Full Issue, Vol. 64 No. 4

Western North American Naturalist

Volume 64 Number 4 Article 20

10-29-2004

Follow this and additional works at: https://scholarsarchive.byu.edu/wnan

Recommended Citation (2004) "Full Issue, Vol. 64 No. 4," Western North American Naturalist: Vol. 64 : No. 4 , Article 20. Available at: https://scholarsarchive.byu.edu/wnan/vol64/iss4/20

This Full Issue is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Western North American Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Western North American Naturalist 64(4), © 2004, pp. 421–432

EVALUATING THE GEOGRAPHIC DISTRIBUTION OF PLANTS IN UTAH FROM THE ATLAS OF VASCULAR PLANTS OF UTAH

R. Douglas Ramsey1,2 and Leila Shultz1

ABSTRACT.—Locations of 73,219 vascular plant vouchers representing 2438 species were digitized from the Atlas of the Vascular Plants of Utah (Albee et al. 1988). Source maps consist of 1:6,000,000-scale shaded relief maps of Utah with points representing collection locations by species. Location points, representing 1 or more specimens, were transposed onto these maps from the approximately 400,000 herbarium records of 3 major universities and federal land management agencies. These source maps were digitized into an ARC/Info™ database in order to reproduce the atlas in digital form. Analysis of all locations revealed a mapping bias of the original authors to avoid placing sample locations on county boundaries and over major river corridors. A comparison between ecoregions and elevation showed that the Col- orado Plateau and Wasatch/Uinta Mountains have the highest species diversity, and that areas of low elevation (1000–2000 m) have the highest number of unique species in the state. Further, species richness is related to elevation and to ecoregion boundaries.

Key words: GIS, vascular plants, Utah, ecoregions, species richness.

Biogeographers study the geographical dis- species in Utah. The database represents vouch- tribution of plants and animals. Combining er collections from 3 major research universi- geographic information systems (GIS) and bio- ties (Brigham Young University, University of geography offers a powerful tool to help under- Utah, Utah State University) and from 3 gov- stand the geographic distribution of life forms. ernment agencies (Bureau of Land Manage- This combination is used extensively by bio- ment, U.S. Park Service, U.S. Forest Service). geographers to understand the geography of Original maps, stored in the University of Utah taxa and to evaluate scale dependencies (Neil- Garrett Herbarium archives, show individual son and Marks 1994, Nichol 1994, Stoms 1994, data points color-coded by herbaria in which Ramsey et al. 1995, and Shultz et al. 1998). A specimens are stored. Seven years were re- key component to any biogeographical analy- quired to compile this atlas. It is the most sis using GIS is the availability of accurate complete source for spatial distribution of information describing the spatial distribution individual plant locations for Utah. of plants and animals. Information about the Specimen vouchers examined by the authors geographic distribution of individual plant were generated over more than a century by a species is not readily available. This lack of host of individuals. The authors critically ex- information limits biogeographers to the study amined approximately 400,000 specimens rep- of distribution of vegetation types or to a small resenting 2822 native and introduced species, group of individual taxa. In this paper we use and they mapped multiple sample locations of an extensive database of individual species loca- 2438 species. Where the same species was tions for Utah to evaluate plant collection dis- collected in approximately the same area, a tribution and species richness within the state. single dot represented multiple collections, Our objective is to understand if databases and where the location of a specific voucher like this can provide accurate biogeographical was in question, the voucher was not mapped. information in an ecological context and pro- Further, plants that were collected from only a vide an understanding of sampling bias. single location, usually considered rare, were The Atlas of the Vascular Plants of Utah not mapped. These plants represent an addi- (Albee et al. 1988) is a compilation of herbar- tional 384 taxa and are listed in the appendix ium collection locations for 2438 vascular plant of the published atlas. Including the unmapped

1Department of Forest, Range, and Wildlife Sciences, Utah State University, Logan, UT 84322-5230. 2Corresponding author.

421 422 WESTERN NORTH AMERICAN NATURALIST [Volume 64 species, there were 2822 species cataloged in instructed to digitize the center of the mapped Utah when the atlas was published in 1988. sampling point. Each species distribution map This paper deals only with the 2438 mapped was individually digitized and stored in its species. own GIS layer. Attribute information consists The authors of the Atlas of the Vascular of genus, species, lowermost elevation limit, Plants of Utah used a shaded relief map of and uppermost elevation limit for each species Utah scaled at approximately 1:6,000,000 to as stated in the atlas. All taxa are named locate vouchers. Accompanying each map is according to standard Natural Resources Con- the species’ scientific name, authority, com- servation Service (NRCS; SCS 1993) genus/ mon name, brief habitat description, growth species acronyms and are stored in individual habit, indigenous status, blooming time, and family workspaces. elevation range. The purpose of publishing In order to spatially evaluate sampling and the atlas was to document areas that have mapping bias, all points for each species are been sampled for individual species and eluci- combined in family-wide databases and in a date biogeographic patterns, to identify sam- collection-wide coverage. Family-wide data- pling gaps, and to direct future activities in bases contain all digitized points for each those areas with little or no sampling. species within that family. The collection-wide This database, in its original published form, coverage contains all points digitized for the represents an important body of work depict- entire atlas. Each point maintains its original ing distribution of plant species throughout attribution in all coverages. the state. The atlas can be used by ecologists, The spatial distribution of voucher speci- evolutionary botanists, morphologists, physiol- mens was evaluated according to ecoregions, ogists, reproductive biologists, and others to elevation, and a 649-km2 hexagonal map tessel- understand the distribution of plants along lat- lation of the state to determine landscape level itudinal, elevational, and ecoregional bound- representation of samples and to help evaluate aries. As a GIS database, the atlas can be used species richness. The Utah portion of the in conjunction with other ecological data sets national ecoregion map produced by Omernik to provide biogeographical information, poten- (1987) was used to delineate ecologically dis- tial range distribution, and sampling adequacy. tinct zones in the state. There are 5 Omernik While the atlas is an important piece of ecoregions that fall within the state bound- work and is voluminous in its coverage of plant aries: Northern Great Basin (NGB), Southern distribution, like all geographic data sets it Great Basin (SGB, which in Utah represents makes assumptions and has limitations that the eastern extension of the Mojave Desert preclude certain types of analysis. This paper floristic province), Colorado Plateau (CP), will detail the process of converting the atlas Wasatch-Uinta Mountains (WUM), and Wyo- to digital form, provide an understanding of ming Plateau (WP). An elevation zone map the limitations of using small scale (large area) was generated from a statewide, 30-m resolu- databases, and evaluate sampling adequacy tion, digital elevation model and was catego- and species distribution across ecoregion and rized into 500-m elevation zones. Elevation elevation boundaries. zones began at 500 m (msl) and terminated with a zone of elevation above the 3500-m METHODS (msl) mark (Fig. 1). The Environmental Pro- tection Agency (EPA) Emap-based, hexagonal The atlas was digitized over the span of 1 sampling frame (Carr et al. 1992) was used to year by 3 student technicians. To reduce digi- evaluate species richness (as a function of the tizing variation between technicians, a menu collections) across the state. interface was created using ARC/Info™ menu and Arc Macro Language tools. Control points RESULTS were positioned at each corner of a state boundary map to reference each distribution The atlas provides collection locations for map to the same GIS state boundary layer. 2438 species, representing 117 families. A Technicians worked together to maintain con- total of 73,219 sample points describe the dis- sistency with the database and were each tribution of species. Figure 2 shows the digital 2004] DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH 423

Fig. 1. Elevation zone map of Utah using 500-m increments. 424 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 2. Digital distribution map of sego lily (Calochortus nuttallii T. & G.) from the Atlas of the Vascular Plants of Utah. 2004] DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH 425

Fig. 3. Distribution of sampling points for the collection-wide database. version of the sego lily distribution map. Figure tion shows that there was no obvious bias in 3 shows the distribution of sampling points for mapping of points for the individual species the collection-wide coverage. (Fig. 2). However, when the collection-wide While there are areas within the state where database is displayed (Fig. 3), there is a decided a species does exist but has not been docu- mapping bias. When all 73,219 points are mented, cursory examination of point distribu- simultaneously displayed, political boundaries 426 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 4. Distribution of sampling points for the collection-wide database with ecoregions superimposed. 2004] DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH 427

Table 1. Proportion of Utah composed of each of the 5 ecoregions and percent of samples and species found in each ecoregion. Ecoregion Percent of state Percent of samples Percent of species Colorado Plateau 41.05 42.30 84.17 Northern Great Basin 38.91 21.21 67.06 Wasatch/Uinta Mountains 18.11 35.56 75.96 Southern Great Basin 0.22 0.53 14.40 Wyoming Plateau 1.71 1.75 26.05

Table 2. Proportion of Utah composed of 500-m elevation zones and percent of samples and species found in each zone. Elevation zone Percent of state Percent of samples Percent of species 500–1000 m 0.49 0.81 16.78 1000–1500 m 29.63 17.12 69.85 1500–2000 m 41.14 36.17 86.10 2000–2500 m 19.85 26.93 79.04 2500–3000 m 6.84 13.85 62.80 3000–3500 m 1.96 4.99 39.13 >3500 m 0.08 0.13 3.65

of individual counties in Utah and some river which was its intended purpose. The authors corridors are easily discernible. All 3 authors were comfortable with this bias, knowing that (Albee, Shultz, and Goodrich) seem to have most older herbarium records could not be had a mapping bias to avoid placing sample positioned more accurately and that attempt- locations on top of county boundaries. The ing to do so would be misleading. original shaded relief map used to located Plotting collection points onto ecoregion and voucher specimens contained county bound- elevation zone maps of the state depicts the aries, water bodies, and major water courses distribution of samples and species along major (the original map is similar to the map base of ecological gradients. Table 1 shows the rela- Fig. 2) to aid in locating sample points. These tionship among 5 ecoregions in the state, their map features apparently caused the mapping proportional size, the proportion of samples bias when the authors chose not to put sam- allocated to each, and the proportion of indi- pling points on county boundaries. Conversa- vidual species found in each. By comparing tions with 1 of the original authors (Shultz) each ecoregion, we found that all areas have a indicate that the authors intentionally posi- higher sampling frequency than the propor- tioned points within county boundary lines to tion of land they occupy in the state, except avoid confusion regarding voucher location. for the Northern Great Basin, which covers This bias is also consistent with standard car- 38.9% of the land area but has only 21% of the tographic practice not to directly overlay fea- samples (Fig. 4). This difference may be attrib- tures or feature names on a map, preventing utable to 2 factors: (1) the large tracts of confusion by the map reader. nonaccessible Department of Defense (DOD) Another limitation to the atlas is the size of lands and (2) the large area covered by mud- the original points used to show sample loca- flats and water in the Great Salt Lake desert, tions. The dot size represents an area approxi- which effectively lower sample density and mately 10 km in diameter on the ground, species richness of this area. The 3 largest of effectively setting a minimum resolution of the 5 main water bodies in the state reside in approximately 78 km2. This, in addition to the this ecoregion. When DOD-owned land and error of positioning points, makes the atlas a the area covered by water are removed from the general representation of sample location, analysis, the available sampling area decreases 428 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 5. Estimation of species richness relative to an EPA 649-km2 hexagon sampling frame. Values in each hexagon refer to the number of individual species collected within that hexagon. 2004] DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH 429 to 32.3% of the state, still well above the 21.2% include the Wasatch Front and the southwest- of samples allocated to it. This analysis sug- ern corner of the state where 4 different eco- gests that the Northern Great Basin is under- regions meet (SGB, NGB, WUM, and CP). represented when compared with the other ecoregions. DISCUSSION The ecoregion with the highest sampling relative to area is the Southern Great Basin The Atlas of the Vascular Plants of Utah is 1 with a 2.41 ratio percent sample to percent of 2 publications of its type in the nation area. This area, albeit small, juxtaposes 3 eco- (Albee et al. 1988). With the goal of portraying regions (Northern Great Basin, Colorado Pla- biological limits of individual species within teau, and Southern Great Basin). The Southern the state, presentation of information at this Great Basin is followed by the Wasatch/Uinta scale provides taxonomists with a guide indi- Mountains with a sampling ratio of 1.96. This cating where species have been sampled, thus analysis shows a disproportionately high amount guiding future collection efforts. However, in of sampling in the latter 2 ecoregions and a its original individual map form, it fails to relative paucity of sampling in the Northern show areas of the state that have had little or Great Basin. no sampling as a whole. It is difficult to evalu- Elevation distribution of samples is as ex- ate overall sample density by examining 2438 pected and follows closely the elevation distri- individual distribution maps. By placing the bution of ecoregions (Table 2). In general, the atlas into a GIS we are able to identify those highest sample density is found at the 3000– areas lacking in overall sample density. 3500 m zone with a sampling ratio of 2.55. Distribution maps in the atlas are intended The lowest sample density is at the 1000–1500 to show where samples have been collected, m elevation zone that coincides predomi- not the potential distribution of individual nantly with the Northern Great Basin and par- species. Therefore, lack of samples for a partially with the Colorado Plateau. Collection ticular species in one part of the state with density is high at the 500–1000 m zone and appropriate habitat does not indicate that the decreases in the 1000–1500 m zone. Sample plant does not grow there, but simply that it density then increases with elevation to the was never collected there. However, since these maximum at the 3000–3500 m zone and drops data were compiled from over 400,000 voucher again above 3500 m. specimens collected by many private, state, A comparison between ecoregions/eleva- and federal agencies for more than 100 years, tion and number of individual species found the atlas is one of the most complete sources within each area of Utah shows that the South- of habitat information for individual species ern Great Basin and Wyoming Plateau have available. Biogeographical analysis of species the greatest number of species-to-area ratios distributions can be carried out with a certain followed by the Wasatch/Uinta Mountains, Colo- level of confidence. Such analysis, however, rado Plateau, and finally Northern Great Basin should be limited to the scale of the informa- (Table 1). tion and not extrapolated to finer levels of res- Along the elevation gradient, the highest olution. species-to-area ratio is above 3500 m (Table 2). The digital form of the atlas can be used to In general, as elevation increases, species-to- evaluate species distribution and limits within area ratio increases with the exception of the and between available ecoregion and elevation 500–1000 m zone, which has the next to high- delineations. Species richness between and est species-to-area ratio. This relationship with within ecoregions, elevation, or independent elevation agrees with the findings of other sampling frames (i.e., hexagons) can be carried investigators (Gough et al. 1994, Woods et al. out to help evaluate biodiversity. However, as 1994, Benayas 1995). in the case of the hexagons, while a relation- The estimation of species richness across ship between ecoregion and elevation bound- the state using the hexagon tessellation also aries and species richness is intuitive and, for depicts some interesting patterns (Fig. 5). the most part, correct, sampling bias may also be Areas of highest species richness generally fall a determinant of species richness distribution. between ecoregional boundaries and in moun- According to the hexagon sampling frame, the tainous areas. These areas of highest richness areas of highest species richness occur in 430 WESTERN NORTH AMERICAN NATURALIST [Volume 64 Elevation zone (m) 48 154 235 329 407 0 a ______SimilarSimilarSimilarSimilar 402Similar 361 402Similar 255 1574Similar 174 361 1297Similar 1574 80 968 255 1725 2 1297 521 409 1334 1725 174 27 968 1660 801 1445 1334 80 521 1445 54 2059 911 801 2 1915 911 27 76 902 54 902 409 1660 1527 76 83 2095 83 949 1915 88 88 1527 89 949 89 able 3. Cross-comparison of species associations between elevation zones and the remainder state. T 7 of 409 species occur at the 500-m elevation zone that do not 1000-m zone. Elevation zone (m) 500–1000 1000–1500 1500–2000 2000–2500 2500–3000 3000–3500 >3500 All 500–1000 Dissimilar 7 a 1000–15001500–2000 Dissimilar2000–2500 Dissimilar2500–3000 1301 Dissimilar3000–3500 1738 Dissimilar>3500 1672 DissimilarAll 525 1357 Dissimilar 630 129 874 563 Dissimilar 202 87 406 433 197 374 2025 735 153 62 86 765 731 1182 482 43 35 1298 335 1676 1016 52 13 2045 507 629 43 1851 6 40 903 1448 12 866 1 1480 4 2345 5 0 2004] DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH 431 northeastern Utah County (734 species) and can give a relatively accurate estimate of habi- northeastern Salt Lake County (705 species). tat preference by species. Further, with the These areas coincide with the locations of the transferral of the published atlas to a spatial state’s 2 largest universities, Brigham Young database format, voucher specimens collected University and the University of Utah, and of after the publication of the atlas can be more the 2 largest herbaria. An area in northeastern easily included as well as updates to nomen- Cache County also shows a high amount of clature and habit. With the increased use of species richness (537 species), which may also GIS and GPS by taxonomists, future collections be a function of the location of Utah State Uni- can answer more site-specific questions and versity. Whether this distribution of species improve habitat preference models for indi- richness is a function of environmental gradi- vidual species (Aitken 1998). Relatively inex- ents or sampling bias or both remains to be pensive (hundreds of dollars) GPS receivers, determined. We presume that it is a function digital topographic maps, and road maps can of both, considering the large number of sam- easily be purchased. Collectors can manually ples mapped. record GPS-derived geographic positions, in a Patterns of endemism within ecological simplified form, onto samples collected in the zones can also be calculated. Such analyses field; this provides an improved location de- can provide species lists of individual taxa that scription. If collectors desire more complex are restricted to a particular region of the state. and automated methods, more expensive and More detailed analyses can be made from the robust GPS receivers coupled with data dic- digital format than from the published format tionaries and GIS software can record speci- (Shultz 1993). An example is the intersection men locations with automatically derived bio- of the elevation zone map with the combined physical parameters (elevation, slope, aspect, 73,219-sample database. Table 3 shows a com- climate, soil type, etc.). parison between each elevation zone and the rest of the state. The 1000–1500 m and 1500– ACKNOWLEDGMENTS 2000 m zones show the highest level of endemism, with 43 species found only at the We acknowledge the hard work of geogra- 1000–1500 m zone and 40 species found only phy undergraduate students at Utah State Uni- versity who helped develop this database. We in the 1500–2000 m zone. A cross-comparison also acknowledge Bonnie Banner, manager of of all elevation ranges shows a similar trend. the GIS Teaching Laboratory at USU, who spent From this analysis we see how similar 1 eleva- much time organizing students and maintain- tion zone is to another. To properly interpret ing quality control. this table, the reader must understand that while certain species assemblages may be dis- LITERATURE CITED similar between 2 adjacent elevation zones, 1 or more of those species may occur at another, AITKEN, M.A. 1998. Predictive modeling of rare plant nonadjacent, arbitrary elevation zone. Sampling habitat in the eastern Great Basin. Master’s thesis, adequacy by species in the atlas precludes this Forest Ecology, Utah State University, Logan. ALBEE, B.J., L.M. SHULTZ, AND S. GOODRICH. 1988. Atlas assumption. However, this table can be used of the vascular plants of Utah. Utah Museum of Nat- to describe species distribution similarity ural History, Salt Lake City. 670 pp. along an elevation gradient for the entire state. BENAYAS, J.M.R. 1995. Patterns of diversity in the strata of As one moves from 1 elevation zone to another, boreal montane forest in British Columbia. Journal of Vegetation Science 6:95–98. the level of similarity decreases. CARR, D.B., A.R. OLSEN, AND D. WHITE. 1992. Hexagon Voucher specimens used to generate the mosaic maps for display of univariate and bivariate atlas were collected before the common use of geographical data. Cartography and Geographic Infor- GIS and global positioning systems (GPS). mation Systems 19:228–236, 271. GOUGH, L., J.B. GRACE, AND K.L. TAYLOR. 1994. The rela- Descriptions for collection locations are gen- tionship between species richness and community eral at best, precluding site-specific informa- biomass—the importance of environmental variables. tion from being gathered at large scales (small Oikos 70:271–279. area). However, with multiple samples per NEILSON, R.P., AND D. MARKS. 1994. A global perspective of regional vegetation and hydrologic sensitivities species, modal information collected from gen- from climate-change. Journal of Vegetation Science eral maps depicting biophysical parameters 5:715–730. 432 WESTERN NORTH AMERICAN NATURALIST [Volume 64

NICHOL, J.E. 1994. An examination of tropical rain-forest tronically. Pages 259–273 in C.-I. Peng and P.P. microclimate using GIS modeling. Global Ecology Lowry II, editors, Rare, threatened, and endangered and Biogeography Letters 4:69–78. floras of Asia and the Pacific Rim. Institute of Botany, OMERNIK, J.M. 1987. Ecoregions of the conterminous Academia Sinica Monograph Series 16. United States. Annals of the Association of American SOIL CONSERVATION SERVICE. 1993. Utah list of scientific Geographers 77:118–125. and common plant names, with plant symbols and RAMSEY, R.D., A. FALCONER, AND J.R. JENSEN. 1995. The habit. USDA, NRCS, Salt Lake City, UT. 195 pp. relationship between NOAA-AVHRR NDVI and eco- STOMS, D.M. 1994. Scale dependence of species richness regions in Utah. Remote Sensing of Environment maps. Professional Geographer 46:346–358. 53:188–198. WOODS, K.D., AND C.V. COGBILL. 1994. Upland old-growth SHULTZ, L.M. 1993. Patterns of endemism in the Utah forests of Adirondack Park, New York, USA. Natural flora. Pages 249–263 in R. Sivinski and K. Lightfoot, Areas Journal 14:241–257. editors, Southwestern rare and endangered plants. New Mexico Department of Forestry and Resources Received 20 October 2003 Conservation Division, Miscellaneous Publication 2, Accepted 2 April 2004 Santa Fe. SHULTZ, L.M., N. MORIN, AND R.D. RAMSEY. 1998. Floris- tics in North America: tracking rare species elec- Western North American Naturalist 64(4), © 2004, pp. 433–438

THE MALE OF CULICOIDES REEVESI WIRTH, WITH A REDESCRIPTION OF THE FEMALE AND NEW SEASONAL ACTIVITY, DISTRIBUTION, AND BITING RECORDS (DIPTERA: CERATOPOGONIDAE)

William L. Grogan, Jr.1, Gustavo R. Spinelli2, Robert A. Phillips3, and David L. Woodward4

ABSTRACT.—The previously unknown male of the biting midge, Culicoides reevesi Wirth, is described and illustrated; the female is also redescribed and this species is reassigned to the leoni group. Previously known from California, Arizona, and New Mexico, C. reevesi is recorded for the 1st time from Utah (new record). Females of this aggressive, hematophagous species were collected while biting humans during evening crepuscular periods in California. Females exhibited a strong attraction to CO2 traps, and seasonal surveillance demonstrated that host-seeking occurred from late May until mid-October in both California and Utah. Small numbers of males were also collected in CO2 traps; however, both sexes showed little attraction to ultraviolet and incandescent light traps.

Key words: Diptera, Ceratopogonidae, Culicoides reevesi, biting midge, leoni group, nearctic, seasonal abundance.

When Willis W. Wirth described the biting did not mention the very short flagellomeres 9 midge, Culicoides reevesi, as a new species in and 10, a major diagnostic feature of this species, his classic monograph The Heleidae of Califor- but did so later (Wirth and Blanton 1956), as nia (Wirth 1952), the type series consisted of did Atchley. Atchley noted, “This is a man-bit- only females, from which he illustrated the ing species previously reported as attacking wing. Wirth originally considered this species man in California. In New Mexico it has been to be a member of the debilipalpis group due to collected biting man at three localities in the its “nearly bare wings and one spermatheca.” western portion of (t)he state. My own records Soon thereafter, Wirth and Blanton (1956) re- indicate that this species is probably crepuscu- viewed 5 related species in the debilipalpis lar, attacking at approximately sunset” (Atchley group of the subgenus Oecacta Poey, which 1967:1003). More recently, Wirth et al. (1985) they termed the leoni group: C. benarrochi assigned this species to the subgenus Haemato- Ortiz and Mirsa, from Venezuela; C. fieldi Wirth myidium Goeldi and provided an excellent and Blanton, from Honduras; C. glabellus Wirth photograph of the female wing from a speci- and Blanton, from Panama; C. leoni Barbosa, men from Needles, San Bernardino County, from Ecuador; and C. reevesi. Wirth and Blan- California. ton (1956, 1959) characterized the leoni group Recent collections of C. reevesi in Lake mainly by their small size (female wing length County, California, by DLW and Grand County, 0.63–0.75 mm), wings with sparse macrotrichia Utah, by RAP with light and CO2 traps have and extensive pale spots, antennal flagellum yielded several males and numerous addi- with sensilla coeloconica on flagellomeres 1, 5, tional females of this species. In this article we or 6–8, flagellomere 9 shorter than flagellomere describe and illustrate the previously unknown 8, a single spermatheca, and unique features male, once again describe and illustrate the of the male genitalia. female, and describe new distribution, season- Wirth (1965) subsequently reported C. al activity, and biting records of this very small, reevesi from Arizona and New Mexico, and but fierce, man-biting species. Atchley (1967) redescribed the female and We employ the updated terminology of Cer- illustrated the wing, flagellomeres 7–11, palpus, atopogonidae as presented by Downes and hind tibial spur, and spermatheca. Interest- Wirth (1981). Measurements and ratios are pre- ingly, in his original description Wirth (1952) sented as mean (minimum–maximum values,

1Department of Biological Sciences, Salisbury University, Salisbury, MD 21801. 2Departamento Cientifico de Entomologia, Museo de La Plata, Paseo del Bosque s/n, 1900 La Plata, Argentina. 3Moab Mosquito Abatement District, Box 142, Moab, UT 84532. 4Lake County Vector Control District, 410 Esplanade, Lakeport, CA 95453.

433 434 WESTERN NORTH AMERICAN NATURALIST [Volume 64 and number of specimens) measured. All aedeagus broader than long with high basal specimens examined were mounted on micro- arch (0.55 of total length) and pair of small scope slides in phenol-balsam following the subapical lateral processes, tip truncate. procedures of Wirth and Marston (1968). FEMALE.—Wing length 0.75 (0.65–0.84, n Voucher specimens are in the synoptic collec- = 25) mm; breadth 0.38 (0.32–0.43, n = 25) tions of biting midges of WLG at Salisbury mm. Head: dark brown. Eyes narrowly sepa- University, Maryland; RAP at the Moab Mos- rated by distance equal to diameter of 1–1.5 quito Abatement District in Moab, Utah; and ommatidia, with short pubescence. Antennal DLW at the Lake County Vector Control Dis- flagellum (Fig. 1A) light brown with flagellom- trict in Lakeport, California; additional eres 1–10 vasiform, 11–13 more elongated, voucher specimens will be deposited in the 9–10 distinctly shorter than 8 or 11; multiple U.S. National Museum of Natural History, sensilla coeloconica on flagellomere 1, usually Washington, DC (USNM), Museo de La Plata, a single sensilla on flagellomeres 6–8; total fla- Argentina (MLPA), University of California, gellum length (minus interflagellar spaces) 0.45 Riverside Museum of Insects (UCRC), and (0.38–0.52, n = 24) mm, lengths of flagellom- the Canadian National Collection of Insects, eres (in micrometers) of a typical specimen Ottawa (CNCI). 46-26-28-33-31-33-33-33-26-26-41-42-51; Culicoides reevesi Wirth antennal ratio 0.70 (0.63–0.77, n = 24). Palpus (Figs. 1–8) (Fig. 2) slightly paler than flagellum; 3rd segment moderately swollen with large, deep, Culicoides reevesi Wirth, 1952:193 (female, subapical pit with smaller diameter opening, California, fig. wing). pit with numerous capitate sensilla; palpal Culicoides reevesi: Wirth and Blanton, 1956:51 ratio 1.99 (1.70–2.30, n = 24). Proboscis long; (assigned to leoni group of subgenus proboscis/head ratio 0.90 (0.75–1.00, n = 24); Oecacta; redescription; figs. wing, palpus, mandible with 16–19 fine teeth. Thorax: dark hind tibial comb, spermatheca, scutum and brown, scutal pattern not apparent in slide- scutellum); Wirth and Blanton, 1959:426 mounted specimens. Legs brown; tibiae with (leoni group); Atchley, 1967:1002 (leoni group; subbasal pale bands; hind tibial comb with 4 redescription; Arizona, New Mexico; figs. spines, longest nearest spur; proximal 4 hind wing, palpus, hind tibial comb, spermatheca, tarsomeres lack apical spine. Wing (Fig. 3) antennal flagellomeres); Borkent and Wirth, with sparse macrotrichia restricted to distal 1997:80 (in world catalog). portions of posterior veins and apical wing Culicoides (Haematomyidium) reevesi: Wirth margin; 2nd radial cell in dark spot; large pale et al., 1985 (subgeneric reassignment, wing spot over r-m crossvein, poststigmatic area of photo). cell r5, and distal portion of cell m4, with DIAGNOSIS.—A typical species of the leoni smaller pale spots as follows: 1 in subapical group of the subgenus Oecacta as defined by portion of cell r5, 2 in cell m1, distal one sepa- Wirth and Blanton (1956, 1959), distinguished rated from wing margin by distance shorter from other members of the group by the fol- than spot length, 1 on extreme distal portion lowing combination of characters: antennal of cell m2 that abuts wing margin, 2 poorly flagella of both sexes with flagellomeres 9–10 developed in anal cell; costal ratio 0.56 (0.54– distinctly shorter than flagellomeres 8 or 11; 0.59, n = 25). Halter pale. Abdomen: light female with sensilla coeloconica present on brown. Segment 8 a complete ring, heavily flagellomeres 1 and 6–8, third palpal segment sclerotized; sternite 8 (Fig. 5) with 2 subcutic- with moderately deep sensory pit, proboscis/ ular, darkly pigmented, submarginal areas. A head ratio 0.94–1.00, wing with sparse macro- single pyriform spermatheca (Fig. 4) with well- trichia restricted to distal sections of veins and developed, narrow neck, measurements of typi- apical wing margin, poststigmatic pale spot in cal specimen 0.042 × 0.038 mm, neck 0.012 cell m1 separsated from wing margin by dis- mm; a rudimentary 2nd spermatheca and scle- tance shorter than spot length, 1 spermatheca; rotized ring also present. male lacking sensilla coeloconica on distal fla- MALE.—Slightly smaller, similar to female gellomeres, wing characters similar to female, with the following notable sexual differences. parameres with well-defined ventral lobe and Wing length 0.67 (0.65–0.72, n = 7) mm; fine fringe of denticles on distal portions, breadth 0.30 (0.29–0.32, n = 7) mm. Antennal 2004] BIOLOGY OF CULICOIDES REEVESI 435

Figs. 1–8. Culicoides reevesi: 1, antennal flagella: A, female, B, male; 2–5, female structures; 6–8 male structures: 2, palpus; 3, 6, wings; 4, spermatheca plus rudimentary 2nd spermatheca and sclerotized duct; 5, sternite 8; 7, genitalia, parameres removed; 8, parameres. Scales = 0.05 mm. 436 WESTERN NORTH AMERICAN NATURALIST [Volume 64 flagellum (Fig. 1B) with flagellomeres 2–9 fused; Irwin, 1 female (UCRC); Riverside, 10-VII-1984, flagellomere 1 with single sensilla coeloconica, B.A. Federici, biting man, 1 female (UCRC). distal flagellomeres apparently lacking sensilla; New Mexico: Cherry Creek, Pinos Altos, 22- total flagellum length (minus interflagellar VI-1953, W.W. Wirth, biting man, 1 female. spaces) 0.53 (0.48–0.57, n = 8) mm; antennal Utah: Grand County, Moab, CO2 trap, 24-VI- ratio 0.67 (0.64–0.70, n = 8). Third palpal seg- 1999, R.A. Phillips, 4 females; 4 km SE Moab, ment broader; palpal ratio 1.70 (1.50–1.88, n = CDC LT, 7-VIII-2001, R.A. Phillips, 1 male; 4 6). Proboscis shorter; mandible vestigial, with- km SW Moab, CO2 trap, 1-VIII-2000, 1 fe- out teeth. Wing (Fig. 6) with fewer macro- male, CDC LT, 14-IX-2001, 1 female, UVLT, trichia, restricted to apical margin of cell r5, 24-VII-2002, 1 female, R.A. Phillips; 8 km SE on and along distal section of vein M1; proxi- Moab, CO2 trap, 13-VI-2002, R.A. Phillips, 1 mal pale spot in cell m1 straddles vein M1 into female. cell r5, a single distal pale spot on anal cell; TAXONOMIC DISCUSSION.—Wirth et al. (1985) radial cells shorter; costa shorter, costal ratio inexplicably placed this species in the subgenus 0.50 (0.48–0.51, n = 7). Genitalia as in Figures Haematomyidium. We disagree with this re- 7–8. Sternite 9 with broad, shallow caudome- assignment and are convinced that C. reevesi dian excavation, ventral membrane non-spicu- is a typical member of the leoni group in the late; tergite 9 long, tapering slightly distally subgenus Oecacta as defined by Wirth and with stout, triangular apicolateral processes Blanton (1956, 1959), an arrangement with that are slightly curved medially. Gonocoxite which Atchley (1967) concurred. Subsequently, nearly twice as long as broad; ventral root however, Wirth et al. (1988) considered the large, foot-shaped, posterior heel well devel- leoni group unplaced to subgenus. oped, pointed; dorsal root slender; gonostylus Among the other species in the leoni group, slender, slightly curved distally, tip abruptly C. leoni Barbosa is most similar to C. reevesi curved, sharply pointed. Aedeagus broader than but differs in its smaller size (female wing long, length 0.8 of breadth; basal arch 0.55 of length 0.63 mm), sensilla coeloconica on fla- total length; basal arm heavily sclerotized, gellomeres 1 and 5–8, a broader 3rd palpal slender, long, bent proximally with knob-like segment (palpal ratio 1.7), wing pattern with tip; distal portion tapering gradually distally to smaller pale spot on r-m crossvein, poststig- truncate tip with pair of small, slender, sharply matic pale spot not oblique, distal pale spot in pointed lateral subapical processes (these dif- anal cell abutting wing margin, and male para- ficult to discern on some specimens). Parameres meres without ventral lobe and simple, fili- (Fig. 8) separate; proximal portions stout, form tips. heavily sclerotized, with broadly separated SEASONAL ACTIVITY AND BITING HABITS.— basal knobs, nearly straight on distal halves There are no previous reports describing the with well-developed ventral subapical lobes; seasonal activity periods of C. reevesi adults distal portions slender, tapering to finely because prior collections have been almost pointed tips with lateral fringe of very fine entirely limited to females caught while biting denticles. man (Wirth 1952, Wirth and Blanton 1959, DISTRIBUTION.—Southwestern United States. Atchley 1967). We used CDC-style suction Specimens known from Arizona, California, traps (John W. Hock Co., Gainesville, FL), New Mexico, and Utah (new record). modified by removal of the light, but with car- SPECIMENS EXAMINED.—California: Lake bon dioxide released as an attractant (CO2 County, Anderson Marsh St. Pk., Lower Lake, traps), to monitor the seasonal abundance of CO2 trap, 12-VII-2002, 3 males, 31-VIII-2002, host-seeking females. These CO2 traps were 4 females, D.L. Woodward; Konocti Conserva- operated from April through October 1 day tion Camp, Lower Lake, CO2 trap, 3-VII-2002, per week in 2 blue oak (Quercus douglasii D.L. Woodward, 4 females; Lakeport, CO2 trap, Hooker and Arnott) woodlands (425 m eleva- 20-VIII-2000, D. Woodward, 2 males; Lake- tion) near Clear Lake, Lake County, Califor- port, 457 Woodward Way, CO2 trap, 1-VIII- nia, during 2001 and 2002. Traps were also 2002, D. Woodward, 3 females, 9-VIII-2002, 4 employed on 1 or more days per week in sub- females; Lakeport, 39°01′24″N, 122°55′14″W, urban, agricultural, desert, riparian, marsh, CO2 trap, D. Woodward, 3 females, 2 males; and scrub oak (Q. gambelii Nuttall) habitats Riverside County, Deep Creek, 18-V-1964, M.E. (1205–1390 m elevation) near Moab, Grand 2004] BIOLOGY OF CULICOIDES REEVESI 437

Fig. 9. Seasonal abundance of host-seeking females of Culicoides reevesi in Lake County, California, and Grand County, Utah. Mean numbers of females caught in California by operating 4 CO2 traps per week in 2 blue oak (Quercus douglasii) woodlands during 2001 and 2002 are scaled on the left y-axis. Mean numbers of females caught in Utah by operating 2–12 CO2 traps per week in various habitats during 1999 through 2002 are scaled on the right y-axis.

County, Utah, from 1999 through 2002. On use of CO2 as an attractant during an addi- each trap-day traps were hung from tree limbs tional 312 trap-days. Adult C. reevesi showed with their entrances 1.3–1.6 m aboveground, little, if any, attraction to either incandescent baited with dry ice and operated from mid- or ultraviolet light traps as catches for 177 afternoon until the morning of the following light trap-days using incandescent light yielded day. only 4 females and 1 male, and 135 light trap- A total of 8181 female and 7 male C. reevesi days with ultraviolet light produced only 2 adults were caught during 206 CO2 trap-days females. in California, whereas 1701 female and no Additional studies of C. reevesi in oak wood- male C. reevesi were caught during 736 CO2 lands in Lake County, California, supported trap-days in Utah. Only 20% of the sampled Atchley’s (1967) conclusion that females read- habitats in Utah included an overstory of ily bite humans, particularly during the scrub oak, but 84% of the C. reevesi females evening crepuscular period. We used a CO2 caught in CO2 traps were collected from these trap with a time-segregated collection system oak habitats. Although the study sites in the 2 (John W. Hock Co., model 1512 collection states were separated by a distance of 1160 km bottle rotator) over a 10-day sample period and by more than 780 m in altitude, the sea- and collected host-seeking C. reevesi females sonal activity patterns of C. reevesi females (n = 1565) throughout diurnal periods. More were quite similar in both states (Fig. 9). In than 72% of females were caught during the both localities host-seeking females were first interval from 2 hours before sunset until sun- captured in late May, peaked in abundance set. Biting studies conducted during evening during late June, diminished in abundance crepuscular periods in the same habitats re- through July with another activity peak through sulted in the collection of C. reevesi females (n August, and gradually declined to no individu- = 27) that were biting humans on the arms, als by mid-October. hands, forehead, and scalp. In some cases the In addition to surveillance by CO2 traps, all bites of C. reevesi females resulted in raised, Utah habitats were concurrently sampled with red welts on human skin that itched intensely CDC-style light traps operated without the for several days. Although collections of C. 438 WESTERN NORTH AMERICAN NATURALIST [Volume 64 reevesi from humans were not attempted in WIRTH, W.W. 1952. The Heleidae of California. Univer- Grand County, Utah, CO trap catches of host- sity of California Publications in Entomology 9: 2 95–266. seeking females in that locality have been cor- ______. 1965. Family Ceratopogonidae. Pages 121–142 in related with biting midge complaints by local A. Stone, C.W. Sabrosky, W.W. Wirth, R.H. Foote, residents. and J.R. Coulson, editors, A catalog of the Diptera of America north of Mexico. U.S. Department of Agri- culture, Agriculture Research Service, Agriculture ACKNOWLEDGMENTS Handbook 276. WIRTH, W.W., AND F. S . B LANTON. 1956. Studies in Panama We thank Brad Mullens (UCRC) for the Culicoides (Diptera: Heleidae). IX. Two new species loan of specimens. Thanks are also extended related to leoni Barbosa and reevesi Wirth. Bulletin to Art Borkent, Christopher Briand, Boris of the Brooklyn Entomological Society 51:45–52. Kondratieff, and an anonymous reviewer for ______. 1959. Biting midges of the genus Culicoides from Panama (Diptera: Heleidae). Proceedings of the their comments and suggestions of an earlier United States National Museum 109:237–482. draft of the manuscript. WIRTH, W.W., A.L. DYCE, AND B.V. PETERSON. 1985. An atlas of wing photographs, with a summary of the numerical characters of the Nearctic species of Culi- LITERATURE CITED coides (Diptera: Ceratopogonidae). Contributions of the American Entomological Institute 22:1–46. ATCHLEY, W.R. 1967. The Culicoides of New Mexico WIRTH, W.W., A.L. DYCE, AND G.R. SPINELLI. 1988. An (Diptera: Ceratopogonidae). University of Kansas atlas of wing photographs, with a summary of the Science Bulletin 46:937–1020. numerical characters of the neotropical species of BORKENT, A., AND W. W. W IRTH. 1997. World species of Culicoides (Diptera: Ceratopogonidae). Contributions biting midges (Diptera: Ceratopogonidae). Bulletin of the American Entomological Institute 25:1–72. of the American Museum of Natural History 233. WIRTH, W.W., AND N. MARSTON. 1968. A method for mount- 257 pp. ing small insects on microscope slides in Canada bal- DOWNES, A., AND W. W. W IRTH. 1981. Chapter 28. Cerato- sam. Annals of the Entomological Society of America pogonidae. Pages 393–421 in J.F. McAlpine, B.V. 61:783–784. Peterson, G.E. Shewell, H.J. Teskey, J.R. Vockeroth, and D.M. Wood, editors, Manual of Nearctic Diptera. Received 12 August 2003 Volume 1. Agriculture Canada Monograph 27. Accepted 29 January 2004 Western North American Naturalist 64(4), © 2004, pp. 439–444

FIRST PLEISTOCENE JUMPING MOUSE (ZAPUS, ZAPODINAE, RODENTIA) FROM UTAH

Dennis R. Ruez, Jr.,1 and Christopher J. Bell1

ABSTRACT.—Two of the Little Dell Dam fossil localities produced the 1st Pleistocene records of the jumping mouse Zapus from Utah. We describe these teeth in detail and compare their morphology with both extinct and extant jumping mouse taxa. Although it is not possible to confidently assign these specimens to a particular species, the Little Dell Dam fossils are clearly distinct from the only living jumping mouse (Zapus princeps) currently known from Utah. The paracone is attached to the rest of the occlusal surface of the upper 1st and 2nd molars in modern Z. princeps from Utah; the paracone is isolated in the molars from Little Dell Dam. The fossils from Little Dell Dam are the 1st reported records of Pleistocene Zapus west of the Rocky Mountains.

Key words: Little Dell Dam, Utah, Pleistocene, Zapus, Zapodinae.

Pleistocene vertebrates from Utah are mainly from Utah. Here we describe these teeth in limited to isolated fossils from sediments asso- detail and compare them with the known den- ciated with Ice Age Lake Bonneville (Nelson tition of modern and extinct zapodine species. and Madsen 1980, 1987, Miller 1982, Jefferson In Utah there is a single Holocene locality et al. 1994, Gillette and Miller 1999). The fauna with records of Zapus; 2 teeth assigned to Z. from Little Dell Dam Locality 2 (LDD2; Salt princeps were recovered from the Stevens Lake County, UT) consists of at least 14 species Creek fauna (Smith et al. 1999). Zapus prin- of mammals (Gillette et al. 1999), a diversity ceps is the only zapodine that occurs in Utah only exceeded in the Pleistocene of Utah by today (Durrant 1952), although Z. hudsonius Crystal Ball Cave (49 spp.; Heaton 1984, 1985), does extend into eastern Colorado and Wyoming Silver Creek Junction (26 spp.; Miller 1976), (Clark and Stromberg 1987, Fitzgerald et al. Rock Springs Cave (17 spp.; Jefferson et al. 1994). A 3rd species, Z. trionatus, occurs today 1994), and Bechan Cave (17 spp.; Jefferson et in coastal regions of California, Oregon, Wash- al. 1994). These more taxonomically diverse ington, and British Columbia (Hall 1981). assemblages differ from LDD2 in being Ranch- olabrean (late Pleistocene) in age. Based on arvi- MATERIALS AND METHODS coline rodents, the LDD2 fauna was assigned to the Irvingtonian land mammal age (middle Dental terminology follows that employed Pleistocene), making it the first-known, and by Martin (1989) for advanced zapodines and oldest, taxonomically diverse Irvingtonian ver- is supplemented with additional topographic tebrate fauna known from Utah (Gillette et al. names utilized by Wang (1985) for brachyo- 1999). The less diverse Little Dell Dam Local- dont zapodine molars (Fig. 1). Neither Martin ity 1 (LDD1) fauna contains only 7 taxa and (1989) nor Wang (1985) provided names for all was also estimated to be Irvingtonian in age the reentrant folds. Those not labeled by Martin but younger than LDD2 (Gillette et al. 1999). (1989) are provided names (Fig. 1) in a manner Among the LDD2 specimens are 5 com- that is consistent with his conventions. Addi- plete teeth of the jumping mouse Zapus (Zapo- tionally, we follow the usage of entolophid as dinae, Rodentia). Additionally, LDD1 and employed by Reig (1977), which corresponds LDD2 each yielded a single, unidentified, to the hypolophid of Wang (1985) and “hypo- fragmentary zapodine tooth. These 7 specimens lophulid I” of Wood and Wilson (1936). Hypo- constitute the 1st Pleistocene records of Zapus lophulid is retained in this paper for the direct

1Department of Geological Sciences, Jackson School of Geosciences, The University of Texas at Austin, Austin, TX 78712-0254.

439 440 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 1. Occlusal patterns of Zapus molars from Little Dell Dam 2, Utah. The thickness of the black areas indicates the thickness of the enamel expressed on the occlusal surface. A, UMNH VP 12324, LM2; B, UMNH VP 12325, RM2; C, UMNH VP 12326, Lm2; D, UMNH VP 12327, Lm3. For all teeth pictured, anterior is toward the top of the page. Labial is toward the right for A and toward the left for B, C, and D. The dashed line on UMNH VP 12327 indicates the approximate outline of the posterior portion of the tooth if not for the aberrant wear. Abbreviations: ant. = anterior; post. = posterior.

connection between the hypoconid and ento- DESCRIPTIONS conid and is the equivalent of “hypolophulid II” of Wood and Wilson (1936). We treat Zapus The occlusal surface of UMNH VP 12324 as a member of Zapodinae but do not address (LM2) is elongate in outline, with wide open- the taxonomic status of the Zapodidae, namely, ings for the labial reentrant folds (Fig. 1A). whether Dipodidae exclusive of zapodines is There is a prominent anterior cingulum with a monophyletic, because the issue is not ger- small, anteriorly pointing projection on the mane to our paper. labial half. The preparacone fold gradually All fossils are curated in the University of arches posteriorly to connect with the anterior Utah Museum of Natural History (UMNH). mesostyle fold. The lunate paracone is isolated Six were recovered from the lower peat layer and concave anterolabially. The hypoloph ex- of LDD2, and a single partial molar is known hibits variable enamel thicknesses along the from LDD1. The zapodine tooth fragments from traverse from the anterior cingulum to the LDD1 (UMNH VP 12329) and LDD2 (UMNH mesoloph to the hypocone. The protocone is VP 12328) are too incomplete for confident connected to the hypoloph, closing the pre- taxonomic allocation and are not described hypocone fold and forming a large enamel here. Camera lucida drawings were made with lake with variable enamel thickness. The an Olympus S2X9 binocular microscope. Abbre- mesoloph extends out almost directly labially viations: M, upper molar; m, lower molar; L, as a thin finger from the hypoloph. The left; R, right. mesostyle projects laterally farther than the 2004] FIRST PLEISTOCENE ZAPUS FROM UTAH 441 other labial cusps. The posterior mesostyle conid fold is oriented almost directly antero- fold reaches lingually across the tooth, nearly posteriorly and separates the mesoconid from pinching out the thin band of dentine along the posterior cingulum. The entoconid turns the distal portion of the hypoloph, then turns sharply posteriorly and makes a very slight posteriorly and extends to the posterior cingu- connection with the posterior cingulum at the lum, closing the dentine connection between lingual edge of the tooth, separating the pos- the hypocone and posteroloph-metaloph junc- tentoconid fold from the elongated prehypo- ture. Lingual to the metacone and pos- conid fold. The central portion of the posterior teroloph, the postmetacone fold is expanded cingulum is greatly expanded, although it lacks anteroposteriorly. The occlusal surface has a a postcingular enamel lake. The occlusal sur- maximum length of 1.06 mm and width of 0.61 face has a maximum length of 1.10 mm and a mm. width of 0.61 mm. The RM2 (UMNH VP 12325; Fig. 1B) is The anterior edge of the Lm3 (UMNH VP less worn than the LDD2 LM2 (UMNH VP 12327) is nearly straight, with the band of den- 12324). As a result there are 4 noticeable dif- tine within the metalophid of uniform thick- ferences. The prehypocone fold has 2 inter- ness (Fig. 1D). Neither the protoconid nor the ruptions, both more medially placed than in metaconid is expanded. The metaconid and UMNH VP 12324. The paracone does not yet mesostylid connect at the lingual edge of the enclose any dentine and is not completely iso- tooth, enclosing the anterior mesolophid fold lated from the rest of the occlusal pattern. The as an enamel lake. A 2nd, smaller enamel lake posterior mesostyle fold is split into lingually occurs near the mesostylid. The triangular and posteriorly directed portions by a dentine mesostylid is greatly expanded and nearly con- connection from the mesoloph to the metaloph. tacts the small entoconid. Those 2 regions Finally, the postmetacone fold is incompletely bracket a narrow posterior mesolophid fold formed. The occlusal surface has a maximum that penetrates to the mesoconid. The poste- length of 1.06 mm and width of 0.63 mm. rior portion of the tooth exhibits aberrant wear. The LDD2 Lm2 (UMNH VP 12326; Fig. At the occlusal surface a short posterior median 1C) is slender, tapers slightly posteriorly, and fold separates the hypoconid from the entoconid. If this portion of the tooth were worn to lacks enamel lakes in the anterior and poste- the same level as the occlusal surface of the rior cingula. The anterior cingulum is asym- rest of the tooth, no fold would be present. metrical, protruding much farther anteriorly at The occlusal surface has a maximum length of the labial edge. The anteroconid is bulbous 0.49 mm and width of 0.61 mm. and lacks any hint of an anterior median fold. From the position of the metastylid, there is a COMPARISONS AND DISCUSSION posterolabially trending dentine connection to the arcuate metalophid. The metaconid lies at Characters allowing generic determination the end of a posterolingual extension of the among zapodines were discussed by Preble metalophid. At the labial edge of the anterior (1899), Krutzsch (1954), Klingener (1963), and portion of the tooth, the preprotoconid fold Martin (1989). The LDD2 specimens have a opens from a narrow slit between the antero- flat occlusal surface and well-developed conid and protoconid into a small horizontal enamel lophs. Among zapodines these general valley bounded by the metastylid lingually and features are found in the extant Zapus, Napeo- the metalophid posteriorly. The ectolophid zapus, and Eozapus, and in the extinct Plioza- shows a very faint postprotoconid fold labial pus and Javazapus. Of Pliozapus, only the from the position of the anterior mesolophid lower dentition is known. Among the 3 living fold. The mesostylid is slightly expanded com- zapodines, the occlusal pattern complexity of pared with the width of the mesolophid and the upper and lower 1st and 2nd molars of the enamel of the former contacts that of the Zapus is intermediate between the simple pat- metaconid. The result is lingual closure of the tern in Eozapus and the numerous additional anterior mesolophid fold to create an enamel enamel flexures in Napeozapus. In each of the lake. The opening of the posterior mesolophid morphological features discussed below, the fold is only constricted slightly by the expan- LDD2 teeth match the occlusal pattern seen sion of the mesostylid. The narrow prehypo- in Zapus. 442 WESTERN NORTH AMERICAN NATURALIST [Volume 64

