Hazards to pheasants and cottontail rabbitsassociated with zinc phosphide baiting for micro-tine rodents in orchards. EPA Final Report underInteragency Agreement EPA-IAG-D4-0449.

M a t s c h k e , G. H., M. P. MA R S H, AND D. L. OTIS. 1983.Efficacy of zinc phosphide broadcast baiting forcontrolling Richardson’s ground squirrels onrangelands. Journal of Range Management 36:504-506.

Merriam, C. H. 1902. The prairie dog of the Great Plains.Pages 257-270 in Yearbook of the United StatesDepartment of Agriculture. Government PrintingOffice, Washington, D. C

N ICKERSON , M. F., G. E. BRINK, and C. FE D E M A. 1976.Principal range plants of the central and southernRocky Mountains: names and symbols. UnitedStates Department of Agriculture, Forest Service.General Technical Report RM-20, Rocky Moun-tain Forest and Range Experiment Station, FortCollins, Colorado. 121 pp.

PICKWELL, G. A. 1931. The Prairie Horned Lark. Trans-actions of the Missouri Academy of Science(St. Louis) 27: 1-153.

RAYMOND, W. H., AND R. U. King. 1976. Geological map ofthe Badlands National Monument and vicinity,west-central South Dakota. United States Geolog-ical Survey. Map I-934.

ROMESBURG , C. 1981. Randomization test. ResourceEvaluation Newsletter. Pages l-3. TechnicalArticle 1. USDI-BLM, Denver Colorado.

Root, R. B. 1967. The niche exploitation pattern of theBlue-grey Gnatcatcher. Ecological Monographs37: 317-349.

ROTENBERRY, J. T. 1982. Birds in shrubsteppe habitat.Pages 307-309 in D. E. Davis, ed., Handbook ofcensus methods for terrestrial vertebrates. CRCPress, Inc., Boca Raton, Florida.

Rudd, R. L., AND R. E. GENELLY. 1956. Pesticides: theiruse and toxicity in relation to wildlife. CaliforniaFish and Game Bulletin No. 7.

Schenbeck, G. L. 1982. Management of black-tailedprairie dogs on the National Grasslands. Pages207-217 in R. M. Timm and R. J. Johnson, eds.,Fifth Great Plains Wildlife Damage ControlWorkshop Proceedings, 13-15 October 1981,University of Nebraska, Lincoln.

SIEGFRIED, W. R. 1968. The reaction of certain birds torodent baits treated with zinc phosphide. Ostrich39: 197-198.

T i e t j e n , H. P. 1976. Zinc phosphide-its developmentas a control agent for black-tailed prairie dogs.United States Department of Interior, Fish andWildlife Service Special Science Report WildlifeNo. 195. 14 pp.

T i e t j e n , H. P., AND G. H. Matschke. 1982. Aerial pre-baiting for management of prairie dogs with zincphosphide. Journal of Wildlife Management 46:1108-1112.

1991] R o d e n t i c i d e Impacts o n P a s s e r i n e s 309

U r e s k , D. W., R. M. King, A. D. Apa, a n d R. L. L inder . ceedings, Fort Collins, Colorado, GTR RM-154.1986. Efficacy of zinc phosphide and strychnine USDA Forest Service Rocky Mountain Forest andfor black-tailed prairie dog control. Journal of Range Experiment Station.

Range Management 39: 298-299. USDI. 1956. Prairie dogs and their control. United States

U r e s k , D. W. , R. M. King, A. D. Apa, M.S. Deisch, a n dDepartment of the Interior, Fish and Wildlife

5. R. L. L i n d e r . 1988. Rodenticidal effects of zincLeaflet No. 357.4 pp.

Van B r u g g e n , T. 1976. The vascular plants of Southb phosphide and strychnine on nontarget species.

Pages 57-63 in D. W. Uresk, G. L. Schenbeck,Dakota. Iowa State University Press, Ames.

and R. Cefkin, tech. coordinators, Eighth Great Received 1 May 1991Plains Wildlife Damage Control Workshop Pro- Accepted 15 September 1991