The posterior mesostyle fold on the LDD2 men. The LDD2 m3 does not close the poste- M2 fold has a posteriorly directed portion, rior mesolophid fold into an enamel lake as in unlike Eozapus and Javazapus. The prehypo- 4 of 5 m3’s of Z. sykesae, but the opening is cone fold runs anteroposteriorly, whereas a constricted and may close with additional wear. smaller, but more equidimensional, prehypo- Each LDD2 M2 has continuous dentine from cone fold (when present) lies diagonally in the anterior cingulum to the hypocone; this is Eozapus; this fold is absent in Javazapus. 1 of 3 patterns found in Z. sykesae. Labial reentrant folds are not closed by the Zapus sandersi is known from more speci- labial cingulum as in Napeozapus; the anterior mens and more localities than any other ex- and posterior cingula are more pronounced tinct species of the genus (Martin 1989). The than in Napeozapus. The anterior mesostyle anterior cingula of Z. sandersi m2’s have enamel fold and preparacone fold do not extend lin- lakes, unlike the specimen from LDD2. The gually more than halfway across the tooth as anterior mesolophid fold is closed in the LDD2 they do in Eozapus and Javazapus. m2 but is not even constricted in Z. sandersi. The LDD2 m2 has a deep and constricted Zapus sandersi has no indication of a premeta- prehypoconid fold unlike the shallow and conid fold like that present in the LDD2 m2. broad equivalent in Eozapus. The anteroconid The m3 of Z. sandersi has as many as 5 enamel is much simpler and smaller than in Napeoza- lakes (Hibbard et al. 1978; Fig. 4f ) and lacks a pus. The LLD2 specimen lacks the cingulum posterior median fold, while the LDD2 tooth that closes all of the lingual reentrant folds in has 2 enamel lakes and a prominent posterior Napeozapus. The mesostylid is connected to median fold at the occlusal surface. the ectolophid via the mesolophid rather than Zapus burti is known only from the Borchers to the metaconid (and lacking a mesolophid) fauna in Kansas (Hibbard 1941). The Z. burti as in Javazapus. The entolophid is connected m2 does not taper posteriorly as does the to the posterior cingulum at the lingual border LDD2 tooth. The m2 anteroconid lies labial to of tooth unlike in Pliozapus, Javazapus, and the midline in LDD2 but it is centered in Z. Eozapus. The LDD2 m2 has a much larger burti. Variants in the anterior cingulum of Z. anterior cingulum than Pliozapus. burti include the morphology of the LDD2 The LDD2 m3 lacks the preprotoconid fold specimen. The postentoconid fold of Z. burti present in Napeozapus. The LDD2 tooth has 2 extends beyond the midline of the m2 but is lingual reentrant folds of different lengths, almost absent in the LDD2 tooth. Corre- instead of 3 folds of equal lengths as in Eoza- spondingly, the prehypoconid fold penetrates pus. The anterior mesolophid fold of the further lingually on the LDD2 specimen. Also, LDD2 specimen is isolated within a dentine the prehypoconid fold is very broad in Z. burti field and should perhaps be more properly but tightly constricted in LDD2. The metalo- termed an enamel lake instead of a fold (Mar- phid in the holotype of Z. burti does not reach tin 1989). the metaconid as it does in LDD2. The poste- For species-level comparisons we utilized rior mesolophid fold of the Z. burti m3 is the descriptions and published illustrations of much wider than that in LDD2 Zapus. numerous taxa (Wilson 1936, Hibbard 1941, An early species within the genus, Zapus 1951, Krutzsch 1954, Shotwell 1956, Klingener rinkeri is known only from Fox Canyon, Kansas 1963, Hibbard et al. 1978, Martin 1989, 1994). (Hibbard 1951) and was described as possess- Comparisons with many taxa are not possible ing several plesiomorphic characters for Zapus because the upper dentition for several species (Martin 1989). The m2 of Z. rinkeri does not is not known. Additionally, ontogenetic variation taper posteriorly as in the LDD2 specimen. precludes the clear determination of unique The leading edge of the m2 anterior cingulum diagnostic characters in some cases. The LDD2 is straight in Z. rinkeri but slopes posterolin- Zapus molars are unique in being significantly gually in LDD2. The posterior expansion of smaller than other described taxa. the preprotoconid fold in the LDD2 m2 is not The primary difference between the m2 present in Z. rinkeri, but differences in lingual from LDD2 and those from Zapus sykesae folds may be due solely to different ontogenetic (Martin 1989) is the absence of anterior and ages. The m3 of the holotype of Z. rinkeri is posterior enamel lakes in the former speci- too worn to make any valid comparison with 2004] FIRST PLEISTOCENE ZAPUS FROM UTAH 443 the LDD2 tooth. Klingener (1963) also referred the LDD2 specimens to Z. hudsonius and to a maxillary fragment and isolated M2 to Zapus identify the material only as Zapus. rinkeri, but, as with most fossil species of We are currently unable to use the pres- Zapus, there are no published descriptions or ence of Zapus in the Pleistocene of Utah to diagnostic characters. interpret the paleoenvironment. This is not Among the 3 extant species of Zapus, the only because we are unable to identify the fos- morphology of the m1 anteroconid can be sil material to species level, but also because used to separate Z. hudsonius from Z. princeps populations of Zapus, even of the same species, and Z. trionatus (Martin 1994). To a lesser exhibit a wide range of ecological tolerances, extent, the same feature of the m2 is also although most have distributions tied to the claimed to be useful in a similar manner availability of surface moisture (Whitaker 1972). (Klingener 1963). Klingener (1963:253) stated, The fossils from LDD2 are most significant “[In] Z. hudsonius the anteroconid in m1 and because they represent the 1st Pleistocene m2 is separated from the metalophid by the records of Zapus west of the Rocky Moun- confluent preprotoconid and premetaconid tains. folds.” This statement may be in error. None of the Z. hudsonius m2’s illustrated by Klingener ACKNOWLEDGMENTS (1963) have confluent preprotoconid and premetaconid folds, but both of the illustrated Z. We thank S. Sampson and M. Getty at the princeps and Z. trionatus show this morphology. Utah Museum of Natural History and M. Hay- Recognizing the similarities not only in den at the Utah Geological Survey for access dental and cranial characters, but also in bac- to specimens and locality information. E. Ekdale, ula and sperm morphology, Jones (1981) sub- G. Bever, T. Macrini, and C. Jass provided help- sumed Z. trionatus as a subspecies within Z. ful comments on drafts of this manuscript. We princeps. Although this taxonomic change is thank our 2 reviewers, J. Mead and S. Lucas, for not currently used in the neontological litera- their comments and suggestions, and the Utah ture, there are no published characters that Geology Foundation for financial support. permit separation of the dentition of the 2 taxa (Krutzsch 1954, Klingener 1963, Martin 1994). LITERATURE CITED Statements below that refer to the dental mor- CLARK, T.W., AND M.R. STROMBERG. 1987. Mammals in phology of Z. princeps are also applicable to Z. Wyoming. University Press of Kansas, Lawrence. trionatus. 314 pp. Based on the morphology of the M1 and DURRANT, S.J. 1952. The mammals of Utah. University of M2 paracone, Jones (1981) was generally able Kansas Publications, Museum of Natural History to distinguish Zapus princeps from Z. hudso- 6:1–549. FITZGERALD, J.P., C.A. MEANEY, AND D.M. ARMSTRONG. nius. The paracone on individuals of Z. hudso- 1994. Mammals of Colorado. Denver Museum of nius is isolated from the rest of the tooth in all Natural History and University Press of Colorado, populations except those in northern Canada Niwot. 467 pp. GILLETTE, D.D., AND W.E. MILLER. 1999. Catalogue of west of Hudson Bay. The paracone of individ- new Pleistocene mammalian sites and recovered fos- uals of Z. princeps is attached to the rest of the sils from Utah. Pages 523–530 in D.D. Gillette, edi- tooth in populations overlying and near the tor, Vertebrate paleontology in Utah. Utah Geologi- range of Z. hudsonius, but more peripheral cal Survey Miscellaneous Publication 99-1. populations of Z. princeps (Arizona, southern GILLETTE, D.D., C.J. BELL, AND M.C. HAYDEN. 1999. Pre- liminary report on the Little Dell Dam fauna, Salt New Mexico, and Pacific coast) more com- Lake County, Utah (middle Pleistocene, Irvington- monly have an isolated paracone. Jones (1981) ian land mammal age). Pages 495–500 in D.D. Gillette, reported Zapus princeps from Utah with an editor, Vertebrate paleontology in Utah. Utah Geo- attached paracone in all 325 M1’s and 329 of logical Survey Miscellaneous Publication 99-1. HALL, E.R. 1981. The mammals of North America. 2nd edi- 331 M2’s examined. Therefore the morphol- tion. John Wiley & Sons, New York. 1181 pp. ogy of the LDD2 Zapus teeth is distinct from HEATON, T.H. 1984. Preliminary report on the Quaternary the pattern currently dominant in the state. vertebrate fossils from Crystal Ball Cave, Millard Because the modern patterns of Zapus teeth County, Utah. Current Research in the Pleistocene 1:65–67. vary in areas where Z. princeps and Z. hudso- ______. 1985. Quaternary paleontology and paleoecology nius are not sympatric, we are hesitant to refer of Crystal Ball Cave, Millard County, Utah: with 444 WESTERN NORTH AMERICAN NATURALIST [Volume 64

emphasis on mammals and description of a new spe- NELSON, M.E., AND J.H. MADSEN, JR. 1980. A summary of cies of fossil skunk. Great Basin Naturalist 45:337–390. Pleistocene fossil vertebrate localities in the north- HIBBARD, C.W. 1941. The Borchers fauna, a new Pleisto- ern Bonneville Basin of Utah. Pages 97–112 in J.W. cene interglacial fauna from Meade County, Kansas. Gwynn, editor, Great Salt Lake, a scientific, histori- University of Kansas Science Bulletin 38:197–220. cal and economic overview. Utah Geological and ______. 1951. A new jumping mouse from the Upper Plio- Mineral Survey, Utah Department of Natural Re- cene of Kansas. Journal of Mammalogy 32:351–352. sources Bulletin 116. HIBBARD, C.W., R.J. ZAKRZEWSKI, R.E. ESHELMAN, G. ______. 1987. A review of Lake Bonneville shoreline fau- EDMUND, C.D. GRIGGS, AND C. GRIGGS. 1978. Mam- nas (late Pleistocene) of northern Utah. Pages 318– mals from the Kanopolis local fauna, Pleistocene 333 in R.S. Kopp and R.E. Cohenour, editors, Ceno- (Yarmouth) of Ellsworth County, Kansas. Contribu- zoic geology of western Utah—sites for precious tions from the Museum of Paleontology, University metal and hydrocarbon accumulation. Utah Geologi- of Michigan 25:11–44. cal Association Publication 16. JEFFERSON, G.T., W.E. MILLER, M.E. NELSON, AND J.H. PREBLE, E.A. 1899. Revision of the jumping mice of the MADSEN, JR. 1994. Catalogue of the late Quaternary genus Zapus. North American Fauna 15:1–42. vertebrates from Utah. Natural History Museum of REIG, O.A. 1977. A proposed unified nomenclature for the Los Angeles County, Technical Reports 9:1–34. enamelled components of the molar teeth of the JONES, G.S. 1981. The systematics and biology of the genus Cricetidae (Rodentia). Journal of Zoology, London Zapus (Mammalia, Rodentia, Zapodidae) (Canada, 181:227–241. United States). Doctoral dissertation, Indiana State SHOTWELL, J.A. 1956. Hemphillian mammalian assemblage University, Terre Haute. from northwestern Oregon. Bulletin of the Geologi- KLINGENER, D. 1963. Dental evolution of Zapus. Journal cal Society of America 67:717–738. of Mammalogy 44:248–260. SMITH, K.S., R.L. CIFELLI, AND N.J. CZAPLEWSKI. 1999. KRUTZSCH, P. 1954. North American jumping mice (genus An early Holocene, high-altitude vertebrate faunule Zapus). University of Kansas Publications, Museum from central Utah. Pages 537–543 in D.D. Gillette, of Natural History 7:349–472. editor, Vertebrate paleontology in Utah. Utah Geo- MARTIN, R.A. 1989. Early Pleistocene zapodid rodents from logical Survey Miscellaneous Publication 99-1. the Java local fauna of north-central South Dakota. WANG, B.Y. 1985. Zapodidae (Rodentia, Mammalia) from Journal of Vertebrate Paleontology 9:101–109. the Lower Oligocene of Qujing, Yunnan, China. ______. 1994. A preliminary review of dental evolution Mainzer Geowissenschaftliche Mitteilungen 14: and paleogeography in the zapodid rodents, with 345–367. emphasis on Pliocene and Pleistocene taxa. Pages 99– WHITAKER, J.O., JR. 1972. Zapus hudsonius. Mammalian 113 in Y. Tomida, L. Chaunkuei, and T. Setoguchi, Species 11:1–7. editors, Rodent and lagomorph families of Asian ori- WILSON, R.W. 1936. A Pliocene rodent fauna from Smiths gins and diversification. National Science Museum Valley, Nevada. Carnegie Institution of Washington Monographs 8, Tokyo. Publication 473:15–34. MILLER, W.E. 1976. Late Pleistocene vertebrates of the WOOD, A.E., AND R.W. WILSON. 1936. A suggested nomen- Silver Creek local fauna from north central Utah. clature for the cusps of the cheekteeth of rodents. Great Basin Naturalist 36:387–424. Journal of Paleontology 10:388–391. ______. 1982. Pleistocene vertebrates from the deposits of Lake Bonneville, Utah. National Geographic Research Received 8 July 2003 Reports 14:473–478. Accepted 5 December 2003 Western North American Naturalist 64(4), © 2004, pp. 445–453

A RELOCATION TECHNIQUE FOR BLACK-TAILED PRAIRIE DOGS

Kelly A. Roe1 and Christopher M. Roe1

ABSTRACT.—Relocations of black-tailed prairie dogs have occurred both to save individual prairie dogs from urban development and to reestablish populations that have been extirpated. Unfortunately, however, many past relocation efforts rarely exceeded 40% retention. Many factors have contributed to very low retention rates in past relocation efforts including lack of (1) suitable habitat, (2) proper artificial burrow systems, (3) aboveground acclimation cages or pens, and (4) skilled people conducting the relocations. In an attempt to increase prairie dog relocation success, we developed techniques that are easy to implement, promote high retention, and effectively conserve labor, financial resources, and prairie dog populations. We conducted 3 relocations along the Front Range of Colorado in 2001 and 2002. Relocation techniques we developed resulted in at least 46%–92% retention. Our results suggest that a large percentage of prairie dogs can be retained by (1) ensuring that habitat is suitable, (2) using underground nest chambers modeled after natural nest chambers, (3) acclimating prairie dogs to the release site in large retention pens rather than in retention caps or other small acclimation cages (i.e., rabbit hutch), and (4) providing supplemental feed and water ad libitum to the prairie dogs.

Key words: prairie dog, relocation, translocation, black-tailed prairie dog, Cynomys ludovicianus, artificial release.

The black-tailed prairie dog (Cynomys in the face of development or agricultural ludovicianus) is 1 of 5 species of prairie dogs activities, or to reestablish prairie dogs in areas in the United States. Historically, the species where they were extirpated. Unfortunately, sur- existed from Canada to northern Mexico, from vival and retention of relocated prairie dogs the foothills of the Rockies east to the mid- during many past relocation efforts have rarely grass prairie (Fitzgerald et al. 1994). Black- exceeded 40% (Jacquart et al. 1986, McDon- tailed prairie dogs were the most numerous ald 1993, Robinette et al. 1995, Truett and and widespread herbivore in the Great Plains Savage 1998, Truett et al. 2001, Meaney et al. (Barko 1997) and perhaps even North America 2002). (Wuerthner 1997). Various sources report that Many factors have contributed to very low the prairie dog has declined 98% throughout retention rates from past relocation efforts. its range during the past 100 years because of These include the lack of (1) suitable habitat, poisoning campaigns, land use conversion, (2) proper artificial burrow systems, (3) above- plague, and other causes (Whicker and Det- ground acclimation cages or pens, and (4) ling 1988, Miller et al. 1994). skilled people conducting the relocations. One The U.S. Fish and Wildlife Service (USFWS), technique that resulted in as little as 0% reten- after receiving a petition from the National tion involved simply releasing prairie dogs into Wildlife Federation, determined in their 12- augered holes rather than proper artificial bur- month finding that the prairie dog was “war- row systems. Prairie dogs released into augered ranted but precluded” from listing as a threat- holes (with or without acclimation structures) ened species (USFWS 2000). In addition, to seldom remain where they are released (Turner meet minimum conservation standards set forth 1979, Jacquart et al. 1986, Truett and Savage in the Conservation Assessment and Strategy 1998, Truett et al. 2001). Other techniques uti- (Van Pelt 1999) and the Multi-State Conserva- lizing underground artificial nest chambers tion Plan (Luce 2003), some of the 11 states coupled with aboveground acclimation cages within the range of the black-tailed prairie dog still, however, generally result in <50% reten- may need to conduct live relocations. Reloca- tion (Truett et al. 2001, Meaney et al. 2002). tions may be utilized as a management techni- Therefore, it is vitally important that relocation que to ensure no net loss of prairie dog acreage techniques be developed and utilized which

1Roe Ecological Services, Box 438, Green Mountain Falls, CO 80819.

445 446 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Habitat attributes including soil type, vegetation composition, slope, and size of the relocation area of DMOS relocation sites in 2001 and GWROS sites in 2002. Site/Soil type Vegetation Slope Relocation area DMOS Clay-loam1 bluegrass (Poa spp.) 0–5% Phase 1: 8 ha smooth brome (Bromus inermis) adjacent Phase 2: 4 ha adjacent yellow sweet clover (Melilotus officinalis) to phase 1 site

GWROS Cobbly clay loam2 blue grama (Bouteloua gracilis) 0–5% 8 ha Indian rice grass (Achnatherum hymenoides) red three awn (Aristida purpurea) prickly pear (Opuntia polyacantha) common mullein (Verbascum thapsus) field bindweed (Convolvulus arvensis) ragweed (Ambrosia spp.)

1Moreland and Moreland 1975 2Price and Amen 1980 are relatively easy to implement, promote high cm 3 weeks before the phase 1 relocation and retention and survival, and effectively conserve 1 week before the phase 2 relocation. The labor, financial resources, and prairie dog pop- GWROS release site was not mowed because ulations. the vegetation was generally shorter than 20 cm. We describe 3 relocations we conducted on We buried artificial burrow systems at all 3 the Front Range of Colorado in 2001 and 2002. sites. Underground nest chambers were con- These relocations focused on successfully structed from 0.95-cm-thick plywood, and high- establishing prairie dogs rather than compara- density cardboard tubes served as tunnels tively testing techniques. Relocation techniques leading from the nest chambers to the soil sur- described in this paper resulted in 46%–92% face (Fig. 1). We used biodegradable and mod- retention. ifiable tubes to ensure the relocation site remained free of synthetic materials and quickly METHODS resembled a natural prairie dog colony. Using a mini-excavator to dig holes for the Relocation Sites artificial burrow systems, we buried most We conducted 3 prairie dog relocations be- chambers ≥ 1 m deep. Although most natural tween 2001 and 2002. Two relocations were prairie dog nest chambers are 2–3 m deep conducted in June–August 2001 to Davidson (Hoogland 1995), burying artificial nest cham- Mesa Open Space (DMOS) in Louisville, Col- bers at this depth could result in an extreme orado. A 3rd relocation was conducted in June amount of soil and vegetation disturbance. 2002 to Great Western Reservoir Open Space Because prairie dog “listening” chambers are (GWROS) in Broomfield, Colorado. Table 1 approximately 1 m deep (Hoogland 1995) and describes the soil type, vegetation, slope, and provide prairie dogs with adequate protection size of the relocation area for each site. Preda- from predators, inclement weather, and sum- tors such as Red-tailed Hawks (Buteo jama- mer temperature extremes, we determined icensis) and coyotes (Canis latrans) were ob- that 1 m was appropriate to balance resource served on each site before, during, and after disturbance with prairie dog survival. relocations. Aboveground retention caps and pens covered the entrances of artificial burrows during Site Preparation relocations. Retention caps and pens temporar- Black-tailed prairie dogs prefer a visually ily contained the released prairie dogs so they unobstructed habitat, generally shorter than could acclimate to their new surroundings and 20 cm (Clippinger 1989, McDonald 1993, Fitz- would not immediately disperse. We used reten- gerald et al. 1994, Hoogland 1995). Because tion caps, which were enclosed cylinders 1 m2 vegetation on the DMOS site was >20 cm tall, made from 0.635-cm hardware cloth (Fig. 2), we mowed it to a height of approximately 10 during all 3 relocations. Retention pens, made 2004] PRAIRIE DOG RELOCATION TECHNIQUE 447

Fig. 1. Diagram of the artificial burrow system for prairie dogs designed by Roe Ecological Services and installed on DMOS in 2001 and GWROS in 2002. from 2.5-cm chicken wire (Fig. 3), were much determined to be of the same coterie. We larger than retention caps, being 14.5 m2 (3.81 made a map of the trap site and corresponding m long per side). The chicken wire was pur- coterie associations, labeling each trap with a chased in 1.8 × 45.72-m rolls, which were then piece of electrical tape showing the unique cut into three 1.8 m × 15.24-m lengths. To allow coterie identifier of the area within which the each length to be folded into a square pen with trap was placed. a bottom, side, and top, we made two 60-cm Relocation Process slits (cutting from the outer edge inward) at 3.81-m intervals. The bottom was staked to During all relocation efforts we used only the ground with lawn staples, the top was held cage-type live-traps for prairie dog capture. together with plastic zip ties, and the sides Trapped prairie dogs were collected at approx- were kept vertical with metal posts zip-tied to imately 2-hour intervals throughout the day the chicken wire. and moved to a shaded area for processing to reduce likelihood of heat stress. Processing Trap Site Coterie included dusting with insecticide to remove Identification fleas, recording the sex, age class (adult, year- Based on the assumption that prairie dogs’ ling, or juvenile), and coterie identifier, and natural social groupings are important and assigning each individual an artificial burrow should be maintained, we relocated prairie at the release site. Prairie dogs captured dur- dogs in coteries (family groups) as much as ing the GWROS relocation were marked with possible. We determined coterie association Nyanzol-D hair dye with individual patterns. by observing social interactions and daily After processing them, we immediately trans- movement patterns of individual prairie dogs ported all individuals to the relocation site in a on the trap site. Prairie dogs that interacted pickup truck equipped with a ventilated bed amicably and without obvious territorial dis- topper in the live-traps within which they were play and remained within a similar area were trapped to minimize stress and to preserve 448 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 2. Prairie dog retention cap designed by Roe Ecological Services and installed on DMOS in 2001 and GWROS in 2002: a, diagram; b, photo. each individual’s sex classification, coterie iden- After we removed a cap or pen, we attached tifier, and artificial burrow assignment. the water bottle to a post and placed grain and When possible, we released prairie dogs mineral mix at least every 3rd day on top of with their original coterie members into the the mound or soil adjacent to the burrow en- same or adjacent artificial burrows at the trance. This provided an additional food source, release site. We released only 1 adult male or offered an incentive to remain at the original 2 yearling males with females and juveniles release burrow, and facilitated monitoring of from the same coterie. Lone males or addi- relocated prairie dogs. Newly constructed nat- tional males of a coterie were released with ural burrows that showed prairie dog activity males of the same coterie or adjacent coteries also received deposits of grain and mineral as often as possible. We released an average of mix. We continued providing food and water 7.8 prairie dogs per artificial burrow system, for approximately 3 weeks post-release for all with a maximum of 11. relocation efforts. Supplemental Feeding Monitoring and Watering Three methods commonly are used to esti- We attached a water bottle to the side of the mate prairie dog population size and density. retention cap or pen and placed a high-quality Mark-recapture is a reliable but labor-inten- grain and mineral mix inside. This provided sive method (Otis et al. 1978, Fagerstone and access to food and water ad libitum while the Biggins 1986, Menkens et al. 1990). Counts of prairie dogs became acclimated to the new site. plugged and reopened burrows provide an We checked levels of food and water and cap index of prairie dog activity (Knowles 1982, or pen integrity daily. Tietjen and Matschke 1982). Visual counts can 2004] PRAIRIE DOG RELOCATION TECHNIQUE 449

Fig. 3. Prairie dog retention pen designed by Roe Ecological Services and installed on GWROS in June 2002; a, diagram; b, photo. provide a quick estimate of prairie dog popu- with prairie dog density. Therefore, we con- lation density (Fagerstone and Biggins 1986, ducted visual counts for estimation of post- Knowles 1986). We did not have the financial release retention. means or the desire to disturb the newly relo- We observed prairie dogs on the DMOS cated populations to conduct a mark-recapture site from the cab or top of a pick-up truck with study. Population estimation by counts of 10 × 50 binoculars because the site had rela- plugged and reopened burrows was not prac- tively flat terrain and was recently mowed. tical because the number of prairie dogs per Prairie dogs from the GWROS site also were burrow was initially unnaturally high. Further- observed and identified with the use of 10 × more, Lewis et al. (1979) confirmed that the 50 binoculars from a hill approximately 30 m number of burrows is not necessarily correlated above and 100 m away from the release site. 450 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 2. Relocation methods, number of prairie dogs relocated, and percent retention for DMOS phase 1 and phase 2 relocations conducted in 2001 and GWROS relocation conducted in 2002. Number Percent Site Methods used relocated retention DMOS phase 1 50 artificial burrow systems 399 45% 44 nest chambers were 5-sided without bottoms (182 retained) 6 nest chambers were 6-sided with plywood bottoms Retention caps without bottoms 2- to 3-day acclimation period

DMOS phase 2 24 artificial burrow systems 196 62% All 24 nest chambers were 6-sided with plywood (121 retained) bottoms Retention caps with bottoms 5-day acclimation period

GWROS 16 artificial burrow systems 104 89% 3 nest chambers were 5-sided with no bottoms (96 retained) 13 nest chambers were 6-sided with cardboard bottoms 12 retention caps with bottoms 6 retention pens 8-day acclimation period

We mapped the location of individual prairie proximity to the relocation site). Coyote pre- dogs on the GWROS site during observations dation and dispersal rates for all relocations conducted from dawn until dusk for 2 consec- were highest for the first 3 or 4 days post- utive days until we could not identify any release and decreased thereafter. additional unique prairie dogs. Because prairie Retention caps and pens kept most prairie dogs were not marked for the phase 1 and 2 dogs from immediately dispersing. We left re- DMOS relocations, we mapped the location of tention caps and pens in place for 2–3 days every burrow and identified the number and post-release for phase 1 at DMOS and for 5–8 approximate age class of prairie dogs using days for phase 2 on the DMOS and GWROS each burrow. We also conducted these obser- releases. We originally planned to acclimate vations from dawn until dusk for 2 consecutive phase 1 prairie dogs for 5 days. However, we days. For all 3 sites we conducted monitoring did not put bottoms on the retention caps, and activities approximately 2 weeks post-release. within 2–3 days the prairie dogs had dug out. After digging out, most prairie dogs were ob- RESULTS served returning to the shelter of the cap when seeking safety and at night. Some prairie dogs, Table 2 describes methods used, number of however, dispersed and did not return to the prairie dogs relocated, and percent retention artificial burrow. Therefore, we put bottoms for all 3 relocations. It is possible that more on the retention caps made of the same hard- prairie dogs survived from the DMOS phase 2 ware cloth with a 10-cm cutout for the tube, release and dispersed into the adjacent phase and we extended the acclimation period to 5 1 area. However, because individuals were not days for the phase 2 effort. marked, we can base our conclusions only on We acclimated prairie dogs in the caps and the number of prairie dogs observed in the pens for 8 days for the GWROS relocation. By area where they were released. Prairie dogs the end of the 7th day, the prairie dogs had that no longer remained on the relocation sites caused damage to the vegetation and root may have (1) been killed by coyotes or raptors, layer through clipping and digging even (2) died of stress or other unknown factors, or though they had sufficient supplemental food. (3) dispersed away from the designated release Therefore, we removed the caps and pens on area (i.e., DMOS prairie dogs may have moved the 8th day post-release to minimize resource to 1 of 3 existing prairie dog towns in close damage. 2004] PRAIRIE DOG RELOCATION TECHNIQUE 451

DISCUSSION remain and use the chambers once we removed retention caps or pens. One limitation of DMOS phase 1 and 2 Many other relocation efforts past and pres- relocations was that release sites were in the ent use smaller structures not modeled after middle of a recreational open space. Two trails natural nest chambers, which may have lead heavily used by hikers, joggers, dog walkers, and to lower retention levels. For example, efforts cyclists bisected the relocation sites. Farrar et conducted along the Front Range of Colorado al. (1998) reported that relocated prairie dogs commonly used 2 different styles of under- are nearly twice as sensitive to human activity ground nest chambers. One style is a 6-sided as native prairie dogs. The activity on the trails plywood box approximately 23.5 cm high, 18.4 may have affected release site fidelity and sub- cm wide, and 61 cm long (26,376.4 cm3). The sequent survival through frequent interruptions 2nd style is a 20.32-cm-diameter and 122-cm- in foraging and social interactions (Farrar et al. long (39,518 cm3) pressed cardboard caisson 1998). Thus, many prairie dogs may have left tube wired to plywood end caps. The use of the relocation area before they had enough larger, more natural-sized nest chambers may time to become established. reduce crowding and ensure that ample oxy- A 2nd potential limitation was that the vege- gen is available within these chambers to tation composition of DMOS was not suitable allow multiple individuals to spend the night for prairie dog survival and persistence because in a single box without asphyxiating. the primary species were not grazing tolerant The combined cost of construction and in- and grew 30 cm in height (Roe and Roe 2003). stallation of the larger plywood boxes is gener- Similarly, most of GWROS had a large propor- ally $10 more per box than the small plywood tion of annual and noxious weeds, which are boxes or small caisson tubes (Table 3). Typi- not generally considered suitable for prairie cally, however, twice as many prairie dogs can dog habitation (Roe and Roe 2003). A small be relocated and retained in larger plywood portion of GWROS, upon which the relocation boxes. This translates into fewer boxes, less was conducted, supported more species that labor and equipment rental fees per box, and were considered grazing tolerate and maintain a shorter height. This vegetation composition less soil and surface disturbance. Therefore, also may have contributed to the higher level the total cost of artificial burrow systems using of prairie dog retention at GWROS. small plywood or small caisson tubes may be High levels of coyote activity also were up to twice as much as large plywood boxes. considered a limitation for all 3 relocations. Another improvement on the GWROS pro- On several occasions we observed coyotes ject was the creation and use of retention pens hunting at the release sites. Observations of instead of traditional retention caps on 6 of the the release sites 1 and 2 weeks post-release 16 artificial burrows. Retention caps, because showed evidence of coyote use, including scat of their small size, appear to restrict prairie and digging on and around artificial burrows. dog behavior and interaction with their nat- This high level of predator activity also could ural environment. In contrast, retention pens have caused prairie dogs to disperse away allow prairie dogs to interact in a less crowded from the release site. containment area and to investigate a greater Alternatively, the strengths as expressed by portion of their natural environment while the increasing retention of our relocations, being adequately retained. Within the reten- especially the relocation on GWROS, included tion pens prairie dogs are able to consume (1) size of our artificial nest chambers, (2) de- natural vegetation, develop dusting sites, initi- velopment of new techniques including the ate new burrow systems, and interact socially use of retention pens, and (3) supplemental with their coterie and pen mates. The reten- feeding and watering during the relocation tion pens are generally over 4 times as expen- and post-release. Natural prairie dog nest cham- sive as retention caps (Table 3); however, bers are roughly 30 cm high and 50 cm in materials used to construct the pens are re- diameter (Hoogland 1995), or 58,904 cm3 in usable and their use may have resulted in volume. Our artificial underground nest cham- increased retention of relocated prairie dogs. bers are 108,000 cm3. Consequently, we ob- Supplemental feeding and watering was served 5 to 15 prairie dogs freely choosing to also important to prairie dog health and site 452 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 3. Estimated cost comparison of equipment and materials needed for 10 augered holes, 10 artificial burrow systems, and 10 acclimation cages/pens. Costs are based on a unit of 10 because equipment is often rented on a daily basis and typically at least 10 holes can be augered or 10 artificial burrow systems installed in a day. Method Equipment and materials used Estimated cost (per 10)

AUGERED HOLES • 1-day rental of a 2-man, hand-held auger1 $58 OR • 1-day rental of a bobcat with an auger bit (including delivery)1 $320 TOTAL $58–320

ARTIFICIAL BURROW SYSTEMS2 • 1-day rental of a trencher walk-behind unit Small plywood (including delivery)2 $250 (26,376 cm3)• 6-sided plywood nest chamber (including plywood, corner blocks, screws, and glue)3 $66 • Black Corex plastic tubing (3.1 m length)3 $40 TOTAL $356

Small caisson tube • 1-day rental of a trencher walk-behind unit (39,518 cm3) (including delivery)1 $250 OR • 1-day rental of a full-sized backhoe (including delivery)1 $360 • Caisson nest chamber (including caisson tube, wire, and plywood end caps)3 $50 • Black Corex plastic tubing (3.1 m length)3 $40 TOTAL $340–450

Large plywood • 1-day rental of a mini-excavator (including (108,000 cm3) delivery)1 $310 • 5-sided plywood nest chamber (including plywood, corner blocks, 3.5 × 32-mm screws, glue, and cardboard bottom)3 $70 • Cardboard tube (1.8 m length)4 $50 TOTAL $430

ACCLIMATION CAGES/PENS5 • Includes 0.635-cm hardware cloth Retention cap (circumference, top, and bottom), zip ties or wire to hold the cap together, and lawn staples to hold the cap securely to the ground3 TOTAL $175

Retention pen • Includes 2.5-cm chicken wire, U-posts, zip ties to hold the sides together and to the U-posts, and lawn staples to hold the pen securely to the ground3 TOTAL $740

1Based on estimates obtained in December 2003 from NationsRent in Boulder, CO. These estimates do not include fuel, tax, or damage waiver. 2None are reusable and are a recurring cost for each relocation. 3Based on estimates obtained in December 2003 at Home Depot in Louisville, CO. 4Contact Roe Ecological Services to obtain. 5All are reusable and with care should be a 1-time cost.

retention. For all 3 projects, extremely hot and during post-release may help to ensure prairie dry conditions at the trap and release sites dog health. Ample, easily obtained food may made the prairie dogs highly susceptible to allow prairie dogs to spend more time digging malnutrition and dehydration. Even though new burrows, or modifying artificial burrows, prairie dogs naturally acquire most of their and may improve overwinter survival. These water from forage, we frequently observed relocation techniques demonstrate increased prairie dogs drinking from water bottles. Pro- post-release retention and will likely contrib- viding relocated prairie dogs with a high level ute to long-term success of future prairie dog of nutritional supplementation and free water relocations. 2004] PRAIRIE DOG RELOCATION TECHNIQUE 453

ACKNOWLEDGMENTS MEANEY, C., A. RUGGLES, L. WHITTEMORE, N. CLIPPINGER, C.AHRENS, M. REED-ECKERT, D. FERNANDEZ, ET AL. We thank the editors and reviewers for 2002. Monitoring of prairie dogs on Boulder County Open Space. Unpublished report to the Boulder their comments during the review process for County Parks and Open Space Department, Boul- this manuscript, which helped us ensure its der, CO. clarity and quality. We also thank the City of MENKENS, G.E., D.E. BIGGINS, AND S.H. ANDERSON. 1990. Lousiville, Colorado, and the City and County Visual counts as an index of white-tailed prairie dog density. Wildlife Society Bulletin 18:290–296. of Broomfield, Colorado, for funding relocation MILLER, B., G. CEBALLOS, AND R. READING. 1994. The efforts. Finally, we thank Ray Sterner, Kathleen prairie dog and biotic diversity. Conservation Biol- Fagerstone, Gary Witmer, Mark Brennan, and ogy 8:677–681. Dave Seery for providing equipment, informa- MORELAND, D.C., AND R.E. MORELAND. 1975. Soil survey tion, and advice during our early years. of Boulder County area, Colorado. United States Department of Agriculture, Soil Conservation Ser- vice, Washington, DC. LITERATURE CITED OTIS, D.L., K.P. BURNHAM, G.C. WHITE, AND D.R. ANDER- SON. 1978. Statistical inference from capture data on BARKO, V.A. 1997. History of policies concerning the black- closed animal populations. Wildlife Monographs 62: tailed prairie dog: a review. Proceedings of the Okla- 1–135. homa Academy of Science 77:27–33. PRICE, A.B., AND A.E. AMEN. 1980. Soil survey of Golden CLIPPINGER, N.W. 1989. Habitat suitability index models: area, Colorado. United States Department of Agri- black-tailed prairie dog. United States Fish and Wild- culture, Soil Conservation Service, Washington, DC. life Service, Biological Report 82(10.156). Washing- ROBINETTE, K.W., W.F. ANDELT, AND K.P. BURNHAM.1995. ton, DC. Effect of group size on survival of relocated prairie FAGERSTONE, K.A., AND D.E. BIGGINS. 1986. Comparison dogs. Journal of Wildlife Management 59:867–874. of capture-recapture and visual count indices of ROE, K.A., AND C.M. ROE. 2003. Habitat suitability guide- prairie dog densities in black-footed ferret habitat. lines for black-tailed prairie dog relocations. Wildlife Great Basin Naturalist Memoirs 8:94–98. Society Bulletin 31:1246–1253. FARRAR, J.P., K.L. COLEMAN, M. BEKOFF, AND E. STONE. TIETJEN, H.P., AND G.H. MATSCHKE. 1982. Aerial prebaiting 1998. Translocation effects on the behavior of black- for management of prairie dogs with zinc phosphide. tailed prairie dogs (Cynomys ludovicianus). Anthro- Journal of Wildlife Management 46:1108–1112. zoos 11:164–167. TRUETT, J.C., AND T. S AVAGE. 1998. Reintroducing prairie FITZGERALD, J.P., C.A. MEANEY, AND D.M. ARMSTRONG. dogs into desert grasslands. Restoration and Man- 1994. Mammals of Colorado. University Press of Col- agement Notes 16:189–195. orado, Niwot. TRUETT, J.C., J.L.D. DULLUM, M.R. MATCHETT, E. OWENS, HOOGLAND, J.L. 1995. The black-tailed prairie dog: social AND D. SEERY. 2001. Translocating prairie dogs: a life of a burrowing animal. University of Chicago review. Wildlife Society Bulletin 29:863–872. Press, Illinois. TURNER, B. 1979. An evaluation of Utah prairie dog (Cyno- JACQUART, H.C., J.T. FLINDERS, M.P. COFFEEN, AND R.N. mys parvidens) transplant success. Utah Division of HASENYAGER. 1986. Prescriptive transplanting and Wildlife Resources, Salt Lake City. monitoring of Utah prairie dog (Cynomys parvidens) UNITED STATES FISH AND WILDLIFE SERVICE (USFWS). populations. Utah Division of Wildlife Resources, 2000. Twelve-month administrative finding for a Salt Lake City. petition to list the black-tailed prairie dog as threat- KNOWLES, C.J. 1982. Habitat affinity, populations and con- ened. Federal Register, 04 February 2000, 65(24): trol of black-tailed prairie dogs on the Charles M. 5476–5488. Russell National Wildlife Refuge. Doctoral disserta- VAN PELT, W.E. 1999. The black-tailed prairie dog conser- tion, University of Montana, Missoula. vation assessment and strategy—fifth draft. Nongame ______. 1986. Some relationships of black-tailed prairie and Endangered Wildlife Program, Arizona Game dogs to livestock grazing. Great Basin Naturalist 46: and Fish Department, Phoenix. 198–203. WHICKER, A.D., AND J.K. DETLING. 1988. Ecological con- LEWIS, J.C., E.H. MCILVAIN, R. MCVICKERS, AND B. PETER- sequences of prairie dog disturbances. Bioscience SON. 1979. Techniques used to establish and limit 38:778–785. prairie dog towns. Proceedings of the Oklahoma WUERTHNER, G. 1997. Viewpoint: the black-tailed prairie Academy of Science 59:27–30. dog—headed for extinction? Journal of Range Man- LUCE, R.J. 2003. A multi-state conservation plan for the agement 50:459–466. black-tailed prairie dog, Cynomys ludovicianus, in the United States—an addendum to the Black-tailed Received 19 May 2003 Prairie Dog Conservation Assessment and Strategy, Accepted 20 April 2004 November 3, 1999. Prairie Dog Conservation Team, Sierra Vista, AZ. MCDONALD, K.P. 1993. Analysis of the Utah prairie dog recovery program, 1972–1992. Utah Division of Wildlife Resources Publication 93-16, Cedar City. Western North American Naturalist 64(4), © 2004, pp. 454–464

A REVIEW OF THE ADULT ANAGAPETUS (TRICHOPTERA: GLOSSOSOMATIDAE)

Dave Ruiter1

ABSTRACT.—Seven Anagapetus species [aisha Denning, bernea Ross, chandleri Ross, debilis (Ross), hoodi Ross, schmidi (Levanidova), thirza Denning] have been described from eastern Russia and western North America. Anagape- tus thirza Denning is proposed as a synonym of A. chandleri Ross. Anagapetus schmidi is returned to Glossosoma. Descriptions, distributions, and figures are provided for both males and females of the remaining 5 Anagapetus species.

Key words: Trichoptera, Glossosomatidae, Anagapetus, Glossosoma schmidi.

Anagapetus (Ross 1938) was created as a gapetus or Glossosoma based on the characters subgenus of Agapetus Curtis for the male of Morse and Yang (1993) used to separate the 2 Anagapetus debilis (Ross 1938), primarily be- genera. Morse and Yang (1993) suggested the cause of the lack of specialized male sec- larval forelegs originating near the middle of ondary sexual characters typical of Agapetus. the pronotum (see Wiggins 1996) were auta- Without additional rationale, Ross (1947) treated pomorphic for the Anagapetini; the G. schmidi Anagapetus as a distinct genus. Ross (1951) larval forelegs originate near the apex of the provided rationale to support Anagapetus as pronotum. Morse and Yang (1993) listed 3 the most primitive glossosomatid and listed additional adult homologues to support tribal several characters to support this conclusion. separation: the alignment of forewing veins Ross (1956) elevated Anagapetus to tribal sta- SR, R4+5, R5 (straight in Glossosomatinae; see tus (Anagapetini vs. Agapetini and Glossoso- Ross 1956, Schmid 1980, 1998); male forewing matini within the Glossosomatinae) based on callosity (present in Glossosomatinae; see Ross the presence of the short transverse suture that 1956, Schmid 1980, 1998); and male inferior divides the mesepisternum (Fig. 1). Schmid appendage angled downward (present in Anaga- (1962) suggested that Anagapetus was, in fact, petini; see Ross 1956). Glossosoma schmidi has a highly specialized group of species and that the linear vein alignment but lacks any modifi- Ross’s tribal elevation was unsupportable. cation of the male forewing anal area typical Schmid (1980, 1998) reduced the 10 subgen- of Glossosoma. The G. schmidi male inferior era (Ross 1956) of Glossosoma to 6 and moved appendage (Levanidova 1979) is angled down- Anagapetus to a 7th subgenus of Glossosoma ward (a suggested autapomorphy for Anagape- based on the opinion that the suture separat- tini); however, the structure is shaped quite ing the upper and lower parts of the epister- differently (see Ross 1956, Levanidova 1979). num “is not constant and is insignificant.” Morse Morse and Yang (2004) provide additional and Yang (1993) provided additional summary evaluation of the homologies to support the of the confusion and suggested 4 characters Glossosoma subgenera, including an additional (see discussion below) that support separation character to support separation of Glossoso- of Anagapetini from Glossosomatini. matini from Anagapetini: the unguitractor plate Glossosoma (Anagapetus) schmidi Levanidova of Martynov (1927). Glossosoma schmidi pos- (1979), an eastern Russian species described sesses this structure (Levanidova 1979), but it within the subgenus Anagapetus, presents prob- is also present on Anagapetus. The unguitractor lems for Morse and Yang’s (1993, 2004) Glos- plate size in Anagapetus seems to vary, with sosomatinae phylogeny. Glossosoma schmidi that of A. chandleri Ross being very small, or cannot be readily separated from either Ana- absent, while that of A. hoodi Ross is as large

16260 South Grant Street, Centennial, CO 80121.

454 2004] REVIEW OF ADULT ANAGAPETUS 455

Figs. 1–4. Comparison of Anagapetus hoodi and Glossosoma schmidi: (1) Anagapetus hoodi mesepisternum; (2) Glos- sosoma schmidi mesepisternum; (3) Anagapetus hoodi head, pronotum, mesonotum; (4) Glossosoma schmidi head, pronotum, mesonotum. as those of Glossosoma. Based on this, it is not that Schmid’s (1980, 1998) removal of tribal clear that any characters support monophyly status for Anagapetus is appropriate. Details of for Anagapetus, other than the downward- the Glossosoma females need to be evaluated turned male inferior appendages (which Morse in detail in hopes of clarifying these issues. and Yang [2004] conclude are actually exten- Anagapetus adults can be separated from sions of the 9th sternum), and the male apico- Glossosoma by the downward-turned inferior lateral club hairs on the 9th segment. In either case, G. schmidi does not appear to belong appendages and lateral tuft of club hairs on within Anagapetus, and it is not clear to which the male 9th segment. Anagapetus larvae can Glossosoma subgenus (Morse and Yang 1993, be distinguished by the foreleg originating 2004) G. schmidi belongs. It would appear near midlength of pronotum (Wiggins 1996). 456 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 5. Anagapetus aisha: (a) male genitalia, lateral view; (b) male genitalia, dorsal view; (c) female genitalia, lateral view; (d) female segment 8, dorsal view; (e) vaginal apparatus, ventral view.

Ross (1956) created 3 groups within Ana- be readily separated by the characters discussed, gapetus (chandleri, debilis Ross, and bernea- and figured, below. hoodi), primarily based on a conclusion that A. Segment 8 of Anagapetus females is a tubular debilis lacked the peculiar, pale club hairs on structure composed of a distinctly sclerotized the apico-lateral margin of segment 9. These distal portion while the cephalic sclerotized club hairs are present on all the Anagapetus portion merges imperceptibly into a membra- species, including A. debilis. They are best nous area attached to segment 7. The relative seen from the dorsal or ventral view, as they shape of the distal portion is useful for species are located on a distinct lateral hump. (This determinations. The vaginal apparatus is very hump with club hairs is apparently lacking in similar in all species, except in A. hoodi. How- Glossosoma, including G. schmidi.) The club ever, minor differences in the apex of the vagi- hairs, while quite stout, can be rubbed off. nal apparatus of the species are consistent with- And the hump can be small, which may have in the material examined. led to Ross’s conclusions. No clear characters Distribution of the material examined during were discovered that would separate the species this study is presented in Figures 10 and 11 into distinct groups, although the species can and can be obtained directly from the author. 2004] REVIEW OF ADULT ANAGAPETUS 457

Fig. 6. Anagapetus bernea: (a) male genitalia, lateral view; (b) male segment 10, dorsal view; (c) female genitalia, lateral view; (d) female segment 8, dorsal view; (e) vaginal apparatus, ventral view.

Previously published records, particularly from Fifth segment with small flaps covering lateral California, should be reviewed to confirm the glands, no apparent differences between species. identifications. Both sexes with short, stout sternal projection on segment 6, larger sternal projection on seg- Anagapetus Ross 1938 ment 7 of male. Male segment 9 annular, geni- Spurs 2-4-4. Dorsal cephalic warts with a talia lacking superior appendages. Male segment nonsclerotized, open suture extending from 10 divided mesally; lateral portions sclerotized, the mesal wart below the lateral ocelli to the and variously ornamented with fine spines or posterior margin. Mes- and metepisterna each plates at apicoventral corners. Male inferior divided by a short transverse suture, mesotibia appendages long, distally broad, each with api- and tarsa oval in cross section, not flattened. cal incision; with 1 or 2 sclerotized spurs on Mesoscutellar wart a transverse oval, appear- mesal surface, these spurs usually asymmetri- ing as if 2 separate warts have fused mesally. cal on left and right inferior appendages. Phallic 458 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 7. Anagapetus chandleri: (a) male genitalia, lateral view; (b) male segment 10, dorsal view; (c) female genitalia, lateral view; (d) female segment 8, dorsal view; (e) vaginal apparatus, ventral view. structure consisting of S-shaped sclerite re- sclerotized, long, with narrowed shaft before tracted within segment 9; caudal portion highly expanded oval apex. membranous, which may allow the S-shaped sclerite to extend from abdomen. Female seg- Anagapetus aisha Denning 1964 ment 8 weakly sclerotized anteriorly, without (Fig. 5) distinct anterior margin, distal margin vari- Anagapetus sp. Erman 1989. ously cleft dorsally, ventrally, or both. Segment 9 extremely extensile, with its anterior margin The holotype is from Kings Canyon National folded within segment 8, for a distance rela- Park, Fresno County, California (6400 feet alti- tively consistent within a particular species. tude). It is not clear where this locality is since Two large L-shaped sclerites located dorsally the road ends at Kanawyers turnaround at within membranous segment 9. Segment 10 about 5000 feet altitude. This is also the only small with 2 apical papillae. Vaginal apparatus specimen of A. aisha examined during this 2004] REVIEW OF ADULT ANAGAPETUS 459

Fig. 8. Anagapetus debilis: (a) male genitalia, lateral view; (b) male segment 10, dorsal view; (c) female genitalia, lateral view; (d) female segment 8, dorsal view; (e) vaginal apparatus, ventral view.

study from south of El Dorado County, Cali- somesad. Clavate hairs are present on the lat- fornia. The male has a combination of a rela- eral margin of segment 9. Denning’s (1964) tively shallow apical cleft on each inferior description is incorrect on 2 points: (1) the appendage and segment 10 is about as long as apicoventral projection of each inferior append- tall, with a row of small sclerotized teeth along age is broken in the type, normal specimens the ventromesal margins. The interior surface having the ventral projection as long as the of the apicodorsal arm of each inferior append- dorsal projection, and (2) clavate hairs are pres- age is covered with a scurf of downward- ent on the lateral margin of segment 9. There directed setae; this scurf extends more than half are other specimens determined by Denning the length of the inferior appendage. The mesal as A. aisha in the Brigham Young University process at the base of each inferior appendage collection that have the complete ventral infe- is very small, apically acute, and directed dor- rior appendage projection and clavate segment 460 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 9. Anagapetus hoodi: (a) male genitalia, lateral view; (b) male segment 10, dorsal view; (c) female genitalia, lateral view; (d) female segment 8, dorsal view; (e) vaginal apparatus, ventral view.

9 setae. Specimens examined during this study aisha, with the Sagehen Creek specimens, I indicate the depth of the apical inferior append- have concluded that the Sagehen Creek popu- age cleft varies from less than one-fourth the lation is A. aisha. inferior appendage length to nearly one-third. The best character to separate A. aisha from Anagapetus bernea Ross 1947 the closely related A. bernea is the long area of (Fig. 6) scurf on the inferior appendage of A. aisha. The holotype is from Oxbow Springs, Hood The A. aisha female has a nearly quadrate River County, Oregon. Paratypes are from distal portion of segment 9, with the distal Stevens Pass, Berne, Washington, on Highway dorsal and ventral apical margins each only 2. Berne is in Chelan County, east of Stevens slightly cleft. Pass. The male has a combination of deeply The Sagehen Creek, Nevada County, Cali- cleft inferior appendages and segment 10 fornia, population of Anagapetus (Erman 1989) about as long as tall. The apex of segment 10 is initiated this paper. After comparison of sev- broadly rounded, usually with a small, sclero- eral other populations, and the holotype of A. tized tooth at the apicoventral corner. Each 2004] REVIEW OF ADULT ANAGAPETUS 461

Fig. 10. Distribution of material examined.

apicodorsal inferior appendage arm is covered Anagapetus chandleri Ross 1951 with a scurf of downward-directed setae. The (Fig. 7) mesal process at the base of each inferior Anagapetus thirza Denning, 1965 (new syonym) appendage is acute at the apex and directed dorsomesad. The holotype is from 2 miles east-southeast The female segment 9 margin is distinctly of Mariposa Grove, Mariposa County, Califor- tapered in lateral view, with the dorsal margin nia (7000 feet altitude). There is a roadside park shorter than the ventral. Both the dorsal and along the access road to Mariposa Grove about ventral distal margins are cleft mesally. 2 air miles southeast of Mariposa Grove, but the 462 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 11. Distribution of material examined.

altitude is 5585 feet. The highest point near The female segment 9 has a very short dis- Mariposa Grove is Wawana Point at about 6800 tal portion, with the dorsal and ventral apical feet. All streams in the area appear to be tribu- margins each deeply cleft. tary to Big Creek in Yosemite National Park. The holotypes of A. chandleri and A. thirza The male has a combination of a relatively were compared and found to be identical. shallow apical cleft in the inferior appendages and segment 10 about as long as tall, with a Anagapetus debilis Ross 1938 distinct, fingerlike ventrolateral sclerotized (Fig. 8) projection. The ventromesal edge of segment 10 may have a row of very fine spines. The The holotype is from Logan River Canyon, dorsal edge of the apicodorsal inferior append- Cache County, Utah. The male is distinguished age arm has a linear row of setae directed dor- by the combination of deeply cleft inferior sodistad. The mesal process at the base of appendages, segment 10 about twice as long each inferior appendage is acute at the apex as tall, and interior surface of apicodorsal arm and directed dorsomesad. of each inferior appendage covered with a 2004] REVIEW OF ADULT ANAGAPETUS 463 scurf of downward-directed setae. The mesal Ecological Entomology (Illinois Natural His- process at the base of each inferior appendage tory Survey); Robert W. Wisseman; and Cheryl is bifid at the apex and curved caudomesad. B. Barr, University of California, Berkeley, also This results in the distinction between the fig- provided material. I sincerely thank all of ures of Ross (1938) and Schmid (1982). Depend- these individuals for material and assistance, ing on position at preservation, the bifid apex which made this study possible. Oliver S. Flint, is more or less visible. Clavate hairs are pre- Jr., provided very necessary review comments sent on the lateral margin of segment 9, not as and discussion. John C. Morse also provided figured by Ross (1938). Some specimens have useful review comments, discussion, and a copy a row of minute spines on the mesal surface of of his Glossosoma subgenera manuscript. These segment 10. comments and materials have greatly improved In lateral view the distal portion of female this paper. segment 9 is longer than tall. The transverse apicoventral margin is nearly linear, and the LITERATURE CITED apicodorsal margin is slightly cleft. DENNING, D.G. 1964. Descriptions of five new Trichop- tera. Pan-Pacific Entomologist 40:241–245. Anagapetus hoodi Ross 1951 ______. 1965. New Trichoptera from United States and (Fig. 9) Mexico. Pan-Pacific Entomologist 41:262–272. ERMAN, N.A. 1989. Species composition, emergence, and The holotype is from North Fork Iron Creek, habitat preferences of Trichoptera of the Sagehen Hood River County, Oregon. The male has a Creek Basin, California, USA. Great Basin Naturalist combination of deeply cleft inferior appendages 49:186–197. LEVANIDOVA, I.M. 1979. Glossosoma (Anagapetus) schmidi and segment 10 about twice as long as tall. sp. n.—a new species and subgenus for Eurasian The dorsal edge of each apicodorsal inferior streams (Trichoptera, Glossosomatidae). Transactions appendage arm has a linear row of setae of the Soviet Entomological Society 61:92–95 (in directed dorsodistad. The mesal process at the Russian). MARTYNOV, A.V. 1927. Contribution of the aquatic entomo- base of each inferior appendage is bifid at the fauna of Turkestan. Annuaire du Musee Zoologique apex and directed dorsad, usually hidden de l’Académie des Sciences de l’URSS 28:162–193. within tergite 10. Clavate hairs are present on MORSE, J.C., AND L. YANG. 1993. Higher classification of the lateral margin of segment 9. the Chinese Glossosomatidae (Trichoptera). Pages 139–148 in Proceedings of the seventh international In lateral view the apical portion of female symposium on Trichoptera, Umea, Sweden, 3–8 segment 9 is longer than tall, and the dorsal August 1992. Backhuts Publishers, Leiden, The and ventral apical margins are each cleft. Netherlands. ______. 2004. The world subgenera of Glossosoma Curtis (Trichoptera: Glossosomatidae), with a revision of ACKNOWLEDGMENTS the Chinese species of Glossosoma subgenera Syna- fophora Martynov and Protoglossa Ross. Proceed- The initiation of this study occurred years ings of the Entomological Society of Washington ago when Elizabeth A. Bergey, Oklahoma Bio- 106:52–73. logical Survey, provided what appeared to be a ROSS, H.H. 1938. Descriptions of Nearctic caddis flies (Trichoptera) with special reference to the Illinois new species of Anagapetus from Sagehen Creek, species. Bulletin of the Illinois Natural History Sur- California. Nancy A. Erman subsequently vey 21:101–183. generously provided additional material from ______. 1947. Descriptions and records of North Ameri- this location. Tatyana Arefina greatly assisted can Trichoptera, with synoptic notes. Transactions of the American Entomological Society 73:125–168. with Russian translation and provision of A. ______. 1951. The caddisfly genus Anagapetus (Trichop- schmidi material. Tatyana Vshivkova also tera: Rhyacophilidae). Pan-Pacific Entomologist 27: assisted with Russian translation and litera- 140–144. ture. Richard W. Baumann, Monte L. Bean ______. 1956. Evolution and classification of the moun- Museum, Brigham Young University; Boris C. tain caddisflies. University of Illinois Press, Urbana. SCHMID, F. 1962. A propos de deux récents ouvrages sur Kondratieff, C.P. Gillette Museum of Arthro- la phylogénie et la zoogéographie des Trichoptères. pod Diversity, Colorado State University; Miscelánea Zoologica, Barcelóna 1:1–27. Nancy E. Adams and Oliver S. Flint, Jr., ______. 1980. Genera des Trichoptères du Canada et des National Museum of Natural History; Keith États adjacents. Les Insectes et Arachnides du Canada. Partie 7. Agriculture Canada Publications 1692. Dabney and Wojciech J. Pulawski, California ______. 1982. Revision des Trichoptères Canadiens: II. Les Academy of Sciences; Colin Favret, Center of Glossosomatidae et Philopotamidae (Annulipalpia). 464 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Mémoires de la Société Entomologique du Canada WIGGINS, G.B. 1996. Larvae of the North American caddis- 122. fly genera (Trichoptera). 2nd edition. University of ______. 1998. The insects and arachnids of Canada, part Toronto Press. 7: genera of the Trichoptera of Canada and adjoining or adjacent United States. NRC Research Press, Na- Received 10 January 2003 tional Research Council of Canada, Ottawa, Ontario. Accepted 31 March 2004 Western North American Naturalist 64(4), © 2004, pp. 465–470

REPRODUCTION AND DEMOGRAPHY OF TOWNSENDIA APRICA (ASTERACEAE), A RARE ENDEMIC OF THE SOUTHERN UTAH PLATEAU

Vincent J. Tepedino1, Sedonia D. Sipes2, and Terry L. Griswold1

ABSTRACT.—Townsendia aprica (Asteraceae: Astereae), a rare pulvinate perennial of the Southern Utah Plateau, was listed as threatened under the U.S. Endangered Species Act in 1985. Here we report on the reproductive biology and pollination of this little-known species and provide an estimate, for a single site-year, of size-specific reproductive effort. Last Chance townsendia appears a short-lived perennial that begins reproducing in its 2nd year (1.5–2.0 cm diameter). Maximum reproductive effort is attained with the 2.5–4.0 cm diameter size class: 38% of these plants produced 84% of the flower heads. Few plants survived past the 4-cm size class. The species is primarily self-incompatible: neither autogamous nor geitonogamous breeding system treatments produced a significant number of achenes. Unlike some populations of some congeners, the Ivie Creek population was not apomictic. Outcrossing is the primary means of reproduction and native solitary bees are the most important pollinators. Paramount are several species in the genus Osmia, and the ground-nesting species Synhalonia fulvitarsis, which nests among the T. aprica plants. Synhalonia fulvitarsis also visits a contemporaneous blooming phlox (P. austromontana), which may facilitate pollination of the rare townsendia. The Townsendia-Phlox-Synhalonia interaction may represent another example of why we must consider communities rather than individual species in our conservation efforts.

Key words: Townsendia aprica, rare plant, reproduction, breeding system, pollinator, conservation, demography, Asteraceae.

Plant taxa may be rare for many anthro- ent-demanding pollinators such as bees to its pogenic or natural reasons (Fiedler and Ahouse flowers (Jain 1976, Kearns et al. 1998, Tepedino 1992). Certain causes of rarity may be miti- 2000). Second, Last Chance townsendia blooms gated through changes in land management in early spring when frequent harsh and change- policies; e.g., altering grazing schedules or able weather presents problems for plants that restricting off-road vehicle use may improve depend on exothermic insects to move their rare plant reproduction, recruitment, and re- pollen (Macnair et al. 1989). Third, some con- covery. Conversely, the plight of other rare generics of T. aprica are apomictic, and Steb- plants, e.g., edaphic specialists whose rarity is bins (1950) and Beaman (1957) have suggested due to disappearance of substrate, may be that such agamic complexes are unlikely to irremediable by changes in management pro- long endure or to give rise to new taxa. Such tocol. Only additional information on rare plant considerations are of general import in eluci- characteristics will enable informed decisions dating correlates of rarity and in providing on how best to allocate scarce resources for specific information needed by land managers. conservation; descriptive studies are a first step Here we describe the breeding system and to realistic management decisions (Schemske identify important flower visitors of this rare et al. 1994). Utah endemic. We also provide an estimate, for Some rare species, such as Townsendia 1 population in a single year, of size-specific re- aprica (Last Chance townsendia), a threatened productive effort as measured by flower number. perennial composite of the Southern Utah Plateau region, may be at risk because of char- MATERIALS AND METHODS acteristics connected with reproduction and Species Description pollination. First, like other rare species, Last Chance townsendia may have difficulty attract- Townsendia aprica was listed as threatened ing density-dependent, high-energy, and nutri- under the U.S. Endangered Species Act in 1985.

1USDA ARS Bee Biology & Systematics Laboratory, Department of Biology, Utah State University, Logan, UT 84322-5310. 2Department of Biology, Utah State University, Logan, UT 84322-5305. Present address: Department of Botany, Southern Illinois University, Carbondale, IL 62901-6509.

465 466 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Thirteen occurrences were initially recorded; florets opened, and 12 were unmanipulated. 38 more have now been discovered in the Plants were marked with a numbered metal same 3 central Utah counties: Emery, Sevier, tag that was nailed into the substrate. Chicken Wayne. The species occurs in small, barren wire cages covered with pliable 1-mm-mesh openings of pinyon-juniper interspersed with nylon excluded all but the smallest insects; salt-desert shrubs in soil substrates that tend cages were nailed into the substrate to anchor to be nonuniform (U.S. Fish and Wildlife Ser- them during the frequent high winds. Mature vice 1993). seed heads were collected and returned to the Last Chance townsendia plants, like other lab, where numbers of ray and disc florets and members of the genus (Reveal 1970, Shultz filled achenes were counted. The Wilcoxon and Holmgren 1980), are acaulescent, pulv- nonparametric paired-rank test was used to inate, and have one to several rosettes, each of compare percent florets with filled achenes. which may support a heterogamous-radiate In 1991 we used plants at Ivie Creek to flower head (Mani and Saravanan 1999). Heads compare achene production from self- and are composed of apricot-colored ray florets outcrossing treatments (geitonogamy and xen- (yellow-orange when fresh, aging to cream or ogamy, respectively). We also replicated our even white) that are female, and disc florets autogamy and agamospermy treatments from that are hermaphrodites. The florets open cen- 1990 to confirm our results. To conduct these tripetally beginning with ray flowers; thus each experiments, we caged 25 plants with exactly head is protogynous. 3 heads per plant and identified them as before. In contrast, hermaphrodite disc florets are In addition, each head was marked with an protandrous, with the initial male stage lasting identifying color of nontoxic paint on the in- from a few hours to a day or more, depending volucral bracts signifying the treatment to be upon the temperature. The anther column is received. All florets, ray and disc, of 2 heads of thigmotactic and aids in pollen presentation. each plant received 1 of the following treat- When touched, the column retracts, thus forcing the egress of pollen like toothpaste from a ments: autogamy/agamospermy (unmanipu- tube. The female stage may last from a few lated), geitonogamy (receptive stigmas polli- hours (if pollinated immediately) to a few days. nated with fresh pollen from other florets on The stigmatic lobes of unisexual ray flowers the same plant), xenogamy (receptive stigmas separate and spread fully, but the lobes of some pollinated with fresh pollen from florets on disc flowers remain attached at the tip (and another plant). Thus, each plant had 1 head in sometimes along their full length) and never each treatment. This procedure was adopted completely separate (personal observations). because of the small size and crowding of the florets and the difficulty of marking and polli- Data Collection nating them individually. Care was taken to Data were collected on the population alternate the sequence of treatments among structure, reproductive biology, and pollina- plants so as to avoid any effect of pollination tion of Last Chance townsendia from 1988 to priority. We did not test for the possibility that 1991 at 2 sites in the large upper Last Chance T. aprica is a pseudogamous apomict because Creek population. In 1989 we constructed a all known apomicts in the genus and most in demographic profile of plants at the Ivie Creek the Asteraceae are diplosporic, while pseudo- site by randomly selecting 200 plants, measur- gamic apomicts are predominantly aposporic ing each cushion at its widest diameter, and (Beaman 1957, Nygren 1967). counting the number of flower heads. With As experimental plants were chosen, we first the data we constructed a size-frequency dis- selected the closest plant with a head at the tribution and related size to reproductive same phenological stage. This uncaged head potential. was marked as an open-pollinated control to In 1990 we used 24 plants from the Ivie assess pollinator-limitation; it was later col- Creek population to determine whether the lected for comparison of seed production with florets were capable of seed production with- the xenogamy treatment of the experimental out receiving outcrossed pollen. We chose only plant. Again we compared percent filled achenes plants with single heads of comparable phe- per floret per head using nonparametric tests: nology. Twelve plants were caged before any the Wilcoxon rank test, the Mann-Whitney test 2004] REPRODUCTION OF A RARE UTAH ENDEMIC 467

Fig. 1. Measures of size and number of flower heads produced per plant of Townsendia aprica at the Ivie Creek study site in central Utah: A, percent frequency distribution of plant size; B, percent frequency distribution of number of heads per plant; C, percent of plants with ≥ 1 flower head by plant size; D, mean number of heads per plant by plant size (solid line; upper s shown); percent of total recorded flower heads produced by each size class (broken line). for unpaired comparisons, or the Friedman 2- with increasing size (Fig. 1d). There was a way ANOV rank test (Conover 1971). highly significant positive correlation between In 1988 and 1989 we made single-day col- mean heads per plant and plant diameter (P = lections from the flowers of populations along 0.003, t = 6.4, df = 6, r2 = 0.91). Most plants upper Last Chance Creek and at Sage Hole, >2 cm produced ≥1 head. At 3 cm, plants Sevier County. In 1991 few collections were averaged >2 heads; the few plants that grew made because of frequent inclement weather to 4 cm produced the most heads per plant, and the demands of conducting breeding sys- averaging 6. Most heads (about 84% of the tem treatments and the pollinator-limitation total produced) were produced by the 38% of experiment. plants in the 2.5–4.0 cm size category (Fig. 1d). Results from preliminary experiments RESULTS showed that achenes are not produced unless Cushions of T. aprica ranged in diameter the receptive stigmas of florets receive pollen from 0.5 cm to 5.0 cm (mode = 2.0 cm, Fig. transported by insects or perhaps by wind; 1a). Plants were concentrated in the smaller i.e., flowers are neither autogamous or agamo- size classes suggesting a young and healthy spermous. In comparison, 42.6% of ray flowers population: over 75% were in the 0.5–2.5 cm and 45.5% of disc flowers on open-pollinated size classes; few exceeded 4.0 cm. heads produced filled achenes. Disc flowers Plants produced up to 10 heads. Although produced significantly more achenes than did the modal head number was 0, most plants ray flowers (Wilcoxon test, P = 0.01) but this (67.5%) produced 1 or more heads (Fig. 1b). was due to the greater average number of disc Plants began producing heads with the 1.5-cm flowers than ray flowers (Wilcoxon test, P < size class (Fig. 1c); flower production increased 0.02; Table 1). 468 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Mean (s) florets, achenes, and percent achenes per floret for ray and disk flower heads of caged and open Townsendia aprica in central Utah. N = 12.

______Ray ______Disk Florets Achenes % Florets Achenes %

OPEN mean 13.0 5.5 42.6 22.6 9.3 45.4 s 2.0 3.6 27.9 7.0 5.0 26.7 min 10 0 0 13 1 4.6 max 18 13 92.9 36 17 77.3 CAGED mean 12.5 0 0 21.3 0 0 s 4.2 — — 9.9 — — min 7 — — 4 — — max 20 — — 35 — —

TABLE 2. Mean (s) florets, achenes, percent achenes per floret, and sample size (N) of Townsendia aprica heads subjected to autogamy/agamospermy (AA), geitonogamy (G), xenogamy (X), or open (O) pollination treatments.

______Ray______Disk Florets Achenes % N Florets Achenes % N AA 11.3 (2.4) 0.9 (1.9) 6.4 (12.9) 19 23.8 (9.5) 1.1 (3.6) 2.4 (7.4) 19 G 12.5 (3.7) 1.0 (2.6) 6.0 (14.7) 20 25.8 (7.1) 0.7 (1.2) 2.3 (4.2) 20 X 13.0 (3.9) 4.6 (3.7) 34.9 (25.2) 20 24.3 (8.6) 12.6 (5.4) 52.3 (15.1) 20 O 11.3 (2.6) 3.1 (2.5) 26.2 (16.7) 11 25.8 (9.3) 11.4 (7.0) 41.5 (20.0) 11

Our 2nd breeding system experiment com- years (Table 3). The genus Osmia was repre- pared reproductive success of self- and cross- sented by 9 species, all in small numbers. pollinated florets, geitonogamy and xenogamy Most abundant were members of the antho- treatments, respectively. Our results for both phorid species Synhalonia fulvitarsis, which ray and disc flowers showed that outcrossed we found nesting in the ground among the T. pollen produces maximal achene production: aprica plants at Ivie Creek. In 1991 many the xenogamy treatments had significantly more S. fulvitarsis and other bee species, as greater percent seed set than did the geitonog- well as assorted dipterans, were seen on the amy and autogamy/agamospermy treatments, flowers than we were able to capture (see and there was no significant difference be- Methods). tween the latter 2 (Table 2; Friedman tests, both P < 0.0001). Also, we found no difference DISCUSSION between disc and ray flowers in percent seed set due to autogamy/agamospermy or geitonog- Although it is uncommon for size and age amy (Wilcoxon tests, both P > 0.15), but in the to be closely correlated in plant species (Hutch- xenogamy treatment, disc flowers produced ings 1986), our data suggest that T. aprica significantly greater percent achenes than did defies this rule (Fig. 1). Townsendia aprica rays (Wilcoxon test, P = 0.04). appears a short-lived perennial that begins We also examined the likelihood that achene blooming with a single head in the 2nd year as production in T. aprica was limited by a dearth diameter size of 1.5–2.0 cm diameter is reached. of pollinators. Comparisons of percent seed No plant <1.5 cm diameter flowered, and rel- set between open-pollinated heads and exper- atively few of the 2-cm modal size class pro- imentally cross-pollinated heads revealed no duced flower heads. Because maximum repro- significant differences for ray or disc flowers duction is not achieved until later in life, it is (Table 2; Mann-Whitney tests, both P > 0.10). essential that populations be managed to We captured 15 species of bees in 4 fami- ensure the persistence of plants to the larger lies visiting the flowers of T. aprica in 3 site- size classes. 2004] REPRODUCTION OF A RARE UTAH ENDEMIC 469

TABLE 3. Bees captured on the flowers of Last Chance townsendia in 3 years at the Ivie Creek (5/2/89, 1991) and Sage Hole (5/12/88) sites in central Utah. Ivie Creek sites noted by IC, Sage Hole by SG. Species also collected foraging on Phlox austromontana noted by *. Species that nest in wood substrates noted by W. Family Species

APIDAE MEGACHILIDAE (continued) Ceratina nanulaIC, SH, W O. kincaidiiSH, W Nomada sp.IC* O. pusillaIC Synhalonia fulvitarsis IC* O. sanrafaelaeIC, W* HALICTIDAE O. sladeniSH* Lasioglossum (Evylaeus) sp.IC* O. vandykeiSH* MEGACHILIDAE O. (Acanthosmioides) sp.IC Dioxys pomonaeIC O. (Melanosmia) sp.SH* Osmia coloradensisIC, W Stelis pavoninaIC O. gaudiosaIC, SH, W*

Townsendia aprica illustrates how determin- heads in the 2nd year. Stiff winds may move ing method(s) of reproduction of rare plants some pollen, but pollen is unlikely to adhere can aid conservation objectives. Stebbins (1950) to receptive stigmas because wind also coats argued that plant species that reproduced the stigmatic surface with dirt. Hand-applied agamically were more likely to be evolutionary pollen would not adhere to dirt-covered stig- dead ends. Obligate apomixis is known from at mas. The habit of the ray petals folding over least 12 of 21 species of Townsendia (Beaman and covering the disc flowers at night and in 1957). As a perennial of higher elevations inclement weather may partially shield stigmas (>2000 m), T. aprica also fits an apomictic pro- from clogging with dirt and from receiving file. All known Townsendia taxa with apomic- pollen. tic populations also have sexual populations. If The prime movers of T. aprica pollen are Stebbins (1950) is correct, then protecting native solitary bees (Table 3). Native bees are apomictic populations of T. aprica would be of also important to the success of many other slightly lower priority than protecting sexual rare plant species in the western United States populations. Thus, the Ivie Creek population (Tepedino 2000). The activity of these insects merits conservation priority. Until the mode of is inhibited by inclement spring weather. Cold, reproduction is examined for other T. aprica rain, snow, and wind can greatly restrict the populations, land managers should assume that time that bees have to forage on and pollinate all populations reproduce sexually and accord flowers. Inclement weather also kills pollen, them equal status. Finally, we do not propose the slows development, and may vitiate female re- sacrifice of any extant apomictic populations. production (Frankel and Galun 1977). For ex- All populations of an imperiled species are of ample, studies in 1991 began on 11 May. Incle- value; apomicts, for example, are superior colo- ment weather during 5 of the initial 9 days nizers because they can reproduce unassisted. prohibited pollinator activity and many florets While we found no evidence for apomicty likely expired without producing achenes. in the Ivie Creek population, we commonly The most abundant species to visit the flow- found unsplit stigmas in disc flowers. These ers was the early spring bee Synhalonia fulvi- disc flowers, particularly at the center of the tarsis, which also nested in the ground among capitulum, may be functional males, and heads T. aprica and Phlox austromontana plants. Nine may be developing toward functional monoecy. species of mason/leafcutter bees in the genus This phenomenon is known in other composites Osmia were also captured and observed. At where it is usually indicated by a notched style least 4 of these Osmia species, and Ceratina (Burtt 1977, Mani and Saravanan 1999). Other nanula, nest in holes in wood (Hurd 1979, F.D. populations should be surveyed for this character. Parker unpublished); their numbers might be Our breeding system experiments showed augmented by the use of trap-nests (Krombein that the Ivie Creek population of T. aprica is 1967). Supplying trap-nests would be espe- predominantly outcrossed: achene production cially advisable if achene production were lim- in all selfing treatments was very low and may ited by pollinator scarcity. Although we found have resulted from slight contamination of some no indication of pollinator limitation (Table 2), 470 WESTERN NORTH AMERICAN NATURALIST [Volume 64 this could quickly change because bees are HURD, P.D., JR. 1979. Superfamily Apoidea. Pages 1741– readily and subtly victimized by human activ- 2209 in K.V. Krombein, P.D. Hurd, Jr., D.R. Smith, and B.D. Burks, editors, Catalog of Hymenoptera in ity. Then, we might need to take steps to pre- America north of Mexico. Smithsonian Institution serve bees along with the plants they service Press, Washington, DC. (Tepedino et al. 1997). HUTCHINGS, M.J. 1986. The structure of plant popula- Of especial interest was the commonness tions. Pages 97–136 in M.J. Crawley, editor, Plant with which bees that visited T. aprica also vis- ecology. Blackwell Scientific Publications, Boston. JAIN, S.K. 1976. The evolution of inbreeding in plants. ited the contemporaneously blooming Phlox Annual Review of Ecology and Systematics 7:469–495. austromontana (several species of Osmia, S. KEARNS, C.A., D.W. INOUYE, AND N.M. WASER. 1998. fulvitarsis; Table 3). Interestingly, S. fulvitarsis, Endangered mutualisms: the conservation of plant- although usually a species that visits a variety pollinator interactions. Annual Review of Ecology and Systematics 29:83–112. of flowering plant species (Hurd 1979), has also KROMBEIN, K.V. 1967. Trap-nesting wasps and bees: life been found closely associated with another histories, nests, and associates. Smithsonian Press, species of Phlox in Wyoming (Tepedino 1979). Washington, DC. 570 pp. Perhaps the presence of P. austromontana MACNAIR, M.R., V.E. MACNAIR, AND B.E. MARTIN. 1989. facilitates (Thomson 1978, Rathcke 1983) the Adaptive speciation in Mimulus: an ecological comparison of M. cupriphilus with its presumed progeni- pollination of the rare T. aprica as has been tor M. guttatus. New Phytologist 112:269–279. suggested for other rare plant species (Geer et MANI, M.S., AND J.M. SARAVANAN. 1999. Pollination ecol- al. 1995). Studies should investigate the asso- ogy and evolution in Compositae (Asteraceae). Sci- ciation between these 2 plant species and S. ence Publishers, Enfield, NH. 166 pp. NYGREN, A. 1967. Apomixis in the angiosperms. Encyclo- fulvitarsis. For example, is T. aprica achene pedia of Plant Physiology 18:551–596. production higher in the presence of P. austro- RATHCKE, B. 1983. Competition and facilitation among montana, and can this be related to increased plants for pollination. Pages 305–329 in L. Real, edi- visitation by S. fulvitarsis? tor, Pollination biology. Academic Press, New York. REVEAL, J.L. 1970. A revision of the Utah species of Town- sendia (Compositae). Great Basin Naturalist 30:23–52. ACKNOWLEDGMENTS SCHEMSKE, D.W., B.C. HUSBAND, M.H. RUCKELSHAUS, C. GOODWILLIE, I. PARKER, AND J.G. BISHOP. 1994. Eval- We thank Bill Bowlin for field help, U.S. uating approaches to the conservation of rare and Forest Service personnel for all-around coop- endangered plants. Ecology 75:584–606. eration, “Olli” and “H-less” Messinger and SHULTZ, L.M., AND A.H. HOLMGREN. 1980. A new species Dianne Arnold for help with tables and figure, of Townsendia from northern Arizona. Brittonia 32: and Linda Jennings, Leila Shultz, and Jeanette 144–147. STEBBINS, G.L. 1950. Variation and evolution in plants. Whitton for making helpful comments about Columbia University Press, New York. 643 pp. our presentation. TEPEDINO, V.J. 1979. Bee visitation of Phlox bryoides (Polemoniaceae). Great Basin Naturalist 39:197–198. LITERATURE CITED ______. 2000. The reproductive biology of rare rangeland plants and their vulnerability to insecticides. Pages 1– BEAMAN, J.H. 1957. The systematics and evolution of 10 in G.L. Cunningham and M.W. Sampson, techni- Townsendia (Compositae). Contributions from the cal coordinators, Grasshopper integrated pest man- Gray Herbarium, Harvard University 183:1–157. agement user handbook. Volume 3.5. USDA APHIS BURTT, B.L. 1977. Aspects of diversification in the capitu- Technical Bulletin 1809, Washington, DC. lum. Pages 41–59 in V. H. Heywood and B.L. Turner, TEPEDINO, V.J., S.D. SIPES, J.L. BARNES, AND L.L. HICKER- editors, The biology and chemistry of the Composi- SON. 1997. The need for “extended care” in conser- tae. Volume 1. Academic Press, New York. vation: examples from studies of rare plants in the CONOVER, W.J. 1971. Practical nonparametric statistics. western United States. Pages 245–248 in K.W. Rich- John Wiley & Sons, New York. 462 pp. ards, editor, Proceedings of the 7th International FIEDLER, P.L., AND J.J. AHOUSE. 1992. Hierarchies of Symposium on Pollination. Acta Horticultura 437. cause: toward an understanding of rarity in vascular Leiden, The Netherlands. plant species. Pages 23–47 in P.L. Fiedler and S.K. THOMSON, J.D. 1978. Effect of stand composition on insect Jain, editors, Conservation biology: the theory and visitation in two-species mixtures of Hieracium. practice of nature conservation, preservation and American Midland Naturalist 100:431–440. management. Chapman and Hall, New York. U.S. FISH AND WILDLIFE SERVICE. 1993. Last Chance FRANKEL, R., AND E. GALUN. 1977. Pollination mechanisms, townsendia (Townsendia aprica) recovery plan. U.S. reproduction and plant breeding. Springer-Verlag, Fish and Wildlife Service, Denver, CO. 18 pp. New York. 281 pp. GEER, S.M., V.J. TEPEDINO, T.L. GRISWOLD, AND W.R. Received 17 April 2003 BOWLIN. 1995. Pollinator sharing by three sympatric Accepted 9 March 2004 milkvetches, including the endangered species Astra- galus montii. Great Basin Naturalist 55:19–28. Western North American Naturalist 64(4), © 2004, pp. 471–481

MARMOT DISTRIBUTION AND HABITAT ASSOCIATIONS IN THE GREAT BASIN

Chris H. Floyd1

ABSTRACT.—In this study I describe the distribution and habitat associations of yellow-bellied marmots (Marmota flaviventris) in the Great Basin, compare my findings with those of E.R. Hall during his 1929–1936 survey and later surveys, and discuss potential reasons for changes in marmot distribution over time. I found 62 marmot burrow sites in 18 mountain ranges, mostly in rocky meadows situated on well-drained slopes between 2100 m and 3000 m elevation. Mar- mots were generally found near burrows dug within talus slopes, talus-like rock piles, or clusters of massive boulders. Oceanspray (Holodiscus discolor) was the shrub most commonly associated with occupied rock formations. Marmots were most abundant in the Ruby/East Humboldt Range and were common in the Desatoya, Shoshone, Toiyabe, Toquima, Cherry Creek, Schell Creek, Deep Creek, and Stansbury Ranges. Marmots appeared to be uncommon in the Monitor Range and rare in the Clan Alpine, Roberts, and Snake Ranges. I was unable to find marmots in the Diamond, Egan, Spruce-Pequop, White Pine, and Oquirrh Ranges, although I located old, weathered marmot scats in all but the latter 2 ranges. Other evidence confirms that marmots do actually occur in the Oquirrh Range, but extensive searches of the White Pine Range, including some of the same rock formations where E.R. Hall collected marmots, revealed no sign of marmots. My distribution data suggest that marmots may have gone extinct in some Great Basin mountain ranges during the last century. These disappearances may represent a natural extinction-recolonization dynamic, but a more alarming possibility is a recent die-off linked to climate change, which is predicted to force montane vegetation zones further upslope, shrinking the habitat of associated faunas. However, marmots in this study were observed as low as 1550 m elevation, indicating an altitudinal flexibility that may allow this species to survive climatic change better than more specialized boreal species such as pikas (Ochotona princeps) and water shrews (Sorex palustris).

Key words: Marmota flaviventris, marmot, Great Basin, climatic change, boreal mammals, talus, Holodiscus discolor.

Climatic change over the next several early 1970s and Lawlor (1998) during 1982– decades is predicted to force boreal life zones 1995, as part of their studies of Great Basin higher into the mountaintops, imperiling the insular biogeography. In this study I investi- resident faunas by fragmenting and constrict- gated distributions of yellow-bellied marmots ing their habitats (Gottfried et al. 1999, Hill et (Marmota flaviventris) and characterized habi- al. 2002, Kullman 2002). Boreal mammals in tat associations of marmots in the Great Basin the Great Basin may face an especially high during 1999–2002. These large rodents, widely extinction risk by being restricted to small distributed throughout western North Amer- islands of montane/alpine habitat surrounded ica, including the Great Basin, Rocky Moun- by arid lowlands (McDonald and Brown 1992). tains, Cascade Range, and Sierra Nevada (Frase To detect altitudinal movements and extinctions, and Hoffman 1980), are most commonly found it is crucial to have accurate ecological information on present-day distributions (Kodric- in montane and alpine meadows in close asso- Brown and Brown 1993). Such data are ciation with rocks. Burrows are dug beneath lacking for most Great Basin mountain ranges rocky outcrops or talus slopes (Svendson 1974, (Rickart 2001). Frase and Hoffmann 1980). Marmots are active E.R. Hall conducted the first wide-ranging only during the season of vegetative growth study of mammalian distributions in the Great and hibernate for approximately 8 months of Basin during 1929–1936, focusing mainly on the year (Frase and Hoffman 1980). My objec- Nevada (Hall 1946). Subsequent surveys focus- tive was to compare my findings to earlier sur- ing on boreal mammals were conducted by veys and discuss explanations for any apparent Brown (1971, 1978) during the late 1960s and shifts in marmot distribution.

1Department of Wildlife, Fish and Conservation Biology, University of California, Davis, CA 95616. Present address: Department of Biology, University of Wisconsin–Eau Claire, Eau Claire, WI 54702.

471 472 WESTERN NORTH AMERICAN NATURALIST [Volume 64

METHODS in each mountain range. In each range except for 2 (the Oquirrh and Roberts, discussed I searched Great Basin mountain ranges for below), I thoroughly explored at least 2 major marmots during July–August 1999, May–August drainages, especially those that were known to 2000 and 2001, and May–June 2002 as part of have marmots or appeared to have potential a genetic study to test the “nonequilibrium” marmot habitat. Marmots tended to sit on model of island biogeography (Brown 1971, prominent rocks and give loud alarm calls at 1978, Lawlor 1998, Floyd 2003). Thus, I focused the moment they detected me, thus making it on the ranges surveyed in earlier studies (Fig. easy to find most occupied marmot burrows 1), except the White Mountains, where marmot within about 100–500 m. Marmot burrows with- distributions are well known (Stallman and out visible marmots were identified by hole Holmes 2002), and the Grant-Quinn Canyon size and presence of marmot scats. Generally, Range, where marmots do not occur (Lawlor 2–8 hours were spent at each “burrow site,” 1998). Following Brown (1971) and Lawlor here defined as a location (usually a slope, ridge, (1998), I refer to the Ruby and East Humboldt or small valley) in which marmots or marmot Ranges as the same mountain complex because burrows were found. Burrows separated by less of their proximity. The authors did the same than 500 m walking distance were considered with the Toiyabe and Shoshone, Toquima and to be within the same burrow site, unless they Monitor, and Egan, Cherry Creek, and Schell occupied slopes in different major drainages Creek Ranges, but I treat these separately (e.g., on opposite sides of a ridge). Burrow sites here because they are geologically and ecolog- were reported as latitude-longitude coordinates ically distinct (Charlet 1996, NVGAP 1996). estimated to the nearest second using a GPS I searched for marmots by walking up major unit or 1:25,000 topographic maps. Elevations drainages (defined as valleys containing pri- were estimated to within 50 m. More detailed mary or secondary branches of a watercourse location descriptions, including information that emerges at the base of a mountain range), on all drainages searched, are available from the from the mouth of the drainage (ca. 1800–2100 author upon request. When there were multi- m elevation) to the highest surrounding peak ple observations from a single burrow site, I or ridgeline, listening and looking for marmots reported the coordinates and elevation of the along the way. Because my initial surveys sug- burrow that would be easiest for another ob- gested that marmots frequently inhabited server to locate and that would lead to neigh- open areas, I attempted to thoroughly explore boring marmots. The minimum-maximum ele- all unforested ridges, slopes, and valley bot- vation range was reported when the burrows toms within view. Targeting open areas did not within a location were distributed across an unduly bias my characterization of marmot altitudinal gradient greater than 100 m. To habitat, however, because reaching them re- characterize marmot habitat, I estimated slope quired me to extensively traverse other habitat aspect (exposure) and described vegetation zones such as riparian areas, sagebrush (Arte- and rock features in the immediate vicinity of misia tridentata) scrublands, pinyon-juniper sites where marmots or scats were found. woodlands (Pinus monophylla and Juniperus osteosperma), and boreal forests (e.g., aspen RESULTS [Populus tremuloides], limber pine [Pinus flexilis], and white fir [Abies concolor]). From the Habitat Associations high points, I was able to comprehensively A total of 62 marmot burrow sites were scan (using 10 × 40 binoculars) the surround- found in 18 mountain ranges (Table 1). Bur- ing landscape for several kilometers around, rows primarily were located in rocky meadows often down to the bottom of the drainage. situated on well-drained slopes within the As was the case with previous studies (Hall montane vegetation zone, which begins at the 1946, Brown 1971, 1978, Lawlor 1998), no upper margin of the pinyon-juniper zone at attempt was made to search every major drain- approximately 2400 m elevation (Charlet 1996, age, nor were all ranges searched with equal NVGAP 1996). Marmots were almost always effort, because my main purpose was to obtain found <10 m from exposed rock formations adequate genetic representation from marmots that were generally of 3 types: (1) stable talus 2004] DISTRIBUTION AND HABITATS OF MARMOTS 473

Fig. 1. Map of Great Basin mountain ranges where yellow-bellied marmot (Marmota flaviventris) distributions and habitat associations were studied. Abbreviations for mountain ranges: CC, Cherry Creek; CL, Clan Alpine; DC, Deep Creek; DE Desatoya; DI, Diamond; EG, Egan; EH, East Humboldt; MT, Monitor; OQ, Oquirrh; RO, Roberts; RU, Ruby; SC, Schell Creek; SH, Shoshone; SN, Snake; SP, Spruce-Pequop; ST, Stansbury; TY, Toiyabe; TQ, Toquima; WP, White Pine. Outlined ranges are those in which marmots were not observed but old marmot scats were found. slopes made up of coarse, angular rocks ≥20 stones, apparently because such soils prevent cm in length, associated with outcrops or crags the construction of workable burrows (Svend- of volcanic or sedimentary strata (Fig. 2); (2) sen 1974). Single boulders and rock outcrops talus-like rock piles derived from glacial deposits without associated talus usually lacked marmots or in situ bedrock erosion (A. Wilcox personal unless there were multiple crevices through- communication; Fig. 3); (3) clusters of massive out the boulder or outcrop. Talus-dwelling boulders, often of intrusive granitic origin, con- marmots generally burrowed within the talus taining multiple deep crevices that were inac- itself and not in the outcrop above. However, cessible to larger animals (Fig 4). marmots frequently used outcrops or promi- Occupied rock formations usually were nent rocks for lookout posts, as evidenced by deeply embedded in the soil and separated the abundance of scats and the fact that mar- from each other by ≥40 m of montane meadow mots when approached generally alarm-called vegetation dominated by mountain sagebrush from the outcrop before fleeing into the talus (A. tridentata vaseyana) and occasionally by below (Hall 1946, CHF personal observation). woodlands (especially aspen and curleaf moun- The actual burrow entrances in talus were sel- tain-mahogany [Cercocarpus ledifolius]). The dom discernable because of the many possible heterogeneous distribution of rock formations entry-exit routes through the jumbled rocks. resulted in a correspondingly patchy distribu- Elevations of marmot locations ranged from tion of marmots. 1560 m to 3500 m; both the mean and median Marmots were almost never found in talus were approximately 2550 m, and about 81% of that was gravelly or composed of small, platy locations were located between 2100 m and 474 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Locations of marmot burrow sites on 18 Great Basin mountain ranges listed west to east among ranges, and south to north within ranges. All ranges are in Nevada except the Deep Creek, Stansbury, and Oquirrh Ranges, which are in Utah. Sites marked with an asterisk (*) are sites where only marmot scats (not marmots themselves) were found. Sightings in the Oquirrh Range were those reported by T. Becker. Where there were multiple observations within a valley, the location of a single representative location is given. The range of elevation is listed where there were multiple localities across an altitudinal gradient of ≤100 m. Mountain Drainage or Latitude Longitude Elevation range County feature (N) (W) (m) Clan Alpine Churchill Mouth of Cherry Creek 39º32′33″ 117º50′23″ 1900 Clan Alpine Churchill Upper Cherry Valley 39º34′41″ 117º56′24″ 2300 Clan Alpine Churchill Cherry Creek / War Canyon road 39º35′18″ 117º54′45″ 2500 Desatoya Lander Carroll Summit / Bald Mountain 39º16′46″ 117º43′41″ 2400 Desatoya Churchill Eastgate 39º18′16″ 117º52′53″ 1560 Desatoya Lander North of Carroll Summit 39º18′36″ 117º44′42″ 2750 Desatoya Lander Haypress Creek 39º19′40″ 117º42′18″ 2500 Desatoya Churchill Topia Creek 39º24′42″ 117º46′06″ 2500 Desatoya Churchill Cedar Creek 39º26′32″ 117º46′45″ 2250 Shoshone Lander North Shoshone Peak 39º09′16″ 117º27′59″ 2500–2600 Shoshone Lander Underdown Canyon* 39º09′51″ 117º27′03″ 2600 Toiyabe Nye Stewart Cr. Trail / Arc Dome area 38º52′07″ 117º20′52″ 3300 Toiyabe Nye N. Twin River* 38º53′06″ 117º18′19″ 2600 Toiyabe Nye San Juan Creek 39º06′02″ 117º14′53″ 2600 Toiyabe Nye Toiyabe Range Pk / Alice Gendron Cr.* 39º06′29″ 117º12′38″ 2150–2700 Toiyabe Lander South of Bob Scott Summit– US 50 39º26′29″ 117º00′16″ 2350 Toiyabe Lander Communications Towers road 39º28′26″ 117º03′04″ 2400 Toiyabe Lander Austin Summit–US 50 (east) 39º28′34″ 117º01′48″ 2100–2350 Toiyabe Lander Austin Summit–US 50 (west) 39º29′22″ 117º02′55″ 2200 Toiyabe Lander Amadour Canyon 39º33′48″ 117º04′15″ 2250 Toquima Nye Antone Creek 38º39′09″ 116º56′00″ 2600–2700 Toquima Nye Jefferson Summit and Shoshone Mt. 38º41′34″ 116º56′45″ 2650–2750 Toquima Nye Mount Jefferson 38º45′05″ 116º55′44″ 3500 Toquima Nye Upper Pine Creek 38º46′59″ 116º55′49″ 3300 Toquima Nye Moores drainage* 38º49′14″ 116º55′06″ 3100 Monitor Nye Big Cottonwood Canyon 38º22′49″ 116º45′05″ 2500–2600 Monitor Nye Table Mountain* 38º47′20″ 116º35′00″ 3000 Monitor Eureka N. Fork Allison Creek / Summit Mt. 39º22′02″ 116º28′23″ 2750–2900 Roberts Eureka Vinini Creek / Roberts Peak 39º51′51″ 116º17′36″ 2700–2800 Diamond White Pine Sadler Canyon* 39º34′37″ 115º48′28″ 2600 Diamond Eureka Threemile Canyon* 39º50′43″ 115º47′58″ 2750

3000 m elevation (Fig 5). For slopes with mar- These oceanspray-free areas usually were found mot locations, aspect (exposure) was signifi- on sparsely vegetated or woodland-covered cantly clumped (Rayleigh test: z = 2.99; N = slopes at lower elevations. 93; P < 0.05), with a mean east-facing (89º) Elderberry (Sambucus cerulea) and moun- direction (Batschelet 1981). tain sagebrush often were present at marmot Almost without exception marmot burrow burrow sites. Plants less commonly present sites were closely associated with oceanspray, included chokecherry (Prunus virginiana), cur- Holodiscus discolor. As is characteristic of this rants (Ribes sp.), wild rose (Rosa woodsii), shrub, H. discolor typically ringed the outer snowberry (Symphoricarpos sp.), perennial boundaries of talus slopes and rock piles or grasses (often Leymus cinereus), and prickly protruded from the interstices of granitic pear (Opuntia polyacantha). boulder piles (Mozingo 1987; Figs. 2–4). Talus Exceptions to the association with ocean- and boulder piles without associated H. dis- spray occurred chiefly in the northern half color generally lacked any sign of marmots. of the Ruby/East Humboldt Range, where 2004] DISTRIBUTION AND HABITATS OF MARMOTS 475

TABLE 1. Continued. Mountain Drainage or Latitude Longitude Elevation range County feature (N) (W) (m) Ruby Elko Harrison Pass* 40º22′41″ 115º30′05″ 2850 Ruby Elko Battle Creek mines / Myers Creek 40º30′19″ 115º23′33″ 2500 Ruby Elko Island Lake 40º36′53″ 115º23′00″ 2750–3200 Ruby Elko Thomas Canyon 40º37′25″ 115º24′31″ 2700 Ruby Elko Upper Lutt’s Creek / Thompson Creek 40º39′05″ 115º20′49″ 2750–3100 Ruby Elko Lamoille Canyon 40º39′39″ 115º26′08″ 2350 Ruby Elko Soldier Creek (upper) 40º45′53″ 115º15′14″ 2750 Ruby Elko Soldier Creek (trailhead) 40º46′34″ 115º19′33″ 2000 Ruby Elko Gardner Creek/Secret Peak 40º48′40″ 115º15′08″ 2550 E. Humboldt Elko Franklin River / N. Ruby Valley Rd. 40º48′45″ 115º08′33″ 1970 E. Humboldt Elko Horse Creek 40º48′54″ 115º06′48″ 2250–2450 E. Humboldt Elko Secret Pass 40º51′51″ 115º15′00″ 1850 E. Humboldt Elko Angel Lake 41º01′32″ 115º05′03″ 2500–2750 Egan White Pine Telegraph Peak* 39º40′39″ 114º54′01″ 2750 Cherry Creek White Pine Goshute Basin 40º03′46″ 114º51′06″ 2000 Cherry Creek White Pine Goshute Basin–continued 40º04′58″ 114º50′50″ 2750 Spruce-Pequop Elko Basco Springs–water tank* 40º31′09″ 114º50′21″ 2850 Schell Creek White Pine Success Summit (S)* 39º15′24″ 114º40′59″ 2750 Schell Creek White Pine Success Summit (N) 39º15′52″ 114º41′18″ 2850 Schell Creek White Pine McCoy Creek* 39º22′30″ 114º33′39″ 2750 Schell Creek White Pine N. Fk. Muncy Canyon 39º37′11″ 114º37′24″ 2350 Schell Creek White Pine Fitzhugh Canyon 39º38′32″ 114º39′00″ 2800 Schell Creek White Pine McMaughn Creek* 39º50′59″ 114º39′54″ 2550 Snake White Pine Baker Cr. Campground / Trailhead–GBNP 38º58′38″ 114º14′37″ 2400 Snake White Pine Ohio Creek* 39º02′38″ 114º22′14″ 2500 Snake White Pine Cottonwood Springs / O’Neal Peak 39º25′32″ 114º17′23″ 2650 Deep Creek Juab, UT Middle Canyon 39º53′11″ 113º51′43″ 2550 Deep Creek Juab, UT Basin Creek / Scotts Basin 39º53′29″ 113º52′20″ 2250–2400 Stansbury Tooele, UT E. Hickman Creek / Morgan Canyon 40º24′55″ 112º34′08″ 2050–2200 Stansbury Tooele, UT S. Willow Creek 40º29′09″ 112º36′26″ 2250–2500 Oquirrh Tooele, UT Mercur Canyon (townsite) 40º19′ 112º12′ 2000

snowbrush (Ceanothus velutinus) largely re- rock outcrop without marmot scats. Among placed oceanspray and big sagebrush as the Great Basin ranges in Nevada, the Ruby/East dominant montane shrub around occupied sites. Humboldt Range has the largest area of mon- In addition, the few marmots that I found on tane shrub cover (Table 2) and an abundance alpine plateaus and talus fields above 3300 m of rocky outcrops, talus, and glacial moraines elevation usually occupied rocky outcrops lack- surrounded by open montane/subalpine mead- ing sizable shrubs. Vegetation associated with ows (Charlet 1996, NVGAP 1996). marmots in this study roughly corresponds Marmots were common throughout the with the montane shrub zone category of the Desatoya Range, though not as abundant as in Nevada Gap Analysis Program (Charlet 1996, the Ruby/East Humboldt Range. The Desatoya NVGAP 1996). is the 2nd smallest mountain range of those surveyed (Table 2) but has 3 well-watered drain- Distribution ages with much marmot habitat (Smith, Topia, Marmots were abundant throughout the and Campbell Creeks). Marmots were common Ruby/East Humboldt Range, where multiple in the Shoshone, Toiyabe, Toquima, Cherry individuals were observed in almost every Creek, Schell Creek, Deep Creek, and Stans- drainage surveyed and it was rare to find a bury Ranges. Each of these ranges had at least 476 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 2. Rocky outcrops and talus studded with oceanspray (Holodiscus discolor) and surrounded by meadows Fig. 3. Talus field ringed with oceanspray (Holodiscus dominated by mountain sagebrush (Artemesia tridentata discolor) above Cottonwood Springs Canyon, below O’Neal vaseyana), at ca. 2600 m elevation in Sadler Canyon in the Peak, at ca. 2650 m elevation in the Snake Range, NV. Diamond Range, NV. shrub cover comprises just 8.5% of the range; 2 drainages with 15 or more marmot burrows. only the Clan Alpine range has proportionally Marmots seemed to be uncommon in the Moni- less montane shrub cover (Table 2). I found tor Range, where I found burrows in only 2 only a single drainage with extensive marmot drainages, each of which appeared to have habitat, in the northern half of the range in the fewer than 10 marmots. basin above Cottonwood Springs (Table 1), Marmots were rare in the Clan Alpine, where there are several hectares of limestone Roberts, and Snake Ranges. I found evidence outcrops and talus fields (Fig 3). I observed of no more than 10 marmots in each range and only 1 marmot there, though dried scats were believe that total abundance may be less than abundant among some of the outcroppings. 20 individuals in each, because I searched Marmots appeared to be equally rare in the almost every open area to be found. The rarity southern half of the Snake Range, mostly en- of marmots in the Clan Alpine and Roberts compassed by Great Basin National Park. I Ranges may be a consequence of each having found only 3 marmots in the park, all living in only a single drainage containing substantial the roadbed of the Baker Creek road, near the marmot habitat. Furthermore, these ranges Baker Creek campground and trailhead. support relatively little montane shrub cover Extensive searches of other drainages in and among the surveyed ranges (Table 2). Com- around the park turned up only old, weath- pared to the nearby (smaller) Desatoya Range, ered scats in a single location (Ohio Creek; the Clan Alpine Range has less than half the Table 1), supporting the opinion of park offi- area of montane shrub cover (Table 2). cials that the Baker Creek and nearby Lehman Although the Snake Range is one of the Creek campgrounds are probably the only largest ranges in the Great Basin, montane places in the park that support marmots (K. 2004] DISTRIBUTION AND HABITATS OF MARMOTS 477

Fig. 4. Rock outcrops embedded in meadows with mountain sagebrush (Artemesia tridentata vaseyana), oceanspray (Holodiscus discolor), and Great Basin wild rye (Leymus cinereus) in Antone Creek at ca. 2600 m elevation in the Toquima Range.

Heister personal communication). Surprising- 2002) revealed only weathered scats on a small ly, there was no sign of marmots in the large cluster of rock outcrops and talus comprising patches of subalpine meadows at the base of what appeared to be the only suitable marmot the moraines above the park’s high-elevation habitat there. During 2003, in a valley in the lakes (>3200 m), habitat that often supports northern part of the Diamond Range (Table 1), marmots in other parts of their range (Frase I found another burrow site with abundant and Hoffman 1980). Hall (1981) reported a marmot scat, all of which looked ≤2 years old. marmot at 3700 m elevation above Treasure A local rancher, who worked as a trapper in Lake, but he was unable to find marmots any- the Diamond Range for 40 years, related that where else in the Snake Range (E.R Hall un- he had never observed marmots there (G. Par- published field notes). The fact that Hall ex- man personal communication). tensively trapped the Baker Creek drainage Hall (1946) documented the only marmots and, except for the Treasure Lake sighting and ever reported for the White Pine Range, where a single unconfirmed account of a marmot in he collected 2 individuals and found plentiful the Hendry’s Creek drainage, reported no evi- marmot scats in the abandoned mining town dence of marmots anywhere in the range, sug- of Hamilton (E.R. Hall unpublished field notes). gests that marmots have been rare in the Snake I searched this same area, including some of Range for most of the last century. A local the same rock piles reported by Hall, and found rancher whose family has grazed cattle in the no marmots or scat, which is noteworthy be- Snake Range for several decades affirmed that cause of the presence of numerous rocky out- the Baker Creek campground was the only crops and talus studded with oceanspray. A location where he had seen M. flaviventris (D. search of the southern half of the White Pine Eldridge personal communication). Range around Duckwater Peak, the highest, I was unable to find marmots in the Dia- wettest, and most ecologically diverse portion mond, White Pine, Egan, Spruce-Pequop, and of the range, revealed no sign of marmots and Oquirrh Ranges, although I located old scats very little suitable habitat, except for a small in all ranges except the White Pine and Oquirrh. patch of meadows and talus at the base of the Grayson and Livingston (1993) reported the east slope of Current Mountain. 1st known record of M. flaviventris in the Dia- I can make only cautious generalizations mond Range, in Sadler Canyon. My own exten- about marmots in the Egan, Spruce-Pequop, sive explorations of this canyon (1999–2000, and Oquirrh Ranges because my investigations 478 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 5. Distribution of elevations (m above sea level) of 62 locations where yellow-bellied marmots (Marmota flaviventris) or marmot scats were found on 18 Great Basin mountain ranges during 1999–2002. there were not as extensive as elsewhere. The had not been found in the Oquirrh Range as Egan Range is rich in montane shrub habitat of 1955 but suspected that marmots were (Table 2), but the combination of montane there undetected. I found no marmots there as meadows and suitable rock formations appears well, but apparently they are locally common to be rare. I explored only the Spruce Moun- near the ghost town of Mercur, in upper Mer- tain portion (the highest, least arid part) of the cur Valley, and are considered a pest because Spruce-Pequop Range and found no sign of they chew through pipes and tailing pond lin- marmots except for a few weathered scats on a ers (T. Becker personal communication). Min- dry ridge in the southeast portion of Spruce ing activity in the Oquirrh Range may have Mountain. Marmots were not observed in the created marmot habitat by clearing woodlands Spruce-Pequop Range during Brown’s (1971) and heavy brush and by producing extensive study (J. Brown personal communication). fields of artificial talus. Most of the Oquirrh Range has been subject to open-pit mining and is off limits to the DISCUSSION public, preventing a thorough search. Oquirrh is reputedly a Goshute or Paiute word mean- Global climate change models predict that ing “wooded mountain” or “brush mountain” montane and subalpine vegetation zones in (B. Bosworth personal communication), and the Great Basin will creep upslope, reducing the latter is an accurate description of most of habitat area and forcing the resident faunas to the montane zone, where almost all rocky out- shift their distributions accordingly (McDon- crops and talus that I found were covered in ald and Brown 1992, Fleishman et al. 2001). tall (≥2 m) shrubs. Durrant (1952) and Durrant Because of their habitat requirements, how- et al. (1955) commented that M. flaviventris ever, yellow-bellied marmots may be unable to 2004] DISTRIBUTION AND HABITATS OF MARMOTS 479

TABLE 2. Size of 14 Great Basin mountain ranges and amount of montane shrub cover per range, as estimated by NVGAP (1996). Data on ranges in Utah were not available.

______Montane shrub cover Total mountain Absolute Relative Mountain range area (km2) area (km2) area (km2) Cherry Creek 559 150.1 26.9 Clan Alpine 1149 49.7 4.3 Desatoya 705 106.9 15.2 Diamond 760 189.6 24.9 Egan 2295 296.1 12.9 Monitor 3046 604.9 19.9 Roberts 448 93.5 20.9 Ruby / E. Humboldt 2708 935.6 34.6 Schell Creek 2429 329.8 13.6 Shoshone 1852 179.5 9.7 Snake 2109 178.8 8.5 Toiyabe 3126 855.3 27.4 Toquima 1753 409.6 23.4 White Pine 1523 381.4 25.0

adapt by merely shifting upslope. The rock for- desert basin. In addition, radiotelemetry stud- mations that marmots prefer are stationary, ies in the southern Rocky Mountains have and in many ranges they become less common found that marmots will disperse as far as 15 with increasing elevation above the montane km (Van Vuren and Armitage 1994), which is zone (CHF personal observation). In addition, on par with the distances separating several of over the last century pinyon-juniper wood- the mountain ranges in this study. Low-eleva- lands have drastically expanded their range in tion sightings of marmots in this study, cou- the Great Basin, both up- and downslope pled with recent sightings of bushy-tailed (Tausch et al. 1981, Miller and Wigand 1994, woodrats (N. cinerea) well below the montane Wall et al. 2001). Presently occupied talus and zone (Grayson et al. 1996, Grayson and Mad- rock outcrops may become enclosed by pinyon- sen 2000), suggest that some boreal species juniper woodland, precluding the suitability of are not as restricted to high elevations as pre- these areas for marmots. In fact, many of the viously assumed (Brown 1971), supporting lower-elevation rock formations presently sur- Lawlor’s (1998) argument that desert lowlands rounded by pinyon-juniper woodlands likely separating Great Basin mountain ranges may were embedded in montane meadows during not be impermeable barriers to dispersal. cooler portions of the Holocene (Grayson 1993). The apparent present-day absence of mar- Drastic changes in species composition of the mots in the Diamond, White Pine, Egan, and herbaceous understory are also predicted (Gott- Spruce-Pequop Ranges, in combination with fried et al. 1999, Bartlein et al. 2003). evidence of past occurrences (from historical Nonetheless, several observations of M. fla- records or old scats), suggests 1 of 2 possibili- viventris at low elevations suggest that mar- ties: (1) that there are undetected marmots mots may survive climatic change better than persisting in some uninvestigated part of the mountain-dwelling mammals that are more range, or (2) extinctions have occurred in recent strictly limited to the montane/alpine zones times. The 1st possibility cannot be ruled out, (e.g., ermines [Mustela erminea] and pikas as demonstrated by evidence that marmots do [Ochotona princeps]; Beever et al. 2003). Four indeed occur in the Oquirrh Range, after I had marmot sightings in my study were below searched there and found none. However, the 2000 m elevation (Table 1), and the lowest was fact that my extensive investigation of Sadler at 1560 m near Eastgate, Nevada, where Hall Canyon over multiple summers revealed only (1946) also found marmots in 1938. Local resi- desiccated scats, and the fact that I found no dents informed me that marmots at Eastgate marmot sign at all in the White Pine Range, in- forage in rock-bordered pastures that extend cluding some of the same rock formations that from the base of the Desatoya Range into the were occupied in 1938, indicate that marmots 480 WESTERN NORTH AMERICAN NATURALIST [Volume 64 have disappeared from a substantial portion of field notes, and D. Charlet and K. Taylor for the Diamond Range and perhaps all of the assisting with plant identification. T. Lawlor, White Pine Range. These “extinctions” may D.H. Van Vuren, and an anonymous reviewer be part of a natural extinction-recolonization provided very helpful comments on previous dynamic, as was found to occur among yellow- versions of the manuscript. bellied marmot colonies in the Rocky Moun- tains of Colorado, where sites periodically LITERATURE CITED became vacant when the previous years’ residents died and were not replaced by immigra- BARTLEIN, P.J., C. WHITLOCK, AND S.L. SHAFER. 2003. Future climate in the Yellowstone National Park re- tion; some colonies there went extinct more gion and its potential impact on vegetation. Conser- frequently than others, while others have per- vation Biology 11:782–792. sisted for more than 40 years (D.H. Van Vuren BATCHELET, E. 1981. Circular statistics in biology. Acade- and K.B. Armitage personal communication). mic Press, New York. 371 pp. The most extinct-prone sites in the Colorado BEEVER, E.A., P.F. BRUSSARD, AND J. BERGER. 2003. Pat- terns of apparent extirpation among isolated popula- system generally were thought to be suboptimal tions of pikas (Ochotona princeps) in the Great because of higher predation or hibernation Basin. Journal of Mammalogy 84:37–54. mortality rates (DHVV personal communica- BROWN, J.H. 1971. Mammals on mountaintops: nonequi- tion). Habitat in the Diamond and White Pine librium insular biogeography. American Naturalist 105:467–478. Ranges (and other ranges where marmots are ______. 1978. The theory of insular biogeography and the rare) may be suboptimal and thus only occa- distribution of boreal birds and mammals. Great sionally inhabited. Basin Naturalist Memoirs 2:209–227. The altitudinal flexibility of M. flaviventris BROWN, J.H., AND A. KODRIC-BROWN. 1977. Turnover rates may bode well for the species in the presum- in insular biogeography: effect of immigration on extinction. Ecology 58:445–449. ably warmer decades to come, because if arid CHARLET, D.A. 1996. Atlas of Nevada conifers: a phyto- lowlands can be crossed, then declining popu- geographic reference. University of Nevada Press, lations can be rescued (Brown and Kodric- Reno. 320 pp. Brown 1977) and extinct sites recolonized. DURRANT, S.D. 1952. Mammals of Utah: taxonomy and However, the probability of successful cross- distribution. University of Kansas, Museum of Nat- ural History 6:1–549. basin dispersal may decrease as low-elevation DURRANT, S.D., M.R. LEE, AND R.M. HANSEN. 1955. Addi- environments become more arid, eventually tional records and extensions of known ranges of transforming the Great Basin mountain ranges mammals from Utah. University of Kansas, Museum into the strictly isolated islands that Brown of Natural History 9:69–80. FLEISHMAN, E., G.T. AUSTIN, AND D.D. MURPHY. 2001. (1971) originally hypothesized them to be. Biogeography of Great Basin butterflies: revisiting patterns, paradigms, and climate change scenarios. ACKNOWLEDGMENTS Biological Journal of the Linnean Society 74:501–515. FLOYD, C.H. 2003. Ecological genetics of dispersal and This work was supported by funds from the mating system in populations of yellow bellied mar- American Museum of Natural History, Ameri- mots (Marmota flaviventris). Doctoral dissertation, University of California, Davis. can Society of Mammalogists, Sigma Xi, and FRASE, B.A., AND R.S. HOFFMANN. 1980. Marmota flavi- UC Davis Jastro Shield Research Awards. My ventris. Mammalian Species 135:1–8. sincere appreciation to R. Miller, T. Platt, S. GOTTFRIED, M., H. PAULI, K. REITER, AND G. GRABHERR. Coppeto, and D.H. Van Vuren for assisting me 1999. A fine-scaled predictive model for changes in species distribution patterns of high mountain plants in the field, and to the following individuals induced by climate warming. Diversity and Distri- who provided valuable information or other butions 5:241–251. assistance: C. Baughman, L. Crabtree, J. Casey, GRAYSON, D.K. 1993. The desert’s past: a natural prehis- K. Heister, K.B. Armitage, G. Parman, E. Beever, tory of the Great Basin. Smithsonian Institution Press, A. Wilcox, T. Becker, J. Brown, D. Stoner, B. Washington, DC. 356 pp. GRAYSON, D.K., AND S.D. LIVINGSTON. 1993. Missing mam- Bosworth, V. Bakker, T. Lawlor, P. Bradley, as mals on Great Basin mountains: Holocene extinc- well as the Nevada Department of Wildlife, tions and inadequate knowledge. Conservation Biol- Utah Division of Wildlife Resources, USDA ogy 7:527–432. National Forest Service, National Park Ser- GRAYSON, D.K., S.D. LIVINGSTON, E. RICKART, AND M.W. SHAVER III. 1996. Biogeographical significance of vice, and the Bureau of Land Management. I low-elevation records for Neotoma cinerea from the thank C. Conroy and the UC Berkeley Museum northern Bonneville Basin, Utah. Great Basin Natu- of Vertebrate Zoology for access to E.R. Hall’s ralist 56:191–196. 2004] DISTRIBUTION AND HABITATS OF MARMOTS 481

GRAYSON, D.K., AND D.B. MADSEN. 2000. Biogeographic NVGAP. 1996. Nevada Gap GIS database. Utah Coopera- implications of recent low-elevation recolonization tive Fish and Wildlife Research Unit, Utah State by Neotoma cinerea in the Great Basin. Journal of University, Logan. (http://www.brrc.unr.edu/mtn/ Mammalogy 81: 1100–1105. html/ref/gapref.html) HALL, E.R. 1946. Mammals of Nevada. University of Cali- RICKART, E.A. 2001. Elevational diversity gradients, bio- fornia Press, Berkeley. 710 pp. geography and the structure of montane mammal ______. 1981. The mammals of North America. Volume 1. communities in the intermountain region of North 2nd edition. John Wiley and Sons, New York. 606 + America. Global Ecology and Biogeography 10: 90 pp. 77–100. HILL, J.K., C.D. THOMAS, R. FOX, M.G. TELFER, S.G. STALLMAN E.L., AND W. G. H OLMES. 2002. Selective for- WILLIS, J. ASHER, AND B. HUNTLEY. 2002. Responses aging and food distribution of high elevation yellow- of butterflies to twentieth century climate warming: bellied marmots (Marmota flaviventris). Journal of implications for future ranges. Proceedings of the Mammalogy 83:576–584. Royal Society of London Series B 269:2163–2171. SVENDSEN, G.E. 1974. Behavioral and environmental fac- KODRIC-BROWN, A., AND J.H. BROWN. 1993. Incomplete tors in the spatial distribution and population dynam- data sets in community ecology and biogeography: a ics of a yellow-bellied marmot population. Ecology cautionary tale. Ecological Applications 3:736–742. 55:760–771. KULLMAN, L. 2002. Rapid recent range-margin rise of tree TAUSCH, R.J., N.E. WEST, AND A.A. NABI. 1981. Tree age and shrub species in the Swedish Scandes. Journal and dominance patterns in Great Basin pinyon- of Ecology 90:68–77. juniper woodlands. Journal of Range Management LAWLOR, T.E. 1998. Biogeography of Great Basin mammals: 34:259–264. paradigm lost? Journal of Mammalogy 79:1111–1130. VAN VUREN, D., AND K.B. ARMITAGE. 1994. Survival of dis- MCDONALD, K.A., AND J.H. BROWN. 1992. Using montane persing and philopatric yellow-bellied marmots: what mammals to model extinctions due to global change. is the cost of dispersal? Oikos 69:179–181. Conservation Biology 6:409–415. WALL, T.G., R.F. MILLER, AND T.J. S VEJCAR. 2001. Juniper MILLER, R.F., AND P. E . W IGAND. 1994. Holocene changes encroachment into aspen in the northwest Great in semiarid pinyon-juniper woodlands. BioScience Basin. Journal of Range Management 54:691–698. 44:465–474. MOZINGO, H. 1987. Shrubs of the Great Basin: a natural Received 1 April 2003 history. University of Nevada Press, Reno. 342 pp. Accepted 15 December 2003 Western North American Naturalist 64(4), © 2004, pp 482–491

DYNAMICS OF A DWARF BEAR-POPPY (ARCTOMECON HUMILIS) POPULATION OVER A SIXTEEN-YEAR PERIOD

K.T. Harper1,2 and Renée Van Buren3

ABSTRACT.—A population of the dwarf bear-poppy (Arctomecon humilis Coville, Papaveraceae) at Red Bluff, Wash- ington County, Utah, was monitored twice annually between 1987 and 2002. This is a narrowly endemic, gypsophilous species that has been formally listed as endangered since 1979. During the 16 years of observation, density of this species has fluctuated between 3 and 1336 individuals on the 0.07-ha monitoring plot. Moderate to large recruitments of seedlings occurred in 1992, 1995, and 2001. Seedling recruitments from a large, long-lived seed bank are triggered by abundant precipitation during the February–April period. At least 5.0 cm of rainfall is required during that interval to produce any seedlings. Seedlings experienced considerable mortality in the 1st few months of life in all observed cases. The average seedling initiated in the very large recruitment event of 1992 survived for only 2.6 years. Seedlings in that cohort that were alive 1 year after germination had an average longevity of 4.6 years. None of the seedlings that emerged in 1992 were still alive in October 2002. Mortality in this species was poorly correlated with fluctuations in precipitation or temperature. No epidemics of parasites or herbivores were observed. Mortality in the species appears to be caused by a variety of factors acting over a cohort’s lifetime.

Key words: Arctomecon humilis, demography, longevity, germination, seed bank.

Arctomecon humilis Coville is a moderately 1st few hours after anthesis. The abundant short-lived perennial herb endemic to gypsif- pollen is quickly shed or collected by foraging erous soils. It occurs only on soils derived from bees, but stigmas may remain receptive to the Moenkopi Formation, a deep-water marine pollen for another 48 hours. Petals begin to deposit of lower Triassic age. The formation is wither within 72 hours but remain attached at widely exposed in Washington County, extreme the base of developing fruits and assist in seed southwestern Utah. The species is a low-grow- dispersal. Individual plants may produce as ing, evergreen plant that exists as light green many as 400 flowers in a single season (Nelson rosettes during the nonflowering season. The 1989, Nelson and Welsh 1993), but the aver- plants form low (<5 cm high) mounds consist- age plant produces only a few dozen flowers. ing of dead, persistent leaves and 1 to many Dwarf bear-poppy are restricted to habitats short, leafy stem-tips arising from a branched that are relatively barren of other herbaceous caudex. Caudexes top an unbranched, deep plants. The substrate weathers to form gently (8–15 dm) taproot. Leaves bear a sparse to rolling mounds that are smooth, rock-free, and moderately dense vesture of hirtellous to long only sparsely covered by shrubs that might (<5 mm), pilose hairs. Leaves are cuneate to interfere with vehicular traffic. The few shrubs obovate and tipped by 3 or 4 triangular teeth. that do occur belong primarily to the species Each tooth is terminated by an acerose hair. Lepidium fremontii (Fremont’s pepperplant), The fancied similarity of such leaves to the Atriplex confertifolia (shadscale), Ephedra tor- foot of a bear is the source for the common reyana (Torrey’s ephedra), Hymenoclea salsola name (bearclaw-poppy) for this species. (burrowbrush), and Lycium andersonii (Ander- At flowering, few-flowered (1–5 flowers) son’s lycium). Herbs are uncommon on this inflorescences rise to the height of 1.5–2.5 dm parent material, but lichens, mosses, and cyano- and produce white-petaled flowers that may bacteria do form heavy covers (25%–100%) on be up to 5.0 cm in diameter. Numerous bright sites where the poppy grows. This nonvascular yellow stamens surround the pistil during the plant cover probably contributes significantly

1Professor Emeritus, Integrative Biology, Brigham Young University, Provo, UT 84602. 2Botanical Scholar in Residence, Utah Valley State College, Orem, UT 84058. 3Department of Biology, Utah Valley State College, Orem, UT 84058.

482 2004] ARCTOMECON HUMILIS DEMOGRAPHY 483 to both erosion control and the availability of that station (National Oceanographic and biologically usable nitrogen (Nelson and Harper Atmospheric Administration 2000). The mon- 1991, Evans and Lange 2001). itoring site occurs on land managed by the Arctomecon humilis, like its congener, A. Bureau of Land Management (BLM), U.S. De- californica, exhibits large year-to-year fluctua- partment of the Interior. tions in population density (Nelson and Welsh The site was closed to motorized off-road 1993). A large and persistent soil seed bank of vehicles in 1988, but access to the site by non- long-lived seeds is maintained by both species motorized vehicles has continued to be legal (Meyer 1987, Nelson 1989). Since climatic con- to the present time. Bike trails are marked and ditions conducive to germination of the seeds managers have attempted to confine traffic to of these species often occur at intervals longer marked trails. That effort has been largely suc- than longevity of most individuals in their cessful. populations, dramatic cycles in abundance are The site was partially fenced in the mid- common (Nelson and Welsh 1993). 1980s in a joint action by BLM and the Nature Arctomecon humilis was formally listed as Conservancy. The BLM completed a fence endangered in 1979 under provisions of the around the entire Red Bluff population in 2002. Endangered Species Act of 1973. Among the An initial monitoring site was established in reasons cited for listing the species were (1) its 1987 by Nelson (1989). She placed a perma- restriction to a single geological formation, (2) nent grid over an area of 2.1 ha. At that time the threats to the species’ habitat from rapid the area supported only 400 living plants of urbanization, and (3) the uncontrolled use of the dwarf bear-claw poppy. The grid consisted the habitat by off-road vehicles (U.S. Fish and of 100-m2 units (0.01 ha, each 10 m × 10 m). The Wildlife Service 1979). grid units were marked at their corners by The recovery plan for this species recom- steel bars (4.5 dm long and 0.95 cm in diame- mended that demographic studies be initiated ter). Individual plants were marked at their to help managers separate changes in popula- base by a 2.5-cm-diameter aluminum disk, each tion density due to variations in climate from uniquely numbered. Each disk was tied to a declines caused by human activities (Ander- zinc-coated nail, 8 cm in length, with aluminum son and England 1985). This paper presents wire. To minimize vandalism on markers, when the results of monitoring the density of the numbers were not being read, we folded the species between 1987 and 2002 at the Red tags parallel to the anchoring nail and pushed Bluff site near Bloomington, Utah. both out of sight into the soil. Metal detectors were used to relocate marker tags. STUDY SITE AND METHODS The monitoring site was visited in May and late September or early October each year. At Population dynamics of the dwarf bear- each visit we recorded rosette diameter of each poppy are based on long-term observations of plant as well as general health of the plant and a population at a site northwest of Blooming- any evidence of recent dieback in the rosette. ton, Washington County, Utah. The study plot When the visit was late enough in May, the is in the northwestern quarter of section 9, number of flowers initiated and whether polli- Township 43 South, Range 16 West. Elevation nation had resulted in formation of a capsule at the site varies from about 845 m to 855 m were noted. When the spring visit was too early above sea level (~2770–2805 feet). Geologic to provide a realistic estimate of total flowers parent material at the site is the Shinabkaib produced by plants that year, we nevertheless Member of the Moenkopi Formation, a fine- noted whether a plant would flower that year textured, sedimentary deposit (Cook 1960). Soils as indicated by the presence of buds. at the study site have a pH of 7.2–7.5 and an Unusually favorable conditions for germi- average gypsum content of 39.3% by weight nation of seedlings occurred in late winter and (Nelson and Harper 1991). Average precipita- early spring of 1992. Such a large crop of seed- tion is reported as 207 mm (8.16 inches) per lings had emerged by early May 1992 that we year at the St. George weather station, which found it impossible to monitor all seedlings on is located approximately 5.0 km (3.0 miles) Nelson’s original grid network. Since we felt it from the study site. Annual mean monthly expedient that a sizable cohort of seedlings be temperature is reported as 16.8°C (62.3°F) for tracked from emergence to death, we limited 484 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Population dynamics of the dwarf bearclaw-poppy on a 0.07-ha monitoring plot near Bloomington, Washing- ton County, UT. Number of living plants was determined in May and late September or early October of the year of concern. Mortality is based on deaths among established plants and seedlings discovered on the plot during the 12-month period preceding the mid-October date of the year of concern. The number of recruits still living when plants were per- manently tagged in 1992 is noted in parentheses. Living plants in Recruits during % mortality Year mid-October year during year 1987 13 0 1988 9 4 47 1989 7 1 30 1990 4 0 43 1991 3 1 40 1992 400 1333 (398) 70 1993 195 36 55 1994 159 1 19 1995 175 132 40 1996 122 0 30 1997 97 0 20 1998 86 20 27 1999 81 7 13 2000 61 0 25 2001 17 107 87 2002 5 0 71

our observation to 0.07 ha (7 units of 0.01 ha area lished plant density during the 16-year period. each). In May 1992 we marked 1333 seedlings Numbers were very low from 1988 through on that area with wooden toothpicks. By late early 1992. The recruitment pulse of 1992 de- September 1992 only 398 individuals survived clined steadily from that date to late 2000. on the 0.07-ha site. Those seedlings were per- Severe drought in that and subsequent years manently marked as well as all new seedlings (2001 and 2002) precipitously reduced the that subsequently emerged on the plot over population to very low numbers again. the period 1992–2002. The report that follows is based primarily upon the seedling cohort Climatic Conditions marked in September 1992. Precipitation and temperature at the St. George weather station are reported in Table RESULTS 2. Water-year (1 Oct.–30 Sept.) precipitation varied widely over the 16 years of record, Plant Density ranging from a high of 331 mm per year in Since the size of the monitoring plot was 1995 to a low of 53 mm in water-year 2002. changed in 1992 to deal with the large num- Late winter–early spring (Feb.–April) precipi- ber of seedlings that emerged in that year, we tation is well correlated [r = + 0.685, P = have gone back to raw field data to determine 0.0034; Y = –72.6 + 16.5 X, where Y repre- the number of established plants and recruits sents number of new seedlings surviving to that occurred on that smaller monitoring plot October and X represents precipitation (re- in years prior to 1992. Data in Table 1 extend ported in tenths of cm) falling at St. George back to the initiation of the monitoring plot in during the period 1 Feb.–30 Apr. of the year of 1987. Results show that this population has concern] with seedling recruits that survived experienced only a single large recruitment the growing season. This seasonal rainfall is event (1992) during the 16 years of observa- also highly variable among years, ranging from tion. In 4 other years, at least 1 seedling per 5 m to 155 mm. Neither autumn (1 Oct.–31 100-m2 grid unit survived long enough to be Dec.) or total water-year precipitation nor mean tagged during our October inventory (1993, monthly temperature is significantly correla- 1995, 1998, and 2001). Nevertheless, the pop- ted with seedling recruitment in the dwarf ulation has shown but a single “hump” in estab- bearclaw-poppy. 2004] ARCTOMECON HUMILIS DEMOGRAPHY 485

TABLE 2. Climatic conditions recorded at the St. George weather station for each year of concern for this study. Annual precipitation is reported for the “water-year” (October 1 of previous year through September 30 of year of concern), since that period seems more likely to influence plant performance than calendar-year precipitation and temperature. All data are from the National Climatic Data Center, Asheville, NC. Water-year Feb.–April Mean monthly precipitation precipitation temperature (°C) Year (mm) (mm) (water-year) 1987 189 54 17.3 1988 290 87 17.0 1989 153 31 17.6 1990 121 52 17.6 1991 116 55 16.3 1992 225 155 16.4 1993 319 92 17.2 1994 197 94 18.6 1995 331 131 17.8 1996 114 34 19.3 1997 327 25 18.1 1998 311 131 17.2 1999 214 37 17.7 2000 130 64 19.4 2001 191 88 18.3 2002 53 5 16.5

Percent mortality in this population is not sig- Reproductive Events nificantly correlated with any environmental Flowering in this cohort began during the variable considered (Tables 1, 2). Mortality in 2nd growing season of life, when 23.5% of the this species varies greatly from year to year individuals produced at least 1 flower. Plants (13%–87% of plants alive at the end of the pre- that did flower produced, on average, slightly ceding growing season), but it appears to be less that 4 flowers per individual (Fig. 2). Dur- dependent on a variety of factors rather than 1 ing the 3rd growing season, about 62% of the or 2 triggering conditions. We observed no living members of the 1992 cohort flowered, evidence of epidemic outbreaks of either insect producing slightly more than 23 flowers per or microbial pathogens or herbivores of any plant. In the next growing season, almost all type during the course of study. However, lar- members of the cohort flowered. The average vae of an unknown butterfly species did destroy plant produced slightly over 34 flowers during many flower buds during spring 1990. this 4th growing season. Growth Rates The monitoring population was not always in full flower at the time of the spring inven- Growth of the individual plants in the 1992 tory; thus, the number of flowers per individ- cohort is reported in Figure 1. Interestingly, ual was not determined in some years. The average plant size for this cohort reached a number of flowers per plant was not determined maximum value at the end of the 7th growing for survivors of the 1992 cohort in 1996, 1997, season and declined thereafter. Plant diameter 1999, and 2000. The 1998 season was cool, and averaged about 22 cm at the end of the 7th the average number of flowers per individual growing season. The few plants remaining alive reported in Figure 2 (about 19) is an underesti- (4) at the end of their 10th growing season had mate of the total flowers produced per plant, an average diameter of slightly less than 12 cm since we sampled well before the plants reached (Fig. 1), but variation in size among surviving full flower. The number of flowers per individ- plants was great. No members of the 1992 ual reported for 2001 is probably a good esti- cohort were alive at the end of the 2002 grow- mate. Older plants vary greatly in size and ing season. The decline in average diameter of vigor, and both of those variables markedly plants during the last 3 years of life is attribut- impact flower production. able to progressive death of the several stem Pollination success (fruit set) was carefully tips produced by healthy plants. evaluated for the monitored population in the 486 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 1. Rosette diameter averages for all living members of the 1992 seedling cohort of A. humilis. Diameter values were taken annually in late September or early October. Bars represent average diameter of surviving members of the 1992 cohort in each year. The error bar atop each average diameter bar represents the standard error for that mean. The number of individual survivors represented in each annual mean is shown above the diameter bar. years 1997 and 2000 (Table 3). Since precipita- 82%) is typical of results obtained for 2 other tion in those years varied widely above (in 1997) populations of this species growing elsewhere and below (in 2000) the long-term average for in Washington County in year 2000. Seedfill the site, observed results probably bracket the results (84.5%–52.8%) also are similar to those long-term pollination and seedset success for obtained for the other 2 populations in 2000. this population. Our results published elsewhere We observed an average of 34 ovules per ovary (Harper et al. 2000) suggest that the high pol- in the 3 populations considered by Harper et lination success reported in Table 3 (~94%– al. (2000). If one assumes a worst case scenario 2004] ARCTOMECON HUMILIS DEMOGRAPHY 487

Fig. 2. Average number of flowers produced per flowering individual of the 1992 seedling cohort of A. humilis in a given year. The spring sample was sometimes too early to provide an acceptable estimate of the number of flowers produced per plant. for percentage pollination and seedfill (i.e., 82% bird granivores. Only 6.6% of the seeds pre- pollination and 52.8% of the ovules becoming sented could not be accounted for at the con- filled seeds per fruit) and the flowering per- clusion of the 2-week presentation period, centage and average number of flowers per age which coincided with natural seed dispersal in class reported above (Fig. 2 and text relative the Red Bluff population. Since the seeds thereto) for the 1992 cohort and their progeny, were displayed in glass saucers, some of the one would predict that just slightly over lost seeds could simply have been blown from 250,000 filled seeds would be produced over the dishes by recurrent winds that swept the the 10-year period spanning the lifetime of site during the study period. any individual belonging to the 1992 cohort. Ants appear to be the principal dispersal Although one must assume that some seed agent. Seeds of this species showed an average predation would occur in any population and weight of 2.0 mg based upon results from 7 some seeds would be expected to lose viability lots of 50 seeds. Each seed has a narrow, elon- while in the seed bank, such losses are appar- gate aril of white tissue running along the ently not large. Nelson and Harper (1987) pre- hilum edge of the seed. Seeds are jet black sented seeds of this species in a manner that and the seed coat is hard. Ants are attracted to made them readily available to rodent and the arils and nibble at them, but seeds are 488 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 3. Pollination and seedfill results for the Red occurred in the 1st growing season. Losses Bluff monitoring population of Arctomecon humilis in 1997 were also steeper during the 2nd year of life and 2000. than during subsequent years. From 1993 to ______Year 2002, when the last members of the 1992 seed- 1997 2000 ling cohort died, mortality progressed slowly and at a rather steady rate (Fig. 3). Percentage of average water-year precipitation 159.4 63.4 The average seedlings of the initial 1992 Pollination (%) 93.9 82.0 cohort had a longevity of only 2.6 years. Seed- Seedfill (%) 84.5 52.8 lings of that cohort that were still alive 1 year after germination had an average longevity of 4.6 years. eventually abandoned between the point of DISCUSSION discovery and the nest site. Thus, ants appear not to damage the viability of seeds and often Our study demonstrates the crucial role of deposit them below the soil surface and in the seed bank in persistence of bearclaw- places that are probably rarely discovered by poppy populations. Without a reserve of long- larger seed predators. lived seeds, the species would probably dis- Since 303 seedlings of the dwarf bearclaw- appear because of the long periods between poppy emerged during the period 1998–2001 climatic events that are capable of triggering and survived long enough to be located and significant recruitment. Data in Table 1 show tagged on 1 of the 2 annual inventory visits each that sizable recruitment events occurred in year, we know that the accumulating seed bank 1995 and 2001, but that those seedlings were was depleted of some seeds by natural germ- almost completely lost during their 1st year of ination events. Such depletion was almost cer- life. Survival of the species thus requires that tainly small, however. Assuming only 1% sur- many seeds remain dormant in the seed bank vival of new seedlings, only approximately even when conditions are suitable for germi- 30,000 seeds would have been lost to germina- nation. Fortunately, bearclaw-poppy seeds are tion. dormant at and, for an unknown duration (but Nelson and Harper (1987) used the tetra- probably more than 1 year), after dispersal, zolium test to evaluate viability of seeds from because of immature embryos (personal com- herbarium specimens and seeds extracted munication, Nancy Vivrette, Ranson Seed from the soil seed bank. They found that most Laboratory, Santa Barbara, CA). Nelson and filled seeds taken from a 10-year-old herbar- Harper (1987) observed that bearclaw-poppy ium specimen were still viable. They observed seeds from the seed bank showed great varia- a 64% viability of seeds from the seed banks at tion in embryo size: some were fully mature, the monitoring site. Assuming 250,000 seeds while most ranged in size from that observable to have been produced by plants of the 1992 in seeds ready for dispersal through all grada- cohort and their progeny over the 10-year tions to mature size. Accordingly, it seems period of life of any member of that cohort, likely that some seeds are dormant at the time one would predict a possible loss of 19,800 of any germination event. seeds to predators, 30,000 to germination, and Our data suggest that the 1992 cohort had 36% of the remaining seeds may have become rebuilt a large seed bank during its lifetime. nonviable while in the seed bank. Nevertheless, Even if it is assumed that the previous seed that would still leave slightly over 160,000 bank was totally depleted in the large germi- viable seeds in the seed bank. Since those nation event of 1992, by 2002 when the last seeds would be spread over 0.07 ha, there may member of that cohort died, the seed bank have been as many as 286 viable seeds per m2 should have been large even when conserva- in the soil of the monitoring site at the end of tive estimating criteria are used. the 2002 growing season. Nelson (1989) estimated over 800 seeds per 1.0 m2 in the seed bank extracted from 0.6-m- Plant Longevity radius circular plots centered on large, dead Survivorship for the 1992 cohort is shown poppy plants. Our estimate of 286 seeds per in Figure 3. Most mortality among that cohort 1.0 m2 at the Red Bluff monitoring site thus 2004] ARCTOMECON HUMILIS DEMOGRAPHY 489

Fig. 3. Survivorship curve for 1333 individual seedlings of A. humilis that appeared in early May 1992. Percentage survival values are based on individuals of the cohort that were alive in late September or early October of each year, except for the 100% value, which is based on individuals observed in early May 1992. No member of the cohort survived to October 2002. seems reasonable, especially since it is based interplant distances average 3.5 m and 12.0 m on total plot area, not just the area around (only about 119 and 10 plants per acre), flower mature plants. In 2000 we (Harper et al. 2000) pollination could be expected to decline to estimated that the average distance between about 90% and 64% and seedfill would proba- bearclaw-poppy plants at the Red Bluff site bly drop to about 50% and 35%, respectively. was 1.66 m. In light of such an interplay between popu- The recovery plan for this species (Ander- lation density and seed production, human son and England 1985) makes no recommen- disturbances that severely reduce poppy pop- dations concerning density of the plant on ulations below the naturally high densities sites where it occurs. Although there seems that exist for a few years after a major recruit- little that managers could do to increase popu- ment event are likely to drastically reduce seed lation density beyond that observed naturally production. Obviously, low population density after a major recruitment event, various kinds markedly reduces effectiveness of the insect of human-caused disturbance can reduce den- pollinators that service bearclaw-poppy. It sity of bearclaw-poppy populations far below seems likely that human disturbances that cre- the number of plants successfully recruited. ate dust could cause fouling of stigmatic sur- Therefore, following any large, episodic repro- faces and reduce seed set, but studies of such ductive event, managers must ask the question, relationships have not been done. How many plants per unit area are desirable? Large variations in density of the bearclaw- Harper et al. (2000) demonstrated that inter- poppy through time observed at our monitor- plant distance has a significant effect on pollina- ing site are apparently common for this species tion success and, especially, on the percentage as well as for the California bear-poppy (Meyer of ovules that become filled seeds. Their data 1987, Nelson and Welsh 1993). A sizable liter- showed that when interplant distances aver- ature suggests that such large differences in age 1.5 m (about 568 plants per acre), percent- population density that recur through time age flower pollination averaged slightly over have a strong influence on genetic characteris- 92% and seedfill averaged about 67%. When tics of a species (Ellstrand et al. 1978, Thomson 490 WESTERN NORTH AMERICAN NATURALIST [Volume 64

1981, Handel 1983, Schmitt 1983, Kunin 1997, bearclaw-poppy that ranged from small- to Harper et al. 2000). These and other references large-flowering individuals. It is not clear suggest that dense populations attract numerous from their data whether the size differences pollinators and receive such efficient pollina- reflected age or growth rate differences, but it tor services that fruit and seed set are greatly is possible that observed size differences among enhanced. Species that produce seeds that are a single seedling cohort are related to genetic long-lived in the soil (as the Arctomecon species differences. do) may thus build up large seed banks when their populations are dense. Ellstrand et al. ACKNOWLEDGMENTS (1978) suggest that seed banks produced by such local, dense populations are likely to be We are indebted to Deanna R. Nelson for deficient in heterozygotes. However, when original design of the monitoring program and populations are sparse, wide outcrosses can be for much of our knowledge about bearclaw- expected (Ellstrand et al. 1978, Handel 1983). poppy and its seed bank. Janet Cooper, Zachary Seeds entering the seed bank at such times Aanderud, Diane Carter, and Donna Barnes should be more genetically diverse at any have provided assistance with collection of field given location. Species such as Arctomecon monitoring data. Zachary Aanderud gave invalu- humilis, which experience large variations in able help in laboratory processing of seedfill population density and produce large, long- data in both 1997 and 2000. We are grateful to lived seed banks, should thus retain consider- those people for competent assistance. This able genetic diversity even though average work has been financed by the BLM primarily, population size may not be large. with consistent contributions in space and Our results (Harper et al. 2000) for Arc- supplies from both BYU and UVSC. tomecon humilis reproductive success at local density values, which varied from high to very LITERATURE CITED sparse, support Kunin’s (1997) conclusion that ALLPHIN, L., M.D. WINDHAM, AND K.T. HARPER. 1998. population density (number of individuals per Genetic diversity and gene flow in the endangered unit area) is a more important variable than dwarf bear-poppy, Arctomecon humilis (Papaveraceae). population size per se in plant conservation American Journal of Botany 85:1251–1261. biology. Population size (number of individu- ANDERSON, J., AND J.L. ENGLAND. 1985. Dwarf bear-claw poppy, Arctomecon humilis Coville: recovery plan. als in a population) is a prominent variable in Prepared by U.S. Fish and Wildlife Service, Region 6, all considerations of rare species management Denver, CO. 26 pp. and probability of local extinction (Ricklefs COOK, E.F. 1960. Geologic atlas of Utah: Washington 1997), while population density is rarely men- County. Utah Geological and Mineral Survey Bul- tioned. Kunin’s (1997) results and those of letin 70. 119 pp. ELLSTRAND, N.C., A.M. TORRES, AND D.A. LEVIN. 1978. Harper et al. (2000) demonstrate that popula- Density and the rate of apparent outcrossing in tion density deserves more attention. Helianthus annuus (Asteraceae). Systemic Botany Finally, we call attention to the large size 3:403–407. difference observed among seedlings of the EVANS, R.D., AND O.L. LANGE. 2001. Biological soil crusts and ecosystem nitrogen and carbon dynamics. Pages 1992 cohort. We have elsewhere documented 263–279 in J. Belnap and O.L. Lange, editors, Bio- (Harper and Van Buren 1996) that size differ- logical soil crusts: structure, function, and manage- ences persisted for at least 4 years and had ment. Ecological Studies Series 150. Springer-Ver- large impacts on mortality rates (i.e., greater lag, Berlin. HANDEL, S.N. 1983. Pollination ecology, plant population mortality among the smaller seedlings) among structure and gene flow. Pages 163–211 in L. Real, the several dozen seedlings initially assigned editor, Pollination biology. Academic Press, Inc., to each of the 2 categories in October 1992. London. Also, fewer small seedlings flowered each year HARPER, K.T., AND R. VAN BUREN. 1996. Arctomecon humilis Coville. Monitoring report for 1996. Report to during the first 4 years of life, and the smaller Endangered Plant Specialist, Bureau of Land Man- plants consistently produced fewer flowers agement, U.S. Department of the Interior, Richfield, than larger members of their cohort. Meyer UT. 11 pp. (1987) observed similar persistent differences HARPER, K.T., R. VAN BUREN, AND Z.T. AANDERUD. 2000. in size of seedlings of A. californica. Allphin et The influence of interplant distance and the number of flowers on seed set in dwarf bear-poppy (Arctome- al. (1998) reported progressively more het- con humilis). Pages 105–109 in J. Maschinski and L. erozygotes in samples of plants of the dwarf Holter, editors, Southwestern rare and endangered 2004] ARCTOMECON HUMILIS DEMOGRAPHY 491

plants: proceedings of the third conference. U.S. De- poppy. Report to the Utah Native Plant Society. 134 partment of Agriculture, Forest Service, Rocky Moun- pp. tain Research Station, Proceedings RMRS-P-23. ______. 1991. Site characteristics and habitat requirements KUNIN, W.E. 1997. Population size and density effect in of the endangered dwarf bear-claw poppy (Arctome- pollination: pollinator foraging and plant reproduc- con humilis Coville, Papaveraceae). Great Basin Nat- tive success in experimental arrays of Brassica kaber. uralist 51:167–175. Journal of Ecology 85:225–234. NELSON, D.R., AND S.L. WELSH. 1993. Taxonomic revision MEYER, S.E. 1987. Life history of Arctomecon californica, of Arctomecon Torr. and Frem. Rhodora 95:197–213. a Mohave Desert endemic perennial herb. Unpub- RICKLEFS, R.E. 1997. The economy of nature. 4th edition. lished manuscript, available from author, Shrub Sci- W.H. Freeman and Co., New York. ences Laboratory, U.S. Forest Service, Provo, UT. SCHMITT, J. 1983. Density-dependent pollinator foraging, NATIONAL OCEANOGRAPHIC AND ATMOSPHERIC ADMINIS- flowering phenology, and temporal pollen dispersal TRATION. 2000. Utah climatological annual summaries patterns in Linanthus bicolor. Evolution 37:1247–1257. 1961–1990. NOAA—Climatic Summaries, NCDD. THOMSON, J.D. 1981. Spatial and temporal components of Asheville, NC. resource assessment by flower-feeding insects. Jour- NELSON, D.R. 1989. Site characteristics and habitat require- nal of Animal Ecology 50:49–59. ments of the endangered dwarf bear-claw poppy U.S. FISH AND WILDLIFE SERVICE. 1979. Determination (Arctomecon humilis Coville, Papaveraceae) and that Arctomecon humilis is an endangered species. demographic and seed bank biology of Arctomecon Federal Register 44(216):64250–64252. humilis (Papaveraceae), a short-lived perennial. Master’s thesis, Brigham Young University, Provo, Received 3 October 2003 UT. 53 pp. Accepted 12 February 2004 NELSON, D.R., AND K.T. HARPER. 1987. Project summary: ecological study of the endangered dwarf bearclaw- Western North American Naturalist 64(4), © 2004, pp. 492–496

EVALUATING THE USE OF MORPHOMETRIC MEASUREMENTS FROM MUSEUM SPECIMENS FOR SEX DETERMINATION IN MOUNTAIN PLOVERS (CHARADRIUS MONTANUS)

William M. Iko1, Stephen J. Dinsmore2, and Fritz L. Knopf1

ABSTRACT.—The Mountain Plover (Charadrius montanus) is a shorebird species endemic to the dry, terrestrial ecosystems of the Great Plains and southwestern United States. Breeding Bird Survey data suggest that Mountain Plover populations have declined by >60% in the last 30 years. A better understanding of the population dynamics of the Mountain Plover is important in determining future management goals for this species. However, this effort is hampered by the inability to determine the sex of Mountain Plovers accurately under field conditions. In an effort to develop a simple method for sexing plovers in the hand, we measured external morphometric characteristics from 190 museum specimens of adult Mountain Plovers in alternate (breeding) plumage. Logistic regression and discriminant function analyses were performed on 10 external morphometric measurements (lengths of unflattened wing chord, 10th primary, central rectrix, outer rectrix, total head length, exposed culmen, culmen, bill depth, bill width, and tarsus). The results of these analyses indicated that Mountain Plover sexes were similar for all measures except culmen length. How- ever, further analysis determined that culmen length accurately predicted sex in less than two-thirds of the specimens, suggesting that this measure is a poor predictor of sex in Mountain Plovers. Structurally, Mountain Plovers appear to be nearly identical between the sexes, and other methods of sexing birds (e.g., plumage characteristics, behavioral observations, or molecular markers) should be further assessed for devising a simple method for sexing Mountain Plovers under field conditions.

Key words: Mountain Plover, Charadrius montanus, sex determination, morphometric measurements, museum specimens.

The Mountain Plover (Charadrius montanus) Endangered Species Act (U.S. Department of is a shorebird species endemic to the dry, ter- the Interior 2002). restrial ecosystems of the Great Plains and the Despite concern over its declining popula- southwestern United States (Knopf 1996). A tions, little is known about the population species of the shortgrass prairie ecosystem, dynamics of this species, particularly its mating plovers use open, relatively flat, arid environ- system (Graul 1973a, Jehl and Murray 1986). ments and prefer intensively grazed grassland Earlier studies describing breeding behavior habitats (Knopf and Miller 1994, Knopf 1996, of Mountain Plovers indicated that this species Knopf and Rupert 1999). During the last 3 is monogamous, with the female laying an ini- decades, breeding populations of the Moun- tial clutch for the male to incubate and then a tain Plover have become increasingly isolated 2nd clutch to incubate herself (Graul 1973a, as its native shortgrass prairie habitat has been Graul 1976, McCaffery et al. 1984). However, converted to agriculture and urban develop- sex determination in these past studies has ment (Samson and Knopf 1996, Knopf and relied on direct observations of courtship be- Rupert 1999). Recent analyses of North Amer- havior and copulations (Graul 1973a), egg lay- ican Breeding Bird Survey data indicate that ing by females (Graul 1973a, SJD personal ob- Mountain Plovers have declined by 63% in the servation), or internal examination of collected last 30 years (U.S. Department of the Interior plovers (FLK unpublished data). Compilation 2002). Population declines of the Mountain of detailed population data, such as sex ratios Plover have made it a species of concern of breeding populations and differential mor- throughout its current range and proposed for tality or migration among the sexes, has been listing as a threatened species under the U.S. hampered by the inability to determine the sex

1U.S. Geological Survey, Fort Collins Science Center, 2150 Centre Avenue, Bldg. C, Fort Collins, CO 80526-8118. 2U.S. Fish and Wildlife Service, Colorado Management Assistance Office, 755 Parfet Street, Suite 496, Lakewood, CO 80215. Present address: Department of Wildlife and Fisheries, Mississippi State University, Box 9690/257 Thompson Hall, Mississippi State, MI 39762.

492 2004] SEX DETERMINATION OF MOUNTAIN PLOVERS 493 of breeding Mountain Plovers accurately in the lengths (to nearest 0.1 mm) of unflattened wing field. As an initial step in addressing this issue, chord; outer or 10th primary; central rectrix; we evaluated the use of external morphometric outer rectrix; total head (from the posterior measures from museum specimens to develop end of the occipital crest of the skull to the a simple technique for determining the sex of anterior end of the upper mandible); exposed Mountain Plovers in the field. culmen (from the base of the upper mandible at the beginning of the feather tracts to the METHODS anterior end of the upper mandible); culmen (from the anterior end of the nares to the ante- The goal of this study was to develop a sim- rior end of the upper mandible); bill depth ple mathematical equation for sexing Moun- (measured at the anterior end of the gonys); tain Plovers in the hand using linear morpho- bill width (measured at the anterior end of the metric measurements. However, because of gonys); and tarsus (from the intertarsal joint to the difficulty in capturing and collecting such the distal end of the last leg scale before the measurements from live Mountain Plovers toe emerges). We chose these external measures (WMI personal observation), we chose first to because of their common use among bird ban- evaluate a morphometric sexing criteria using ders and because differences between the sexes museum specimens. After an extensive search would most likely be evident in these features of museum collections, we located 611 Moun- (Prater et al. 1977, Hayman et al. 1986, Pyle tain Plover specimens, which we believe rep- 1997). Body mass recorded on museum labels resent nearly all of the Mountain Plover speci- was not used because too few specimens had mens in North America (see Acknowledgments). this information (17 of 190 specimens). We chose to limit our analysis to specimens in To develop a simple mathematical equation adult breeding (alternate) plumage where sex to predict the sex of Mountain Plovers, we was identified by the specimen preparator, first tested for sex differences using a univari- because of our uncertainty of age-related dif- ate approach with a logistic regression model ferences in soft tissue measurements and the using the logit link function in the PROC potential misidentification of sex by museum GENMOD procedure in SAS (SAS Institute preparators due to atrophy of the gonadal tis- 1990). Sex was the dependent variable in our sues during the nonbreeding season. For the regression models and followed a binomial purpose of this analysis, we have assumed that distribution. The importance of each measure- the sex indicated on the museum labels of the ment for explaining sex was tested using a chi- remaining adult breeding plumage specimens square test with 1 degree of freedom (df). For is correct. To reduce observer bias, only one morphometric variables showing significant researcher (WMI) collected morphometric mea- differences between sexes, we used the inter- surements. Because the condition of each speci- cept and regression coefficient to obtain a men varied, a complete set of measurements logistic regression equation to predict the sex could not always be obtained. All efforts were of our museum specimen data set. For these made by the researcher to make the same mor- analyses we used α = 0.05 as the level of sta- phometric measurement at the same anatomi- tistical significance. We also used a multivari- cal location on each study specimen used. ate approach with a stepwise discriminant When such measures could not be replicated function analysis to predict sex in Mountain on a given study specimen, or a complete set Plovers in accordance with other morphomet- of measures could not be obtained, the mor- ric studies (Brennan et al. 1984, Johnstone and phometric measurements from that specimen Niven 1989). Using the same 10 variables were eliminated from our statistical analyses. listed above, we attempted to classify plovers After eliminating winter (basic) plumaged as male or female using the Fisher discrimi- birds, specimens whose museum labels indi- nant function analysis in PROC DISCRIM in cated “juvenile” or “sex unknown,” and speci- SAS (SAS Institute 1990). A stepwise discrimi- mens with an incomplete set of measurements, nant function analysis was used to identify we had a total sample of 190 known-sex plover which subset of the 10 variables was most use- specimens (males, n = 112; females, n = 78) in ful for discriminating between male and female alternate plumage for inclusion in our analysis. plovers. We then used the Fisher discriminant The measurements we collected were the functions for all 10 variables from DISCRIM 494 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Univariate measurements (x– ± s) between female (n = 78) and male (n = 112) adult Mountain Plover museum specimens in breeding plumage. Females Male Variable (mm) (mm) χ2 P > χ2 Unflattened wing chord 147.2 ± 4.1 147.5 ± 4.4 0.1 0.69 10th primary length 103.3 ± 4.6 102.4 ± 8.2 0.66 0.42 Central rectrix length 61.2 ± 3.5 61.6 ± 4.1 0.41 0.52 Outer rectrix length 59.5 ± 4.1 59.6 ± 3.9 0.05 0.82 Total head length 53.9 ± 1.7 53.3 ± 3.7 1.41 0.24 Exposed culmen length 21.4 ± 1.1 21.1 ± 1.1 2.58 0.11 Culmen length 13.9 ± 0.9 13.6 ± 0.9 4.70 0.03 Bill depth 4.4 ± 0.3 4.4 ± 0.3 0.72 0.40 Bill width 4.3 ± 0.3 4.3 ± 0.4 0.02 0.88 Tarsus length 39.8 ± 1.4 39.6 ± 1.5 1.09 0.30

to generate predictive equations of group (male of our Mountain Plover specimens (Table 2). or female) affiliation. For these analyses we Using this procedure, we were able to cor- used α = 0.15 as the probability of entry into, rectly classify 62% of our birds (n = 117), with and of remaining in, the model in PROC DIS- most correctly classified specimens being male CRIM and α = 0.05 as the level of statistical (91%). Most of the misclassifications were of significance for all other tests. females that were classified as males (81%). The discriminant function analysis produced RESULTS similar results using Fisher discriminant functions for predicting sex and a model using all We first regressed each of the 10 variables 10 variables (Table 3). Of the 10 morphometric separately with sex as the response variable. variables, the Fisher discriminant function Of the 10 measures we analyzed, only culmen analysis found only culmen length (F1,188, P = length (χ2 = 4.70, df = 1, P = 0.03) showed a 0.03) as a significant predictor of sex. Simi- significant difference between the sexes (Table larly, in the STEPDISC discriminant function 1). None of the remaining measurements dif- analysis, only culmen met the criteria for reten- fered by sex in Mountain Plovers. We further tion in the model. We used these equations to tested our ability to predict the sex of all birds classify each plover as male or female, as above in our sample using the regression equation (Table 2). Discriminant function analysis was with only culmen length. Using the intercept able to correctly classify only 63% of our birds and regression coefficient obtained when we (n = 119), with a pattern of misclassification regressed culmen on sex, we calculated a pre- similar to that found with the logistic regres- dicted sex for each of the 190 birds (Table 2). sion results (74% misclassification for females We predicted sex using the regression equa- and 12% misclassification for males). tion: DISCUSSION

Xbeta = 5.4353 – 0.3691 * Culmen The results of our study indicate that for 9 of 10 morphometric measurements we col- Xbeta values were then mapped into the logit- lected, Mountain Plover body sizes were simi- link function of: lar between the sexes, suggesting that they are a monomorphic species. Culmen length, as in Sex = ______1 other studies (Skeel 1982, Brennan et al. 1984, 1 + e–Xbeta Jehl and Murray 1986, Sandercock 1998), was the most useful of the 10 measures, but only where if Sex > 0.5, the bird was classified as a successfully identified the sex of 62%–63% of male, and if Sex < 0.5, the bird was classified our museum specimens. Mean values for cul- as a female. Although we found significant dif- men length between male and female speci- ferences in culmen length between sexes, this mens, although significantly different, were measure alone was a poor predictor of the sex slight, potentially negating the usefulness of 2004] SEX DETERMINATION OF MOUNTAIN PLOVERS 495

TABLE 2. Comparison of known to predicted sex of Mountain Plover specimens in breeding plumage based on culmen length (n = 190). The 1st value is percentage estimate from the logistic regression (LR) model and the 2nd value is the percentage estimate from the discriminant function (DF) analysis.

______Predicted sex ______Female ______Male Total Known sex LR (n) DF (n) LR (n) DF (n) n Female 19% (15) 36% (20) 81% (63) 74% (58) 78 Male 9% (10) 12% (13) 91% (102) 88% (99) 112 Total 12% (25) 17% (33) 87% (165) 83% (157) 190a aOverall number of correct classifications (logistic regression: 117 of 190 or 62%; discriminant function analysis: 119 of 190 or 63%).

TABLE 3. Fisher discriminant function analysis for pre- 1993). However, incorporation of other field dicting sex of Mountain Plover specimens in breeding measures from live-caught Mountain Plovers, plumage based on a 10-variable model (n = 190). The model includes measures of unflattened wing chord, 10th such as body mass, may lead to a more useful primary, central rectrix, outer rectrix, total head length, sex determination criterion. Johnstone and exposed culmen, culmen, bill depth, bill width, and tarsus Niven (1989) demonstrated in their field study length. on Grey-faced Petrels (Pterodroma macroptera Variable Female Male gouldi) that a discriminant function classification formula based on bill depth alone yielded Constant –958.52 –953.34 Unflattened wing chord 7.57 7.63 a predictability rate of 63% and 65% for males 10th primary length –1.18 –1.22 and females, respectively; however, by incor- Central rectrix length –0.26 –0.22 porating body mass into this formula, accuracy Outer rectrix length 0.16 0.15 rose to 93% and 91% for males and females, Total head length 2.26 2.23 Exposed culmen length 7.43 7.37 respectively. Unfortunately, few of the museum Culmen length 3.60 3.32 specimens used in our analyses had reliable Bill depth 15.16 14.64 body mass data recorded on museum labels Bill width 11.10 11.28 (only 17 of the 190 specimens). Also, body mass Tarsus length 12.16 12.10 measures from live birds should be used with caution as mass can fluctuate due to several variables, including reproductive status (such as presence of eggs in the oviducts of females), this single measure alone in predicting the sex migratory condition (such as fat deposition), of Mountain Plovers. The results of our logis- and general physiological condition of the tic regression analysis suggest that the value of bird. However, the incorporation of body mass culmen length as a predictor of sex in Moun- measures in this study will have to depend on tain Plovers is questionable. Our discriminant future field studies involving live-capture and function analysis yielded results similar to those measurement of Mountain Plovers. found using logistic regression, further strength- Other sex determination criteria, such as ening claims that the Mountain Plover is a field plumage characteristics, should also be considered for sexing Mountain Plovers. Field monomorphic species. studies on this species have indicated that The use of museum specimens in sex dis- plovers suspected of being males tend to have crimination can be helpful, especially when a brighter alternate plumage, including a more studying avian species that are in decline or distinct head pattern and brighter rufous or are restricted in their handling or collection. orange color on the neck and mantle (Graul However, the use of linear morphometric 1973b, Jehl and Murray 1986, SJD personal measurements solely from museum specimens observation). If such observational data of for field application warrants some caution. dichromatism in this species could develop Linear morphometric measurements from standardized measurements of plumage color- museum specimens can vary due to feather ation, a potential field methodology for sexing wear, museum preparation and storage, and Mountain Plovers could be established. Other potential shrinkage (Johnston 1990, Winker sexing criteria, such as genetic markers (Kahn 496 WESTERN NORTH AMERICAN NATURALIST [Volume 64 et al. 1998, Dinsmore et al. 2002) obtained from DINSMORE, S.J., G.C. WHITE, AND F. L . K NOPF. 2002. Ad- blood or feather samples, may prove even more vanced techniques for modeling avian nest survival. Ecology 83:3476–3488. effective as these techniques are improved and GRAUL, W.D. 1973a. Adaptive aspect of the Mountain become more widely available. Plover social system. Living Bird 12:69–94. ______. 1973b. Possible functions of head and breast mark- ings in Charadriinae. Wilson Bulletin 85:59–70. ACKNOWLEDGMENTS ______. 1976. Food fluctuations and multiple clutches in the Mountain Plover. Auk 93:166–167. We are greatly indebted to C. Preston and HAYMAN, P., J. MARCHANT, AND T. P RATER. 1986. Shorebirds: B. Alther, Denver Museum of Nature and Sci- an identification guide to waders of the world. ence, for the use of their museum facilities Houghton Mifflin Co., Boston. 412 pp. JEHL, J.R., JR., AND B.G. MURRAY, JR. 1986. The evolution and maintenance of museum loans. We also of normal and reverse sexual size dimorphism in thank the museum curators who took time to shorebirds and other birds. Pages 1–86 in R.F. John- respond to our initial inquiries, especially the ston, editor, Current ornithology. Volume 3. Plenum following: R. Sloss (American Museum of Nat- Press, New York and London. ural History); K. Cebra (California Academy of JOHNSTON, R.F.1990. Variation in size and shape in pigeons Columba livia. Wilson Bulletin 102:213–235. Sciences); R. Panza (Carnegie Museum of Nat- JOHNSTONE, R.M., AND B.E. NIVEN. 1989. Sexing Grey- ural History); M. Hennen (Chicago Academy of faced Petrels by discriminant analysis of measure- Sciences); E. Merritt (Cincinnati Museum of ments. Notornis 36:261–265. Natural History); D. Willard (Field Museum of KAHN, N.W., J. ST. JOHN, AND T.W. Q UINN. 1998. Chromo- some-specific intron size differences in the avian Natural History); R. Browning (National Muse- CHD gene provide an efficient method for sex iden- um of Natural History); K. Garrett (Los Ange- tification in birds. Auk 115:1074–1078. les County Museum); J. Dick (Royal Ontario KNOPF, F.L. 1996. Mountain Plover (Charadrius montanus). Museum); R. McKernan (San Bernardino In: A. Poole and F. Gill, editors, The birds of North America, No. 211. Academy of Natural Sciences, County Museum); P. Unitt (San Diego Natural Philadelphia, PA, and American Ornithologists’ Union, History Museum); K. Fahy (Santa Barbara Washington, DC. Museum of Natural History); C. Sumida and F. KNOPF, F.L., AND B. MILLER. 1994. Charadrius montanus— Kinoshita (Western Foundation of Vertebrate montane, grassland, or bare-ground plover? Auk Zoology); C. Cicero (Museum of Vertebrate 111:504–506. KNOPF, F.L., AND J.R. RUPERT. 1999. Use of cultivated Zoology, University of California, Berkeley); fields by breeding Mountain Plovers in Colorado. R. Cole (University of California, Davis); R. Studies in Avian Biology 19:81–86. Humphrey (University of Colorado Museum); MCCAFFERY, B.J., T.A. SORDAHL, AND P. Z AHLER. 1984. R. Paynter, Jr. (Museum of Comparative Zool- Behavioral ecology of the Mountain Plover in northeastern Colorado. Wader Study Group Bulletin ogy, Harvard University); M. Robbins (Natural 40:18–21. History Museum, University of Kansas); J. Hin- PRATER, A.J., J.H. MARCHANT, AND J. VUORINEN. 1977. Guide shaw (Museum of Zoology, University of Michi- to the identification and ageing of Holarctic waders. gan); R. Zink and J. Klicka (Bell Museum of British Trust for Ornithology Guide 17. British Trust for Ornithology, Tring, Herts. Natural History, University of Minnesota); R. PYLE, P. 1997. Identification guide to North American Dickerman (Museum of Southwestern Biol- birds. Part 1. Slate Creek Press, Bolinas, CA. 732 pp. ogy, University of New Mexico); J. Hafner and SAMSON, F.B., AND F. L . K NOPF. 1996. Prairie conservation: J. Northern (Moore Laboratory of Zoology, preserving North America’s most endangered ecosystem. Island Press, Washington, DC. Occidental College); J. Braun (Oklahoma Muse- SANDERCOCK, B.K. 1998. Assortative mating and sexual um of Natural History, University of Okla- size dimorphism in Western and Semipalmated homa); F. Sibley (Peabody Museum of Natural Sandpipers. Auk 115:786–791. History, Yale University). We thank C. Ramotnik SAS INSTITUTE, INC. 1990. SAS/STAT user’s guide. Ver- for assistance in locating museum specimens sion 6. 4th edition. SAS Institute, Inc., Cary, NC. SKEEL, M.A. 1982. Sex determination of adult Whimbrels. and collections and B. Roelle, R. Stendell, G. Journal of Field Ornithology. 53:414–416. White, P. Pyle, N. Woffinden, and B. Sandercock U.S. DEPARTMENT OF THE INTERIOR. 2002. Endangered for their input to the manuscript preparation. and threatened wildlife and plants: threatened status and special regulation for the Mountain Plover. Fed- eral Register 67 (234):72396–72407. LITERATURE CITED WINKER, K. 1993. Specimen shrinkage in Tennessee War- blers and “Traill’s” Flycatchers. Journal of Field Orni- BRENNAN, L.A., J.B. BUCHANAN, C.T. SCHICK, S.G. HERMAN, thology 64:331–336. AND T.M. J OHNSON. 1984. Sex determination of Dun- lins in winter plumage. Journal of Field Ornithology Received 28 July 2003 55:343–348. Accepted 17 November 2003 Western North American Naturalist 64(4), © 2004, pp. 497–502

LIFE CYCLE OF THE NET-SPINNING CADDISFLY, CHEUMATOPSYCHE ANALIS (BANKS) (TRICHOPTERA: HYDROPSYCHIDAE), IN TWO SMALL FRONT RANGE URBAN STREAMS, FORT COLLINS, COLORADO

Robert E. Zuellig1, Boris C. Kondratieff2, and Richard A. Thorp2

ABSTRACT.—The life cycle of Cheumatopsyche analis (Banks) is described from 2 urban streams located in north central Colorado. Larvae were collected every 15 days during spring, summer, and fall and every 30 days in winter from August 1999 through August 2000. A total of 8652 larvae were collected, and head capsules were measured to the nearest 0.01 mm. The life cycle of C. analis at both sites appears to be partially bivoltine with a long, almost continuous period of recruitment after peak emergence in June. Some offspring from the early generation that emerges in May probably emerge as adults in September, while the remainder of the population overwinter as instars III through V. Life cycle patterns for C. analis in these 2 urban streams were most similar to other populations of Cheumatopsyche from Vir- ginia.

Key words: life cycle, Cheumatopsyche analis, Hydropsychidae, urban stream.

The filter-feeding caddisflies of the family have described life cycles of aquatic insects Hydropsychidae are well-known members of from urban systems (Paul and Meyer 2001). In most stream communities throughout North this paper we describe the life cycle of C. America (Wiggins 1996). Often they comprise analis from 2 small urban streams located in a substantial portion of aquatic insect biomass Fort Collins, Colorado. Both streams are influ- in streams, especially in relation to secondary enced by stormwater runoff and other typical production (Benke and Wallace 1980, Wallace stream alterations associated with urban devel- and Merritt 1980, Parker and Voshell 1983, opment, thus providing a unique opportunity Smock and Roeding 1986, Bowles and Allen to examine the life history of this species in an 1991). Larvae of the hydropsychid genus urban lotic system. Cheumatopsyche are common in warm-water streams with moderate to high organic loads METHODS (Wallace 1975, MacKay 1986). One species, C. Study Area analis (Banks), occurs over much of southern Canada and the U.S. (Gordon 1974). This The study area is located in north central species tolerates a broad range of lotic habitats Colorado within the city limits of Fort Collins, and stream temperatures (Gordon and Wallace which is situated in the northwest central por- 1975, MacKay 1986, Kondratieff et al. 1997). tion of the South Platte River basin (Fig. 1). Fort Along the eastern Front Range of the Rocky Collins is a moderately sized urban area (ca. Mountains of Colorado, C. analis is often the 119,000 residents) that sits approximately 6.5 km dominant net-spinning caddisfly, especially in from the eastern edge of the Rocky Mountains. the transition zone where streams meet the Dennehy et al. (1993) provides a detailed Great Plains. In fact, C. analis is often the only description of the environmental setting of the caddisfly species present in many regional small study area and the surrounding South Platte streams (Zuellig 2001). Herrmann et al. (1986) River basin. Typically, soils consist of cobbly also reported C. analis to be a particularly com- clay loam, clay loam, or gravelly sandy clay. The mon species in Colorado. depth to bedrock is usually >1.5 m (Dennehy Despite the recent focus on studying the et al. 1993). Annual precipitation is <38 cm. ecology of urban streams, very few studies Several authors have documented changes that

1Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523. 2Department of Bioagricultural Sciences and Pest Management, Colorado State University, Fort Collins, CO 80523.

497 498 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 1. Spring Creek and Mail Creek site locations in relation to the Cache la Poudre River, Larimer County, CO, and to the South Platte River basin. have occurred across the landscape that began River (Fig. 1). The Spring Creek and Mail Creek in the early 1800s (Eschner et al. 1983, Wohl basins are relatively small, draining a total 2001). area of 5800 ha and 648 ha, respectively. Both We chose 1 site on each of 2 perennial streams are in close proximity to each other streams, Mail Creek and Spring Creek, for and both sites are at similar elevations. The analysis as part of a larger study that investi- Spring Creek site was located at a city park in gated the effects of urban stream habitat on the lower part of the basin (40°33′59″N, aquatic communities (Zuellig 2001). Both 105°02′10″W, elevation 1515 m). The Mail streams are tributaries to the Cache la Poudre Creek site was located at a small business park 2004] LIFE CYCLE OF CHEUMATOPSYCHE ANALIS (BANKS) 499 in the upper part of the basin (40°31′23″N, were deposited in the C.P. Gillette Museum of 105°04′52″W, elevation 1547 m). Both streams Arthropod Diversity (CSUC), Colorado State have been incorporated into extensive irriga- University, Fort Collins, Colorado. tion canal networks, which began in the 1860s Data Analysis (Eschner et al. 1983). Water from the Cache la Poudre River and other adjacent basins is trans- Larval instars were determined by plotting ported laterally across several basins into both size frequency distributions of head capsule streams. The main purpose for these lateral width to obtain width ranges for instars I water transfers is to provide water for irrigated through V for each site. Larvae were assigned agriculture on the eastern plains of Colorado. instar size, and head capsule width was plot- Both sites are subject to stormwater runoff ted to visually assess instar variability between and other typical practices associated with sites. We also compared mean head capsule urbanization such as channelization and ripar- width from these populations with means of ian zone removal. other populations from the literature. Size frequency distribution diagrams were constructed Sample Collecting and for each site based on the relative percent of Processing the total number of individuals at each instar We collected benthic macroinvertebrate collected each month. samples at each site from riffle areas using a kick net with an opening of 25 × 45 cm RESULTS equipped with a 500-µm mesh net. The kick net was placed on the streambed, and the sub- Keys for species identification of immature strate was disturbed and rubbed clean of any Cheumatopsyche larvae are currently unavail- invertebrates for a period of 1 minute by a col- able; however, all adults collected from the lector moving in an upstream direction. Sam- stream margins during this study were veri- ples were collected from August 1999 through fied as C. analis. Therefore, we are confident August 2000 every 15 days during spring, sum- the larvae measured in this study were those mer, and fall and once every 30 days during of C. analis. Head width measurements from the winter. Because of excessive ice cover at 8652 larvae were used to determine the distribution of the mean and the range of head cap- both sites, we did not collect samples in Janu- sule width across instars at both sites (Table 1). ary. Samples were fixed in the field using a 5% Variability in head capsule width at both sites formaldehyde solution and transported to the increased as larvae increased in size. Mean laboratory for processing. head capsule width of instars I through III At the laboratory, stream samples were µ from these 2 urban streams was similar to others washed through a 500- m sieve and were then reported in the literature; however, mean head spread evenly into a white enamel pan. All capsule width of instars IV and V from these 2 Cheumatopsyche larvae were removed from urban populations was often slightly larger than each sample, placed into storage vials, and other populations. preserved in 80% ethanol. We measured head The life cycle of C. analis in these 2 urban capsule width using a dissecting microscope streams was strikingly similar, appearing to be equipped with an ocular micrometer calibrated partially bivoltine with a long period of recruit- to the nearest 0.01 mm. ment from June through September (Fig. 2). We also collected adult caddisflies along Pupae were first observed in April and were stream margins at each site using a sweep net, present until September except for a 2-week and these were preserved in 80% ethanol. Adult period in June in both streams. The presence specimens and mature pupae were identified of pupae in both streams from late April in the laboratory following Gordon (1974) and through September (Fig. 2) coincides with sent to David E. Ruiter for verification. Gordon Herrmann et al. (1986) and with C. analis (1974) considered the name analis as a nomina records at CSUC from northern Colorado and dubia and resurrected the name pettiti for C. southern Wyoming streams. Large numbers of analis (Flint and England 2003). Recently, adults were observed along stream margins Flint and England (2003) placed C. pettiti as a during May and June, which we estimate as junior synonym to C. analis. Adult specimens peak emergence for the early generation. Some 500 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Cheumatopsyche analis (Banks) head capsule measurement summary for Spring Creek and Mail Creek, City of Fort Collins, Larimer County, CO. (s) = standard deviation of mean head capsule width in mm. Number of Mean head width Range of head width Instar measurements (mm) (s) (mm)

SPRING CREEK I450.21 (0.01) 0.17–0.21 II 141 0.26 (0.01) 0.26–0.30 III 854 0.41 (0.02) 0.34–0.47 IV 1442 0.65 (0.04) 0.51–0.77 V 1636 1.00 (0.06) 0.81–1.40

MAIL CREEK I610.21 (0.01) 0.17–0.21 II 247 0.26 (0.01) 0.26–0.30 III 1388 0.41 (0.02) 0.34–0.47 IV 1578 0.64 (0.03) 0.51–0.77 V 1260 0.98 (0.05) 0.81–1.37

of the pupae present in late August are proba- streams along the Front Range of Colorado bly offspring from individuals that emerged at indicate emergence from May through August. the beginning of emergence in May. The re- The variety of life cycles and emergence mainder that did not emerge in September patterns of C. analis agrees with earlier de- probably overwintered as instars IV or V. The scriptions of the adaptability of this species to timing of the presence of large numbers of tolerate a wide variety of environmental con- adults seemed to be identical at both sites even ditions. The plasticity of C. analis has appar- though most of the Spring Creek population ently allowed this species to become well overwintered as instars IV or V; whereas, the adapted to the Front Range urban streams of majority of the Mail Creek population over- Colorado, even though these streams have wintered as instar III (Fig. 2). Recruitment of been subjected to over 175 years of distur- early instar individuals was nearly continuous bance (Wohl 2001). Cheumatopsyche analis in both populations. often dominates caddisfly communities in small streams of this region (Zuellig 2001). Along DISCUSSION with the Hydropsyche bronta group and H. Several studies report a wide range of life occidentalis and several species of Simulium, histories and emergence patterns for C. analis especially S. vittatum complex, C. analis is often (as pettiti). MacKay (1986) and Morin and the dominant filter-feeding taxa present across Harper (1986) report a univoltine life cycle in the plains portion of the transition zone (Zuel- Minnesota and Montreal, whereas Sanchez lig 2001, CSUC records). Since C. analis is and Hendricks (1997) document a bivoltine among the few taxa occupying the filter-feed- and partial bivoltine life cycle for this species ing trophic niche in small Front Range streams, in a southwestern Virginia stream. In contrast, it probably plays an important role in the Kondratieff et al. (1997) report continuous energy flow of these systems. Unfortunately, generations in a nonseasonal Hawaiian stream. no historical caddisfly records are available for Univoltine populations emerged from June small Front Range streams of Colorado prior through August in Minnesota (MacKay 1986), to the initiation of irrigated agriculture (ca. and May through August in Montreal (Morin 1860). It would be interesting to know if C. and Harper 1986). Floyd and Schuster (1990) analis was historically the dominant caddisfly report the presence of C. analis adults from species in the eastern portion of the transition April through August in Kentucky, and Herr- between mountains and plains or if it is domi- mann et al. (1986) report emergence from late nant today because of the series of widespread April through early October in Colorado. Col- stream modifications that have occurred across lection records from CSUC of C. analis from this region. For example, C. analis occurs, but 2004] LIFE CYCLE OF CHEUMATOPSYCHE ANALIS (BANKS) 501

Fig. 2. Life cycle of Cheumatopsyche analis (Banks) collected from August 1999 to August 2000 from 2 urban streams, Spring Creek (A) and Mail Creek (B), Fort Collins, CO. Width of bar = 100% of the individuals collected during the corresponding month. is not common, in Chief Creek, which is a rel- U.S. and Canada. The populations from these atively undisturbed small stream in eastern 2 urban streams most closely match the popu- Colorado, Yuma County (BCK personal obser- lations of C. analis described by Sanchez and vation). Hendricks (1997) from southwestern Virginia. The life cycle of C. analis from these 2 Apparently, the adaptability of C. analis has urban streams appears to be partially bivoltine allowed this species to proliferate in these with an extended period of emergence from modified urban streams as well as in many dif- an early generation in May and June followed ferent types of streams across its broad geographic distribution. by a partial generation in August that provides an almost continuous source of recruitment. ACKNOWLEDGMENTS Life cycle and emergence length and timing were variable when compared with other pop- We thank David E. Ruiter, who graciously ulations reported from different parts of the verified pupae and adult identifications and 502 WESTERN NORTH AMERICAN NATURALIST [Volume 64 reviewed an earlier version of this manuscript. MACKAY, R.J. 1986. Life cycles of Hydropsyche riola, H. Additionally, we thank Joe Nicholson and Tad slossonae and Cheumatopsyche pettiti (Trichoptera: Hydropsyche) in a spring-fed stream in Minnesota. Stout for helping process samples. American Midland Naturalist 115:19–24. MORIN, A., AND P. H ARPER. 1986. Phenology and micro- LITERATURE CITED distribution of adults and larvae of filter-feeding Tri- choptera in lower Laurentian Lake outlet (Quebec). BENKE, A.C., AND J.B. WALLACE. 1980. Trophic basis of Archives of Hydrobiologia 108:167–183. production among net-spinning caddisflies in a PARKER, C.R., AND J.R. VOSHELL. 1983. Production of fil- southern Appalachian stream. Ecology 61:108–118. ter-feeding Trichoptera in an impounded and a free- BOWLES, D.E., AND R.T. ALLEN. 1991. Secondary produc- flowing river. Canadian Journal of Zoology 61:70–87. tion of net-spinning caddisflies (Trichoptera: Curvi- PAUL, M.J., AND J.L. MEYER. 2001. Streams in the urban palpia) in an Ozark stream. Journal of Freshwater landscape. Annual Review of Ecological Systems 32: Ecology 6:93–100. 333–365. DENNEHY, K.F., D.W. LITKE, C.M. TATE, AND J.S. HEINY. SANCHEZ, R.M., AND A.C. HENDRICKS. 1997. Life history 1993. South Platte River basin—Colorado, Nebraska, and secondary production of Cheumatopsyche spp. and Wyoming. Water Resources Bulletin 29:647–683. in a small Appalachian stream with two different ESCHNER, T.R., R.F. HADLEY, AND K.D. CROWLEY. 1983. land uses on its watershed. Hydrobiologica 354: Hydrologic and morphologic changes of the Platte 127–139. River basin in Colorado, Wyoming, and Nebraska: a SMOCK, L.A., AND C.E. ROEDING. 1986. The trophic basis historical perspective. U.S. Geological Survey Pro- of production of the macroinvertebrate community fessional Paper 1277-A. of a southeastern USA blackwater stream. Holarctic FLINT, O.S., JR., AND R.A. ENGLAND. 2003. A reassignment Ecology 9:165–174. and new state records of Trichoptera occurring in WALLACE, J.B. 1975. Food partitioning in net-spinning Tri- Hawai’i with discussion on origins and potential eco- choptera larvae: Hydropsyche venularis, Cheumato- logical impacts. Records of the Hawaii Biological psyche etrona, and Macronema zebratum (Hydropsy- Survey for 2001–2002. Bishop Museum Occasional chidae). Annals of the American Entomological Society Papers 73:31–40. 68:463–472. FLOYD, M.A., AND G.A. SCHUSTER. 1990. The caddisflies WALLACE, J.B., AND R.W. MERRITT. 1980. Filter-feeding (Insecta: Trichoptera) of the Buck Creek system, ecology of aquatic insects. Annual Review of Ento- Pulaski County, Kentucky. Transactions of the Ken- mology 25:103–132. tucky Academy of Sciences 51:127–134. WIGGINS, G.B. 1996. Larvae of the North American caddis- GORDON, A.E. 1974. A synopsis and phylogenetic outline fly genera (Trichoptera). 2nd edition. University of of the Nearctic members of Cheumatopsyche. Pro- Toronto Press, Toronto, Canada. ceedings of the Academy of Natural Sciences of WOHL, E.E. 2001. Virtual rivers: lessons from the moun- Philadelphia 126:117–160. tain rivers of the Colorado Front Range. Yale Univer- GORDON, A.E., AND J.B. WALLACE. 1975. Distribution of sity Press, New Haven, CT, and London. the family Hydropsychidae (Trichoptera) in the Savan- ZUELLIG, R.E. 2001. Macroinvertebrate and fish commu- nah River basin of North Carolina, South Carolina nities along the Front Range of Colorado and their and Georgia. Hydrobiologia 46:405–423. relationship to habitat in the urban environment, HERRMANN, S.J., D.E. RUITER, AND J.D. UNZICKER. 1986. Master’s thesis, Colorado State University, Fort Distribution and records of Colorado Trichoptera. Collins. Southwestern Naturalist 31:421–457. KONDRATIEFF, B.C., R.J. BISHOP, AND A.M. BRASHER. 1997. Received 25 November 2003 The life cycle of an introduced caddisfly, Cheumato- Accepted 31 March 2004 psyche pettiti (Banks) (Trichoptera: Hydropsychidae), in Waikolu Stream, Molokai, Hawaii. Hydrobiologia 350:81–85. Western North American Naturalist 64(4), © 2004, pp. 503–513

NATIVE PLANT COMMUNITY COMPOSITION AND GENETIC DIVERSITY ASSOCIATED WITH LONG-TERM WEED INVASIONS

Brian A. Mealor1, Ann L. Hild1, and Nancy L. Shaw2

ABSTRACT.—Many studies have assessed genetic changes in exotic plant species in their native and introduced ranges, but none have focused on genetic variation in native plant species in response to exotic invasion. We examine characteristics of native plant communities within and outside old (>25 year) invasions of Acroptilon repens (Russian knapweed) and Cardaria draba (hoary cress). We also document genetic variability of 4 native grass populations (Hes- perostipa comata [needle and thread], Achnatherum hymenoides [Indian ricegrass], Sporobolus airoides [alkali sacaton], and Poa secunda [Sandberg bluegrass]) from 2 areas: adjacent to and within weed invasions. Native plant species richness and diversity did not differ between invaded and noninvaded areas. Inter-simple sequence repeat (ISSR) analysis of individual native perennial grasses of each of the 4 species suggests that populations exposed to long-term coexistence with exotics may differ from adjacent noninvaded populations. We suggest that future research efforts should focus on intraspecific diversity of native plant species to identify possible candidates for restoration following weed control.

Key words: native grasses, invasive plants, diversity, ISSR, Acroptilon repens, Cardaria draba, Achnatherum hymenoides, Hesperostipa comata, Sporobolus airoides, Poa secunda.

Exotic species often limit the ability of with restoration ecology in an effort to charac- restoration ecologists to return native species terize interactions between native plants and to disturbed sites. Presence of exotic species their new exotic neighbors. This study describes may be a primary impetus for restoration a new approach for combating reentry of exotic activities. Exotics may become established fol- weeds into restoration plantings. lowing disturbances and often leave viable Much literature suggests that invasion by seed banks long after control efforts (D’Anto- exotic species may decrease native species di- nio and Meyerson 2002). Revegetation with versity or alter nutrient cycles, thereby affect- competitive plants after chemical or mechani- ing community diversity (Elton 1958, Gordon cal treatments is needed to limit reestablish- 1998, Levine and D’Antonio 1999, Levine ment of exotic invaders (Sheley and Stivers 2000). It is often assumed that exotics form 1999, Whitson 1999). Because of the increas- monocultures and exclude native populations ing demand for local native species in restora- originally present on a site (Watson 1980). Even tion, we must consider the genetic integrity of so, intrapopulation variability in native plants our native seed sources (Belnap 1995, Linhart has been documented over very small dis- 1995, Roundy et al. 1997, Jones and Johnson tances (Linhart 1995, McGuinnies et al. 1998, 1998). Yet it is unknown whether exotic inva- Hild et al. 1999) and over short time periods sions alter the genetic makeup of native plant (Snaydon and Davies 1982, Al-Hiyaly et al. populations by reducing native abundance. 1988, 1989). Linhart and Grant (1996) assert No study has considered the impact of long- that “genetic differentiation could be observed term weed invasions on the genetic variability over distances of even a few meters, and, of remaining on-site native populations. whenever localized selection was sufficiently Restoration ecologists are concerned with intense, even a few centimeters.” These stud- constructing communities that are resistant to ies investigated local differentiation of native invasion by exotic species, even though our plant populations in response to changing knowledge of invasive species ecology has not environmental conditions, but not to the pres- been well integrated into restoration and man- ence of exotic weeds. We examine differences agement (D’Antonio and Meyerson 2002). In in native plant populations within weed inva- this paper we unite invasive species ecology sions and in adjacent, noninvaded areas. We

1Department of Renewable Resources, University of Wyoming, Box 3354, Laramie, WY 82071-3354. 2USDA Forest Service, Rocky Mountain Research Station, 344 E. Front St., Boise, ID 83702.

503 504 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 1. Descriptive characteristics of 4 study sites invaded either by Acroptilon repens or Cardaria draba in Wyoming and Idaho Annual Mean Age of Exotic Exotic species Elevation precipitation temperature invasiona density Soil Site (m) (cm) (ºC) (years) (stems ⋅ m–2) texture Cardaria draba Bancroft, ID 1740 38 6.9 25 276 silt loam Dubois, WY 2210 24 5.0 25 45 loamy sand Greybull, WY 1190 18 7.1 35 57 silty clay Rock Springs, WY 2010 23 6.0 30 201 silty clay loam Centaurea repens Greybull, WY 1190 18 7.1 35 51 silty clay Mud Lake, ID 1460 21 6.7 30 80 loamy sand Rogerson, ID 1520 27 8.8 90 20 loam Rock Springs, WY 2000 23 6.0 30 35 sandy clay loam aAge of invasion was based on observations of local landowners and managers.

ask if long-term coexistence with exotic neigh- bers as resistant, remnant individuals. Then we bors may, over time, select for a unique subset use an anonymous genetic fingerprint analysis of native plant genotypes. to determine if genotypic characteristics of Invasive species can increase in competi- natives from within differ from those of indi- tive ability (Blossey and Notzold 1995), grow viduals immediately outside the invasion. larger (Siemann and Rogers 2001), hybridize (Ellstrand and Schierenbeck 2000), and even MATERIALS AND METHODS change their modes of reproduction (Amsel- lam et al. 2001) after introduction into new In spring 2001 we contacted county weed ranges. This suggests that new environmental and pest agents, landowners, and federal agency conditions (including interactions with neigh- staff to find extensive invasions of Acroptilon boring natives) can shape an exotic species repens and Cardaria draba in Wyoming and over time (Mooney and Cleland 2001). How- Idaho. We selected these 2 weed species beever, no evidence has documented changes in cause they are highly competitive perennial native plant populations after exposure to an plants that are widely dispersed throughout exotic. Our studies focus upon the influence of the Intermountain West and are reported to invading species on the genetic diversity of form dense, monospecific stands (Watson 1980, native plant populations. We suggest that the Sheley and Stivers 1999). inherent variability within native populations We assessed more than 20 potential field may allow for their perpetuation via competi- sites, evaluating them for age of invasion, size of tive selection by exotic neighbors. Given suffi- the invaded area (>25 m2), lack of cultivation cient time, exotic neighbors may select for history, and presence of nearby noninvaded native populations that exhibit combining abil- native community. We selected 4 replicate sites ity (Harper 1977, Aarssen 1983, Aarssen and for each weed species in Wyoming and Idaho Turkington 1985, Goodwin et al. 1999, Call- (Table 1). The Greybull, Wyoming, field site away and Aschehoug 2000). was included in the analysis for both Acrop- To test these ideas, we examined species tilon and Cardaria because both exotics were presence and abundance in old (>25 years) present at high densities. All field locations sites invaded by Acroptilon repens L. (Centau- were in severe drought during 2001 and 2002 rea repens, Russian knapweed) and Cardaria (NOAA 2003). draba L. (Desv.) (Lepidium draba L. DC., hoary In 2001 and 2002, to distinguish invaded cress, whitetop) in Wyoming and Idaho. We from noninvaded positions at each field site, document species richness and abundance of we sampled exotic weed densities in 0.25-m2 natives in old invasions of Acroptilon and Car- rectangular quadrats along 2 parallel transects daria to examine the hypothesis that native (Fig. 1). Areas with high densities of Acroptilon species may remain in invasions in small num- repens or Cardaria draba were considered 2004] NATIVE PLANTS AND LONG-TERM INVASION 505

Fig. 1. Field sampling layout at all field locations. Target exotic plant densities were determined in 0.25-m2 quadrats along 2 parallel transects (T1, T2) to distinguish invaded from noninvaded positions. Species richness was recorded, and seed and leaf tissue samples were collected within the large 25-m2 areas. Individual species abundance was determined in five 1-m2 plots within each position. Sampling plots were separated by no more than 20 m. invaded, and areas with no target exotics that sites, we collected plant foliage from Poa se- were located within 20 m of high weed densi- cunda Presl., Hesperostipa comata Roem. and ties were considered noninvaded. In several Schult., and Sporobolus airoides from Acrop- instances noninvaded areas contained sporadic tilon invasions and Achnatherum hymenoides individuals of exotic plants (<1 m2). In each of (Trin. and Rupr.) Barkworth and Sporobolus 2 positions (invaded, noninvaded) we defined airoides from Cardaria invasions. These native 25-m2 sampling areas separated by <20 m grasses represent a range of reproductive sharing the same slope, aspect, and soil prop- modes from primarily inbreeding to obligate erties. Within each sampling area we estab- outcrossers. Sporobolus airoides is self-polli- lished five 1-m2 quadrats and counted grass nating and Achnatherum hymenoides is con- culm and weed stem densities within each sidered mostly selfing (Jones and Nielson quadrat. Species richness was compiled over 1989), while Hesperostipa comata and Poa se- the entire 25-m2 sampling area, combining the cunda are mixed pollinators (Fryxell 1957, Lar- 2001 and 2002 growing season samples. son et al. 2001). Drought conditions restricted To evaluate whether native diversity and the native species available for genetic analy- abundance were reduced inside invasions, we sis because many plants had senesced very documented native and total species density early in the growing season. Foliage samples (culms ⋅ m–2), species richness, and diversity were collected between 28 May and 6 June (H′; Shannon and Weaver 1949) at each site. 2002. After collection, samples were placed in Richness and diversity of the 2 positions silica gel, stored in a cooler during transport, (treatments) and 4 field sites (replicates) were and frozen to –20°C until DNA extraction. compared within each weed species using Only fresh, actively photosynthesizing leaf ma- Student’s paired t-test statistic (SAS Institute terial was collected for inter-simple sequence 1999). repeat (ISSR) analysis. ISSR analysis is a procedure used to assess Native Plant Population genetic variation in plants by presence or Genotypic Variation in absence of recognizable bands of DNA frag- the Two Positions ments (Deshpande et al. 2001, Stenstrom et al. To characterize the genetic makeup of native 2001, Tuthill and Brown 2003). ISSR markers plant populations within the 2 positions, we are based on microsatellite loci initially devel- sampled native plant foliage for anonymous oped by Zietkiewicz et al. (1994) and Gupta genetic fingerprinting. From the same study et al. (1994). Primers are designed to match 506 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 2. Summary of ISSR primers used to assess genetic variation in 4 native grass species (ACHY = Achnatherum hymenoides, HECO = Hesperostipa comata, POSE = Poa secunda, and SPAI = Sporobolus airoides). Primera Sequence ACHY HECO POSE SPAI

807 (AG)8T Good Good Good Good 813 (CT)8TFair Fair Fair Fair 815 (CT)8G Good Good Good Good 823 (TC)8C Good Good Good Good 834 (AG)8CTT Good Good Good Good 836 (AG)8CTA Good n/a n/a n/a 846 (CA)8AGT n/a n/a n/a Good 848 (CA)8AGG Good n/a n/a n/a 857 (AC)8CTG Good n/a Good Good 868 (GAA)6 Good Good Good Good 873 (GACA)4 Good Good n/a Good 881 GGGT(GGGGT)2G Good Good Good Good 891 ACTACGACT(TG)7 Fair Fair Fair Fair 894 TGGTAGCTCTTGATCA(ACTG)5 n/a Poor n/a Poor aPrimers were provided by the University of British Columbia, Nucleic Acid Service-Protein Unit. In the primer sequences (oriented 5′ ➝ 3′), letters within parentheses represent the tandemly repeated segment of the primer. Values of primers (good, fair, poor) were based on quality, quantity, and repeatability of band production within a species. If a primer was not used on a species, it is valued as n/a.

segments of tandemly repeated bases (e.g., seconds at 96°C, 45 seconds at 44°C, and 90 TCTCTC, a microsatellite) with extra base seconds at 72°C, with a final elongation at pairs of randomly chosen sequence on the 72°C for 10 minutes. PCR reactions were end, called “anchors.” These primers are used visualized following electrophoresis on 1.5% singly in polymerase chain reactions (PCR) agarose gels stained with ethidium bromide at and attach only where there is a matching micro- a standard time of 75 minutes. We used a Bio- satellite locus and adjacent sequence match- Rad Versadoc system to take direct digital pho- ing the anchor. The PCR reactions amplify tos of the gels. Gels were scored directly from segments between microsatellites to produce digital photos for presence (1) or absence (0) of multiple banding patterns that have been used obvious, repeatable bands (present in 2 sepa- for genomic fingerprinting, differentiation be- rate runs). We used 9 primers that produced tween crop cultivars, exploring population from 36 to 118 scorable, informative bands for structure (Zhou et al. 1999, Deshpande et al. each species. 2001, Tuthill and Brown 2003), elucidating Banding patterns were used to calculate phylogenetic relationships (Wolfe et al. 1998), percent polymorphic loci (P) and genetic diversity within populations (H ) for each position. constructing genomic maps, and documenting s We calculated coefficient of population differ- interspecific hybridization (Wolfe and Liston entiation (G ) and Nei’s unbiased genetic dis- 1998). ISSR methods also show great promise st tance (D) to compare positions. All genetic as a method of examining relationships among statistics were calculated using Popgen32 (Yeh individuals within populations and between et al. 1997) and treating positions as popula- populations (Meekins et al. 2001). ISSR primers tions. A locus was designated polymorphic if it used in this study are listed in Table 2. contained 2 or more alleles within a population. DNA was extracted from leaves using stan- The percentage of polymorphic loci (P) within dard 2X CTAB protocol (Doyle and Doyle a population provides a measure of genetic 1987). The 20-µL PCR reactions consisted of variability. Heterozygosity within populations 1X buffer (Sigma; 10 mM Tris-HCl, pH 8.3, 50 (Hs) is a measure of genetic variation that is mM KCl, 1.5.M MgCl2, 0.0001% gelatin), addi- useful for genes of different ploidy levels and tional 50 nmol MgCl2, 8 nmol dNTPs, 6 nmol in different organisms with different repro- primer, 10 µg BSA, 3 µL saturated betaine, 1 ductive systems (Nei 1987). This measure is unit Taq polymerase, and 1 µL DNA. Thermal more applicable for comparing across species cycling parameters were initial denaturing for that differ reproductively, as do our species. 2 minutes at 96°C, followed by 40 cycles of 30 The coefficient of population differentiation 2004] NATIVE PLANTS AND LONG-TERM INVASION 507

TABLE 3. Plant species richness and diversity (H′) in invaded and noninvaded positions in four sites for each weed species. Exotic species Site Native Native Total Total Position richness diversity richness diversity Cardaria draba Bancroft, ID Invaded 7 0.08 13 0.55 Noninvaded 4 0.06 11 0.35 Dubois, WY Invaded 9 0.46 13 0.52 Noninvaded 11 0.65 14 0.75 Greybull, WY Invaded 5 0.49 9 0.59 Noninvaded 10 0.30 14 0.44 Rock Springs, WY Invaded 10 0.25 13 0.73 Noninvaded 12 0.54 19 0.42 ± Mean sx– Invaded 6.3 ± 0.88 0.38 ± 0.05 11.0 ± 1.2 0.45 ± 0.03 Noninvaded 7.7 ± 1.22 0.34 ± 0.06 8.8 ± 1.07 0.51 ± 0.05

Acroptilon repens Greybull, WY Invaded 5 0.49 9 0.59 Noninvaded 10 0.30 14 0.44 Mud Lake, ID Invaded 11 0.30 14 0.44 Noninvaded 12 0.35 14 0.61 Rogerson, ID Invaded 3 0.08 6 0.49 noninvaded 5 0.19 6 0.41 Rock Springs, WY Invaded 10 0.63 14 0.59 Noninvaded 12 0.56 18 0.66 ± Mean sx– Invaded 6.4 ± 0.95 0.33 ± 0.07 10.1 ± 0.91 0.52 ± 0.04 Noninvaded 7.0 ± 0.62 0.35 ± 0.08 9.7 ± 0.71 0.55 ± 0.05

(Gst) measures the amount of total gene diver- RESULTS sity attributed to differences between positions. Nei’s unbiased genetic distance (D) is a Field Study statistical measure that provides a standard- Interestingly, species richness, native species ized scale for the quantification of genetic dif- diversity, and total species diversity did not ferences (Hoelzel and Dover 1991). differ between the 2 positions (invaded and To better understand patterns of variation noninvaded) across the 4 replicate sites for within and between populations, we performed either exotic species (P > 0.05), even though unweighted pair-group mean analysis (UPGMA) the exotics had been on site for ≥25 years on matrices of similarity calculated from all possible pairwise combinations of individuals (Table 3). However, abundance (culm density) using Dice’s (1945) coefficient of similarity of Achnatherum hymenoides (t1,3 = –5.14, P = (NTSYS-pc; Rohlf 1989). We ran 1000 bootstrap 0.007), Poa secunda (t1,3 = 3.12, P = 0.035), searches for each dendrogram to include con- and Sporobolus airoides (t1,3 = 6.34, P = fidence intervals. Dice’s coefficient emphasizes 0.003) was greater outside invasions of both shared traits (ISSR fragments in this study) exotic species. Abundance of Hesperostipa among individuals and ignores shared trait comata did not differ between positions for absence. either exotic species. 508 WESTERN NORTH AMERICAN NATURALIST [Volume 64

a TABLE 4. Genetic statistics ( P, H s, Gst, and D) of 4 grass species derived from invaded and noninvaded positions. Coefficient of population differentiation (Gst) and Nei’s unbiased genetic distance (D) were calculated treating positions as populations, hence the genetic distance values report between positions at each site. Species Position P Hs Gst D Hesperostipa comata Invaded (n = 10) 55.3 0.203 0.353 0.257 Noninvaded (n = 10) 46.8 0.156

Sporobolus airoides Invaded (n = 6) 47.6 0.203 0.121 0.122 Noninvaded (n = 6) 69.1 0.259

Poa secunda Invaded (n = 3) 51.4 0.230 0.324 0.446 Noninvaded (n = 3) 48.6 0.252

Achnatherum hymenoides Invaded (n = 5) 61.0 0.243 0.303 0.276 Noninvaded (n = 5) 60.2 0.234 a Percentage polymorphism (P); heterozygosity within population (Hs); coefficient of population differentiation (Gst); Nei’s unbiased genetic distance between populations (D).

Genetic Variation rated into 2 distinct, yet fairly similar (0.74), Within Positions clusters according to position (Fig. 2). Sporo- Polymorphism in the Sporobolus popula- bolus airoides individuals were more inter- tions was reduced in the invaded position (P spersed by position, but still clustered within = 47.6%; Table 4) relative to the noninvaded positions (Fig. 3). Poa secunda and Achnatherum position (P = 69.1%; Table 4). The remaining hymenoides individuals were less similar over- native species were more polymorphic in all. Poa individuals were clustered with no invaded than in noninvaded positions. Mean apparent pattern (Fig. 4), while Achnatherum heterozygosity for all populations studied was individuals formed 3 fairly distinct clusters, 2 of which seemed to separate by position (Fig. 5). 0.223, and Hs values varied among the 4 species, the most obvious difference in heterozygosity being observed in Hesperostipa comata. DISCUSSION Genetic Variation Contrary to popular views of invasions (sensu Between Positions monospecific stands), our results demonstrate For each species examined, the greatest por- that native species remain within weed inva- tion of variation was within positions. For ex- sions (although in reduced abundance). Per- haps Acroptilon repens and Cardaria draba do ample, for Sporobolus, the Gst (0.121) indicates that the proportion of genetic diversity between not change ecosystem attributes enough to ex- positions contributed 12.1% of the total ge- clude native species. However, it is important netic diversity in the individuals we sampled. to note that our field sites had no history of Hesperostipa, Achnatherum, and Poa varied cultivation and may have lacked sufficient dis- more between positions than did Sporobolus turbance of the native community to result in (Table 4). The greatest genetic distance was monospecific stands. observed between Poa secunda positions (D = There were no clear correlations between 0.446; Table 4), and D varied according to breeding system and genetic variability. Dif- species. Sporobolus airoides differed least (D ferences in Hs are more subtle but follow the = 0.122) between positions of the 4 species same pattern as the polymorphism results. sampled. When comparing the 4 species, our Gst The UPGMA cluster analysis revealed in- results are mixed. While 3 of 4 species have teresting patterns specific to each species. relatively high portions of variability attrib- Individuals of Hesperostipa comata were sepa- uted to positions, they do not follow modes of 2004] NATIVE PLANTS AND LONG-TERM INVASION 509

Fig. 2. UPGMA cluster analysis using Dice’s coefficient of similarity (NTSYS-pc) of individual native grass plants of Hesperostipa comata from an Acroptilon invasion near Mud Lake, ID. Dendrogram is based on presence and absence of ISSR bands within plants derived from the 2 positions (invaded ■, noninvaded ●) relative to exotics. Bootstrap confidence levels are indicated for clusters present in the 50% majority-rule consensus of 1000 UPGMA searches.

Fig. 3. UPGMA cluster analysis using Dice’s coefficient of similarity (NTSYS-pc) of individual native grass plants of Sporobolus airoides from an Acroptilon and Cardaria mixed invasion near Greybull, WY. Dendrogram is based on presence and absence of ISSR bands within plants derived from the 2 positions (invaded ■, noninvaded ●) relative to exotics. Bootstrap confidence levels are indicated for clusters present in the 50% majority-rule consensus of 1000 UPGMA searches. 510 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 4. UPGMA cluster analysis using Dice’s coefficient of similarity (NTSYS-pc) of individual native grass plants of Poa secunda from an Acroptilon invasion near Rock Springs, WY. Dendrogram is based on presence and absence of ISSR bands within plants derived from the 2 positions (invaded ■, noninvaded ●) relative to exotics. Bootstrap confidence levels are indicated for clusters present in the 50% majority-rule consensus of 1000 UPGMA searches.

Fig. 5. UPGMA cluster analysis using Dice’s coefficient of similarity (NTSYS-pc) of individual native grass plants of Achnatherum hymenoides from a Cardaria invasion near Dubois, WY. Dendrogram is based on presence and absence of ISSR bands within plants derived from the 2 positions (invaded ■, noninvaded ●) relative to exotics. Bootstrap confidence levels are indicated for clusters present in the 50% majority-rule consensus of 1000 UPGMA searches. 2004] NATIVE PLANTS AND LONG-TERM INVASION 511 reproduction. For example, Sporobolus airoides 1997). We know that exotic plant invasions can is a self-pollinator but has the lowest value drastically alter the environment in which they (this may be attributed to the longevity of the occur (Gordon 1998). Subsequent study is species). Genetic distance (D) results for Sporo- needed to ascertain the probable sources of bolus are comparable to Gst, and no clear pat- the genetic differences we observed and to tern emerges among the 4 species. However, document adaptive traits in native grass popula- both D and Gst results demonstrate that there tions that may contribute to native success are differences between the 2 positions within against weed invasions. each species. Our genetic data (Fig. 2, Table 4) parallel Our collections within species are limited other studies of population differentiation and to individuals from single sites, a restriction microdifferentiation in natural populations that may be important for finding differences (Thompson 1999, Gitzendanner and Soltis 2000, at the neighborhood scale (sensu Aarssen 1983). Deshpande et al. 2001). Differentiation between Had we included additional control populations positions for Poa secunda (Gst) is high when spatially separated from those we sampled, we compared with previous studies (Larson et al. could have addressed genetic differences due 2001), but we sampled fewer individuals from to spatial separation alone. There is little infor- each position, which may have elevated this mation available on the genetic variability of statistic in our results. Our sampling was neg- native grasses within local populations, and atively affected by drought throughout the subsequent study is currently underway to study, and increased sample sizes would have document such patterns. been preferred. Our genetic data suggest that population We began by asking if native plant individ- differentiation is occurring within weed inva- uals remain on sites after long-term invasion sions. Divergence in our populations may not by exotics. They did in our study, although in be solely attributed to prolonged coexistence reduced abundance. We next asked if con- with exotic species. Such differentiation could specifics differed between invaded and nonin- result from founder effects and genetic drift in vaded areas. The genotypic differences observed addition to natural selection within exotic in the native species we sampled suggest that invasions. Given the complexity of natural sys- native communities may have the ability to tems, it is likely that multiple mechanisms of respond to exotic invasion. Native species rich- evolution come into play. Reduced abundance ness and diversity were not significantly dif- of some natives within invasions indicates pop- ferent in invaded areas even though each exotic ulation bottlenecks. It is difficult to discern be- species of interest was the dominant species in tween neutral forces (i.e., founder effects, drift) that community. and selection in a study such as this. Future Our results represent an initial attempt to study is required to assess whether the genetic document intraspecific genetic variation of differences we observed translate into biologi- native species in old weed invasions, which may cally meaningful advantage against exotics. The hold important implications for restoration separation of selection from neutral forces ecology. There are many approaches to exam- should be addressed and suggests an impor- ine suitable species sources for restoration tant avenue of research. Consequently, we can- (Linhart 1995, Roundy et al. 1997, Jones and not say that the genetic differences we noted Johnson 1998), but few studies examine native are the results of selection for greater compet- genotypic variability associated with invasive itive ability with weeds. Because we do not exotics. There is a lack of evidence document- know the longevity of these species, it is possi- ing local-scale ties of intraspecific diversity to ble that natives may have simply persisted with- ecosystem function (Belnap 1995, Linhart 1995, in the invasions. Even so, such relict native Madritch and Hunter 2002, Partel 2002). individuals would have been subject to the Because competitive, or tolerant, native geno- novel selective environment of the weed inva- types may be developing with prolonged ex- sion. If Acroptilon repens and Cardaria draba posure to exotic neighbors, we suggest that have altered the growing conditions on our knowledge of post-invasion native plant popu- sites, they could have facilitated divergence in lation dynamics is critical to returning ecosys- the native populations through selection, rather tem function following exotic plant invasions. than neutral forces, in a short time (Whitlock Community invasibility may be affected by 512 WESTERN NORTH AMERICAN NATURALIST [Volume 64 resource fluctuation (Davis et al. 2000), in- AMSELLAM, L., J.L. NOYER, AND M. HOSSAERT-MCKEY. creased competitive ability (Blossey and Notzold 2001. Evidence for a switch in the reproductive biology of Rubus alceifolius (Rosaceae) towards apomixis, 1995), and diversity (Levine and D’Antonio between its native range and its area of introduction. 1999). Remnant native genotypes may be the American Journal of Botany 88:2243–2251. basis of community resistance, and mechanisms BELNAP, L. 1995. Genetic integrity: why do we care? An leading to native species coexistence with in- overview of the issues. Pages 265–266 in B.A. Roundy, vasive exotics should be investigated more com- E.D. McArthur, J.S. Haley, and D.K. Mann, compilers, Proceedings of the wildland shrub and arid land pletely. Thus far, research on native populations’ restoration symposium. USDA, Forest Service, Gen- response to invasion has been overlooked. eral Technical Report INT-315, Ogden, UT. Finally, we have not documented competi- BLOSSEY, B., AND R. NOTZOLD. 1995. Evolution of in- tive success of our native populations. Both creased competitive ability in invasive nonindigeneous exotic species are best controlled by revegeta- species: a hypothesis. Journal of Ecology 83:887–889. BOTTOMS, R.M., AND T.D. W HITSON. 1998. A systems tion with competitive grasses following chemi- approach for the management of Russian knapweed cal or mechanical control methods (Bottoms (Centaurea repens). Weed Technology 12:363–366. and Whitson 1998, Sheley and Stivers 1999, CALLAWAY, R.M., AND E.T. ASCHEHOUG. 2000. Invasive Whitson 1999). Future greenhouse and field plants versus their new and old neighbors: a mechanism for exotic invasion. Science 290:521–523. studies will allow us to document relative D’ANTONIO, C.M., AND L.A. MEYERSON. 2002. Exotic plant competitive abilities of natives from invaded species as problems and solutions in ecological restor- areas and to identify traits that may contribute ation: a synthesis. Restoration Ecology 10:703–713. to competitive success against exotic weeds. DAVIS, M.A., J.P. GRIME, AND K. THOMPSON. 2000. Fluctu- Such research efforts examining intraspecific ating resources in plant communities: a general theory of invasibility. Journal of Ecology 88:528–534. genotypes may identify a promising opportu- DESHPANDE, A.U., G.S. APTE, R.A. BAHULIKAR, M.D. LAGU, nity for revegetation following weed control. B.G. KULKARNI, H.S. SURESH, N.P. SINGH, ET AL. 2001. Genetic diversity across natural populations of ACKNOWLEDGMENTS three montane plant species from the Western Ghats, India, revealed by inter-simple sequence repeats. Molecular Ecology 10:2397–2408. We thank the many landowners and man- DICE, L.R. 1945. Measures of the amount of ecologic agers who cooperated on this project, D.E. association between species. Ecology 26:297–302. Tuthill and G.K. Brown for genetic laboratory DOYLE, J.J., AND J.L. DOYLE. 1987. A rapid DNA isolation facilities and assistance, and D. Kazmer and S. procedure for small quantities of fresh leaf tissue. Miller for early input on this project. This work Phytochemical Bulletin 19:11–15. ELLSTRAND, N.C., AND K.A. SCHIERENBECK. 2000. Hybrid- was supported by USDA-NRI Seed Grant ization as a stimulus for the evolution of invasiveness 2001-35311-09846. We appreciate the field in plants. Proceedings of the National Academy of assistance provided by students in the Univer- Science 97:7043–7050. sity of Wyoming Shrubland Ecology Lab and ELTON, C.S. 1958. The ecology of invasions by animals and plants. Methuen, London. the staff at the Rocky Mountain Research Sta- FRYXELL, P.A.1957. Mode of reproduction of higher plants. tion. Special thanks to C.L. Kinter, R.K. Peet, Botanical Review 23:135–233. E.D. McArthur, and several anonymous review- GITZENDANNER, M.A., AND P. S . S OLTIS. 2000. Patterns of ers for reading earlier versions of this manu- genetic variation in rare and widespread plant con- script. geners. American Journal of Botany 87:783–792. GOODWIN, J.R., P.S. DOESCHER, L.E. EDDLEMAN, AND D.R. ZOBEL. 1999. Persistence of Idaho fescue on de- LITERATURE CITED graded sagebrush-steppe. Journal of Range Manage- ment 52:187–198. AARSSEN, L.W. 1983. Ecological combining ability and GORDON, D.R. 1998. Effects of invasive, non-indigeneous competitive combining ability in plants: toward a plant species on ecosystem processes: lessons from general evolutionary theory of coexistence in systems Florida. Ecological Applications 8:975–989. of competition. American Naturalist 122:707–731. GUPTA, M., Y.-S. CHYI, J. ROMERO-SEVERSON, AND J.L. AARSSEN, L.W., AND R. TURKINGTON. 1985. Competitive OWEN. 1994. Amplification of DNA markers from relations among species from pastures of different evolutionarily diverse genomes using single primers ages. Canadian Journal of Botany 63:2319–2325. of simple-sequence repeats. Theoretical and Applied AL-HIYALY, S.A.K., T. MCNEILLY, AND A.D. BRADSHAW. 1988. Genetics 89:998–1006. The effects of zinc contamination from electricity HARPER, H.L. 1977. Population biology of plants. Acade- pylons: evolution in a replicated situation. New Phy- mic Press, New York. tologist 110:571–580. HILD, A.L., B. CHRISTENSEN, AND A. MAIER. 1999. Wyoming ______. 1989. The effect of zinc contamination from elec- big sagebrush seedling growth and maternal plant tricity pylons: contrasting patterns of evolution in five stand position. In: E.D. McArthur, W.K. Ostler, and grass species. New Phytologist 111:183–190. C.L. Wambolt, compilers, Proceedings: shrubland 2004] NATIVE PLANTS AND LONG-TERM INVASION 513

ecotones. USDA, Forest Service Proceedings RMRS SHANNON, C.E., AND W. W EAVER, 1949. The mathematical P-11, Ogden, UT. theory of communication. University of Illinois, HOELZEL, A.R., AND G.A. DOVER. 1991. Molecular genetic Urbana. ecology. Oxford University Press, Oxford, United SHELEY, R., AND J. STIVERS. 1999. Whitetop. In: R.L. Sheley Kingdom. and J.K. Petroff, editors, Biology and management of JONES, T.A., AND D.A. JOHNSON. 1998. Integrating genetic noxious rangeland weeds. Oregon State University concepts into planning rangeland seedings. Journal Press, Corvallis. of Range Management 51:594–606. SIEMANN, E., AND W.E. ROGERS. 2001. Genetic differences JONES, T.A., AND D.C. NIELSON. 1989. Self-compatibility in in growth of an invasive tree species. Ecology Let- ‘Paloma’ Indian ricegrass. Journal of Range Manage- ters 4:514–518. ment 42:187–190. SNAYDON, R.W., AND T.M. D AVIES. 1982. Rapid divergence LARSON, S.R., B.L. WALDRON, S.B. MONSEN, L. ST. JOHN, of plant populations in response to recent changes in A.J. PALAZZO, C.L. MCCRACKEN, AND R.D. HARRISON. soil conditions. Evolution 36:289–297. 2001. AFLP variation in agamospermous and dioe- STENSTROM, A., B.O. JONSSON, I.S. JONSDOTTIR, T. FAGER- cious bluegrasses of Western North America. Crop STROMS, AND M. AUGNER. 2001. Genetic variation Science 41:1300–1305. and clonal diversity in four clonal sedges (Carex) LEVINE, J.M. 2000. Species diversity and biological inva- along the Arctic coast of Eurasia. Molecular Ecology sions: relating local process to community pattern. 10:497–513. Science 288:852–854. THOMPSON, J.D. 1999. Population differentiation in Medi- LEVINE, J.M., AND C.M. D’ANTONIO. 1999. Elton revisited: terranean plants: insights into colonization history a review of evidence linking diversity and invasibil- and the evolution and conservation of endemic ity. Oikos 87:15–26. species. Heredity 82:229–236. LINHART, Y.B. 1995. Restoration, vegetation, and the im- TUTHILL, D.E., AND G.K. BROWN. 2003. Inter-simple se- portance of genetic and evolutionary perspectives. quence repeat (ISSR) variation in three populations Pages 271–287 in B.A. Roundy, E.D. McArthur, J.S. of Gaura neomexicana spp. coloradoensis (Ona- Haley, and D.K. Mann, compilers, Proceedings of graceae), F.E. Warren Air Force Base, Cheyenne, the wildland shrub and arid land restoration sympo- Wyoming. Western North American Naturalist 63: sium. USDA, Forest Service, General Technical 251–257. Report INT-315, Ogden, UT. WATSON, A.K. 1980. The biology of Canadian weeds. 43. LINHART, Y.B., AND M.C. GRANT. 1996. Evolutionary sig- Acroptilon (Centaurea) repens (L.) DC. Canadian nificance of local genetic differentiation in plants. Journal of Plant Science 60:993–1004. Annual Review of Ecology and Systematics 27: WHITLOCK, M.C. 1997. Founder effects and peak shifts 237–277. without genetic drift: adaptive peak shifts occur eas- MADRITCH, M.D., AND M.D. HUNTER. 2002. Phenotypic ily when environments fluctuate slightly. Evolution diversity influences ecosystem functioning in an oak 51:1044–1048. sandhills community. Ecology 83:2084–2090. WHITSON, T.D. 1999. Russian knapweed. In: R.L. Sheley MCGUINNIES, W.J., W.A. LAYCOCK, T. TSUCHIYA, C.M. YON- and J.K. Petroff, editors, Biology and management of KER, AND D.A. EDMUNDS. 1998. Variability within a noxious rangeland weeds. Oregon State University native stand of blue grama. Journal of Range Man- Press, Corvallis. agement 41:391–395. WOLFE, A.D., AND A. LISTON. 1998. Contributions of PCR- MEEKINS, J.F., H.E. BALLARD, JR., AND B.C. MCCARTHY. based methods to plant systematics and evolutionary 2001. Genetic variation and molecular biogeography biology. Pages 43–86 in D.E. Soltis, P.S. Soltis, and of a North American invasive plant species (Alliaria J.J. Doyle, editors, Molecular systematics of plants, petiolata, Brassicaceae). International Journal of Plant II. Kluwer, Boston, MA. Science 162:161–169. WOLFE, A.D., Q.-Y. XIANG, AND S.R. KEPHART. 1998. Diploid MOONEY, H.A., AND E.E. CLELAND. 2001. The evolution- hybrid speciation in Penstemon (Scrophulariaceae). ary impact of invasive species. Proceedings of the Proceedings of the National Academy of Science. National Academy of Science 98:5446–5451. 95:5112–5115. NOAA. 2003. http://www.ncdc.noaa.gov/oa/climate/station- YEH, F.C., R.C. YANG, T. BOYLE, Z.H. YE, AND J.X. MAO. locator.html. 1997. POPGENE, the user-friendly shareware for NEI, M. 1987. Molecular evolutionary genetics. Columbia population genetic analysis. Molecular Biology and University Press, New York. Biotechnology Centre, University of Alberta, Canada. PARTEL, M. 2002. Local plant diversity patterns and evolu- ZHOU, Z.H., M. MIWA, AND T. H OGETSU. 1999. Analysis of tionary history at the regional scale. Ecology 83: genetic structure of a Suillus grevillei population in a 2361–2366. Larix kiaemferi stand by polymorphism of inter-sim- ROHLF, F.J. 1989. NTSYS-pc numerical taxonomy and mul- ple sequence repeat (ISSR). New Phytologist 144: tivariate analysis system. Exeter Publishing, Setauket, 55–63. New York. ZIETKIEWICZ, E., A. RAFALSKI, AND D. LABUDA. 1994. ROUNDY, B.A., N.L. SHAW, AND D.T. BOOTH. 1997. Using Genome fingerprinting by simple sequence repeat native seeds on rangelands. Pages 1–8 in N.L. Shaw (SSR)–anchored polymerase chain reaction amplifi- and B.A. Roundy, compilers, Proceedings: using seeds cation. Genomics 20:176–183. of native species on rangelands. USDA, Forest Ser- vice, General Technical Report INT-GTR-372, Rapid Received 16 July 2003 City, SD. Accepted 13 April 2004 SAS INSTITUTE. 1999. SAS/STAT user’s guide, release 8.0. Cary, NC. Western North American Naturalist 64(4), © 2004, pp. 514–517

FLEAS (SIPHONAPTERA: CERATOPHYLLIDAE, CTENOPHTHALMIDAE) FROM RODENTS IN FIVE SOUTHWESTERN STATES

Glenn E. Haas1, Nixon Wilson2, and Chris T. McAllister3

ABSTRACT.—Nine species of fleas were collected incidental to a survey of rodents for molecular studies in Arizona, Colorado, New Mexico, Texas, and Utah, south of latitude 38°N. Three species were parasites of woodrats, Neotoma spp.: Orchopeas neotomae Augustson was confirmed for Texas, and the distribution patterns of O. agilis (Rothschild) and O. schisintus (Jordan) were more clearly defined. Four species were parasites of mice, Peromyscus spp.: Aetheca wagneri (Baker) was a new flea for P. gratus Merriam, the distribution of O. leucopus (Baker) was extended to far west Texas, Plusaetis sibynus (Jordan) was new for Utah and N. lepida, and the range of Stenoponia americana (Baker) was extended west of the Continental Divide in New Mexico. Other species included Foxella ignota (Baker) and Meringis dipodomys Kohls.

Key words: fleas, rodents, geographic distribution.

DeWalt et al. (1993) and Planz et al. (1996) Ceratophyllidae trapped rodents for molecular studies in the Aetheca wagneri (Baker) American Southwest, south of 38ºN in Arizona, Colorado, New Mexico, Texas, and Utah, be- NEW MEXICO: Catron Co., Luna, 2.4 km N, tween 1988 and 1991. T.S. DeWalt/T.A. Sprad- 1.2 km E, 2195 m: 1 , Peromyscus gratus ling (TSD/TAS) and J.V. Planz (JVP) also col- Merriam (TAS 323), 18 Mar. 1990. lected 129 fleas of 9 species, from 10 species This is one of the most common, wide-rang- of rodents. Herein, we report some changes in ing fleas of mice, especially white-footed/deer distribution, 2 new host records, and confir- mice of the genus Peromyscus (Johnson 1961, mation for 2 species, 1 each in Texas and Utah. Haddow et al. 1983, Holland 1985). Distribu- tion maps show that records decrease sharply MATERIALS AND METHODS southward toward the Mexican border, and our collection from the vicinity of Luna is Fleas were preserved in 70% ethanol until another in the county with the southernmost processed in the laboratory. Specimens were records in the state (Johnson 1961, Haddow et prepared using 10% KOH, graduated alcohols, al. 1983). Osgood’s mouse, P. gratus, is a new oil of cloves, and xylene and were then host for A. wagneri. mounted in Canada balsam on microscope slides. Specimens are deposited in the per- Foxella ignota (Baker) sonal collections of Haas and Wilson and in COLORADO: Conejos Co., Antonito, 1.6 km the Monte L. Bean Life Science Museum, N, 24.2 km W, 2652 m: 5, 12, Tho- Brigham Young University, Provo, Utah. momys talpoides (Richardson) (JVP 1662), 16 Aug. 1988. RESULTS AND DISCUSSION This complex group of fleas is wide ranging and common on pocket gophers, Thomomys Nine taxa of fleas were identified: 7 are spp. Haddow et al. (1983) mapped a void in members of the family Ceratophyllidae and 2 half of Colorado, most of Arizona and New of the family Ctenophthalmidae. The list of Mexico, and all of Texas. Jordan and Roth- species is presented below and annotated with schild (1915) and Campos (1971) listed several distributional and ecological information. records for Colorado, all from the western half

1557 California Ave., PMB 7, Boulder City, NV 89005-2796. 2Department of Biology, University of Northern Iowa, Cedar Falls, IA 50614-0421. 3Department of Biology, Texas A&M University–Texarkana, Texarkana, TX 75505-5518.

514 2004] SOUTHWESTERN RODENT FLEAS 515 of the state. Our record from the northern Orchopeas neotomae Augustson pocket gopher, T. talpoides, extends the range COLORADO: Archuleta Co., same data as of this species 2 counties further east along its JVP 1668, O. agilis above, 1. NEW MEXICO: southern border. Catron Co., Luna: 1, 1, Neotoma mexicana There are 11 named taxa of F. ignota, with Baird (JVP 1784), 14 May 1989. TEXAS: Jeff much confusion in application of the names. Davis Co., Kent (in Culberson Co.), 13.2 km S, No attempt was made to assign our specimens 12.6 km E (Long X Ranch), 1713 m: 4, to a possible subspecies. 8, N. mexicana (JVP 1932), 11 Mar. 1991. This is a parasite of Neotoma spp., primar- Orchopeas agilis (Rothschild) ily the Mexican woodrat, N. mexicana. It was ARIZONA: Mohave Co., Kingman, 9 km S, described from Grand Canyon National Park, 30 km W (near Oatman), 914 m: 2, 4, Coconino Co., Arizona (Augustson 1943a). Neotoma albigula Hartley (JVP 1872), 12 Mar. Originally its distribution was considered to 1990. COLORADO: Archuleta Co., Chromo, 2.4 be centered in the Southwest in the states of km S, 3.2 km W, 2286 m: 4, Neotoma Arizona, New Mexico, and Utah (Lewis 1975, cinerea (Ord) (JVP 1668), 22 Aug. 1988; Cone- Haddow et al. 1983). More recently, however, jos Co., Antonito, 19.3 km W (Rio Grande Na- peripheral records have been reported from tional Forest), 2652 m: 11, 9, N. cinerea northern Colorado (Campos et al. 1985) and (JVP 1659), 16 Aug. 1988; Antonito, 22.5 km central Idaho (Baird and Saunders 1992), but W (Rio Grande National Forest): 16, 28, Lewis (2000) questioned the latter record based N. cinerea (JVP 1663, 1664), 16 Aug. 1988. on a single female. Our records herein include UTAH: Washington Co., St. George, 5.1 km N, southwestern Colorado and west Texas. Hub- 2.7 km W (W of Hwy 18 and Halfway Wash), bard (1947) had secondary information that O. 1128 m: 1, Neotoma lepida Thomas (JVP neotomae occurred in west Texas, and Lewis 1878), 15 Mar. 1990. (2000) concurred, but ours is the 1st definitive Orchopeas agilis has the widest range of record with supporting collection data. As the 7 members of the Orchopeas sexdentatus emphasized by Stark (1958), this species and group commonly infesting woodrats, Neotoma O. sexdentatus (= O. agilis) can be collected spp. (Lewis 2000). Records are numerous from from the same host, as was the case with JVP the southern Rocky Mountains westward to 1668 from Colorado. the Mojave Desert, where O. agilis can be classified as the northern homolog of Orcho- Orchopeas schisintus (Jordan) peas schisintus (see account to follow). The ARIZONA: Yuma Co., Quartzsite (La Paz bushy-tailed woodrat, N. cinerea, and desert Co.), 50 km S, 7 km E: 7, 6, Neotoma woodrat, N. lepida, are common hosts of O. devia Goldman (JVP 1870), 12 Mar. 1990. agilis in the Southwest. Jordan (1929) originally described this arid- land woodrat flea from Cochise Co., Arizona, Orchopeas leucopus (Baker) as a subspecies of Ceratophyllus sexdentatus TEXAS: El Paso Co., El Paso, 13.7 km N, Baker (see Lewis 2000). Distribution records 45.1 km E, 1204 m: 2, 1, Peromyscus are now known for 6 southern Arizona coun- maniculatus blandus Osgood (JVP 1956), 18 ties and Mexico. Augustson (1943b) and Hub- Apr. 1991; El Paso, 40.5 km NE (Hueco Tanks bard (1947) listed O. sexdentatus schisintus State Park), 960 m: 1, 1, Peromyscus diffi- from Alvarado Mine, Yavapai Co., and Organ cilis (J.A. Allen) (TSD 417), 18 Apr. 1991. Pipe Cactus National Monument and Tucson, This is a common flea of Peromyscus spp. in Pima Co., respectively. Stark (1970) reported eastern states but collected less frequently in O. sexdentatus (probably schisintus) from south- the West (see Benton 1980, Haddow et al. ern Arizona. Jellison and Senger (1976) listed 1983). Eads (1950), Jellison and Senger (1976), specimens of O. sexdentatus (probably schisin- and Richerson et al. (1992) each listed 2 col- tus) collected in the 1930s in Pima Co. and the lections and McAllister and Wilson (1992) 1 Plomosa Mountains (then Yuma Co., now La from Texas. Only those of Richerson et al. Paz Co.). Hall et al. (1989) also recorded O. (1992) were from west Texas. sexdentatus (probably schisintus) from the Ajo 516 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Mountains, Organ Pipe Cactus National Mon- is prevalent in Washington Co. in the vicinity ument (Pima Co.). Lewis (2000) cited material of St. George and other arid locales (Stark from Maricopa and Yuma counties and from 1958). Elsewhere in its range in Arizona, Cali- Mexico. fornia, and Nevada, the species occurs on a wide variety of kangaroo rats and associated Plusaetis sibynus (Jordan) rodents (Eads et al. 1987). UTAH: Washington Co.: same data as JVP 1878, O. agilis above, 1. Stenoponia americana (Baker) This is a mouse flea on Peromyscus spp., NEW MEXICO: Catron Co.: same data as TAS originally described from Paradise in the Chir- 323, A. wagneri above, 1. icahua Mountains, southeastern Arizona (Jor- This is a large flea of murid rodents, partic- dan 1925). Hubbard (1947) collected 1 speci- ularly Peromyscus spp. In New Mexico only a men from a deer mouse, P. maniculatus (Wag- few specimens have been collected, mostly in ner), in the Grand Canyon, northwestern Ari- the north central region, east of the Continen- zona. The U.S. Public Health Service collected tal Divide (Williams and Hoff 1951, Stark 2 specimens from a cactus mouse, P. eremicus 1958, Haas et al. 1973). Our new record near (Baird), 19.3 km north of Pinos Altos, Grant Luna is over 42 km west of the Continental Co., New Mexico, in 1951 (T.M. Howard cor- Divide and only 8.7 km east of Arizona, which respondence). These 3 localities are the only currently lacks records. ones mapped for the United States by Had- dow et al. (1983). These authors described the ACKNOWLEDGMENTS habitat as mixed secondary forest and scrub terrain at 1700 m. Our record near St. George We thank J.V. Planz and field assistants for is from the Mohave Desert scrub biome (Turner the collection of southwestern fleas, T.A. 1982). Stark (1958) anticipated occurrence of Spradling for access to her field catalog, R.E. P. sibynus in Utah by including it in his check- Lewis for identification of some of the Orcho- list and illustrated key to the fleas of the state. peas spp. from woodrats, T.M. Howard for Interestingly, our Utah record would appear data on specimens of Plusaetis sibynus in the extralimital as Augustson and Durham (1961) Rothschild Collection of Fleas, British Museum failed to find the species among 35 species (Natural History), and the Arizona Depart- (1434 specimens) of fleas from a large survey ment of Mines and Mineral Resources for north of Lake Mead and the Colorado River in location of the Alvarado Mine. northwestern Arizona. Neotoma lepida is recorded as a host for the LITERATURE CITED 1st time. Peromyscus spp. sometimes nest in AUGUSTSON, G.F. 1943a. A new subspecies of Orchopeas Neotoma stick houses (Hoffmeister 1986, Haas sexdentatus (Baker) (Siphonaptera: Dolichopsyllidae). unpublished data), and that ecology may Bulletin of the Southern California Academy of Sci- account for this accidental host relationship. ence 42:49–51. Smit (1983) recognized 2 subspecies of P. ______. 1943b. Preliminary records and discussion of some species of Siphonaptera from the Pacific South- sibynus, with P. s. jordani (Barrera) known only west. Bulletin of the Southern California Academy of from Mexico. Our specimen from Utah could be Science 42:69–89. either P. s. sibynus (Jordan) or an undescribed AUGUSTSON, G.F., AND F. E . D URHAM. 1961. Records of fleas subspecies. Haddow et al. (1983) stated that (Siphonaptera) from northwestern Arizona. Bulletin other subspecies likely exist between the areas of the Southern California Academy of Science 60: 100–105. where these two occur. BAIRD, C.R., AND R.C. SAUNDERS. 1992. An annotated checklist of the fleas of Idaho (Siphonaptera). Idaho Ctenophthalmidae Agricultural Experiment Station Research Bulletin Meringis dipodomys Kohls 148. 34 pp. BENTON, A.H. 1980. An atlas of the fleas of the eastern UTAH: Washington Co.: same data as JVP United States. Marginal Media, Fredonia, NY. 157 pp. 1878, O. agilis above, 1, Chaetodipus penicil- CAMPOS, E.G. 1971. The Siphonaptera of Colorado. Mas- latus Woodhouse (JVP 1879). ter’s thesis, Colorado State University, Fort Collins. 274 pp. The species is normally a flea of Merriam’s CAMPOS, E.G., G.O. MAUPIN, A.M. BARNES, AND R.B. EADS. kangaroo rat, Dipodomys merriami Mearns, and 1985. Seasonal occurrence of fleas (Siphonaptera) on 2004] SOUTHWESTERN RODENT FLEAS 517

rodents in a foothills habitat in Larimer County, Col- JORDAN, K., AND N.C. ROTHSCHILD. 1915. Contributions orado, USA. Journal of Medical Entomology 22: to our knowledge of American Siphonaptera. Ecto- 266–270. parasites 1:45–60. DEWALT, T.S., E.G. ZIMMERMAN, AND J.V. PLANZ. 1993. LEWIS, R.E. 1975. Notes on the geographical distribution Mitochondrial-DNA phylogeny of species of the and host preferences in the order Siphonaptera, part boylii and truei groups of the genus Peromyscus. 6. Ceratophyllidae. Journal of Medical Entomology Journal of Mammalogy 74:352–362. 11:658–676. EADS, R.B. 1950. The fleas of Texas. Texas State Health ______. 2000 (2001). A taxonomic review of the North Department, Austin. 85 pp. American genus Orchopeas Jordan, 1933 (Siphonap- EADS, R.B., E.G. CAMPOS, AND G.O. MAUPIN. 1987. A tera: Ceratophyllidae). Journal of Vector Ecology review of the genus Meringis (Siphonaptera: Hystri- 25:164–189. chopsyllidae). Journal of Medical Entomology 24: MCALLISTER, C.T., AND N. WILSON. 1992. Two additions 467–476. to the flea (Siphonaptera: Hystrichopsyllidae, Lepto- HAAS, G.E., R.P. MARTIN, M. SWICKARD, AND B.E. MILLER. psyllidae) fauna of Texas. Texas Journal of Science 1973. Siphonaptera-mammal relationships in north- 44:245–247. central New Mexico. Journal of Medical Entomol- PLANZ, J.V., E.G. ZIMMERMAN, T.A. SPRADLING, AND D.R. ogy 10:281–289. AKINS. 1996. Molecular phylogeny of the Neotoma HADDOW, J., R. TRAUB, AND M. ROTHSCHILD. 1983. Distri- floridana species group. Journal of Mammalogy 77: bution of ceratophyllid fleas and notes on their 519–535. hosts. Pages 42–163 + 151 maps in R. Traub, M. RICHERSON, J.V., J.F. SCUDDAY, AND S.P. TABOR. 1992. An Rothschild, and J.F. Haddow, The Rothschild collec- ectoparasite survey of mammals in Brewster County, tion of fleas. The Ceratophyllidae: key to the genera Texas, 1982–1985. Southwestern Entomologist 17: and host relationships. Privately published by M. 7–15. Rothschild and R. Traub (distributed by Academic SMIT, F.G.A.M. 1983. Key to the genera and subgenera of Press, Inc., London). Ceratophyllidae. Pages 1–36 + 205 figs. and 90 HALL, W.E., C.A. OLSON, AND T.R. VAN DEVENDER. 1989. plates (by J. Navarro and R. Traub) in R. Traub, M. Late Quaternary and modern arthropods from the Rothschild, and J.F. Haddow, The Rothschild collec- Ajo Mountains of southwestern Arizona. Pan-Pacific tion of fleas. The Ceratophyllidae: key to the genera Entomologist 65:322–347. and host relationships. Privately published by M. HOFFMEISTER, D.F. 1986. Mammals of Arizona. University Rothschild and R. Traub (distributed by Academic of Arizona Press, Tucson, and Arizona Game and Press, Inc., London). Fish Department. 602 pp. STARK, H.E. 1958 (1959). The Siphonaptera of Utah. U.S. HOLLAND, G.P. 1985. The fleas of Canada, Alaska and Department of Health, Education, and Welfare, Greenland (Siphonaptera). Memoirs of the Entomo- Public Health Service, Atlanta, GA. 239 pp. logical Society of Canada 130. 631 pp. ______. 1970. A revision of the flea genus Thrassis Jordan HUBBARD, C.A. 1947. Fleas of western North America. 1933 (Siphonaptera: Ceratophyllidae) with observa- Iowa State College Press, Ames. 533 pp. tions on ecology and relationship to plague. University JELLISON, W.L., AND C.M. SENGER. 1976. Fleas of western of California Publications in Entomology 53. 184 pp. North America except Montana in the Rocky Moun- TURNER, R.M. 1982. Mohave desertscrub. Pages 157–168 tain Laboratory collection. Pages 55–136 in H.C. in D.E. Brown, editor, Biotic communities of the Taylor, Jr., and J. Clark, editors, Papers in honor of American Southwest—United States and Mexico. Jerry Flora. Western Washington State College, Desert Plants 4. Bellingham. WILLIAMS, L.A., AND C.C. HOFF. 1951. Fleas from the JOHNSON, P.T. 1961. A revision of the species of Monopsyl- Upper Sonoran Zone near Albuquerque, N. Mex. lus Kolenati in North America (Siphonaptera, Cer- Proceedings of the United States National Museum atophyllidae). USDA Technical Bulletin 1227. 69 pp. 101:305–313. JORDAN, K. 1925. New Siphonaptera. Novitates Zoologi- cae 32:96–112. Received 28 May 2003 ______. 1929. Notes on North American fleas. Novitates Accepted 24 November 2003 Zoologicae 35:28–39. Western North American Naturalist 64(4), © 2004, pp. 518–531

LEAFHOPPER ASSEMBLAGES ON NATIVE AND RESEEDED GRASSLANDS IN SOUTHWESTERN MONTANA

James A. Bess1,3, Kevin M. O’Neill1, and William P. Kemp2

ABSTRACT.—Using sweep samples, we surveyed leafhoppers (Homoptera: Cicadellidae) on grassland sites in the Gal- latin Valley of Montana during 1988 and 1991. We sampled 12 sites representing 2 habitat types defined by their dominant plant species in an undisturbed state (Stipa comata / Bouteloua gracilis and Festuca idahoensis / Agropyron spicatum). At half of the sites the native plant communities were present, whereas the remainder had been reseeded with either Agropyron spicatum (to replace the S. comata / B. gracilis assemblage) or Bromus inermis (to replace the F. idahoensis / A. spicatum assemblage). We found at least 66 species of leafhoppers among 44,428 adults collected. Seven taxa comprised 83% of all individuals collected: Doratura stylata (26%), Ceratagallia spp. (18%), Endria inimica (17%), Orocastus perpusillus (7%), Sorhoanus spp. (6%), Athysanella spp. (5%), and Psammotettix lividellus (4%). Sites with similar vegetation had broadly similar leafhopper assemblages, and assemblages differed most between the relatively xeric Stipa comata / Bouteloua gracilis sites and the more mesic sites dominated by Bromus inermis. The composition of a leafhopper assemblage at a site tended to be more similar to those on noncontiguous sites with the same overall vegetation than to those on contiguous sites with different vegetation. These patterns are likely related to the fact that many Cicadellidae are host specialists. In fact, variation in abundance of some of the most common leafhopper taxa on our sites was correlated with the percent cover of their known host plants. Our analyses of the leafhopper assemblages generally support the contention that terrestrial plant associations are among the more useful indicators of insect community composition.

Key words: biological diversity, grassland, habitat type, prairie, Homoptera, Cicadellidae.

In species richness and density of individu- and to maintain populations there may also be als, leafhoppers (Homoptera: Cicadellidae) are influenced by climate (Whitcomb et al. 1987a), among the dominant herbivorous insects on plant phenology (Whitcomb et al. 1994), and temperate grasslands (Waloff and Solomon grazing (Morris 1973, Brown et al. 1992, Kruess 1972, Cherrill and Rushton 1993, Whitcomb and Tscharntke 2002). Plant nitrogen levels, et al. 1994). Many leafhopper species are host- physical structure and age of habitats, and specific. For example, on mixed-grass prairie proximity to adjacent grasslands have also been in Kansas, 54% of the leafhopper species were identified as potential correlates of the compo- specialists on a single grass genus or on a small sition of leafhopper assemblages (Prestidge set of grass genera, whereas 7% were special- and McNeill 1983, Denno and Roderick 1990, ists on particular forbs (Whitcomb et al. 1986). Morris 1990a, 1990b, Brown et al. 1992). In Thus, many researchers have concluded that addition, a plant assemblage variable that is a the distribution of leafhoppers is correlated good predictor of the overall leafhopper with the distribution of their host plants (e.g., assemblage may not predict the distribution of Whitcomb et al. 1986, 1987a, Cherrill and rare species of concern to conservation biolo- Rushton 1993, Hicks and Whitcomb 1993, gists (Panzer and Schwartz 1998). 1996, Novotny 1994). The focus of our study was leafhopper Nevertheless, studies of grassland leafhopper assemblages on grasslands of the Gallatin Valley assemblages in North America indicate that of southwestern Montana. The grasslands in this distributions of many species do not complete- valley include at least 5 native “habitat types” ly encompass that of their host plants (Whit- (Mueggler and Stewart 1980), although many comb et al. 1986, 1987a, 1987b, 1994, Hicks sites have been reseeded with nonnative grasses and Whitcomb 1996). The ability of leafhop- (Kemp et al. 1990a). Our specific objective pers to colonize sites containing host plants was to answer several questions concerning

1Department of Entomology, Montana State University, Bozeman, MT 59717. 2USDA-ARS, Bee Biology and Systematics Laboratory, Utah State University, Logan, UT 84322-5310. 3Present address: OTIS Enterprises, 13501 S. 750 W., Wanatah, IN 46390.

518 2004] GRASSLAND LEAFHOPPERS IN MONTANA 519 the relationship of leafhopper and plant distri- blages.” All but one of the native sites were bution. Do different vegetation assemblages in paired with contiguous sites with reseeded the valley harbor different leafhopper assem- vegetation, and so they are given identical blages? Is the composition of a leafhopper numbers to designate adjoining sites: SB1/AM1, assemblage at a site similar to those on con- SB3/AM3, FA1/BM1, FA2/BM2, and FA3/BM3. tiguous sites with different vegetation or to Because SB2 and AM4 were 17 km apart, they those on noncontiguous sites with the same are given different number designations. overall vegetation? Does the richness and di- Leafhopper Sampling versity of leafhopper assemblages vary with corresponding plant variables? Does the abun- To sample leafhoppers with differing sea- dance of leafhopper species correlate with personal phenologies, we sampled 3 times in 1988 cent cover of their host plants? (26 May–2 June, 18–21 July, and 22–25 August) and 4 times in 1991 (18–27 June, 16–18 July, MATERIALS AND METHODS 6–8 August, and 13–16 September). The spread of dates in each sampling period was caused Study Sites and by travel distances between sites (up to 50 km) Vegetation Assemblages and the fact that we did not sample on cool, We sampled 12 sites in southwestern Mon- overcast days. Sweep samples taken on each tana, USA, between 45°42′ and 45°56′ north date consisted of a set of 200 sweeps taken latitude and 111°00′ and 111°37′ west longi- with a 40-cm-diameter sweep net along a lin- tude (see map in O’Neill et al. 2001 for area ear transect, each sweep traversing a horizon- containing sites). At each site we determined tal arc of 180° (parallel to the ground). Blocker percent cover of grasses, forbs, and bare ground and Reed (1976) found that suction sampling (methods and general results described in tended to capture greater numbers of leafhop- Kemp et al. 1990a, 1990b; see Table 1 for list per individuals than did sweep sampling, but of common plant species). The sites repre- the numbers of leafhopper species collected sented 2 habitat types named for their domi- were similar using the 2 methods. However, it nant plant species in an undisturbed state is likely that sweep sampling will underesti- (Mueggler and Stewart 1980). The 6 sites at mate the abundance of species that live close the western side of the valley are within the to the soil surface, a bias that we assumed was Stipa comata / Bouteloua gracilis habitat type, equal in all vegetation assemblages. a relatively xeric grassland type occurring During sorting of samples we identified and where annual precipitation is typically 20–35 counted only adults. When possible, all leaf- cm. The 6 sites at the eastern side of the val- hoppers were identified to species, but a few ley, near the foothills of the Bridger Moun- taxa were difficult to identify beyond genus tains, are within the Festuca idahoensis / Agro- because of similarities in adult features or pyron spicatum habitat type, a relatively mesic because of taxonomic problems. We identified grassland type occurring in areas generally 44 species (Table 2) that included 67.8% of the receiving 35–50 cm annual precipitation. Half leafhoppers collected. Another 4 uncommon of the sites in each habitat type still have their taxa (<0.2% of the leafhoppers) identified to native plant communities, and so are referred genus were each apparently represented by a to here as SB (Stipa / Bouteloua) and FA (Fes- single species: Dikrella sp., Empoasca sp., Lae- tuca / Agropyron). However, 3 sites in each vicephalus sp., and Lonatura sp. Nine taxa that habitat type were reseeded as part of range included 32.0% of the leafhoppers collected management programs in the 1960s. Three SB are listed as “unresolved” genera: Athysanella sites were reseeded with crested wheatgrass spp., Ceratagallia spp., Cloanthanus spp., Dikra- (Agropyron cristatum) and (at 1 site) alfalfa neura spp., Elymana spp., Hecalus spp., Scle- (Medicago sativa), and are referred to here as roracus spp., Sorhoanus spp., and Texananus AM. Three FA sites were reseeded with smooth spp. In 5 of these genera we identified 14 brome (Bromus inermis) and (at 2 sites) alfalfa; species from male specimens (Table 2). We do these are referred to as BM. We refer to each not include species-level counts for these taxa sampling location as a “site” and sets of sites because females were difficult to identify to with similar vegetation as “vegetation assem- species and we could not always associate 520 WESTERN NORTH AMERICAN NATURALIST [Volume 64 0.03 5.3 11.3 26.8 41.0 0.4 — ——— ——— 20.1 57.3 17.6 19.0 11.011.2 — — — — — — on at least 1 site, (2) less common species that 3.5 1.25 0.1 — — — 0.03 0.9 2.7 2.2 12.3 ommon grass and forb at each site are highlighted in bold. 8.3 —————— —————— —————— —————— 12.2 21.9 23.0 ——————0.1 ercent cover at each site P 0.03 — 0.7 0.8 0.3 0.1 4.1 0.5 ———————— 0.03 0.7 0.4 4.45 1.43 2.1 — — — ————————— ————————— 31.1 49.1 29.3 37.3 4.6 —————————— ————3.10.4———— 2.1 0.1 0.7 ——— —————— ——————————6.4 ——————2.31.6———— ——————0.6 —————— ——————0.5 ————————— ——————2.13.46.3——— ——————7.5 ——————2.46.1 ————0.2———— ———— SB1 SB2 SB3 AM1 AM3 AM4 FA1 FA2 FA3 BM1 BM2 BM3 23.951.0 10.8 50.2 13.8 62.4 16.3 31.7 35.3 51.0 34.0 30.4 6.8 31.320.4 2.7 35.5 10.6 4.4 35.2 25.1 8.4 23.5 35.3 1.6 68.0 34.0 79.20 6.3 0.9 30.6 30.6 49.4 18.4 1.1 6.0 36.2 36.0 26.9 17.1 ______(Host) L. Stapf — 0.03 — 0.05 0.2 0.05 1.5 0.9 0.7 0.7 0.2 — (Pursh) 0.4 — 0.3 — — — 4.9 8.2 6.0 — — — ydb. 4.0 0.1 (L.) (Hook.) illd. — Elmer L. R (H.B.K.) W Pursh L. L. 1.1 0.7 7.8 1.5 — 1.7 0.5 — — — — — L. Muhl. 1.3 — 0.2 0.4 asey 1.9 2.0 2.1 — 0.3 sp. Leyss in. & Rupr. rin. V T Nutt. 0.4 7.5 4.9 L. Tr L. Gaertn. sp. — — 1. Percent cover of the major grass and forb species at each site. Species listed include (1) those that occurred >5% 1. Percent (Pursh) Nutt. Beauv. Beauv. Lag. 10.0Lag. 3.7 20.4 — 0.4 0.8 icia americana upinus sericeus ycopodium ABLE estuca idahoensis oa pratensis oa sandbergii T Agropyron cristatum Alyssum desertorum Antennaria Agropyron intermedium Artemisia frigida Balsamorhiza sagittata Stipa comata Agropyron smithii Agropyron spicatum Bouteloua gracilis Stipa viridula Cerastium arvense Bromis tectorum Carex filifolia F Koeleria cristata P P Bromis inermis L L Medicago sativa Melilotus officinalis Microsteris gracilis V ARE GROUND LL GRASSES LL FORBS were the most common forb at a site, and (3) several other species referred to in the text. Percent cover values for the most c Percent were the most common forb at a site, and (3) several other species referred to in text. B A A 2004] GRASSLAND LEAFHOPPERS IN MONTANA 521

TABLE 2. Taxa of leafhoppers collected at Gallatin Valley sites. Taxon number refers to that used in Figures 1 and 4. Taxon Taxon number Taxon number Taxon 1 Amblysellus grex (Oman) 30 Forcipata loca (DeLong and Caldwell) 2 Attenuipyga platyrhynchus minor (Osborn) 31 Frigartus frigidus (Ball) 3 Athysanella spp.1 32 Hardya dentata (Osborn and Ball) 4 Athysanus argentarius Metcalf 33 Hebecephalus rostratus Beamer and Tuthill 5 Auridius auratus (Gillette and Baker) 34 Hecalus spp. Stål5 6 Auridius helvus (DeLong) 35 Idiodonus aurantiacus (Provancher) 7 Auridius ordinatus (Ball) 36 Laevicephalus sp. DeLong 8 Balclutha neglectus (DeLong and Davidson) 37 Latalus missellus (Ball) 9 Balclutha punctata (Thunberg) 38 Lonatura sp. Osborn and Ball 10 Ballana veruta Van Duzee 39 Macrosteles quadrilineatus (Forbes) 11 Ceratagallia spp.2 40 Mesamia coloradensis (Gillette and Baker) 12 Chlorotettix unicolor (Fitch) 41 Mocuellus caprillus Ross and Hamilton 13 Cloanthanus spp.3 42 Neocoelidia tumidifrons Gillette and Baker 14 Colladonus geminatus (Van Duzee) 43 Norvellina seminuda (Say) 15 Colladonus montanus (Van Duzee) 44 Orocastus labeculus DeLong 16 Commellus sexvittatus (Van Duzee) 45 Orocastus perpusillus (Ball and DeLong) 17 Cuerna striata (Walker) 46 Paraphlepsius occidentalis (Baker) 18 Deltocephalus valens Beamer and Tuthill 47 Pinumius areatus (Stål) 19 Dikraneura spp. Hardy4 48 Prairiana cinerea (Uhler) 20 Dikrella sp. 49 Prairiana subta Ball 21 Diplocolenus configuratus (Uhler) 50 Psammotettix lividellus (Zettinger) 22 Doratura stylata (Boheman) 51 Rosenus cruciatus (Osborn and Ball) 23 Elymana circius Hamilton 52 Scleroracus spp.6 24 Elymana spp. 53 Sorhoanus spp.7 25 Empoasca sp. 54 Stenometopielus cookei (Gillette) 26 Endria inimica (Say) 55 Streptanus confinis (Reuter) 27 Endria rotunda (Beamer) 56 Texananus spp.8 28 Flexamia abbreviata (Crumb) 57 Xerophloea viridis (Fabricius) 29 Flexamia flexulosa (Ball)

1Includes Athysanella acuticauda Baker, A. attenuata Baker, A. occidentalis Baker, A. robusta (Osborn), A. sinuata (Osborn), A. terebrans (Gillette and Baker), A. utahna (Osborn). 2Includes Ceratagallia (Aceratagallia) uhleri (van Duzee). 3None identified to species. 4Dikraneura shoshone DeLong and Caldwell. 5Includes Hecalus major (Osborn), H. viridis (Uhler). 6None identified to species. 7Includes Sorhoanus debilis (Uhler), S. flavovirens (Gillette and Baker), S. orientalis (DeLong and Davidson). 8None identified to species.

them with males in samples. Some samples, in on the ordination plot reflects increased simi- fact, contained only female specimens of par- larity of assemblages from those sites. In a ticular taxa. species-based DCA, decreasing distance between taxa indicates that they tend to occur at Statistical Analyses the same sites. To restrict analysis to a single We conducted chi-square goodness-of-fit taxonomic level, the DCA did not include the tests to compare, among sites within the same unresolved genera, whose abundances were vegetation assemblage, the total number of examined using Kruskal-Wallis analyses. Using individuals collected in most common taxa; Spearmann’s rank correlations, we examined expected values were based on the null hy- the relationships of leafhopper taxon richness pothesis of equal numbers of individuals per (number of species plus number of unresolved site. We used detrended correspondence genera) to plant species richness, and leafhop- analysis (DCA) to examine patterns of similar- per taxon diversity to plant species diversity. ity in leafhopper communities among sites Using counts for each leafhopper taxon and (CANOCO version 3.12). DCA is an ordina- total cover values for plant species, we calcu- tion technique that summarizes counts from a lated diversities for each site using Hill’s #2 site × taxon matrix (ter Braak 1988). In a site- diversity index (Hill 1973), which has low sen- based DCA, decreasing distance between sites sitivity to small sample sizes (Magurran 1988). 522 WESTERN NORTH AMERICAN NATURALIST [Volume 64

TABLE 3. Total number of leafhoppers collected at each site in 1988 and 1991 and total richness and diversity values for leafhoppers and plants. Plant Number of Number Leafhopper Plant diversity leafhoppers collected leafhopper diversity species (Hill’s #2 ______taxa (Hill’s #2 Sites richness index) 1988 1991 collected index) SB1 21 3.58 1277 1652 29 4.66 SB2 16 2.21 2365 1267 27 2.40 SB3 19 4.20 705 1458 22 3.28 AM1 7 2.09 1674 1139 25 1.47 AM3 15 1.36 271 654 20 4.77 AM4 6 1.15 506 273 19 2.83 FA132 14.78 713 1213 28 5.56 FA227 7.89 479 953 27 4.27 FA330 7.05 602 710 34 4.46 BM1 8 2.43 2119 1045 19 2.00 BM2 11 1.44 1055 5282 26 2.97 BM3 11 2.91 5332 11,684 30 4.66

The index ranges upward from 0, a value of 0 grass cover did not differ among assemblages occurring in assemblages consisting of 1 species (P = 0.41). (Ludwig and Reynolds 1988). We also used Overall Leafhopper Fauna Spearmann’s correlations to test the hypothesis that variation in abundance of particular Seven leafhopper taxa constituted 83% of taxa correlates with percent cover of known or the 44,428 adult leafhoppers collected (Table 2): suspected host plants. Although the literature Doratura stylata (26%), Ceratagallia spp. (18%), contains numerous host records for leafhop- Endria inimica (17%), Orocastus perpusillus pers, we analyzed only species for which at (7%), Sorhoanus spp. (6%), Athysanella spp. least 100 individuals were collected and which (5%), and Psammotettix lividellus (4%). Nine- were found on >50% of the 12 sites. teen taxa were represented by fewer than 10 specimens each. We found no differences among RESULTS vegetation assemblages in the number of leafhoppers collected in 1988 (Kruskal-Wallis Plant Assemblages test, P = 0.15) or 1991 (P = 0.08; Table 3). Among the 58 species of plants collected, 3 Leafhoppers in Stipa comata/ occurred at more than half of the sites, Stipa Bouteloua gracilis Assemblages comata, Alyssum desertorum, and Microsteris gracilis, although neither of the latter 2 had The 8724 leafhoppers collected on SB sites high cover values on any site (Table 1). Among included 28 species plus 6 of the unresolved the 36 species found on 3 or fewer sites, 28 genera (Fig. 1). Six taxa made up >90% of the occurred in only 1 type of vegetation assem- leafhoppers collected: Orocastus perpusillus blage: 6 from SB, 16 from FA, and 3 each from (33%), Athysanella (16%), Ceratagallia (22%), AM and BM. Overall, the 4 vegetation assem- Orocastus labeculus (9%), Mocuellus caprillus blages differed in bare ground cover (Kruskal- (6%), and Sorhoanus spp. (5%). Relative abun- Wallis test, P = 0.03), AM and BM sites hav- dance of all these taxa varied among SB sites ing 5 of the 6 highest values, and in forb cover (Fig. 2). Seven, mostly uncommon, species (P = 0.06), FA sites having the 3 highest values. were collected only on SB sites: Colladonus The assemblages also differed in plant species montanus (8 individuals), Endria rotunda (3), richness (P = 0.02) and in plant species diver- Flexamia abbreviata (5), F. flexulosa (100), Pin- sity (P = 0.03; Table 3), the highest values in umius areatus (2), Prairiana subta (2), and Strep- both comparisons being from FA sites. Total tanus confinis (1). 2004] GRASSLAND LEAFHOPPERS IN MONTANA 523

Fig. 1. Abundances of each species of leafhopper (both years combined). Numbers on x-axis refer to numbers given each species in Table 2. Taxa on SB sites were ranked from most to least abundant, and this order was used for AM assemblages; taxa on FA sites were ranked from most to least abundant, and this order was used for BM assemblages. Open bars are taxa consisting of 1 species; black bars indicate the unresolved genera.

Leafhoppers in Agropyron cristatum / but not SB sites was P. lividellus (661 vs. 71 Medicago sativa Assemblages individuals). The 4517 leafhoppers collected on AM sites Leafhoppers in Festuca idahoensis / included 26 species plus 6 of the unresolved Agropyron spicatum Assemblages genera (Fig. 1). Three taxa made up 74% of the leafhoppers: Ceratagallia spp. (57%), P. lividel- The 4670 leafhoppers collected on FA sites lus (15%), and Athysanella spp. (13%). Vari- included 33 species plus all 9 unresolved gen- ability among sites was striking for all 3 taxa era (Fig. 1). Members of 3 taxa made up 54% (Fig. 2). No other taxon comprised more that of the leafhoppers collected (Fig 1): Sorhoanus 4% of the leafhoppers collected on AM sites. (41%), Ceratagallia (13%), and Chlorotettix uni- Ballana veruta and Norvellina seminuda (2 color (7%). All 3 varied in abundance among specimens each) were collected only on AM FA sites (Fig. 2). The next most common taxa sites. were Dikraneura spp. (4%), P. lividellus (4%), Comparing AM leafhoppers with those on and O. labeculus (4%). Four species were SB sites (which contain the native vegetation found only on FA sites: Dikrella sp. (2 speci- formerly on AM sites), 13 taxa found on SB mens), Hardya dentata (4), Idiodonus auranti- sites were absent from AM sites, whereas AM acus (2), and Lonatura sp. (18). sites contained 11 taxa absent from SB sites. Comparison of the abundance rankings of Other notable differences included the rela- the species collected in SB and FA (the 2 native tive rarity on AM sites of O. perpusillus (17 vs. assemblages) indicates that 9 species present 2862 individuals on SB), O. labecullus (71 vs. on SB sites were absent on FA sites, most 789 individuals), and M. caprillus (8 vs. 491 notably F. flexulosa and Stenometopielus cookei. individuals). The only taxon common on AM Conversely, we collected 15 species on FA sites 524 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 2. Total number of individuals of most common taxa collected at each site. For each species/vegetation assemblage, we conducted a chi-square value goodness-of-fit test, the expected values being based on the null hypothesis of equal numbers of individuals per site. All chi-square values were significant at P < 0.001. 2004] GRASSLAND LEAFHOPPERS IN MONTANA 525 not found on SB sites, the most abundant be- etation assemblages, indicating that they shared ing E. inimica and D. stylata. Orocastus per- relatively similar leafhopper assemblages. pusillus, the most common leafhopper identi- Unlike other BM sites, BM1 had few D. sty- fied to species on SB sites (2862 specimens), lata and E. inimica. BM1 also shared high was the 8th most abundant species on FA sites numbers of Ceratagallia spp. with AM1 (which (139). FA sites had among the highest values were not abundant on AM3 or AM4), although for leafhopper taxon richness and diversity that genus was not included in the DCA. In (Table 3). the DCA for 1988, axis 1 accounted for 27.7% of variance in the data, whereas axis 2 Leafhoppers in Bromus inermis / accounted for 12.6%. In the 1991 site-based Medicago sativa Assemblages DCA, sites with similar vegetation were less The 26,517 leafhoppers collected on BM closely associated than in 1988, although some sites included 24 species and all 9 unresolved separation by vegetation type was still evident, genera (Fig. 1). Two species comprised 72% of particularly along axis 1. Axis 1 accounted for the leafhoppers on BM (Fig. 1): D. stylata 39.5% of variance in the data, whereas axis 2 (44%) and E. inimica (28%). Some of the rela- accounted for 7.2%. In contrast to 1988, there tively less abundant taxa occurred in numbers were several instances in 1991 in which native comparable to common taxa on other vegeta- sites clustered fairly closely with their con- tion assemblages: Ceratagallia spp., Amblysel- tiguous reseeded sites: SB3 was more closely lus grex, Diplocolenus configuratus, P. lividel- associated with AM3 than with other SB sites, lus, and Dikraneura spp., which together made and BM1 was more closely associated with up another 21% of the leafhoppers on BM sites. FA1 than with other BM sites. In both years All 7 taxa varied in abundance among sites the separation among SB sites appeared to be (Fig. 2). Athysanus argentarius (135 specimens) due to higher abundance of the certain taxa on was collected only on BM sites. 1 site, but not the other 2: M. caprillus on Comparing BM leafhoppers with those on SB1, Orocastus spp. on SB2, and F. flexulosa FA sites (which contain the native vegetation and Athysanella spp. on SB3. Similarly, com- formerly on BM sites), 10 species were pres- pared to FA1 and FA2, FA3 had the lowest ent on FA but absent from BM (e.g., Atten- numbers of C. unicolor and Macrosteles quadri- uipyga platyrhynchus), whereas BM sites con- lineatus in both years and of Elymana circius tained only 2 species absent from FA (A. grex in 1991. and A. argentarius; Fig. 1). The most striking We observed similar patterns when we com- differences were in abundance of D. stylata pared the species-based DCA for the 2 years and E. inimica on contiguous sites. We col- (Fig. 4). For 1988, two clusters of species occur lected 10,371 individuals of these species on on opposite ends of axis 1. On the left side is a BM3 but only 83 on FA3; a similar comparison cluster of 11 species. For 10 of these, >50% of for BM2 and FA2 was 4934 vs. 33 specimens. the specimens were collected on SB sites; the It is also clear that the SB and BM sites dif- exception was Hebecephalus rostratus (species fered considerably. The 3 most common taxa 33 on the plot), 49% of which were found on identified to species on SB sites (O. perpusil- SB sites. Cuerna striata (species 17), below lus, O. labeculus, and M. caprillus), as well as and to the right of this cluster, was found pri- Athysanella spp., were rare to absent on BM marily on SB sites (54% of specimens), but it sites. BM sites, on the other hand, were domi- was nearly as common on AM1 (37%). On the nated by 2 species not present on SB sites (D. right side of the plot is a tight cluster of 8 stylata and E. inimica). species for which 96%–100% of the specimens were collected on BM sites. Among the 5 spe- Ordination Analyses cies between these 2 clusters were C. unicolor In the 1988 site-based DCA (Fig. 3), SB (species 12, 96% FA) and Balclutha punctata and FA native habitats formed fairly discrete (species 9, 99% BM). The remaining 3 species clusters separated from the reseeded forms of were present in sizable proportions in more each native assemblage. AM3 and AM4 also than 1 vegetation assemblage: Colladonus gem- form a pair, as do BM2 and BM3. However, inatus (species 14, 35% BM but present in all AM1 and BM1 are closer to each other on the 4 assemblages), M. quadrilineatus (species 39, plot than to other sites in their respective veg- 66% BM, 34% FA), and P. lividellus (species 526 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 3. Site-based detrended correspondence analysis. SB = ●; FA = ■; AM = ●; BM = ■.

50, 51% AM, 36% BM, 12% FA). On the 1991 Distribution of the species plot, discrete clusters of species were Unresolved Genera much less apparent, although species showing an affinity for SB sites were again on the left In both years numbers of Athysanella spp. side of the plot, whereas those most common and Sorhoanus spp. differed among sites, on BM sites were on the right side. In most Athysanella being most abundant on SB sites cases these were the same species in those and Sorhoanus spp. on FA sites (Fig. 5). Nei- locations in the 1988 ordination. The more widespread species C. geminatus, M. quadrilineatus, ther Ceratagallia spp. nor Dikraneura spp. ex- and P. lividellus were again in the middle of hibited differences in 1988, but Dikraneura spp. the plot. were more abundant on BM sites in 1991. 2004] GRASSLAND LEAFHOPPERS IN MONTANA 527

Fig. 4. Species-based detrended correspondence analysis. Analyses included all species of grasshoppers (46 in 1988 and 37 in 1991, but not the unresolved genera), but only those for which at least 20 individuals were collected are included on the plot. Symbols assigned based on which vegetation assemblage yielded >50% of the specimens collected for a given species: SB = ●; FA = ■; AM = ●; BM = ■; no vegetation assemblage had more than 50% of the specimens = ∆.

Plant-Leafhopper dance and percent cover of reported hosts Correlations (Table 4). Leafhopper taxon richness was significantly DISCUSSION correlated with plant species richness (Spear- mann correlation, r = 0.65, P = 0.02, Table 3), Our results are consistent with the hypoth- but leafhopper diversity was not correlated esis that the distribution of leafhopper species with plant diversity (r = 0.44, P = 0.14). For and the composition of local assemblages are all 8 of the relatively widespread and abun- related to the composition of local plant assem- dant leafhopper taxa that we examined, we blages (DeLong 1965, Whitcomb 1987, Whit- found significant correlations between abun- comb et al. 1987a, 1987b, 1994; but see Brown 528 WESTERN NORTH AMERICAN NATURALIST [Volume 64

the ordinations. Two taxa, C. unicolor (92%) and Sorhoanus spp. (67%), were found mainly on FA sites, but none were located predominantly on AM sites. Of 14 taxa collected in only a single type of vegetation assemblage, only F. flexulosa (100 specimens) and A. argentarius (135) were represented by >18 individuals. There are perhaps several reasons why we did not observe better separation among sites with different vegetation assemblages in the site-based DCA. First, only a few leafhoppers showed their greatest abundance on FA sites, and none did on AM sites. Second, there was considerable variation within vegetation assemblages in the abundance of some of the common species. Third, the abundant unresolved genera Athysanella, Ceratagallia, and Sorhoanus Fig. 5. Number of individuals of the 4 major unresolved were not included in the analysis. If they had genera collected in each of the vegetation assemblages in been identified to species, differences among 1988 and 1991 (N = 4). P-values refer to results of habitats may have been more apparent. For ex- Kruskal-Wallis tests. ample, Ceratagallia feeds on legumes (Hamil- ton 1998) and was abundant in SB (1894 speci- et al. 1992). The greatest differences occurred mens), AM (2578), and BM (2716) assemblages. between the relatively xeric SB sites in the The distribution of legumes varied among western part of the valley and the more mesic, assemblages. Melilotus officinalis was found higher-elevation BM sites in the east near the only on the 3 SB sites and 1 AM site, whereas Bridger Mountains. Leafhopper assemblages at Vicia americana was the most common forb on a site tended to be more similar to those on the 2 other AM sites, and Lupinus sericeus noncontiguous sites with the same overall veg- was found only on the 3 BM sites. Medicago etation than to those on contiguous sites with sativa was absent from SB and FA sites, and different vegetation. However, there were also Astragalus spp. were not found on BM sites. differences in leafhopper assemblages when Thus, depending on host preferences, the we compared sites within the same vegetation species of Ceratagallia in the Gallatin Valley assemblage. This is not surprising because may have consistently sorted themselves among each site had its own unique frequency distri- vegetation assemblages. However, this remains bution of plant cover values despite similari- to be examined. A similar situation may occur ties in dominant grass species. for Athysanella, whose host preferences may Among 29 leafhopper taxa for which we extend to particular species within grass gen- found >50 specimens, 16 occurred primarily era (Whitcomb et al. 1986, 1994). within 1 type of vegetation assemblage (defined Although leafhopper species frequencies here as having more than two-thirds of the differed between SB and AM sites, reseeding individuals collected in 1 type of assemblage). the native SB sites to create A. cristatum– Six occurred mainly on SB sites: F. flexulosa dominated pastures apparently did not result (100% of all specimens), S. cookei (98%), O. in habitats with a unique array of dominant perpusillus (95%), M. caprillus (80%), O. labecul- leafhopper species. All species present on SB lus (73%), and Auridius helvus (68%). Another sites but not AM sites (and vice versa) con- 8 were found mainly on BM sites: A. argentarius sisted of <20 individuals in our samples. Only (100%), A. grex (99.8%), L. missellus (99.6%), P. lividellus appeared to benefit from reseed- D. stylata (99%), E. inimica (99%), D. configu- ing to A. cristatum, although the Stipa-feeding ratus (98%), B. punctata (97%), and E. circius Orocastus apparently declined following the (87%). Together, these 14 species constituted change. However, the creation of Bromus iner- >88% of the leafhoppers identified to species, mis pastures that also contained greater cover so contrasts between these 2 groups likely in- of Poa pratensis was associated with a major fluenced the positioning of SB and BM sites in shift in the dominant leafhopper species. This 2004] GRASSLAND LEAFHOPPERS IN MONTANA 529

TABLE 4. Spearmann’s correlation (r) between leafhopper taxon abundance and cover of known or suspected host plants; significance levels: P > 0.05, *P < 0.05, **P < 0.001. Data on leafhopper abundance were combined for 1988 and 1991. Number of Number of sites on which Leafhopper Plant taxon r individuals taxon found Source of host records Athysanella spp. Bouteloua gracilis 0.76** 2,231 12 Whitcomb et al. 1994a Ceratagallia spp. Fabaceae 0.66* 7,781 12 DeLong 1948, Hamilton 1998 Diplocolenus configuratus Poa spp. 0.89** 746 7 DeLong 1948 Doratura stylata Poa spp. 0.92** 11,720 7 DeLong 1948 Endria inimica Poa spp. 0.96** 7,655 7 Bess personal observation Mocuellus caprillus Agropyron smithii 0.59* 613 9 K. Hamilton and R. Whitcomb personal communication Orocastus labeculus Stipa comata 0.94** 1,089 9 K. Hamilton personal communication O. perpusillus Stipa comata 0.95** 3,018 11 K. Hamilton personal communication aIncludes records for many Athysanella on Bouteloua spp., although none on the list were collected at our sites.

was particularly evident in numbers of D. sty- grass was significantly correlated with Athysa- lata and E. inimica which, overall, were nearly nella spp. abundance. Flexamia flexulosa 100 times higher on BM than on FA sites. occurred only on the 3 SB sites where B. gra- Although there were also considerably greater cilis was common. The abundance of Cerata- numbers of A. grex, A. argentarius, D. configu- gallia spp. increased with percent cover of ratus, and L. missellus on BM sites, only C. Fabaceae. Ceratagallia were most abundant on unicolor and Sorhoanus spp. showed an appre- SB1, SB3, and AM1, all of which had high ciable decline relative to FA sites. Across the Melilotus officinalis (yellow sweetclover) cover, Gallatin Valley leafhopper diversity may actu- and BM1 with its abundant Medicago sativa ally be higher at the present due to the exis- (alfalfa). tence of both native and reseeded grasslands. As noted by Brown et al. (1992) in their The relationship of leafhopper distribution study of British grasslands, much variability in to plant distribution is also evident when we leafhopper community composition is left un- examined species with known host plant pref- explained by host plant availability. This could erences. Abundances of D. configuratus, D. perhaps be related to the predominance of stylata, and E. inimica were positively corre- cool-season grasses on British grasslands, and lated with percent cover of Poa. All 3 exhibit also applies to northern U.S. grasslands ited their greatest abundance on BM sites, (Whitcomb et al. 1986). In addition, abiotic where Poa pratensis was the 1st or 2nd most conditions differed among our sites, the FA abundant grass. The abundance of Mocuellus and BM sites (1450–1700 m elevation) being caprillus, which feeds on Agropyron smithii, cooler and receiving about 30% more precipi- correlated with percent cover of that species, tation than the SB and AM sites (1300–1400 m; which was most common on SB sites. Several Kemp et al. 1990a). Although climate cannot other genera most common on SB sites were account for differences in leafhopper assem- likely feeding on the 2 grasses that defined blages between contiguous SB/AM or FA/BM that vegetation assemblage. Both species of sites, there may have been microclimatic dif- Orocastus were most abundant here and are ferences driven by variation in vegetation reportedly Stipa feeders. Whitcomb et al. (1986, height, density, and patchiness. In fact, Brown 1994) indicate that many Athysanella feed on et al. (1992) found that leafhopper assemblages Bouteloua, though their leafhopper species list were affected by grazing-induced changes in does not overlap with ours. They also indicate plant architecture. that some Athysanella feed on grasses of the Urbanization, cultivation, grazing, reseed- genera Aristida, Buchloe, Hilaria, and Sporo- ing, and invasion of exotic weeds have frag- bolus, but these 4 genera did not occur on our mented and diversified the grasslands within sites. Restricting our analysis to Bouteloua the Gallatin Valley. Habitat fragmentation also gracilis, we found that percent cover of this isolates small areas from patches of similar 530 WESTERN NORTH AMERICAN NATURALIST [Volume 64 vegetation that form potential sources of mi- many Cicadellidae are host specialists (Whit- grants. Thus, the distribution of leafhoppers in comb et al. 1994), while most grassland Acridi- the valley may be influenced by distances be- dae are generalists that, at most, restrict their tween patches of suitable habitat, combined feeding to grasses vs. forbs (Mulkern et al. 1969). with limitations on the dispersal of leafhoppers (Whitcomb et al. 1987a, Brown et al. 1992, ACKNOWLEDGMENTS Stiling 1994). However, recency of introduction was clearly not always a limiting factor. We thank Derek Blocker and Ron Panzer Athysanus argentarius and D. stylata were com- for assistance in determination of leafhopper mon on BM sites and uncommon or absent in species, Marni Rolston for assistance with data other vegetation assemblages. Both A. argentar- analysis, and Jeffrey Holmes for collecting the ius (Beirne 1956) and D. stylata (Hamilton 1983) vegetation data. Robert Whitcomb and K.G.A. have spread rapidly westward following their Hamilton provided valuable comments on the introduction to eastern North America in the manuscript and provided information on leaf- last century. This large-scale range expansion hopper host records. Karen Bess, Matthew suggests that their limited distribution in the Lavin, Noel Pavlovic, Bret Olson, Catherine valley is not due to poor dispersal capabilities. Seibert, and David Wachter also provided One concern in studies of this type is that a assistance. This work was conducted with sup- strong effect of habitat area on insect species port from the USDA-ARS Rangeland Insect richness may confound correlations between Laboratory, the Montana State University plant variables and insect richness. However, Mountain Research Center, and the Montana in an analysis of leafhopper diversity on tall- Agricultural Experiment Station. grass prairie remnants in the midwestern U.S., Panzer and Schwartz (1998) found that, when LITERATURE CITED the effect of habitat area was removed, plant BEIRNE, B. 1956. Leafhoppers (Homoptera: Cicadellidae) species richness and generic richness remained of Canada and Alaska. Canadian Entomologist 138 as significant predictors of leafhopper species (supplement 2):1–180. richness. Kruess and Tscharntke (2000) noted BLOCKER, H., AND R. REED. 1976. Leafhopper populations of a tallgrass prairie (Homoptera: Cicadellidae): col- that fragmentation of agricultural landscapes lecting procedures and population estimates. Journal caused greater reductions in parasitoid than of the Kansas Entomological Society 49:145–154. herbivore diversity. BROWN, V., C. GIBSON, AND J. KATHIRIHAMBY. 1992. Commu- We have now examined the distribution of nity organisation in leafhoppers. Oikos 65:97–106. CHERRILL, A., AND S. RUSHTON. 1993. The Auchenorrhyn- both leafhoppers and grasshoppers (Kemp et cha of an unimproved moorland in northern Eng- al. 1990a, 1990b, 2002, Wachter et al. 1998) land. Ecological Entomology 18:95–103. across different vegetation assemblages in the DELONG, D. 1948. The leafhoppers, or Cicadellidae, of Illinois (Eurymelinae-Balcluthinae). Bulletin of the Gallatin Valley. A comparison of the distribution Illinois Natural History Survey 24:1–376. patterns of the 2 groups is of interest because ______. 1965. Ecological aspects of North American leaf- these 2 families include many of the important hoppers and their role in agriculture. Bulletin of the insect herbivores on grasslands. Variation in Entomological Society of America 11:9–25. DENNO, R.F., AND G.K. RODERICK. 1990. Population biol- plant cover apparently goes further in explain- ogy of planthoppers. Annual Review of Entomology ing leafhopper distribution than it does grass- 35:489–520. hopper distribution. The cover of known host HAMILTON, K.G.A. 1983. Introduced and native leafhop- plants explained much of the variation in pers common to the Old and New Worlds (Rhyn- chota: Homoptera: Cicadellidae). Canadian Ento- abundance of the 8 taxa we examined, taxa mologist 115:473–511. that made up the vast majority of leafhoppers ______. 1998. The species of the North American leafhop- collected. In contrast, when examined in fields pers Ceratagallia Kirkaldy and Aceratagallia Kirkaldy containing both SB and AM patches, variation (Rhynchota: Homoptera: Cicadellidae). Canadian Entomologist 130:427–490. in the abundance of major grass species ex- HICKS, A., AND R. WHITCOMB. 1993. The genus Laevi- plained just 5%–28% of the variation in grass- cephalus (Homoptera: Cicadellidae): hosts, biogeog- hopper numbers among plots (Kemp et al. 2002). raphy, and three new species. Proceedings of the This could be due partly to the different spa- Entomological Society of Washington 95:481–487. ______. 1996. Diversity of the leafhopper (Homoptera: tial scales at which the families were studied. Cicadellidae) fauna of northern Chihuahuan grass- However, it is also likely due to the fact that lands, with emphasis on gypsum grasslands and 2004] GRASSLAND LEAFHOPPERS IN MONTANA 531

description of a new species of Athvsanella (Cicadel- PANZER, R.D., AND M.W. SCHWARTZ. 1998. Effectiveness of lidae: Deltocephalinae). Proceedings of the Entomo- a vegetation-based approach to insect conservation. logical Society of Washington 98:145–157. Conservation Biology 12:693–702. HILL,M.O. 1973. Diversity and evenness: a unifying nota- PRESTIDGE, R., AND S. MCNEILL. 1983. Auchenorrhyn- tion and its consequences. Ecology 54:427–432. cha–host plant interactions: leafhoppers and grasses. KEMP, W.P., S. HARVEY, AND K.M. O’NEILL. 1990a. Patterns Ecological Entomology 8:331–339. of vegetation and grasshopper community composi- STILING, P.D. 1994. Interspecific interactions and commu- tion. Oecologia 83:299–308. nity structure in planthoppers and leafhoppers. ______. 1990b. Habitat and insect biology revisited: the Pages 449–516 in R.E. Denno and T.J. Perfect, edi- search for patterns. American Entomologist 36: tors, Planthoppers: their ecology and management. 44–49. Chapman and Hall, New York. KEMP, W.P., K.M. O’NEILL, M.M. CIGLIANO, AND S. TOR- TER BRAAK, C.J.F. 1988. CANOCO: an extension of DEC- RUSIO. 2002. Field-scale variation in plant and grass- ORANA to analyze species-environment relation- hopper communities: a GIS-based assessment. Trans- ships. Vegetatio 75:159–160. actions in GIS 6:115–133. WACHTER, D., K.M. O’NEILL, AND W. P. K EMP. 1998. Grass- KRUESS, A., AND T. T SCHARNTKE. 2002. Contrasting re- hopper (Orthoptera: Acrididae) communities on an sponses of plant and insect diversity to variation in elevational gradient in southwestern Montana. Jour- grazing intensity. Biological Conservation 106: nal of the Kansas Entomological Society 71:35–43. 293–302. WALOFF, N., AND M. SOLOMON. 1973. Leafhoppers (Auche- LUDWIG, J., AND J. REYNOLDS. 1988. Statistical ecology: a norrhyncha) of acidic grasslands. Journal of Applied primer on methods and computing. Wiley Inter- Ecology 10:189–212. science. 337 pp. WHITCOMB, R.A. 1987. North American forests and grass- MAGURRAN, A.E. 1988. Ecological diversity and its mea- lands: biotic conservation. Pages 163–176 in D.A. surement. Princeton University Press, Princeton, NJ. Saunders, G.W. Arnold, A.A. Burbidge, and A.J.M. MORRIS, M. 1973. The effects of seasonal grazing on the Hopkins, editors, Nature conservation: the role of Heteroptera and Auchenorrhyncha (Hemiptera) of remnants of native vegetation. Beatty and Sons, Sur- chalk grassland. Journal of Applied Ecology 10: rey, Great Britain. 761–780. WHITCOMB, R., A. HICKS, D. LYNN, H. BLOCKER, AND J. ______. 1990a. The Hemiptera of two sown calcareous KRAMER. 1986. Host specificity: a major mechanism grasslands: I. Colonization and early succession. enhancing insect diversity in grasslands. Proceed- Journal of Applied Ecology 27:367–378. ings of the 10th North American Prairie Conference, ______. 1990b. The Hemiptera of two sown calcareous Denton, TX. grasslands: III. Comparisons with Auchenorrhyncha WHITCOMB, R., A. HICKS, D. LYNN, K. ALLRED, AND H. faunas of other grasslands. Journal of Applied Ecol- BLOCKER. 1987a. Geographic variation in host rela- ogy 27:394–409. tionships of leafhoppers (Homoptera: Cicadellidae) MUEGGLER, W., AND W. S TEWART. 1980. Grassland and in North American grasslands. Pages 293–325 in shrubland habitat types of western Montana. U.S. Proceedings of the 2nd International Workshop on Department of Agriculture, Forest Service GTR Leafhoppers and Planthoppers of Economic Impor- INT-66. tance, Provo, UT. MULKERN, G.B., K.P. PRUESS, H. KNUTSON, A.F. HAGEN, WHITCOMB, R., J. KRAMER, M. COAN, AND A. HICKS. 1987b. J.B. CAMPBELL, AND J.D. LAMBLEY. 1969. Food habits Ecology and evolution of leafhopper-grass host rela- and preferences of grassland grasshoppers of the tionships in North American grasslands. Current Top- north central Great Plains. North Dakota Agricultural ics in Vector Research 4:125–182. Experiment Station Bulletin 481:1–32. WHITCOMB, R., A. HICKS, H. BLOCKER, AND D. LYNN. 1994. NOVOTNY, V. 1994. Association of polyphagy in leafhop- Biogeography of leafhopper specialists of the short- pers (Auchenorrhyncha, Hemiptera) with unpredict- grass prairie. American Entomologist 40:19–35. able environments. Oikos 70:223–232. O’NEILL, K.M., W.P. KEMP, C. SEIBERT, M.G. ROLSTON, J.A. Received 21 February 2003 BESS, AND T.K. PHILIPS. 2001. Natural enemy assem- Accepted 12 April 2004 blages on native and reseeded grasslands in southwest Montana: a family-level analysis. Western North American Naturalist 61:195–203. Western North American Naturalist 64(4), © 2004, pp. 532–537

THE CENTIPEDE GENUS ARTHRORHABDUS POCOCK, 1891, IN THE WESTERN HEMISPHERE: POTENTIAL OCCURRENCE OF A. PYGMAEUS (POCOCK, 1895) IN BELIZE (SCOLOPENDROMORPHA: SCOLOPENDRIDAE: SCOLOPENDRINAE)

Rowland M. Shelley1 and Amazonas Chagas Junior2

ABSTRACT.—The scolopendrid centipede genus Arthrorhabdus Pocock, 1891, comprises 6 species: A. formosus Pocock, 1891, the type species, occurring in southern Africa (Mpumalanga, Free State, Western and Northern Cape Provinces, South Africa, and southern Namibia); A. somalus Manfredi, 1933, in Somalia and Yemen; A. jonesii Verhoeff, 1938, from southern India (Kerala Province); A. mjobergi Kraepelin, 1916, and A. paucispinus Koch, 1984, in Australia (Western and South Australia, Northern Territory, and Queensland); and A. pygmaeus (Pocock, 1895), in the south central and southwestern United States, Mexico, and, potentially, Belize. This sporadic occurrence suggests that the genus is polyphyletic, and the monotypic synonym, Arthrorhabdinus Verhoeff, 1907, is available for pygmaeus, which is not referrable to another established genus. Arthrorhabdus spinifer (Kraepelin, 1903), known only from Belém, Pará State, Brazil, is transferred to Rhoda Meinert, 1886, thereby constituting a new combination. Sixteen new localities are reported for A. pygmaeus, 14 in Mexico and 2 in the U.S.; a specimen from Belize, intercepted in quarantine in Miami, suggests occurrence in this country. The 2 U.S. sites, in Cameron County, Texas, and Pima County, Arizona, extend the generic and specific ranges around 400 miles (640 km) to the southeast and west, respectively. In Mexico, A. pygmaeus ranges southward through the mainland, possibly excluding the Yucatan Peninsula, and also inhabits the southern half of Baja California Sur (BCS). Its apparent absence from the rest of the Baja peninsula suggests that the BCS populations may result from rafting across the Gulf of California from Sinaloa, where the species occurs.

Key words: Arthrorhabdus, A. pygmaeus, A. spinifer, Arizona, Texas, Baja California Sur, Mexico, Belize.

This contribution is the 2nd in a series of lines on centipedes have been established by planned collaborative publications on the Neo- amateur enthusiasts; and a number of people tropical scolopendromorph centipede fauna, maintain live scolopendrids in their homes as the 1st being a study of Mexican representa- part of a hobby in arthropod husbandry. The tives of Newportia Gervais, 1847 (Scolopocryp- time has arrived to capitalize on this base level topidae: Newportiinae) (Chagas and Shelley of enthusiasm and to develop publications that 2003). The objective is to produce compre- will facilitate a degree of expertise among both hensive treatments and consolidated literature professionals and this lay audience. Shelley’s reviews of taxa in regions of the Neotropics to monograph (2002) illuminated U.S. and Cana- supersede the piecemeal accounts and scat- dian scolopendromorphs, but the diversity to tered records that exist today in scores of mis- the south is vastly greater, and this fauna war- cellaneous publications. It is virtually impossi- rants attention to elevate its knowledge to a ble for students even to retrieve this polyglot comparable level. literature, much less decipher anatomical fea- We address herein the genus Arthrorhab- tures and distinguish taxa, and yet substantive dus Pocock, 1891. This is 1 of 3 in the Scolo- contributions are essential for knowledge of pendrinae recorded from 5 continents. The this fauna to advance. Beyond professional others are Scolopendra L., 1758, which inhabits myriapodologists, considerable interest in scolo- all 6 populated continents and, though often pendromorphs, especially the large scolopen- introduced, most oceanic islands between 35° drids, exists among the general public. Scolo- north and south latitudes, and Cormocephalus pendromorphs are marketed at pet stores in Newport, 1845, occurring in South America, the United States; internet websites and chat Central America, Asia, Australia, and on islands

1Research Laboratory, North Carolina State Museum of Natural Sciences, 4301 Reedy Creek Road, Raleigh, NC 27607 USA. 2Departamento de Invertebrados, Laboratório de Aracnologia, Museu Nacional/UFRJ, Quinta da Boa Vista, s/número, São Cristóvão, CEP-20940-040, Rio de Janeiro, Brazil.

532 2004] CENTIPEDE GENUS ARTHRORHABDUS IN WESTERN HEMISPHERE 533 in the Caribbean Sea and western Pacific lin (1914), Bücherl (1939, 1941, 1974), and Ocean. Arthrorhabdus currently comprises 7 Schileyko (2002). We borrowed the holotype species, 1 each in North America, South Amer- from the Zoologisches Museum, Hamburg, ica, and southern Asia (India), and 2 apiece in Germany, and found that the 1st tergite over- Africa (1 also occurring on the Arabian Penin- laps the cephalic plate, which has a partial sula) and Australia (Attems 1930, Manfredi middorsal suture, that all tarsi possess spurs, 1933, Verhoeff 1938, Lawrence 1955a, 1955b, and that the 1st tarsi are shorter than the 2nd. 1959, 1975, Chelazzi 1977, Koch 1983, 1984, These features are characteristic of Rhoda Khanna 2001). The type species is A. formosus Meinert, 1886, and we therefore propose R. Pocock, 1891 (= A. interveniens Porat, 1893), spinifer (Kraepelin, 1903), comb. nov. occurring in Western and Northern Cape Pro- With this action, pygmaeus, whose type vinces, Free State, Mpumalanga, South Africa, locality is Amula, Guerrero, Mexico, becomes and the Great Namaland region of southern the only known representative of Arthrorhab- Namibia (Pocock 1891, Porat 1893, Kraepelin dus in the Western Hemisphere. It was de- 1903, Attems 1930, Lawrence 1955a, 1955b, scribed in Scolopendra by Pocock (1895), 1959, 1975). The other African species, A. transferred to Arthrorhabdus by Kraepelin somalus Manfredi, 1933, occurs in Somalia, (1903), and is 1 of 5 scolopendrids inhabiting over 3300 miles (5280 km) to the northeast, the Central Plains of the U.S. and the Chi- and directly across the Gulf of Aden in Yemen huahuan Desert of the U.S. and Mexico. The on the Arabian Peninsula (Manfredi 1933). others are Hemiscolopendra marginata (Say, Kraepelin (1916) described A. mjobergi from 1821), unknown at present from Mexico, and 3 Kimberley Division, Western Australia, which species of Scolopendra that occur in both coun- record was reiterated by Chamberlin (1920). tries: S. viridis Say, 1821; S. polymorpha Wood, Koch (1984) described A. paucispinus, which 1861; and S. heros Girard, 1853 (Shelley 2002). is widespread in the interior of Western Aus- Based on samples then available, Shelley (2002) tralia, and reported that A. mjobergi occurs characterized the range in the U.S. as a gener- throughout this state and also in South Aus- ally triangular area in west Texas, extending tralia, Northern Territory, and Queensland. from the Red River in Hardeman County at Verhoeff (1938) proposed A. (Trachycormoceph- the base of the panhandle, to Val Verde County alus) jonesii for an individual from Trivandrum on the Rio Grande at the top of the “bend,” to (= Travancore), in the Ponmudi Hills of Kerala Hudspeth County east of El Paso. Pocock (1895) Province, at the southern tip of India, and also reported A. pygmaeus from San Diego, Khanna (2001) subsequently cited it without Duval County, west of Corpus Christi, a rec- the subgenus because Lewis (1986) demon- ord that Shelley (2002) considered dubious be- strated that Trachycormocephalus Kraepelin, cause it lies some 250 miles (400 km) southeast 1903, is a junior synonym of Scolopendra. of the then otherwise coherent range. Lewis further suggested that Arthrorhabdus South of the U.S., A. pygmaeus ranges some may be another synonym of Scolopendra, and 1235 miles (1976 km) to the vicinity of Tehuan- the numerical analyses of Australian scolopen- tepec, Oaxaca, Mexico. An individual from drids by Koch and Colless (1986) showed that Belize that was intercepted in quarantine in A. mjobergi fits with this genus in many Miami indicates that the distribution may respects. extend into this country, around 1380 miles The 2 American species are A. pygmaeus (2210 km) south of the northernmost locality (Pocock, 1895), occurring in southeastern Ari- in Texas (Shelley 2002), but we hesitate to cite zona and central/southern Texas, USA, Baja Belize definitely based solely on this speci- California Sur (BCS) and mainland Mexico men. Mexican records extend southward from (unknown from the Yucatan Peninsula), and, the U.S. border primarily through the central potentially, Belize (Shelley 2002, plus records states and along the Pacific Coast (Fig. 1), and cited herein); and A. spinifer (Kraepelin, 1903), none are available from the Yucatan Peninsula known only from Belém, Pará State, Brazil. or Chihuahua, Sonora, or Baja California Norté The latter was described in Cupipes Kohlrausch, (BCN). Documented sites in west Texas and 1878, and transferred to Arthrorhabdus by Arizona mandate occurrence in Chihuahua and Attems (1930); subsequently, it has been re- the eastern half of Sonora; however, A. pyg- ported from Belém and/or Pará by Chamber- maeus may truly be absent from the Yucatan 534 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 1. Distribution of Arthrorhabdus pygmaeus. The dot in Belize, denoted by the arrow, indicates potential occurrence in this country and is placed centrally because the exact locality of the intercepted specimen is unknown. and BCN. The centipede has been collected dae), Pereira et al. (1999) implied that rafting repeatedly in the southern half of BCS, and was operative in the dispersal of Pectiniunguis we think its occurrence there may be attribut- halirrhytus (Crabill, 1959), which has been able to rafting from Sinaloa, directly across the found on beaches in seaweed and “beach drift” Gulf of California, with subsequent dispersion in the lower Keys of Florida, and in Cozumel northward in the Baja Peninsula that appar- and Quintana Roo, Mexico. ently has not yet reached BCN. The breadth As noted by Shelley (2002), A. pygmaeus of the Gulf in this area varies from 110 to 190 possesses small spines on the 1st tarsi of leg miles (176–304 km), plausible distances for pairs 1–20 and long, narrow tooth plates; addi- rafting events involving centipedes. Shelley tionally, the 1st tergite lacks the transverse, (2002) postulated that the Florida populations procurved sulcus, and the cephalic plate and of Scolopendra alternans Leach, 1813, arose 1st tergite abut, instead of either overlapping by rafting across the Straits of Florida from the other (see Shelley 2002:42, figs. 66–68). It the Greater Antilles in general and Cuba in can be readily distinguished from representa- particular, and Shelley and Kiser (2000) sug- tives of the sympatric scolopendrine genera gested that rafting from the coasts of Ecuador Scolopendra and Hemiscolopendra Kraepelin, and Peru, a much greater distance than that 1903, in which the cephalic plate overlaps the postulated here, is the most likely explanation 1st tergite, and the latter exhibits a conspicuous for the occurrence of S. galapagoensis Boll- procurved sulcus. While examining unsorted man, 1889, in the Galápagos Archipelago. In and unidentified scolopendromorphs in the the order Geophilomorpha (family Schendyli- following U.S. repositories, we discovered 17 2004] CENTIPEDE GENUS ARTHRORHABDUS IN WESTERN HEMISPHERE 535 additional samples of A. pygmaeus that we (1960), which were deleted along with 1 from detail below to augment the localities in Shel- California (Chamberlin 1911) by Shelley (2002). ley (2002); the number of specimens follows By spanning New Mexico, the record also cor- the institutional acronym, and the overall dis- roborates Crabill’s statement (1960) that A. tribution is depicted in Figure 1. Repository pygmaeus occurs in this state. The Cameron acronyms are AMNH—American Museum of County locality, at the southern tip of Texas Natural History, New York; NMNH—National near the Gulf of Mexico, expands the ranges Museum of Natural History, Smithsonian Insti- the same distance southeastward and lends tution, Washington, DC; and TMM—Texas credence to Pocock’s record (1895) from Duval Memorial Museum, Austin. County, some 130 miles (208 km) to the north- USA.—Arizona: Pima Co., Tucson, Santa northwest, which we show as a definite site in Catalina Mountains, 5 August 1948, G.E. Ball, Figure 1. The centipede should therefore be H.E. Evans (NMNH 1) (first definite state rec- expected along the entire southern border of ord). Texas: Cameron Co., Brownsville, 9 De- Texas and potentially may be discovered in the cember 1911, collector unknown (NMNH 1). panhandle per se and southwestern Okla- MEXICO.—Baja California Sur: canyon homa. The distribution from Texas westward near San José de Comondu, 15 February 1966, may parallel that of S. heros (see Shelley 2002: V. Roth (AMNH 4). San José del Cabo, 18 fig. 42) by extending across southern New March 1945, collector unknown (NMNH 1). Mexico, south of high elevations in the Sacra- La Laguna, 4 November 1944, Correo (NMNH mento, Capitan, and Black Mountains, and 4). Hidalgo: 7 miles (11.2 km) SE Zimapan, El spreading northward to the east and west, Tablon, 19 August 1964, J. & W. Ivie (NMNH along the Pecos and Gila Rivers, respectively. 1). Morelos: Puente de Ixtla, 10 August 1943, Generic definitions in the Scolopendridae C. Bolivar, F. Bonet, B. Osario (NMNH 2). need critical assessment. Classical revisions Nayarit: Jesus Maria, 23 June 1955, B. Malkin are in order but beyond our scope at present (NMNH 1) (new state record). Nuevo León: and ideally should incorporate molecular tech- 10 miles (16 km) N Bustamente, 26 Septem- niques. Although 6 component species are ber 1964, J.R. Reddell (TMM 1). Monterrey, 3 presently assigned to Arthrorhabdus, true con- January 1950, S. & D. Mulaik (NMNH 2). generic status has not been demonstrated, and Oaxaca: Suchinxtapec, 21 March 1966, G.E. the sporadic distribution pattern, with Gond- Ball, D.R. Whitehead (NMNH 1). Sinaloa: wanian (African, Indian, and Australian species) Mazatlan, 22 August 1962, G.E. Ball (NMNH and Laurasian elements (A. pygmaeus), sug- 4). 30 miles (48 km) NE Jimenez, Matamoros gests that the genus is polyphyletic. While the Rd., precise location, date, and collector un- types and other specimens must be examined known (NMNH 2). [The vial label places this for a definitive conclusion, published illustra- site in Sinaloa, but map and web searches do tions provide clues to this query, and Pocock’s not show towns by these names.] Tamaulipas: illustrations (1891: figs. 1a–d) of A. formosus, 19 miles (30.4 km) S Ciudad Victoria, 28 De- the type species, seem reasonably similar to cember 1947, collector unknown (NMNH 5); conditions in A. pygmaeus such that the 2 14 miles (22.4 km) S Ciudad Victoria, 8 Janu- could conceivably be congeneric despite the ary 1950, S. & D. Mulaik (NMNH 8); 12 miles geographic hiatus; the size of the spines on the (19.2 km) SE Ciuded Victoria, 14 August 1964, 1st tarsi and the articulation between the J. & W. Ivie (NMNH 1) (new state record). cephalic plate and 1st tergite appear espe- BELIZE.—Locality unknown: discovered in cially close to those in A. pygmaeus. However, bromeliads from Belize by an unknown inspec- the tooth plates of A. somalus, which are short, tor during quarantine in Miami, Florida, on 3 broad, and compatible with Scolopendra (Man- January 1961 (NMNH 1) (potential new coun- fredi 1933: fig. 1), are markedly different from try record). the long, narrow structures in A. pygmaeus, The Arizona locality expands the generic and the coxopleurae of A. paucispinus are not and specific ranges in the U.S. some 400 miles extended and have linear caudal margins (Koch (640 km) westward from Hudspeth County, 1984: figs. 2, 4), whereas they project strongly Texas, and confirms the general records from caudad in A. pygmaeus and are armed with the state by Chamberlin (1911) and Crabill apical and lateral spines. From these and other 536 WESTERN NORTH AMERICAN NATURALIST [Volume 64 differences among published illustrations of pendromorpha: Scolopocryptopidae: Newportiinae). the putative congeners, we believe that com- Zootaxa 379:1–20. CHAMBERLIN, R.V. 1911. The Chilopoda of California II. prehensive evaluation of Arthrorhabdus will Pomona College Journal of Entomology and Zoology result in partitioning, and the monotypic syn- 3:470–479. onym, Arthrorhabdinus Verhoeff, 1907, is ______. 1914. The Stanford expedition to Brazil, 1911, available for pygmaeus. Arthrorhabdus, as rep- John C. Branner, director. The Chilopoda of Brazil. resented by A. formosus, may indeed be a syn- Bulletin of the Museum of Comparative Zoology 58(3):151–221. onym of Scolopendra as Lewis (1986) suggests, ______. 1920. The Myriopoda of the Australian region. but Arthrorhabdinus is clearly distinct because Bulletin of the Museum of Comparative Zoology 64: the longer and narrower tooth plates of pyg- 1–269. maeus and different articulation between the CHELAZZI, L. 1977. Some Scolopendridae centipedes from Somalia (Chilopoda Scolopendromorpha). Monitore cephalic plate and 1st tergite are unmistak- Zoologico Italiano, N.S. Supplemento 9(4):69–84. able. These differences are equivalent to, if CRABILL, R.E. 1960. A new American genus of cryptopid not greater than, accepted distinctions between centipede, with an annotated key to the scolopen- such anatomically similar scolopendrine gen- dromorph genera from America north of Mexico. era as Scolopendra and Hemiscolopendra, and Proceedings of the United States National Museum 111:1–15. Cormocephalus and Rhoda. We therefore be- KHANNA, V. 2001. A check-list of the Indian species of the lieve that pygmaeus is incompatible with Sco- centipedes (Chilopoda: Scolopendromorpha). Annals lopendra and requires a different genus; whether of Forestry 9:199–219. the taxon is properly Arthrorhabdus or Arthro- KOCH, L.E. 1983. Revision of the Australian centipedes of the genus Cormocephalus Newport (Chilopoda: Scolo- rhabdinus remains to be determined. pendridae: Scolopendrinae). Australian Journal of Zoology 31:799–833. ACKNOWLEDGMENTS ______. 1984. Australian species of the centipede genus Arthrorhabdus Pocock (Chilopoda: Scolopendridae: We thank H. Dastych (Zoologisches Muse- Scolopendrinae). Journal of Natural History 18: 363–368. um, Hamburg, Germany) for loaning the holo- KOCH, L.E., AND D.H. COLLESS. 1986. Numerical taxon- type of Cupipes spinifer, and L. Prendini, J. omy of Australian species of nine genera of scolo- Coddington, and J.R. Reddell for loaning sam- pendrid centipedes (Chilopoda: Scolopendridae). ples from the AMNH, NMNH, and TMM, Australian Journal of Zoology 34:87–105. KRAEPELIN, K. 1903. Revision der Scolopendriden. Mittei- respectively. J.G.E. Lewis provided valuable lungen aus dem naturhistorischen Museum in Ham- advice and directed us to additional references. burg 20:1–276. This research was supported in part by a “mini- ______. 1916. Results of Dr. E. Mjöberg’s Swedish expe- PEET” grant from the Society for Systematic dition to Australia 1910–1913. 4. Scolopendriden und Biology to the 2nd author, which enabled him Skorpione. Arkiv für Zoologi 10(2):1–43. LAWRENCE, R.F. 1955a. A revision of the centipedes (Chi- to visit the U.S. in 2003 and collaborate and lopoda) of Natal and Zululand. Annals of the Natal study with the 1st author. It was also supported Museum 8(2):121–174. by a Collections Study Grant from the AMNH, ______. 1955b. Chilopoda. South African animal life. Results which funded travel by the 2nd author to that of the Lund University expedition in 1950–1951, 2:4–56. institution. ______. 1959. A collection of Arachnida and Myriopoda from the Transvaal Museum. Annals of the Transvaal LITERATURE CITED Museum 23(4):363–386. ______. 1975. The Chilopoda of South West Africa. Cim- ATTEMS, C. 1930. Myriapoda 2. Scolopendromorpha. Das bebasia, series A, 4(2):36–45. Tierreich 54:1–308. LEWIS, J.G.E. 1986. The genus Trachycormocephalus a BÜCHERL, W. 1939. Os Quilopodos do Brasil. Memorias junior synonym of Scolopendra, with remarks on the do Instituto Butantan 13:43–362. validity of other genera of the tribe Scolopendrini ______. 1941. Catálogo dos Quilopodos da Zona Neotropi- (Chilopoda: Scolopendromorpha). Journal of Natural cal. Memorias do Instituto Butantan 15:251–372. History 20:1083–1088. ______. 1974. Die Scolopendromorpha der Neotropis- MANFREDI, P. 1933. Miriapodi della Somalia italiana. Atti chen Region. Symposia of the Zoological Society of della Societa Italiana di Scienze Naturali 72:275–284. London 32:99–133. PEREIRA, L.A., A. MINELLI, AND D. FODDAI. 1999. Pectini- CHAGAS, A., AND R.M. SHELLEY. 2003. The centipede genus unguis bollmani n. sp., from the coralline island Cayo Newportia Gervais, 1847, in Mexico: description of a Sombrero (Venezuela), with notes on P. halirrhytus new troglomorphic species; redescription of N. sabina Crabill, 1959 (Chilopoda: Geophilomorpha: Schendyl- Chamberlin, 1942; revival of N. azteca Humbert & idae). Studies on Neotropical Fauna and Environ- Saussure, 1869; and a summary of the fauna (Scolo- ment 34:176–185. 2004] CENTIPEDE GENUS ARTHRORHABDUS IN WESTERN HEMISPHERE 537

POCOCK, R.I. 1891. Notes on the synonymy of some SHELLEY, R.M. 2002. A synopsis of the North American species of Scolopendridae, with descriptions of new centipedes of the order Scolopendromorpha (Chilo- genera and species of the group. Annals and Maga- poda). Virginia Museum of Natural History Memoir zine of Natural History, series 6, 7:51–68, 221–231. 5:1–108. ______. 1895–1910. Chilopoda and Diplopoda. Biologia SHELLEY, R.M., AND S.B. KISER. 2000. Neotype designa- Centrali-Americana. 217 pp. [Fascicle in which Sco- tion and a diagnostic account for the centipede, lopendra pygmaea is described was distributed in Scolopendra gigantea L., 1758, with an account of S. December 1895.] galapagoensis Bollman, 1889 (Chilopoda Scolopen- PORAT, C.O. 1893. Myriopoder frän Vest-Och Syd-Afrika. dromorpha Scolopendridae). Tropical Zoology 13: Bihang till Kongl. Svenska Vetenskaps-Akademiens 159–170. Handlingar 18(4):1–51. VERHOEFF, K.W. 1938. Chilopoden-Studien, zur Kenntnis SCHILEYKO, A. 2002. Scolopendromorpha. Pages 479–500 der Epimorphen. Zoologischen Jahrbüchern 71: in J. Adis, editor, Amazonian Arachnida and Myri- 339–388. apoda: identification keys to all classes, orders, families, some genera, and lists of known terrestrial Received 12 December 2003 species. Pensoft Publishers, Sofia, Bulgaria. Accepted 8 March 2004 Western North American Naturalist 64(4), © 2004, pp. 538–543

PROPHYSAON COERULEUM COCKERELL, 1890, BLUE-GRAY TAILDROPPER (GASTROPODA: ARIONIDAE): NEW DISTRIBUTIONAL RECORDS AND REPRODUCTIVE ANATOMY

Kristiina Ovaska1, William P. Leonard2, Lyle Chichester3, Thomas E. Burke4, Lennart Sopuck5, and Jim Baugh6

Key words: Prophysaon coeruleum, distribution, phylogeography, reproductive anatomy, British Columbia, Idaho, Pacific Northwest, Rocky Mountains, Washington.

The genus Prophysaon is composed of 9 ical variation is present in Oregon, and the recognized species of slugs native to temper- name P. coeruleum may apply to a species ate forests of western North America (Turgeon complex (Kelley et al. 1999). Herein we report et al. 1998). Differences in the anatomical on previously unknown populations of P. coeru- characteristics of the distal genitalia have been leum from southwestern British Columbia and used to identify 2 species groups (subgenera; northern Idaho. Pilsbry 1948): Mimetarion, which includes P. We collected P. coeruleum during surveys vanattae, P. obscurum, P. fasciatum, and P. for terrestrial gastropods on Vancouver Island, humile; and Prophysaon, which includes the British Columbia, Canada, and in Washington remaining species (P. andersoni, P. boreale, P. and northern Idaho, United States. On Van- coeruleum, P. dubium, and P. foliolatum). The couver Island we carried out time- and area- genus has a disjunct geographic distribution: constrained surveys along 100-m-long and 1- most species occur along the Pacific Coast m-wide transects in different forested habitats from southeastern Alaska to northern Califor- at 3 sites near Victoria (56 transects searched nia, but a few are present in the Rocky Moun- for a total of 72 person-hours) in March and in tains of northern Idaho and western Montana October–November 2002. In northern Idaho (Smith 1943, Pilsbry 1948, Frest and Johannes we visually searched approximately 30 sites for 2000, Leonard et al. 2003). One species (P. a total of ca. 200 person-hours between April humile) is restricted to the Rocky Mountains and October from 1999 to 2002. In Washington portion of the distribution. we conducted surveys on approximately 300 Prophysaon coeruleum Cockerell (1890) is a sites in the eastern Cascade Mountains and on small, poorly known slug of conservation inter- >130 sites between the Cascade Crest and the est with previously known range from western Pacific Ocean. In addition, one of us (TB) had Washington to northwestern California (Pilsbry access to the results of hundreds of surveys 1948, Kelley et al. 1999). The species has been conducted for “survey and manage” mollusks designated as a “survey and manage” species by the U.S. Forest Service on National Forests under the Northwest Forest Plan (USDA For- within those areas. est Service and USDI Bureau of Land Man- To confirm identification, we examined the agement 1994) due to its close association with reproductive system of the specimens with a older forests, few known localities, and lack of stereo-zoom microscope under 7.5–60X mag- information on its natural history and ecology nification (one specimen from site 2; two spec- (McGraw et al. 2002). Considerable geograph- imens from site 3 in Appendix). Specimens

14180 Clinton Place, Victoria, British Columbia V8Z 6M1, Canada. 2223 Foote Street NW, Olympia, WA 98502, USA. 32805 Greenbriar Boulevard, Wellington, FL 33414, USA. 44715-61st Avenue NE, Olympia, WA 98516, USA. 51759 Colburne Place, Sidney, British Columbia V8L 5A2, Canada. 62018 Dry Creek Road, Ellensburg, WA 98926, USA.

538 2004] NOTES 539

Fig. 1. Distribution map for locality records of Prophysaon coeruleum: solid circles = this study (numbers correspond to those in Appendix); open circles = previous localities (Pilsbry 1948, Branson and Branson 1984). were deposited in collections at the Carnegie several years). The species appears to have a Museum (CM), at the Royal British Columbia very restricted distribution in Canada. Museum (RBCM), and in the personal collec- In Idaho we found P. coeruleum at 2 of tion of T. Burke. Some specimens were sent to approximately 30 sites surveyed. The 2 sites Dr. Thomas Wilke at George Washington Uni- were about 30 km apart in the Couer d’Alene versity to be included in a genetic study. Lake watershed. These observations extend In British Columbia we found P. coeruleum the species’ known range approximately 400 along 2 of the 56 transects searched. These km eastward and represent the 1st records for localities were about 500 m apart on Rocky the Rocky Mountains (Fig. 1). Apparently, the Point Peninsula in the District of Metchosin, population in Idaho is disjunct from the more about 20 km southwest of Victoria on Vancou- western populations, as the intervening Colum- ver Island (records 1–2 in Appendix). These bia Basin consists of arid shrub-steppe habitat records extend the known distribution of the unsuitable for forest-dwelling gastropods (Bar- species approximately 120 km northwest from nosky et al. 1987, Franklin and Dyrness 1988). Washington State and represent the 1st records Brunsfeld et al. (2001) and Layser (1980) pro- from Canada (Fig. 1). Since 1999, two of us (KO vided numerous examples of fungi, plants, and and LS) have carried out surveys for forest- animals that have Pacific coastal and interior dwelling gastropods throughout Vancouver Rocky Mountain distributions, which mirror Island and on the southern coastal mainland of the distribution of western hemlock forest. British Columbia without locating this species Leonard et al. (2003) reported on a previously (about 170 sites surveyed, some intensively over undocumented example of the slug Prophysaon 540 WESTERN NORTH AMERICAN NATURALIST [Volume 64 dubium, Cockerell, from Idaho and cited addi- Externally, our specimens conform to the tional examples of invertebrates that have sim- description in Pilsbry (1948): live animals were ilar disjunct distribution patterns. gray with light blue flecking, the prominence In Washington we found P. coeruleum at 2 of which varied among localities. On all of the sites in the Cispus River watershed and at 1 specimens distinct, longitudinal grooves were site on Pin Creek, a tributary to the lower present on the foot behind the mantle, and the Columbia River. The slugs appeared to be foot margin was narrow with a distinct border most abundant at the Kraus Ridge site in the above (Fig. 2). The extended length of the Cispus watershed (record 5 in Appendix; 11 specimens ranged between 23 mm and 30 mm detections). One of us (TB) has found P. when alive and 12.5 mm and 15 mm after coeruleum at this site on several previous preservation. occasions. The Iron Creek site (record 6 in The genitalia of the new specimens from Appendix) is about 4 km southwest of the Idaho and British Columbia resemble those of Kraus Ridge site and on the opposite side of P. coeruleum from Washington illustrated by the Cispus River. In several visits we have Pilsbry (1948; Figure 378d), except that the found only a single specimen of P. coeruleum penis is much larger; in 1 of the 2 Idaho speci- at this site. There is a 3rd site in this area, mens it is almost as large as the muscular por- approximately 7 km northeast of the Kraus tion of the epiphallus (Fig. 3). Most of the long, Ridge site at 442 m elevation, from which a slender portion of the epiphallus is in a tangle single P. coeruleum was collected on 14 Novem- immediately anterior to the muscular portion. ber 2001. The specimen was identified by Tom The ovotestis is large and consists of approxi- Kogut, Cowlitz Ranger District biologist, but mately 24 egg-shaped lobules, with little pig- was not collected (T. Kogut personal commu- ment. The albumen gland is a small, appendix- nication). The species appears to have a very like structure at the terminus of the common restricted distribution in Washington. duct; it is not clearly demarcated from the com- Surveys by the U.S. Forest Service and U.S. mon duct either by color or texture, at least in Bureau of Land Management found no addi- the preserved animals. The duct of the semi- tional sites in Washington but found several nal receptacle is very long and slender and localities for the species in western Oregon loosely adherent to the common duct. The and also extended its range into northern Cali- vagina is slender and long, as is the entire fornia. However, because similar, undescribed common duct. species have also been found in the southern Interestingly, the genitalia of the new spec- Oregon Cascades (Kelley et al. 1999, USDA imens resemble Pilsbry’s (1948) Figure 376c Forest Service and USDI BLM 2000, Nancy of P. boreale from Oregon. However, his brief Duncan personal communication), there is description of the external morphology does some question whether the California speci- not match these specimens. Pilsbry (1948) mens indeed belong to this species. noted much individual variation in the shape All sites with P. coeruleum were in moist and size of the bulbous muscular portion of forest stands with a deciduous component of the epiphallus of P. coeruleum. Kelley et al. bigleaf maple (Acer macrophyllum), black cot- (1999) suggested that, as currently used, the tonwood (Populus balsamifera), or in 1 case name P. coeruleum might apply to a species quaking aspen (Populus tremuloides; Appendix). complex, particularly in Oregon where much One site was in an old-growth stand (Krause variation exits among specimens from differ- Ridge); the remaining sites were in 2nd-growth ent localities. A forthcoming molecular study stands. Sword ferns (Polystichum munitum), in- of P. coeruelum (T. Wilkie personal communi- dicating moist conditions, were typically pres- cation), including specimens from across the ent. The sites ranged from near sea level to species range, will provide important informa- about 650 m in elevation (Appendix). This tion regarding whether the taxa represents a species is thought to be associated with older polyphyletic group and will be critically impor- forests and moist plant communities (Kelley et tant for developing meaningful conservation al. 1999). Although not restricted to old-growth planning strategies. forests, it appears to require old-growth attributes, such as abundant coarse woody debris Funding for surveys on Vancouver Island and moist forest floor conditions. came from the Canadian Department of 2004] NOTES 541

Fig. 2. Specimens of Prophysaon coeruleum: A, Vancouver Island, British Columbia (locality 1); B, Kootenai County, Idaho (locality 4); C, Cowlitz County, Washington (locality 7); bar = ca. 5 mm. 542 WESTERN NORTH AMERICAN NATURALIST [Volume 64

last glaciation. Volume K-3. Geological Society of America, Boulder, CO. BRANSON, B.A., AND R.M. BRANSON. 1984. Distributional records for terrestrial and freshwater Mollusca from the Cascade and Coast Ranges, Oregon. Veliger 26: 248–257. BRUNSFELD, S.J., J. SULLIVAN, D.E. SOLTIS, AND P. S . S OLTIS. 2001. Comparative phylogeography of northwestern North America: a synthesis. Pages 319–339 in J. Sil- vertown and J. Antonovics, editors, Integrating ecological and evolutionary processes in a spatial context, Blackwell Science, Oxford. COCKERELL, T.D.A. 1890. New northwestern slugs. Nauti- lus 3:111–113. FRANKLIN, J.F., AND C.T. DYRNESS. 1988. Natural vegetation of Oregon and Washington. Oregon State Uni- versity Press, Corvallis. 452 pp. FREST, T.J., AND E.J. JOHANNES. 2000. An annotated checklist of Idaho land and freshwater mollusks. Journal of the Idaho Academy of Science 36:1–51. KELLEY, R., S. DOWLAN, N. DUNCAN, AND T. B URKE. 1999. Field guide to survey and manage terrestrial mollusk species from the Northwest Forest Plan. USDI Bureau of Land Management, Oregon State Office, Portland. 114 pp. Layser, E.F. 1980. Flora of Pend Oreille County, Washing- ton. Washington State University Cooperative Exten- sion, Pullman. 146 pp. LEONARD, W.P., L. CHICHESTER, AND K. OVASKA. 2003. Prophysaon dubium Cockerell, 1890, the papillose taildropper (Gastropoda: Arionidae): distribution and anatomy. Nautilus 117:62–67. Fig. 3. Reproductive system of Prophysaon coeruleum MCGRAW, R., N. DUNCAN, AND E. CAZARES. 2002. Fungi and from Idaho (CM 64922; locality 3 in Appendix): EP = other items consumed by the blue-gray taildropper epiphallus, MB = muscular body of epiphallus, OVTES slug (Prophysaon coeruleum) and the papillose tail- = ovotestis, PE = penis, SPOV = spermoviduct, SR = dropper slug (Prophysaon dubium). Veliger 45:261–264. seminal receptacle, VA = vagina, VD = vas deferens. PILSBRY, H.A. 1948. Land Mollusca of North America (north of Mexico). Volume 2, part 2. Monographs of the Academy of Natural Sciences of Philadelphia (3), Philadelphia. 640 pp. National Defense. We thank Arthur Robinson SMITH, A.G. 1943. Mollusks of the Clearwater Mountains, for initiating these surveys. Barb Behan initi- Idaho. Proceedings of the California Academy of ated and funded the mollusk surveys for the Sciences 23(36):537–554. Lower Cispus Watershed. Tom Wilkie, Nancy TURGEON, D.D., J.F. QUINN, JR., E.V. COAN, F.G. HOCHBERG, W. G. L YONS, P.M. MIKKELSEN, R.J. NEVES, ET AL. 1998. Duncan, and Tom Kogut shared unpublished Common and scientific names of aquatic inverte- information. Robert Forsyth provided review brates from the United States and Canada: mollusks. comments. We thank Tara Chestnut, Brent 2nd edition. American Fisheries Society Special Haddaway, Kelley Jorgensen, Tom Kogut, Publication 26. 526 pp. Megan and Vicki Leonard, Laura Matthias, USDA FOREST SERVICE AND USDI BUREAU OF LAND MAN- AGEMENT. 1994. Record of decision for amendments Pat McQueary, Bill Null, Ana Licano, Noriada to the Forest Service and Bureau of Land Manage- Martinez, Brad Moon, Casey Richart, Robin ment Planning Documents within the range of the Shoal, Corey Strom, and Joan Ziegltrum for Northern Spotted Owl. [Portland, OR]: U.S. Depart- help in the field. ment of Agriculture, Forest Service; U.S. Department of Interior, Bureau of Land Management. ______. 2000. Final supplemental environmental impact LITERATURE CITED statement for amendment to the survey and manage, protection buffer, and other mitigation measures BARNOSKY, C.W., P.M. ANDERSON, AND P. J . B ARTLEIN. 1987. standards and guidelines Volume 1, Chapters 1–4. The northwestern U.S. during deglaciation: vegetation history and paleoclimatic implications. Pages Received 28 May 2003 289–321 in W. F. Ruddiman and H.E. Wright, Jr., edi- Accepted 31 December 2003 tors, North America and adjacent oceans during the 2004] NOTES 543

APPENDIX. Localities for Prophysaon coeruleum exam- 4. Beauty Creek, Kootenai County, Idaho, USA; elevation ined for this study: CM = Carnegie Museum; RBCM = 650 m asl; 47°35.65′N, 116°39.0′W; ca. 60-year-old Royal British Columbia Museum; TEB = personal collec- stand of western redcedar (Thuja plicata) and black tion of Thomas E. Burke. cottonwood with an understory of red-osier dogwood (Cornus stolonifera), vine maple (Acer circinatum), and 1. Rocky Point, Vancouver Island, British Columbia, Can- pacific yew (Taxus brevifolia); 21 April 2002 (1 speci- ada; elevation <50 m above sea level (asl); 48°20.77′N, men collected by W. Leonard, T. Burke, and J. Baugh): 123°34.12′W; ca. 80-year-old stand of bigleaf maple TEB 02-080. (Acer macrophyllum) and Douglas-fir (Pseudotsuga men- ziesii) with sword fern (Polystichum munitum) under- 5. Kraus Ridge, Gifford Pinchot National Forest, Lewis story; 28 October 2002 (1 specimen collected by K. County, Washington, USA; elevation 460 m asl; Ovaska and L. Sopuk): RBCM 003-102-001. 46°25.98′N, 121°58.3′W; over 250-year-old stand of western hemlock, Douglas-fir, and bigleaf maple with 2. Rocky Point, Vancouver Island, British Columbia, Can- sword fern understory; 22 June 1995 (1 specimen ada; elevation <50 m asl; 48°20.44′N, 123°34.04′W); recorded by T. Burke); 26 September 1996 (2 speci- ca. 70-year-old stand of quaking aspen (Populus tremu- mens recorded by T. Burke); 29 May 1996 (2 speci- loides) with sword fern understory at the edge of small mens recorded by T. Burke); 13 November 1997 (5 wetland adjacent to a large, old coniferous stand; 18 specimens recorded by T. Burke); 13 December 2002 November 2002 (5 specimens collected by L. Sopuck (1 specimen collected by W. Leonard and T. Burke): and K. Ovaska): RBCM 003-103-001 (1 specimen); 1 sent to Dr. Thomas Wilkie at George Washington Uni- specimen donated for genetic studies by Dr. T. Wilke, versity for genetic studies. George Washington University, Washington, DC, USA; 2 specimens kept alive in the author’s collection. 6. Iron Creek, Gifford Pinchot National Forest, Lewis County, Washington, USA; elevation 350 m asl; 3. Chatcolet Lake, Benewah County, Idaho, USA; elevation 46°25.55′N, 121°59.05′W; 2nd-growth stand of western 650 m asl; 47°21.13′N, 116°46.68′W; ca. 70-year-old hemlock, Douglas-fir, and bigleaf maple with sword stand of western hemlock (Tsuga heterophylla), black fern understory; 7 October 1998 (1 specimen recorded cottonwood (Populus balsamifera), paper birch (Betula by T. Burke). papyrifera), and chokecherry (Prunus virginiana) with a sparse understory; 20 April 2002 (7 specimens detected, 7. Tributary to Pin Creek, 0.8 km east of Carrolls, Cowlitz 1 collected by W. Leonard, T. Burke, and J. Baugh): County, Washington; elevation 80 m asl; 46°04.58′N, TEB 02-059; 15 September 2002 (6 specimens col- 122°50.93′W; ca. 50-year-old stand of bigleaf maple, lected by W. Leonard, T. Burke, and J. Baugh): TEB western redcedar, and bitter cherry (Prunus emargi- 02-176 (1 specimen), CM 64922 (3 specimens). nata) with sword fern understory; 7 March 2003 (1 specimen collected by W. Leonard): CM 64963. Western North American Naturalist 64(4), © 2004, pp. 544–547

LATRINE USE BY SAN JOAQUIN KIT FOXES (VULPES MACROTIS MUTICA) AND COYOTES (CANIS LATRANS)

Katherine Ralls1 and Deborah A. Smith2

Key words: latrines, feces, scats, scent marking, chemical communication, kit fox, coyote, Vulpes macrotis, Canis latrans.

Scent marking with glandular secretions, Scent marking by kit foxes (V. macrotis) has urine, or feces is common in species of all car- not been well described. Egoscue (1962) re- nivore families. In general, scent marks are ported that he found no evidence that kit foxes not distributed randomly but are placed care- regularly mark “‘sign’ stations such as old fully on visually conspicuous and often tradi- skeletons or natural objects.” O’Farrell (1987) tionally used objects or locations, such as trail said that kit foxes defecate along trails and intersections (Macdonald 1985, Gorman and near dens and that they mark novel objects by Trowbridge 1989). In some species repeated depositing undersized scats on them. He added, defecation at the same site results in accumu- “There is no evidence that they systematically lations of feces (scats) in small areas known as mark either their territorial boundaries or latrines, middens, or spraint piles. Use of ‘sign stations’ with urine or feces.” Recent stud- latrines is particularly well known in Euro- ies of scent marking by coyotes (Canis latrans) pean badgers (Meles meles; Stewart et al. 2001, in Yellowstone National Park (Gese and Ruff 2002), European otters (Lutra lutra; Kruuk 1997, Allen et al. 1999) concluded that urine 1995), brown hyenas (Hyaena brunnea; Gor- marking was the most important form of olfac- man and Mills 1984), and spotted hyenas (Cro- tory communication and that coyotes did not cuta crocuta; Mills and Gorman 1987). use scats to mark territories. Urine marking is generally thought to be However, we have observed many accumu- the most important form of scent marking in lations of kit fox scats in latrines while survey- canids (Kleiman 1966), but accumulations of ing for endangered San Joaquin kit foxes (V. m . scats in latrines have been described in raccoon mutica) at a variety of locations throughout dogs (Nyctereutes procyonoides), golden jackals their range (USFWS 1998). We believe that (Canis aureus), dholes (Cuon alpinus), and foxes also urinated at these latrines because maned wolves (Chrysocyon brachyurus; Mac- we often found accumulations of white salts donald 1985). Ethiopian wolves (Canis simensis) from dried urine and we sometimes found tend to concentrate scats at latrines (Sillero- damp soil or small puddles smelling of urine Zubiri 1998) and gray wolves (Canis lupus) near accumulations of scats in the early morn- sometimes defecate near trail junctions (Peters ing. We also found coyote scats at some of and Mech 1975), which can lead to accumula- these latrines. Coyote scats are generally much tions of scats that could be considered diffuse larger than kit fox scats and also lack the strong, latrines (Vila et al. 1994). Some foxes, such as musky odor characteristic of kit fox scats (B. island foxes (Urocyon littoralis; Laughrin 1977) Cypher personal communication). and gray foxes (U. cinereoargenteus; Trapp 1978) Here, we provide details on latrines discov- use latrines. However, regular use of latrines ered while using trained detection dogs to has not been described for foxes in the genus locate scats of kit foxes in the Lokern Natural Vulpes, although the red fox (V. vulpes) deposits Area near Bakersfield, California. It has been anal gland secretions on some of its scats demonstrated before that dogs trained to locate (White et al. 1989) in addition to frequent scats of target species can provide a successful marking with urine (Macdonald 1979). method of scat recovery (U. Breitenmoser and

1Conservation and Research Center, Smithsonian’s National Zoological Park, Washington, DC 20008. 2Department of Ecosystem Sciences, University of Washington, Seattle, WA 98195.

544 2004] NOTES 545

C. Breitenmoser-Wursten, unpublished data, 1 bottle, 1 sheep carcass, 1 coyote skull, and 2 1984–1994; P. Paquet, unpublished data, 1982– pieces of bone). Ten of the 61 contained coy- 1989; Smith et al. 2001, 2003; Wasser et al. ote scats in addition to fox scats. We also found 2004). The principal habitat types in Lokern 2 latrines containing only coyote scats. are nonnative annual grassland, saltbush scrub, We found kit fox scats at 98 locations in and upper Sonoran subshrub scrub. The climate 2003. Forty-nine locations had only 1 scat, 17 is semiarid with hot, dry summers and cool, had 2, and 32 had 3 or more. Eight of the 32 wet winters. Approximate summer high and latrines marked a conspicuous object (1 metal winter low temperatures are 37°C and 1°C, object, 1 tire, 1 can, 1 fencepost, and 4 power- respectively. Average yearly precipitation is line poles). Four of the latrines also contained 14.5 cm, occurring primarily as winter rains. coyote scats. Three of these contained 23 fox We established a transect system on an area and 3 coyote scats, 6 fox and 6 coyote scats, approximately 7 km2. Transects ran from north and 20 fox and 3 coyote scats, respectively. We to south and were spaced at 400-m intervals. also found 1 latrine containing only coyote The entire transect system covered 56 km. To scats. Although only about one-third (32/93) of locate scats, a detection dog/handler team sys- the locations where scats were found were tematically searched along each transect (see latrines, latrines contained over 75% (242/315) Smith et al. 2003). We searched the entire tran- of the 315 fox scats we found. The largest sect system in January 2002 and again in Janu- number of kit fox scats we found at the 26 ary 2003, completing the searches within 5 latrines where we counted the number of scats and 4 days, respectively. We recorded the per- present was 30 (Fig. 1). pendicular distance in meters from the tran- We measured distance from the transect sect line to each location where a dog found 1 line for all but 2 of the locations where scats or more scats. were found in 2002 (159) and all locations in A latrine is a place where 1 or more indi- 2003 (98). As distance from the transect was viduals defecate repeatedly, but there are no independent of year (2-tailed t test, t = 1.510, standards for how many scats it takes to con- df =255, P = 0.132), we pooled data for both stitute a latrine. For example, Gorman and Mills years, yielding a sample of 257 locations. Scats (1984) reported that brown hyena latrines con- were found at a minimum distance of 0 m tained 5–50 feces, while Begg et al. (2003) from the transect and at a maximum of 38.40 decided that 2 feces were enough to constitute m. The mean distance was 4.8 ± 6.7 m (s). a latrine in honey badgers (Mellivora capen- Accumulations of multiple scats might have sis). Although 2 fox scats near each other might a stronger smell than a single scat, making it mark the beginning of a new latrine site, we easier for dogs to detect latrines than single believe that 2 scats might also be deposited scats. We reasoned that, if this hypothesis were near each other simply by chance or mark a true, latrines would, on average, be found at site where 1 fox had encountered a scat of greater distances from the transect line than another fox and deposited another in response single scats. The mean distance for single scats even though this site was not habitually visited. was 4.3 ± 6.3 m and that for latrines was 5.9 ± Because of these considerations, we defined a 7.3, but there was no significant difference in latrine as an accumulation of at least 3 kit fox the distances recorded for single scats and scats. We recorded whether fox scats were latrines (2-tailed t test, t = 1.801, df = 238, P found alone or in a latrine and whether they = 0.07). were located on or near a conspicuous object. In urban Bakersfield, where coyotes are rare In 2003 we also recorded how many fox and (B. Cypher personal communication), we ob- coyote scats were present at most latrines used served that kit foxes use latrines even when no by both species. coyotes are present. Therefore, use of latrines We found kit fox scats at 161 locations in appears to be a typical feature of kit fox biol- 2002. Eighty-nine locations had only 1 scat, 11 ogy. Although the functions of these kit fox had 2, and 61 had 3 or more, thus meeting our latrines are not known, they likely play some definition of a latrine. Latrines were rarely role in chemical communication. In several located near fox dens: only 2 of the 61 latrines carnivore species where latrine use has been were near dens. Seven of the 61 latrines marked well studied, such as badgers (Stewart et al. a conspicuous object (1 cement object, 1 tire, 2001, 2002) and hyenas (Gorman and Mills 546 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Fig. 1. Number of locations at which various numbers of kit fox scats were found in the Lokern Natural Area in 2003.

1984, Mills and Gorman 1987), latrines serve described with piles of feces left by coyotes on to mark the territories of the resident social a footpath to mark the entrances of descending group. trails through thick brush in coastal California. Coyotes are known to urinate over places where wolves have urinated (Paquet 1991). If We thank the Christensen Fund, the Alter- they tend to urinate and defecate on kit fox natives Research and Development Foundation, scats, this would lead to the accumulation of the Smithsonian Institution Scholarly Studies coyote scats at some kit fox latrines. However, Program, the Friends of the National Zoo, the we found 3 latrines that contained only coyote National Center for Environmental Research, scats, indicating that coyotes use latrines on the EPA STAR program, and the Abbott Fund our study area even if no kit fox scats are pre- for financial support. Alice Whitelaw, Aimee sent. Thus, it is not clear whether joint latrines Hurt, and Mike Smith provided exceptional result from coyotes responding to kit fox scats, field assistance. Brian Cypher and James Mur- kit foxes responding to coyote scats, or both. doch made helpful comments on the draft If we had trained dogs to locate coyote scats manuscript. rather than kit fox scats and had searched a larger geographic area, we likely would have LITERATURE CITED found more latrines that contained only coyote scats. Although several detailed studies of scent ALLEN, J.J., M. BEKOFF, AND R.L. CRABTREE. 1999. An ob- marking in coyotes do not mention latrine use servational study of coyote (Canis latrans) scent- (Bowen and Cowan 1980, Wells and Bekoff marking and territoriality in Yellowstone National 1981, Bekoff and Wells 1986, Gese and Ruff Park. Ethology 105:289–302. BEKOFF, M., AND M.C. WELLS. 1986. Social ecology and 1997, Allen et al. 1999), Camenzind (1978) men- behavior of coyotes. Advances in the Study of tions 2 coyote latrines discovered in Wyoming. Behavior 16:251–238. One was in an empty hay shed and contained BEGG, C.M., J. BEGG, T. DU TOIT, AND M.G.L. MILLS. over 500 scats. The second, under an old 2003. Scent-marking behaviour of the honey badger, Mellivora capensis (Mustelidae), in the southern wooden bridge crossing a dry creek, also con- Kalahari. Animal Behaviour 66:917–929. tained several hundred scats. Both latrines were BOWEN, W.D., AND I.M. COWAN. 1980. Scent marking in in inconspicuous, covered areas and Camen- coyotes. Canadian Journal of Zoology 58:473–480. zind (1978) did not think that they “served ter- BRATTSTROM, B.H. 1999. Trail marking by coyotes, Canis ritorial functions.” However, it seems likely that latrans. Southwestern Naturalist 44:405–406. CAMENZIND, F.J. 1978. Behavioral ecology of coyotes on coyotes do use accumulations of scats for the National Elk Refuge, Jackson, Wyoming. Pages chemical communication under some circum- 267–294 in M. Bekoff, editor, Coyotes: biology, be- stances, as in the example Brattstrom (1999) havior, and management. Academic Press, New York. 2004] NOTES 547

EGOSCUE, H.J. 1962. Ecology and life history of the kit fox SMITH, D., K. RALLS, B. DAVENPORT, B. ADAMS, AND J. in Tooele County, Utah. Ecology 43:481–497. MALDONADO. 2001. Canine assistants for conserva- GESE, E.M., AND R.L. RUFF. 1997. Scent marking by coy- tionists. Science 291:435. otes, Canis latrans: the influence of social and eco- SMITH, D.A., K. RALLS, A. HURT, B. ADAMS, M. PARKER, B. logical factors. Animal Behavior 54:1155–1166. DAVENPORT, M.C. SMITH, AND J.E. MALDONADO. GORMAN, M.L., AND M.G.L. MILLS. 1984. Scent marking 2003. Detection and accuracy rates of dogs trained strategies in hyenas (Mammalia). Journal of Zoology to find scats of San Joaquin kit foxes (Vulpes macrotis 202:535–537. mutica). Animal Conservation 6:339–346. GORMAN, M., AND B.J. TROWBRIDGE. 1989. The role of odor STEWART, P.D., D.W. MACDONALD, C. NEWMAN, AND C.L. in the social lives of carnivores. Pages 57–88 in J.L. CHEESEMAN. 2001. Boundary faeces and matched Gittleman, editor, Carnivore behavior, ecology, and advertisement in the European badger (Meles meles): evolution. Cornell University Press, Ithaca, New York. a potential role in range exclusion. Journal of Zool- KLEIMAN, D. 1966. Scent marking in the Canidae. Sym- ogy, London 255:191–198. posia of Zoological Society of London 18:66–167. STEWART, P.D., D.W. MACDONALD, C. NEWMAN, AND F.H. KRUUK, H. 1995. Wild otters: predation and populations. TATTERSALL. 2002. Behavioral mechanisms of infor- Oxford University Press, Oxford, UK. mation transmission and reception by badgers, Meles LAUGHRIN, L.L. 1977. The island fox: a field study of its meles, at latrines. Animal Behaviour 63:999–1007. behavior and ecology. Doctoral dissertation, Univer- TRAPP, G.R. 1978. Comparative behavioral ecology of the sity of California, Santa Barbara. ringtail (Bassariscus astutus) and gray fox (Urocyon MACDONALD, D.W. 1979. Some observations and field cinereoargenteus) in Southwestern Utah. Carnivore experiments on the urine marking behavior of the 1:3–32. red fox, Vulpes vulpes L. Zeitschrift für Tierpsycho- U.S. FISH AND WILDLIFE SERVICE. 1998. Recovery plan logie 51:1–22. for upland species of the San Joaquin Valley, Califor- ______. 1985. The carnivores: order Carnivora. Pages nia. Region 1, Portland, OR. 319 pp. 619–722 in R.E. Brown and D.W. Macdonald, edi- VILA, C., V. URIOS, AND J. CASTROVIEJO. 1994. Use of faeces tors, Social odors in mammals. Clarendon Press, for scent marking in Iberian wolves (Canis lupus). Oxford, UK. Canadian Journal of Zoology 72:374–377. MILLS, M.G.L., AND M.L. GORMAN. 1987. The scent- WASSER, S.K., B. DAVENPORT, E.R. RAMAGE, K.E. HUNT, marking behavior of the spotted hyaena, Crocuta M. PARKER, C. CLARKE, AND G. STENHOUSE. 2004. crocuta, in the southern Kalahari. Journal of Zoology Scat detection dogs in wildlife research and manage- 212:483–497. ment: application to grizzly and black bears in the O’FARRELL, T.P. 1987. Kit fox. Pages 423–431 in M. Novak, Yellowhead Ecosystem, Alberta, Canada. Canadian J.A. Baker, M.E. Obbard, and B. Malloch, editors, Journal of Zoology 82:475–492. Wild furbearer management and conservation in WELLS, M.C., AND M. BEKOFF. 1981. An observational North America. Ministry of Natural Resources, study of scent marking in coyotes, Canis latrans. Ani- Ontario, Canada. mal Behaviour 39:332–350. PAQUET, P.C. 1991. Scent-marking behavior of sympatric WHITE, P.J., T.J. KREEGER, J.R. TESTER, AND U.S. SEAL. wolves (Canis lupus) and coyotes (C. latrans) in Rid- 1989. Anal-sac secretions deposited with feces by ing Mountain National Park. Canadian Journal of captive red foxes (Vulpes vulpes). Journal of Mam- Zoology 69:1721–1727. malogy 70:814–816. PETERS, R.P., AND L.D. MECH. 1975. Scent-marking in wolves. American Scientist 63:628–637. Received 19 May 2003 SILLERO-ZUBIRI, C., AND D.W. MACDONALD. 1998. Scent- Accepted 7 April 2004 marking and territorial behavior in Ethiopian wolves (Canis simensis). Journal of Zoology 245:351–361. Western North American Naturalist 64(4), © 2004, pp. 548–550

FIRST NESTING RECORD OF BLACK-THROATED GRAY WARBLER (DENDROICA NIGRESCENS) FOR MONTANA

Paul Hendricks1

Key words: Black-throated Gray Warbler, Dendroica nigrescens, breeding, Montana, range expansion.

The Black-throated Gray Warbler (Dendroica Marks personal communication), and most re- nigrescens) in the Rocky Mountain region of cently I observed a singing male on 29–30 North America reaches the northern limit of May 2004 about 0.8 km down-canyon from the its breeding range in central to northwestern nest location described in this paper. Wyoming and southern Idaho (Dunn and Gar- On 3 June 2001, Lisa Hendricks and I dis- rett 1997, Guzy and Lowther 1997, American covered a pair of Black-throated Gray War- Ornithologists’ Union 1998). According to Fit- blers in T9S, R26E, Sec. 3 of Bear Canyon in ton and Scott (1984) and Dorn and Dorn (1990), the Pryor Mountains, Carbon County, Mon- breeding in Wyoming is confirmed or likely in tana (45°05N, 108°31W; 1646 m elevation), Utah juniper ( Juniperus osteosperma) wood- about 29 km northwest of Lovell, Big Horn lands of the southwestern and south central County, Wyoming. The slopes along the part parts of the state and suspected (breeding sea- of Bear Canyon where the warblers were ob- son records) in the Thermopolis area of the served are dominated by Utah juniper and northwest. In Idaho the species is a confirmed limber pine (Pinus flexilis), and the canyon breeder in juniper woodlands across the south bottom proper is dominated by a dense growth and as far north as the Snake River corridor of sagebrush (Artemisia tridentata), other shrubs, east of Idaho Falls (Burleigh 1972, Stephens and narrowleaf cottonwood (Populus angustifo- and Sturts 1998). In Alberta there are a few lia). The pair of warblers was active in the can- breeding season records, but as yet no evidence yon bottom, foraging in medium-sized shrubs of breeding (Guzy and Lowther 1997, Slater (3–4 m tall) of skunkbrush (Rhus trilobata) and and Hudon 2002). mountain maple (Acer glabrum). The male did Breeding of the Black-throated Gray War- not sing, and we observed no behavior that bler in Montana has not been documented, suggested breeding was underway. Jeff Marks although the species has been reported in the visited Bear Canyon on 6 June, drew the male state on 5 prior occasions (Wright 1996), all close with a taped song, and captured him in a but once (15 December) in spring and early mist net on the nearby slopes of Utah juniper summer (25 March–15 June). Here I provide where the male was photographed and released; details of the 1st nesting record (6th state age of the male was later determined from the record) in Montana, 350 km north of the near- photograph as after 2nd year (ASY). The female est area in Wyoming where breeding has been was neither seen nor heard. confirmed near Rock Springs and about 150 On 16 June at the same site, I heard the male km north of the nearest area where breeding warbler singing early in the morning from the is suspected near Thermopolis (Fitton and Scott crowns of the surrounding cottonwoods in the 1984, Dorn and Dorn 1990); the record was canyon bottom, heard the female “chip” nearby, briefly noted previously by Martin (2001). An and found the nest when she returned to it additional sighting (an unpaired adult male and settled to incubate. The nest was an open observed on 23 May 2003 in the Sweetgrass cup anchored within a dense growth of vines Hills, Liberty County) has since been accepted (Clematis ligusticifolia) suspended from the by the Montana Rare Birds Committee (J. sides of a cottonwood snag in the center of the

1Montana Natural Heritage Program, 909 Locust Street, Missoula, MT 59802.

548 2004] NOTES 549 canyon bottom. The nest rim was 3.3 m above- the state; in this juniper-dominated landscape ground and about 0.8 m out from the cotton- deciduous shrubs are especially concentrated wood trunk. I observed the nest for about 90 along streambeds in canyon bottoms. There are minutes, during which the female departed areas of the Pryor Mountains where the slopes the nest after incubation bouts of 20 minutes, lining lower canyons are covered with exten- and was absent 3–5 minutes each time; the sive mixed stands of Utah juniper and limber male never fed her at the nest nor visited the pine, making the region appear much like typ- nest. Both birds concentrated their foraging in ical pinyon-juniper habitat. Indeed, the 1st the deciduous vegetation of the canyon bot- documented breeding in Montana of the Blue- tom within 100 m of the nest. gray Gnatcatcher (Polioptila caerulea), another On 28–29 June, Jeff Marks, John Carlson, bird species closely associated with Utah jun- and I visited Bear Canyon to inspect the nest iper habitat in the Great Basin (Fitton and Scott contents and take some additional measure- 1984, Pavlacky and Anderson 2001), occurred ments. During our stay we neither saw nor during the last decade in the same Montana heard the pair of warblers. It is possible that canyon (Wright 1996). the pair abandoned their nest during one of In Wyoming the breeding range of the the severe snow and rainstorms that passed Black-throated Gray Warbler through the 1970s over the Pryor Mountains that year in late included only the southwestern portions of the June. The nest contained 3 cold eggs (18.9 × state (Cary 1917, McCreary 1939, AOU 1957, 13.0 mm, 18.8 × 13.0 mm, 3rd egg broken) of 1983). Beginning in the 1980s it was consid- whitish base color with fine brown macula- ered locally common in juniper stands of the tions scattered generally uniformly over the Casper area in central Wyoming (Fitton and eggshell. Outside nest measurements were 7.3 Scott 1984, AOU 1998). In Montana ornitho- cm diameter at the rim and 7.0 cm deep; inside logical inventories in Utah juniper habitat of (cup) measurements were 4.7 cm diameter at the Pryor Mountains and Bighorn Canyon the rim and 4.1 cm deep. The nest, anchored National Recreation Area of Carbon County to the vine by only 2 stems, was built of plant during the 1960s and 1980s failed to document fibers and stems with strips of sagebrush bark the presence of Black-throated Gray Warbler and a few feathers woven in; the lining was (DeLap and Thompson 1962, Anderson et al. mostly feathers, with lesser amounts of fine fur 1987). These data, though limited in scope and and guard hairs. We collected the nest and intensity, suggest a recent northward expan- eggs; unfortunately, the 2 intact eggs were sion of the breeding range for this species in broken during attempts to blow them. The egg Wyoming and into Montana, similar to range fragments and nest are deposited in the Philip expansions for several other species elsewhere L. Wright Zoological Museum at the Univer- in the West (Johnson 1994), but especially the sity of Montana, Missoula (UMZM 18328). Blue-gray Gnatcatcher (McCreary 1939, Find- In western North America, particularly in holt 1983, Fitton and Scott 1984, Dorn and the Great Basin, the Black-throated Gray War- Dorn 1990, Wright 1996, Walsh et al. 1998), bler is associated with arid pine and juniper another juniper specialist. Unlike the gnat- habitats, especially pinyon–Utah juniper wood- catcher, there is as yet no evidence that a lands (Fitton and Scott 1984, Dorn and Dorn breeding population of Black-throated Gray 1990, Dunn and Garrett 1997, Guzy and Low- Warbler has become established in Montana ther 1997). In southern Wyoming this warbler (personal observation), although the appear- breeds in areas with greater-than-expected ance of another male in 2004 indicates that nearby availability of pinyon pine, greater stem eventual regular breeding is possible. density of Utah juniper, and greater prevalence Factors leading to the recent appearance of of large shrubs in the understory (Pavlacky and breeding Black-throated Gray Warblers and Anderson 2001), a description that broadly Blue-gray Gnatcatchers in Montana are not matches the general habitat occupied in Mon- known with certainty, but may be related to tana. The southern lower-elevation slopes regional changes in climate and/or land man- (1300–1900 m elevation) of the Pryor Moun- agement practices. Evidence suggests that tains and nearby Bighorn Mountains in Mon- Utah juniper stands in southern Montana and tana have many floral elements typical of Great elsewhere along the margins of the Bighorn Basin vegetation (Dorn 1978, Kratz 1988), in- Basin were quite isolated from major areas of cluding the only Utah juniper woodlands in occurrence to the south and west until the last 550 WESTERN NORTH AMERICAN NATURALIST [Volume 64

2000 years BP (Jackson et al. 2002); distribu- DORN, J.L., AND R.D. DORN. 1990. Wyoming birds. Moun- tion and establishment of juniper may be lim- tain West Publishing, Cheyenne, WY. DORN, R.D. 1978. Great Basin vegetation in Carbon County, ited to thermal belts where less frequent frosts Montana. Madroño 25:105–106. occur (Knight 1994). Thus, the gradual spread DUNN, J.L., AND K.L. GARRETT. 1997. A field guide to the of juniper would be expected in the presence warblers of North America. Houghton Mifflin Com- of an increasingly moderate climate, such as pany, New York. has occurred during the past 100 years, and FINDHOLT, S.L. 1983. First nest records for the Plains Tit- mouse and Blue-gray Gnatcatcher in Wyoming. Great especially since 1976 (Walther et al. 2002, Root Basin Naturalist 43:747–748. et al. 2003). In addition, fire suppression during FITTON, S.D., AND O.L. SCOTT. 1984. Wyoming’s juniper the last century has contributed to the invasion birds. Western Birds 15:85–90. of juniper in many areas of the West (Burkhardt GUZY, M.J., AND P.E. LOWTHER. 1997. Black-throated Gray and Tisdale 1976, Knight 1994), including the Warbler (Dendroica nigrescens). In: A. Poole and F. Gill, editors, The birds of North America, No. 319. Big Horn Basin of Wyoming. The combination The Academy of Natural Sciences, Philadelphia, PA, of warmer climate and fire suppression could and The American Ornithologists’ Union, Washing- have contributed indirectly to the recent north- ton, DC. ward expansion of the breeding distribution of JACKSON, S.T., M.E. LYFORD, AND J.L. BETANCOURT. 2002. A 4000-year record of woodland vegetation from some juniper-specialist birds into Montana Wind River Canyon, central Wyoming. Western through an increase in suitable breeding habi- North American Naturalist 62:405–413. tat over their region of occupancy. JOHNSON, N.K. 1994. Pioneering and natural expansion of breeding distributions in western North American I am grateful for companionship in the field birds. Studies in Avian Biology 15:27–44. KNIGHT, D.H. 1994. Mountains and plains: the ecology of of John Carlson, Lisa Hendricks, and Jeff Marks Wyoming landscapes. Yale University Press, New during various stages of this discovery. I thank Haven, CT. Jeff Marks for sending a copy of his photo- KRATZ, A. 1988. Preliminary descriptions of Great Basin- graph of the male warbler to Peter Pyle, who type vegetation occurring in Carbon County, Montana, made the age determination; the original pho- U.S.A. Proceedings of the Montana Academy of Sci- ences 48:47–55. tograph is deposited in the rare birds file of MARTIN, R. 2001. Northern Great Plains. North American the Montana Audubon Society, Helena. Support Birds 55:447–449. for some of the visits was provided through MCCREARY, O. 1939. Wyoming bird life. Revised edition. the Montana Field Office of The Nature Con- Burgess Publishing Company, Minneapolis, MN. PAVLACKY, D.C., JR., AND S.H. ANDERSON. 2001. Habitat servancy and the Montana Natural Heritage preferences of pinyon-juniper specialists near the Program. Earlier versions of this paper bene- limit of their geographic range. Condor 103:322–331. fited from comments by Kimball Garrett, ROOT, T.L., J.T. PRICE, K.R. HALL, S.H. SCHNEIDER, C. Jocelyn Hudon, and an anonymous reviewer. ROSENZWEIG, AND J.A. POUND. 2003. Fingerprints of global warming on wild animals and plants. Nature 421:57–60. LITERATURE CITED SLATER, A., AND J. HUDON. 2002. Fourth report of the Alberta Bird Record Committee. Alberta Naturalist AMERICAN ORNITHOLOGISTS’ UNION. 1957. Check-list of 32:116–117. North American birds. 5th edition. American Orni- STEPHENS, D.A., AND S.H. STURTS. 1998. Idaho bird distri- thologists’ Union, Washington, DC. bution. 2nd edition. Idaho Museum of Natural His- ______. 1983. Check-list of North American birds. 6th edi- tory Special Publication 13. tion. American Ornithologists’ Union, Washington, DC. WALSH, J.J., A. CRUZ, M.E. BERRY, J.F. CHACE, AND D.M. ______. 1998. Check-list of North American birds. 7th edi- EVANS. 1998. Breeding range expansion of the Blue- tion. American Ornithologists’ Union, Washington, DC. gray Gnatcatcher along the northern Colorado Front ANDERSON, S.H., W.A. HUBERT, C. PATTERSON, A.J. REDDER, Range. Journal of the Colorado Field Ornithologists AND D. DUVALL. 1987. Distribution of vertebrates of 32:166–172. the Bighorn Canyon National Recreation Area. Great WALTHER, G.-R., E. POST, P. CONVEY, A. MENZEL, C. Basin Naturalist 47:512–521. PARMESAN, T.J.C. PEEBEE, J.-M. FROMENTIN, ET AL. BURKHARDT, J.W., AND E.W. TISDALE. 1976. Causes of 2002. Ecological responses to recent climate change. juniper invasion in southwestern Idaho. Ecology 57: Nature 416:389–395. 472–484. WRIGHT, P.L. 1996. Status of rare birds in Montana, with BURLEIGH, T.D. 1972. Birds of Idaho. Caxton Printers, Ltd., comments on known hybrids. Northwestern Natu- Caldwell, ID. ralist 77:57–85. CARY, M. 1917. Life zone investigations in Wyoming. North American Fauna 42. Received 10 May 2003 DELAP, D., AND G. THOMPSON. 1962. Summer birds of the Accepted 29 January 2004 Pryor Mountains. Proceedings of the Montana Acad- emy of Sciences 22:14–20. Western North American Naturalist 64(4), © 2004, pp. 551–553

MINK PREDATION ON RADIO-TAGGED TROUT DURING WINTER IN A LOW-GRADIENT REACH OF A MOUNTAIN STREAM, WYOMING

Jason W. Lindstrom1,2 and Wayne A. Hubert1

Key words: mink, Mustela vison, brook trout, Salvelinus fontinalis, cutthroat trout, Oncorhynchus clarki, predation, natural mortality, stream, winter.

Sources of natural mortality among fluvial river downstream from a large reservoir in salmonids are poorly understood, and contri- Wyoming (Simpkins 1997). Similarly, Jakober butions by predators to natural mortality rates (1995) reported the loss of radio-tagged bull have not been measured. Mink (Mustela vison) trout (Salvelinus confluentes) to mink preda- are effective fish predators (Dunston 1978, tion in late autumn after ice formed in a stream 1983, Linscombe et al. 1982, Eagle and Whit- in Montana. Mountain streams can provide man 1987) that can contribute to natural mor- good habitat for mink and support substantial tality of salmonids in streams (Erlinge 1969, populations, particularly along low-gradient Alexander 1976, Melquist et al. 1981, Whitman segments with abundant willows (Salix spp.) in 1981). Evidence that mink predation is a sig- riparian areas and beaver (Castor canadensis) nificant source of natural mortality among flu- ponds along the stream (Liscombe et al. 1982, vial salmonids has been observed. For example, Eagle and Whitman 1987). Heggenes and Borgstrom (1988) reported that We conducted a study that provided insight mink presence led to a marked increase in into the possible contribution of mink preda- mortality of juvenile Atlantic salmon (Salmo tion to winter mortality of salmonids in moun- salar) and brown trout (Salmo trutta) in streams. tain streams. It was part of a larger study to However, Burgess and Bider (1980) concluded assess habitat use and movements of cutthroat that mink predation on brook trout (Salvelinus trout (Oncorhynchus clarki) and brook trout fontinalis) was not a significant source of mor- from fall through winter (Lindstrom 2003). tality in a stream improved to enhance brook The study was conducted on a 7-km reach of trout. South Cottonwood Creek in the Green River Winter is considered a time of stress among watershed on the Wyoming Range in western fluvial salmonids. Causes of overwinter mortal- Wyoming at an elevation of 2460–2530 m ity are poorly understood, but evidence exists above mean sea level. Mean wetted width of that mortality can be associated with dynamic the study reach was 7 m, mean channel slope ice conditions, starvation, predation, or an inter- was 0.9%, late summer discharge was 0.6–0.8 action of these factors (Simpkins and Hubert m3 ⋅ s–1, and minimum winter discharge was 2000, Simpkins et al. 2000). Predation by mink 0.2–0.3 m3 ⋅ s–1. Beaver were common through- can be a source of winter mortality because out the study reach. Dominant riparian vege- mink remain active during winter (Marshall tation was willow. Conifers and deciduous 1936, Sealander 1943), and fishes may be more trees were absent from the riparian zone, and susceptible to mink predation because low large woody debris was rare in the channel. water temperatures reduce the metabolic rates Habitat improvements had been conducted and abilities of fish to escape attacks (Gerell between 1984 and 1994 on a 2-km segment of 1967). Predation by mink on rainbow trout the stream in the middle of the study reach to (Oncorhynchus mykiss) during winter was ob- stabilize the stream channel and increase pool served during a telemetry study in a regulated habitat (Binns 1999).

1U.S. Geological Survey, Wyoming Cooperative Fish and Wildlife Research Unit, University of Wyoming, Laramie, WY 82071-3166. The Unit is jointly supported by the University of Wyoming, Wyoming Game and Fish Department, Wildlife Management Institute, and U.S. Geological Survey. 2Present address: Tribal Fisheries Department, Salish and Kootenai Tribes of the Flathead Nation, Box 278, Pablo, MT 59855.

551 552 WESTERN NORTH AMERICAN NATURALIST [Volume 64

Table 1. Species, total length, date of last observation of live fish, and date of observation following predation by mink for 9 of 50 cutthroat trout and brook implanted with radio transmitters in South Cottonwood Creek, Wyoming, in autumn 2002 and monitored throughout the winter. Total length Date last Date observed to have Species (mm) observed alive been predated by mink Brook trout 280 11 November 18 November Brook trout 235 18 November 25 November Brook trout 260 18 November 25 November Brook trout 280 25 November 10 December Brook trout 250 18 December 2 January Brook trout 245 2 January 15 January Brook trout 280 15 January 30 January Cutthroat trout 300 15 January 30 January Cutthroat trout 255 12 February 26 February

We captured 25 adult cutthroat trout (mean 17 brook trout losses, we attributed 7 to mink total length [TL] = 333 mm, mean weight = predation, 4 to transmitter failures, 1 to spawn- 406 g) and 25 adult brook trout (mean TL = ing behavior, 1 to movement out of the study 264 mm, mean weight = 236 g) by electrofish- area, and 4 to unknown causes. Among the fish ing during autumn (25 September–1 October identified as losses to mink predation (Table 2002) over the length of the study reach. We 1), the 2 cutthroat trout (255 and 300 mm TL) then surgically implanted the fish with radio were shorter than the mean length (333 mm transmitters (Model F170, Advanced Teleme- TL) of cutthroat trout tagged, but the 7 brook try Systems, Isanti, MN; mean weight = 3.1 g) trout (mean TL = 261) were similar in length using the shielded-needle technique (Ross to those tagged (mean = 264 mm TL). Preda- and Kleiner 1982) and released them into the tion by mink on tagged fish occurred through- pool from which they were captured upon their out the winter from the middle of November recovery (0.5–1.0 hour). Fish were tracked through February, but the first 5 fish lost to from the ground during daylight hours using a mink predation were all brook trout (Table 1). directional loop antenna and scanning receiver We commonly observed mink sign in the ripar- (Challenger R2000, Advanced Telemetry Sys- ian area throughout the study reach in the tems, Isanti, MN), and locations of transmit- vicinity of tagged fish. Fish believed to have ters were determined to within a 2-m radius been killed by mink were found in both beaver (Simpkins and Hubert 1998). Fish were located ponds and lateral scour pools prior to their loss. at 1-week intervals during October and Locations of radio tags from fish identified as November and biweekly from December to having been killed by mink were generally mid-March. within 100 m of the last recorded location of During the study period we attributed fish the fish. On each of 2 sampling dates (25 losses to mink predation, transmitter failures, November and 30 January), 2 fish were deter- movements out of the study reach, spawning mined to have been predated by mink, and in behavior, or unidentified causes. Predation by both cases the fish were found in the same mink was inferred when transmitter locations beaver pond when previously located. were over land and below ground and there Mink predation was an important source of were mink signs (tracks or scat) in the imme- mortality among tagged trout during our study. diate vicinity of the transmitter locations. We believe mink killed at least 8% of tagged Transmitters were considered to have failed if cutthroat trout and 28% of tagged brook trout. weakened signals or slowed pulse frequencies Several factors may have contributed to this had been observed during previous location high level of predation by mink. Three areas of efforts. groundwater inflow prevented surface ice from We tracked 14 cutthroat trout and 8 brook totally covering 3 different 250- to 1000-m trout to the end of the study in mid-March segments of the study reach (Lindstrom 2003). 2003. Among 11 cutthroat trout losses, 6 were This may have enabled enhanced predatory attributed to transmitter failures, 2 to mink behavior by mink in these reaches with open predation, and 3 to unknown causes. Among water, but fish were also lost from segments of 2004] NOTES 553 the stream that had total ice cover throughout GERELL, R. 1967. Food selection in relation to habitat in the study. It is possible that surgical implanta- mink (Mustela vison Schreber) in Sweden. Viltrevy 5:1–38. tion of radio transmitters made the study fish HEGGENES, J., AND R. BORGSTROM. 1988. Effects of mink, more vulnerable to mink predation than other Mustela vison Schreber, predation on cohorts of juve- fish in the stream; however, the body burden nile Atlantic salmon, Salmo salar L., and brown trout, created by the transmitters (maximum 1.6% of S. trutta L., in three small streams. Journal of Fish weight) was substantially less than the 2% max- Biology 33:885–894. JAKOBER, M.J. 1995. Autumn and winter movement and imum recommended for radiotelemetry studies habitat use of resident bull trout and westslope cut- with fish (Winter 1996). Brook trout may have throat trout in Montana. Master’s thesis, Montana been more vulnerable to predation by mink State University, Bozeman. than were cutthroat trout due to their fall LINDSTROM, J.W. 2003. Habitat selection and movements of adult cutthroat trout and brook trout from fall spawning behavior and smaller sizes. Observa- through winter in a meandering alluvial valley stream. tions from our study and other recent studies Master’s thesis, University of Wyoming, Laramie. (Jakober 1995, Simpkins 1997) suggest that LINSCOMBE, G., N. KINLER, AND R.J. AULERICH. 1982. Mink. mink may have a substantial effect on natural Pages 629–643 in J.A. Chapman and G.A. Feldhamer, mortality rates of fluvial salmonids during win- editors, Wild mammals of North America: biology, management, and economics. Johns Hopkins Uni- ter in the Rocky Mountain region. Further re- versity Press, Baltimore, MD. search is needed to identify the extent to which MARSHALL, W.H. 1936. A study of the winter activity of mink predation may contribute to winter mor- the mink. Journal of Mammalogy 17:382–392. tality rates in mountain streams. MELQUIST, W.E., J.S. WHITMAN, AND M.G. HORNOCKER. 1981. Resource partitioning and coexistence of sympatric mink and river otter populations. Pages 187– We thank T. Annear, D. Miller, and T. Wesche 220 in J.A. Chapman and D. Pursley, editors, Proceed- for assistance in planning the study; C. Bar- ings of Worldwide Furbearer Conference, Frostburg, rineau for assistance with data collection; and MD. the Wyoming Game and Fish Department, ROSS, M.J., AND C.F. KLEINER. 1982. Shielded-needle technique for surgically implanting radio-frequency trans- especially H. Sexauer and P. Cavalli, for tech- mitters in fish. Progressive Fish-Culturist 44:41–43. nical assistance and funding the project. SEALANDER, J.A. 1943. Winter food habits of mink in southern Michigan. Journal of Wildlife Management LITERATURE CITED 7:411–417. SIMPKINS, D.G. 1997. Winter movements and habitat use by small rainbow trout in the Big Horn River, below ALEXANDER, G.R. 1976. Diet of vertebrate predators on Boysen Reservoir, Wyoming. Master’s thesis, Univer- trout waters in north central Lower Michigan. Michi- sity of Wyoming, Laramie. gan Department of Natural Resources, Fisheries Re- SIMPKINS, D.G., AND W.A. HUBERT. 1998. A technique for search Report 1839. estimating the accuracy of fish locations identified BINNS, N.A. 1999. A compendium of trout stream habitat by radiotelemetry. Journal of Freshwater Ecology 13: improvement projects done by the Wyoming Game 263–268. and Fish Department, 1953–1998. Wyoming Game ______. 2000. Drifting invertebrates, stomach contents, and and Fish Department, Fish Division, Cheyenne. body conditions of juvenile rainbow trout from fall BURGESS, S.A., AND J.R. BIDER. 1980. Effects of stream through winter in a Wyoming tailwater. Transactions habitat improvements on invertebrates, trout popu- of the American Fisheries Society 129:1176–1184. lations, and mink activity. Journal of Wildlife Man- SIMPKINS, D.G., W.A. HUBERT, AND T.A. WESCHE. 2000. agement 44:871–880. Effects of fall-to-winter changes in habitat and frazil DUNSTON, N. 1978. The fishing strategy of the mink (Mus- ice on the movements and habitat use of juvenile tela vison): time-budgeting of hunting effort? Behav- rainbow trout in a Wyoming tailwater. Transactions ior 67:157–177. of the American Fisheries Society 129:101–118. ______. 1983. Underwater hunting behavior of the mink WHITMAN, J.S. 1981. Ecology of the mink (Mustela vison) (Mustela vison Schreber): an analysis of constraints in west-central Idaho. Master’s thesis, University of on foraging. Acta Zoologica Fennica 174:201–203. Idaho, Moscow. EAGLE, T.C., AND J.S. WHITMAN. 1987. Mink. Pages 615– WINTER, J. 1996. Advances in underwater telemetry. Pages 624 in M. Novak, J.A. Baker, M.E. Obbard, and B. 555–590 in B.R. Murphy and D.W. Willis, editors, Malloch, editors, Wild furbearer management and Fisheries techniques. 2nd edition. American Fisheries conservation in North America. Ontario Ministry of Society, Bethesda, MD. Natural Resources, Ontario, Canada. ERLINGE, S. 1969. Food habits of the otter Lutra lutra L. Received 2 October 2003 and mink Mustela vison Schreber in a trout water in Accepted 20 January 2004 southern Sweden. Oikos 20:1–7.

Western North American Naturalist

INDEX

Volume 64—2004

AUTHOR INDEX

Anderson, Stanley H., 86, 376, Griffin, David J., 59 Lindbloom, Andrew J., 338 385 Griswold, Terry L., 465 Lindstrom, Jason W., 551 Auger, Janene, 166 Grogan, William L., Jr., 433 Lyman, R. Lee, 1 Austin, Jane E., 277 Guenther, Debra, 293 Lynch, Ann M., 7 Gul, Bilquees, 193 Ballard, Warren B., 53 Madden, Elizabeth M., 72 Bartholow, John, 406 Haas, Glenn E., 346, 514 Marshall, Jonathan C., 175 Baugh, Jim, 538 Hamilton, Paul S., 137 Matson, John O., 359 Beckmann, Jon P., 269 Hardy, Paul C., 59 McAllister, Chris T., 514 Belk, Mark C., 409, 413 Harper, K.T., 482 McCafferty, W.P., 101 Bell, Adrian, 413 Harvey, James E., 38 McCreary, Brome, 403 Bell, Christopher J., 439 Heinrich, James, 38 Mealor, Brian A., 503 Benedict, Russell A., 231 Hendricks, Paul, 548 Miles, Scott, 93 Bess, James A., 518 Herzig, Ann L., 249 Mills, Michael D., 409 Black, Hal L., 166 Hild, Ann L., 503 Morrison, Michael L., 59, 202, Bogan, Michael A., 353 Horn, Michael J., 306 259 Brantley, S.L., 155 Hubert, Wayne A., 78, 551 Mote, Kevin, 53 Brooks, Jim L., 135 Huebner, Candace Brindza, Mueller, Gordon A., 135, 306 Brown, Gregory K., 109 131 Murphy, Robert K., 72 Brown, Joshua E., 409 Huebschman, Jeffrey J., 353 Burke, Thomas E., 538 Hufbauer, Ruth A., 324 Nelson, C. Riley, 144 Hughes, Jonathan F., 109 Carriker, Melbourne R., 417 O’Neill, Kevin M., 518 Chagas, Amazonas, Junior, 532 Iko, William M., 492 Ogborn, Gary L., 166 Chichester, Lyle, 538 Ingelfinger, Franz, 385 Ovaska, Kristiina, 538 Christopherson, Kirsten, 202 Cifelli, Richard L., 184 Jacobus, Luke M., 101 Pavlacky, David C., Jr., 376 Crawford, John A., 331 Jiménez-Cruz, Elsa, 175 Pearl, Christopher A., 403 Jordan, D. Jess, 403 Perrins, Christopher, 275 Danley, Robert F., 72 Joyce, Michael P., 78 Phillips, Robert A., 433 Dinsmore, Stephen J., 492 Pokorny, Monica L., 219 Downer, John, 144 Kamler, Jan F., 53 Pope, Michael D., 331 Duralia, Thomas E., 208 Kemp, William P., 518 Ports, Mark A., 145 Khan, M. Ajmal, 193 Potts, Daniel L., 27 Engel, Richard E., 219 Knopf, Fritz L., 492 Pritchett, Clyde L., 166 Evangelista, Paul, 293 Kondratieff, Boris C., 497 Promessi, Rebecca L., 359 Krannitz, Pamela G., 208 Pyle, William H., 277 Favret, Colin, 364 Krausman, Paul R., 259 Flores, Mary, 359 Kucera, James R., 346 Rader, Russell B., 409 Floyd, Chris H., 471 Kuenzi, Amy J., 59 Ralls, Katherine, 544 Kwiatkowski, Matthew A., 137 Ramírez-Bautista, Aurelio, 175 Geluso, Keith, 353 Ramsey, R. Douglas, 421 Genoways, Hugh H., 396 Lackey, Carl W., 269 Reese, Kerry P., 338 Gilliland, Rickey L., 53 Lariviere, Serge, 273 Rissler, Peter H., 45 Goldberg, Stephen R., 141 Lawrence, Rick L., 312 Roberson, Don, 274 Gottesman, Amy B., 259 Lei, Simon A., 125 Roberts, Elizabeth A., 312 Gould, Carol Grant, 273 Leonard, William P., 538 Robertson, Ian, 265 Gould, James L., 273 Lesica, Peter, 93 Roe, Christopher M., 445

556 2003] INDEX 557

Roe, Kelly A., 445 Shultz, Leila, 421 Van Buren, Renée, 482 Roehrs, Zachary P., 396 Simmons, Rulon E., 274 Vinson, Mark R., 131 Rombough, Christopher J., 403 Sipes, Sedonia D., 465 Voegtlin, David J., 364 Rompola, Kevin M., 86 Six, Diana L., 257 Ruez, Dennis R., Jr., 439 Skema, Cynthia, 249 Weber, Darrell J., 193 Ruiter, Dave, 454 Smith, Deborah A., 544 West, George C., 274 Sopuck, Lennart, 538 White, Clayton M., 275, 417 Scoppettone, G. Gary, 38, 45 Spinelli, Gustavo R., 433 White, Jeremy A., 353 Sedgwick, James A., 406 Stewart, Sean, 293 Williams, David G., 27 Shaughnessy, Michael J., Jr., 184 Stohlgren, Thomas J., 293 Wilson, Nixon, 514 Shaw, Nancy L., 503 Sullivan, Brian K., 137 Woodward, David L., 433 Shea, Sean, 45 Svejcar, Tony J., 219 Sheley, Roger L., 219, 312 Zager, Peter, 338 Shelley, Rowland M., 257, 532 Tepedino, Vincent J., 465 Zoellick, Bruce W., 18 Shepherd, U.L., 155 Thorp, Richard A., 497 Zuellig, Robert E., 497

KEY WORD INDEX

Accipiter gentilis, 359 biodiversity, 155 cattle, 45 Achnatherum hymenoides, 503 biogeography, 1, 396 Caurinella idahoensis, 101 Acroptilon repens, 503 biological diversity, 518 Cecidomyiidae, 324 age, 409 biological soil crusts, 293 Centaurea solstitalis, 202, 338 agriculture, 1 biomass, 18, 306 Ceratopogonidae, 433 Alectoris chukar, 338 bird(s), 376 Charadrius montanus, 492 American abundance, 59 chemical communication, 544 Avocet, 277 biting midge, 433 Cheumatopsyche analis, 497 Coot, 277 Bitterroot Mountain Range, chromosome Anagapetus, 454 [Montana], 101 cytology, 109 ant(s), 166 black bears, 166, 269 races, 175 colonies, 166 blackbrush, 125 Chrysothamnus nauseosus, 249 anuran, 403 black-tailed prairie dog, 445 chuckwalla, 137 Aphididae, 364 Black-throated Gray Warbler, Chukar, 338 aquatic invertebrates, 131 548 Cicadellidae, 518 Arceuthobium microcarpum, 7 boreal mammals, 471 Cinara, 364 Arctomecon humilis, 482 Brachylagus idahoensis, 1 cliff chipmunk, 86 Arizona, 259, 532 breeding, 548 climatic change, 471 Sonoran Desert, 137 system, 465 climbing behavior, 406 Artemisia, 324 British Columbia, 538 Coleogyne ramosissima, 125 Arthrorhabdus, 532 Bromus tectorum, 293 Colorado River, [Nevada], 38, pygmaeus, 532 brook trout, 551 45 spinifer, 532 brush mouse, 259 community artificial release, 445 Bufo boreas boreas, 403 composition, 219 Asteraceae, 109, 465 organization, 376 Astereae, 109 California, 202 structure, 155 Atriplex rosea, 193 Beale Air Force Base, 202 competition, 413 Warner Mountains, 359 cone production, 208 badger, 184 Yuba County, 202 conservation, 1, 465 Baja California Sur, 532 Canada Goose, 277 contamination, stocking, 135 bats, 231 canid morphology, 406 corridors, 93 Beale Air Force Base, [Califor- Canis latrans, 45, 544 cottonwood, 93 nia], 202 cannibalism, 403 coyote, 45, 184, 544 beavers, 93 canyon mouse, 86 cryptobiotic crusts, 155 behavior, 10 carbon isotopes, 27 Culicoides reevesi, 433 climbing, 406 Cardaria draba, 503 cutthroat trout, 78, 551 Belize, 532 carnivore, 175 Cynomys ludovicianus, 445 biking, 125 carrion, 45 558 WESTERN NORTH AMERICAN NATURALIST [Volume 63 dabbling ducks, 277 habitat, 86, 259 Mammalia, 396 demography, 465, 482 improvement, 78 marmot, 471 Dendroica nigrescens, 548 management, 72, 277 Marmota, 396 density, 18, 306 type, 518 flaviventris, 471 desert, 59, 269 use, 59, 338, 376 Mexico, 175, 532 streams, 18 halophyte, 193 Baja California Sur, 532 sucker, 45 hantavirus, 259 microarthropods, 155 diet, 413 herbivory, 249 Microtus, 346 dimorphic seeds, 193 Hesperostipa comata 503, migration, 59, 78, 353 Diptera, 433 hibernation, 353 mink, 551 distribution, 306, 538 hiking, 125 Missoula County, [Montana], disturbance, 125 Holodiscus discolor, 471 257 ecology, 7 home range, 53 mixed-grass prairie, 72 diversity, 503 Homoptera, 518 Montana, 257, 548 Dorosoma cepedianum, 135 hybrid zone, 109 Bitterroot Mountain Range, Hydropsychidae, 497 101 ecoregions, 421 Missoula County, 257 Elaeagnus angustifolia, 93 Idaho, 257, 277, 338 montane wetland, 277 Elatobium abietinum, 7 Pacific Northwest, 538 morphometric(s), 364 endangered fish, 38 Snake River, 78 measurements, 492 Ephemerellidae, 101 southwest, 18 motorcycle, 125 Ephemeroptera, 101 IDW, 312 Mountain exotic, 7 information theory, 86 Plover, 492 invasion, 93 insect , 364 Quail, 331 feces, 544 impact, 7 mountain snails, 145 fire, 249, 293 interpolation modeling, 312 movements, 53 prescribed, 72 invasive museum specimens, 492 fish, 413 plant mapping, 312 Mustela vison, 551 survey, 45 plants, 503 myrmecophagy, 166 fleas, 346, 514 species, 7 floral traits, 109 inverse distance weighting, 312 native and translocated quail, food habits, 359 Iotichthys phlegethontis, 409 331 foraging, 413 ISSR, 503 native grasses, 503 forbs, 219 natural enemies hypothesis, 93 fragmentation, 385 juniper, 86 natural gas development, 385 functional group diversity, 219 Utah j. woodland, 376 natural mortality, 551 Juniperus osteosperma, 293 nearctic, 433 gall, 324 Nebraska, 231, 353, 396 geographic distribution(s), 231, Kansas, 396 Great Plains, 353 514 keystone species, 184 nesting ecology, 277 geographical information kit fox, 544 nests, 346 systems. See GIS. Kyle Canyon, [Nevada], 125 Nevada germination, 482 Colorado River, 38, 45 GIS, 421 Lake Tahoe, [Nevada], 269 Kyle Canyon, 125 gizzard shad, 135 largemouth bass, 38, 45 Lake Tahoe, 269 Glossosoma schmidi, 454 larvae, 403 southern, 125 Glossosomatidae, 454 Lasionycteris noctivagans, 353 Spring Mountains, 125 grassland(s), 219, 518 latrines, 544 White River, 38, 45 passerine, 72 leaf nutritional quality, 249 nonnative species, 293 grasses, native, 503 least chub, 409 North American monsoon, 27 gray fox, 406 leoni group, 433 Northern Goshawks, 359 grazing, 1 Lepidium papilliferum, 265 northern Great Plains, 72 Great Basin, 413, 471 Lepidomeda, 38 (North Dakota) biogeography, 145 life “northern population,” 257 Great Plains, [Nebraska], 353 cycle, 497 Nostoc parmeliodides, 101 northern, [North Dakota], 72 history, 38, 409 nutrition, 166 groundwater, 131 Little Dell Dam, [Utah], 439 growth, 137, 409 logistic regression, 86 longevity, 482 2003] INDEX 559 oil development, 385 red fox, 406 southwest Idaho, 18 Oncorhynchus redband trout, 18 spawning, 78 clarki, 551 redd, 78 species clarki bouvieri, 78 regression tree analysis, 293 diversity, 72, 219 mykiss gairdneri, 18 relocation, 269, 445 invasive, 7 Oregon, 331 reproduction, 141, 175, 231, keystone, 184 Oreohelix, 145 353, 465 nonnative, 293 loisae sp. nov, 145 reproductive anatomy, 538 richness, 219, 421 Oreortyx pictus, 331 Reptilia, 175 speckled dace, 45 otolith analysis, 409 restoration activities, 208 spinedace, 38, 45 outcrossing, 265 Rhopalomyia, 324 Sporobolus airoides, 503 riparian, 93 spotted knapweed, 312 parasitism, 324 ecosystems, 27 Spring Mountains, parasitoid, 324 forests, 396 [Nevada], 125 passerines, 385 roads, 385 springfish, 45 pelagic fisheries, 306 Rocky Mountains, 538 Squamata, 175 Peromyscus rodents, 202, 514 stage descriptions, 101 boylii, 259 rotation grazing, 72 stand thinning, 208 crinitus, 86 Russian knapweed, 312 stocking contamination, 135 truei, 86 Russian olive, 93 stream, 551 Phrynosomatidae, 175 shading, 18 phylogeography, 538 sagebrush, 324, 385 urban, 497 Picea salinity, 193 striped bass, 306 engelmannii, 7 Salvelinus fontinalis, 551 Stygobromus, 131 pungens, 7 sampling, 312 succession, 376 Pinus Sandhill Crane, 277 survival rates, 331 edulis, 364 Sauromalus, 137 swift fox, 184 monophylla, 364 scats, 544 sympatry, 109 pinyon mouse, 86 Sceloporus grammicus, 175 Plagopterini, 38 scent marking, 544 talus, 471 plant hybridization, 109 Scolopocryptops gracilis, 257 tamarisk, 93 Pleistocene, 439 sculpin, 45 Tamarix, 93 Poa secunda, 503 seasonal Tamias dorsalis, 86 pollen stainability, 109 abundance, 433 Tantilla hobartsmithi, 141 pollinator, 465 activity, 231, 353 taxonomy, 145 polytypy, 175 seed temperature, 193 ponderosa pine, 208 bank, 482 water, 18 Populus fremontii, 27 germination, 193 Texas, 53, 532 postburn seeding, 293 production, 208 threadfin shad, 306 prairie, 518 selection, 338 Townsendia aprica, 465 mixed-grass, 72 selective foraging, 166 trampling, 125 prairie dog, 445 self-pollination, 265 translocated and native quail, black-tailed, 445 sex determination, 492 331 town, 184 shrub steppe, 385 translocation, 445 predation, 403, 551 silver-haired bat, 353 Trichoptera, 454 prescribed burn, 249 Siphonaptera, 346 Trirhabda lewisii, 249 prescribed fire, 72 skunk, 184 Prophysaon coeruleum, 538 slickspot peppergrass, 265 urban stream, 497 Pseudacris regilla, 403 small mammals, 202 Ursus americanus, 269 pygmy rabbit, 1 Smith’s black-headed snake, Utah juniper woodland, 376 141 Utah, 131, 346, 421, 439 quail, translocated and native, Snake River, [Idaho, Little Dell Dam, 439 331 Wyoming], 78 soil compaction, 125 vascular plants, 421 radio telemetry, 338 solar insolation, 18 vegetation structure, 376 Rana cascadae, 403 Sonoran Desert, [Arizona], 137 vehicle, 125 range expansion, 135, 548 southern Nevada, 125 voles, 346 rare plant, 465 Southwest, 59 Vulpes macrotis, 544 560 WESTERN NORTH AMERICAN NATURALIST [Volume 63

Warner Mountains, [California], winter, 551 yellow starthistle, 202, 338 359 Wyoming Yuba County, [California], 202 Washington, 1, 538 Snake River, 78 water temperature, 18 Zapodinae, 439 waterbirds, 277 xeroriparian, 59 Zapus, 439 White River, [Nevada], 38, 45

TABLE OF CONTENTS Volume 64

No. 1—February 2004

Articles Biogeographic and conservation implications of late Quaternary pygmy rabbits (Brachylagus idahoensis) in eastern Washington ...... R. Lee Lyman 1 Fate and characteristics of Picea damaged by Elatobium abietinum (Walker) (Homoptera: Aphididae) in the White Mountains of Arizona ...... Ann M. Lynch 7 Density and biomass of redband trout relative to stream shading and temperature in southwestern Idaho ...... Bruce W. Zoellick 18 Response of tree ring holocellulose δ13C to moisture availability in Populus fremontii at perennial and intermittent stream reaches ...... Daniel L. Potts and David G. Williams 27 Conservation, status, and life history of the endangered White River spinedace, Lepidomeda albi- vallis (Cyprinidae) ...... G. Gary Scoppettone, James E. Harvey, and James Heinrich 38 A fish survey of the White River, Nevada ...... G. Gary Scoppettone, Peter H. Rissler, and Sean Shea 45 Coyote (Canis latrans) movements relative to cattle (Bos taurus) carcass areas ...... Jan F. Kamler, Warren B. Ballard, Rickey L. Gilliland, and Kevin Mote 53 Occurrence and habitat use of passage neotropical migrants in the Sonoran Desert ...... Paul C. Hardy, David J. Griffin, Amy J. Kuenzi, and Michael L. Morrison 59 Species diversity and habitat of grassland passerines during grazing of a prescribe-burned, mixed- grass prairie ...... Robert F. Danley, Robert K. Murphy, and Elizabeth M. Madden 72 Spawning ecology of finespotted Snake River cutthroat trout in spring streams of the Salt River valley, Wyoming ...... Michael P. Joyce and Wayne A. Hubert 78 Habitat of three rare species of small mammals in juniper woodlands of southwestern Wyoming ...... Kevin M. Rompola and Stanley H. Anderson 86 Beavers indirectly enhance the growth of Russian olive and tamarisk along eastern Montana rivers ...... Peter Lesica and Scott Miles 93 Contribution to the morphology and descriptive biology of Caurinella idahoensis (Ephemeroptera: Ephemerellidae) ...... Luke M. Jacobus and W.P. McCafferty 101 A putative hybrid swarm within Oönopsis foliosa (Asteraceae: Astereae) ...... Jonathan F. Hughes and Gregory K. Brown 109 Soil compaction from human trampling, biking, and off-road motor vehicle activity in a blackbrush (Coleogyne ramosissima) shrubland ...... Simon A. Lei 125 2003] INDEX 561

Notes Groundwater invertebrate fauna of the Bear River Range, Utah ...... Candace Brindza Huebner and Mark R. Vinson 131 Collection of an adult gizzard shad (Dorosoma cepedianum) from the San Juan River, Utah ...... Gordon A. Mueller and Jim L. Brooks 135 Growth in Sonoran Desert populations of the common chuckwalla (Sauromalus obesus) ...... Brian K. Sullivan, Matthew A. Kwiatkowski, and Paul S. Hamilton 137 Reproductive cycle of Smith’s black-headed snake, Tantilla hobartsmithi (Serpentes: Colubridae), in Arizona ...... Stephen R. Goldberg 141 Book Review Weird Nature by John Downer ...... C. Riley Nelson 144

No. 2—April 2004

Articles Biogeographic and taxonomic relationships among the mountain snails (Gastropoda: Oreohelicidae) of the central Great Basin ...... Mark A. Ports 145 Effect of cryptobiotic crust type on microarthropod assemblages in piñon-juniper woodland in central New Mexico ...... S.L. Brantley and U.L. Shepherd 155 Selection of ants by the American black bear (Ursus americanus) ...... Janene Auger, Gary L. Ogborn, Clyde L. Pritchett, and Hal L. Black 166 Comparative life history for populations of the Sceloporus grammicus complex (Squamata: Phrynosomatidae) ...... Aurelio Ramírez-Bautista, Elsa Jiménez-Cruz, and Jonathon C. Marshall 175 Influence of black-tailed prairie dog towns (Cynomys ludovicianus) on carnivore distributions in the Oklahoma Panhandle ...... Michael J. Shaughnessy, Jr., and Richard L. Cifelli 184 Temperature and high salinity effects in germinating dimorphic seeds of Atriplex rosea ...... M. Ajmal Khan, Bilquees Gul, and Darrell J. Weber 193 Influence of yellow starthistle (Centaurea solstitialis) on small mammals in central California ...... Kirsten Christopherson and Michael L. Morrison 202 Cone and seed production in Pinus ponderosa: a review ...... Pamela G. Krannitz and Thomas E. Duralia 208 Plant species diversity in a grassland plant community: evidence for forbs as a critical management consideration ...... Monica L. Pokorny, Roger L. Sheley, Tony J. Svejcar, and Richard E. Engel 219 Reproductive activity and distribution of bats in Nebraska ...... Russell A. Benedict 231 Feeding behavior and performance of a rabbitbrush leaf-beetle (Trirhabda lewisii) feeding on Chrysothamnus nauseosus regrowth after fire ...... Ann L. Herzig and Cynthia Skema 249 Notes Discovery of the centipede Scolopocryptops gracilis Wood in Montana (Scolopendromorpha: Scolopocryptopidae) ...... Rowland M. Shelley and Diana L. Six 257 Habitat use by brush mice (Peromyscus boylii) in southeastern Arizona ...... Amy B. Gottesman, Michael L. Morrison, and Paul R. Krausman 259 Importance of outcrossing for fruit production in slickspot peppergrass, Lepidium papilliferum L. (Brassicaceae) ...... Ian Robertson 265 Are desert basins effective barriers to movements of relocated black bears (Ursus americanus)? ...... Jon P. Beckmann and Carl W. Lackey 269 562 WESTERN NORTH AMERICAN NATURALIST [Volume 63

Book Review The Animal Mind by James L. Gould and Carol Grant Gould ...... Serge Lariviere 273 National Geographic Photography Field Guide: Birds by Rulon E. Simmons ...... Clayton M. White 274 A Birders Guide to Alaska by George C. West ...... Clayton M. White 274 Monterey Birds: Status and Distribution of Birds in Monterey County, California by Don Roberson ...... Clayton M. White 274 Firefly Encyclopedia of Birds by Christopher Perrins ...... Clayton M. White 275

No. 3—August 2004

Articles Nesting ecology of waterbirds at Grays Lake, Idaho ...... Jane E. Austin and William H. Pyle 277 Vegetation response to fire and postburn seeding treatments in juniper woodlands of the Grand Staircase–Escalante National Monument, Utah ...... Paul Evangelista, Thomas J. Stohlgren, Debra Guenther, and Sean Stewart 293 Distribution and abundance of pelagic fish in Lake Powell, Utah, and Lake Mead, Arizona-Nevada ...... Gordon A. Mueller and Michael J. Horn 306 Using sampling and inverse distance weighted modeling for mapping invasive plants ...... Elizabeth A. Roberts, Roger L. Sheley, and Rick L. Lawrence 312 Observations of sagebrush gall morphology and emergence of Rhadopalomyia pomum (Diptera: Cecidomyiidae) and its parasitoids ...... Ruth A. Hufbauer 324 Survival rates of translocated and native Mountain Quail in Oregon ...... Michael D. Pope and John A. Crawford 331 Seasonal habitat use and selection of Chukars in west central Idaho ...... Andrew J. Lindbloom, Kerry P. Reese, and Peter Zager 338 Fleas (Siphonaptera) in nests of voles (Microtus spp.) in montane habitats of three regions of Utah ...... Glenn E. Haas and James R. Kucera 346 Reproduction and seasonal activity of silver-haired bats (Lasionycteris noctivagans) in western Nebraska . . . . Keith Geluso, Jeffrey J. Huebschman, Jeremy A. White, and Michael A. Bogan 353 Diets of nesting Northern Goshawks in the Warner Mountains, California ...... Rebecca L. Promessi, John O. Matson, and Mary Flores 359 Host-based morphometric differentiation in three Cinara species (Insecta: Hemiptera: Aphididae) feeding on Pinus edulis and P. monophylla ...... Colin Favret and David J. Voegtlin 364 Comparative habitat use in a juniper woodland bird community ...... David C. Pavlacky, Jr., ...... and Stanley H. Anderson 376 Passerine response to roads associated with natural gas extraction in a sagebrush steppe habitat ...... Franz Ingelfinger and Stanley Anderson 385 Notes Historical biogeography of the woodchuck (Marmota monax bunkeri) in Nebraska and northern Kansas ...... Zachary P. Roehrs and Hugh H. Genoways 396 Cannibalism and predation by western toad (Bufo boreas) larvae in Oregon, USA ...... D. Jess Jordan, Christopher J. Rombough, Christopher A. Pearl, and Brome McCreary 403 Foxes on a hot tin roof ...... James A. Sedgwick and John Bartholow 406 Age and growth of least chub, Iotichthys phlegethontis, in wild populations ...... Michael D. Mills, Mark C. Belk, Russell B. Rader, and Joshua E. Brown 409 2003] INDEX 563

Diet of the leatherside chub, Snyderichthys copei, in the fall ...... Adrian Bell and Mark C. Belk 413 Book Review Viste Nieve by Melbourne R. Carriker ...... Clayton M. White 417

No. 4—October 2004

Articles Evaluating the geographic distribution of plants in Utah from the Atlas of Vascular Plants of Utah ...... R. Douglas Ramsey and Leila Shultz 421 The male of Culicoides reevesi Worth, with a redescription of the female and new seasonal activity, distribution, and biting records (Diptera: Ceratopogonidae) ...... William L. Grogan, Jr., Gustavo R. Spinelli, Robert A. Phillips, and David L. Woodward 433 First Pleistocene jumping mouse (Zapus, Zapodinae, Rodentia) from Utah ...... Dennis R. Ruez, Jr., and Christopher J. Bell 439 A relocation technique for black-tailed prairie dogs ...... Kelly A. Roe and Christopher M. Roe 445 A review of the adult Anagapetus (Trichoptera: Glossosomatidae) ...... Dave Ruiter 454 Reproduction and demography of Townsendia aprica (Asteraceae), a rare endemic of the Southern Utah Plateau ...... Vincent J. Tepedino, Sedonia D. Sipes, and Terry L. Griswold 465 Marmot distribution and habitat association in the Great Basin ...... Chris H. Floyd 471 Dynamics of a dwarf bear-poppy (Arctomecon humilis) population over a sixteen-year period ...... K.T. Harper and Renée Van Buren 482 Evaluating the use of morphometric measurements from museum specimens for sex determination in Mountain Plovers (Charadrius montanus) ...... William M. Iko, Stephen J. Dinsmore, and Fritz L. Knopf 492 Life cycle of the net-spinning caddisfly, Cheumatopsyche analis (Banks) (Trichoptera: Hydropsy- chidae), in two small Front Range urban streams, Fort Collins, Colorado ...... Robert E. Zuellig, Boris C. Kondratieff, and Richard A. Thorp 497 Native plant community composition and genetic diversity associated with long-term weed invasions ...... Brian A. Mealor, Ann L. Hild, and Nancy L. Shaw 503 Fleas (Siphonaptera: Ceratophyllidae, Ctenophthalmidae) from rodents in five southwestern states ...... Glenn E. Haas, Nixon Wilson, and Chris T. McAllister 514 Leafhopper assemblages on native and reseeded grasslands in southwestern Montana ...... James A. Bess, Kevin M. O’Neill, and William P. Kemp 518 The centipede genus Arthrorhabdus Pocock, 1891, in the Western Hemisphere: potential occurrence of A. pygmaeus (Pocock, 1895) in Belize (Scolopendromorpha: Scolopendridae: Scolopendrinae) ...... Rowland M. Shelley and Amazonas Chagus Junior 532 Notes Prophysaon coeruleum Cockerell, 1880, blue-gray taildropper (Gastropoda: Arionidae): new distributional records and reproductive anatomy ...... Kristiina Ovaska, William P. Leonard, Lyle Chichester, Thomas E. Burke, Lennart Sopuck, and Jim Baugh 538 Latrine use by San Joaquin kit foxes (Vulpes macrotis mutica) and coyotes (Canis latrans) ...... Katherine Ralls and Deborah A. Smith 544 First nesting record of Black-throated Gray Warbler (Dendroica nigrescens) for Montana ...... Paul Hendricks 548 Mink predation on radio-tagged trout during winter in a low-gradient reach of a mountain stream, Wyoming ...... Jason W. Lindstrom and Wayne A. Hubert 551 Index to Volume 64 ...... 555 564 WESTERN NORTH AMERICAN NATURALIST [Volume 63

CONTENTS

(Continued from back cover)

Notes Prophysaon coeruleum Cockerell, 1880, blue-gray taildropper (Gastropoda: Arionidae): new distributional records and reproductive anatomy ...... Kristiina Ovaska, William P. Leonard, Lyle Chichester, Thomas E. Burke, Lennart Sopuck, and Jim Baugh 538 Latrine use by San Joaquin kit foxes (Vulpes macrotis mutica) and coyotes (Canis latrans) ...... Katherine Ralls and Deborah A. Smith 544 First nesting record of Black-throated Gray Warbler (Dendroica nigrescens) for Montana ...... Paul Hendricks 548 Mink predation on radio-tagged trout during winter in a low-gradient reach of a mountain stream, Wyoming...... Jason W. Lindstrom and Wayne A. Hubert 551

Index to Volume 64 ...... 555