HARVARD UNIVERSITY Library of the Museum of Comparative Zoology IE GREAT BASIN NATURALIST ime 49 No. 1 31 January 1989 Brigham Young University MCZ ********* ^\ LIBRARY | JUH 2 9 1 HARVAR UN1VERS1 GREAT BASIN NATURALIST Editor. Stephen L. Wood, Department of Zoology, 290 Life Science Museum, Brigham Young University, Provo, Utah 84602. Associate Editor. James R. Barnes, Department of Zoology, 175 Widtsoe Building, Brigham Young University, Provo, Utah 84602. Dr. Barnes will become the editor on 1 September 1989. Editorial Board. Richard W. Baumann, Chairman, Zoology; Clayton M. White, Zoology; Jerran T. Flinders, Botany and Range Science. All are at Brigham Young University. Ex Officio Editorial Board Members include Clayton S. Huber, Dean, College of Biological and Agricultural Sciences; Norman A. Darais, University Editor, University Publica- tions; Stephen L. Wood, Editor, Great Basin Naturalist. The Great Basin Naturalist was founded in 1939. The journal is a publication of Brigham Young University. Previously unpublished manuscripts in English pertaining to the biological natural history of western North America are accepted. The Great Basin Naturalist Memoirs series was established in 1976 for scholarly works in biological natural history longer than can be accommodated in the parent publication. The Memoirs appears irregularly and bears no geographical restriction in subject matter. Manuscripts for both the Great Basin Naturalist and the Memoirs series will be accepted for publication only after exposure to peer review and approval of the editor. Subscriptions. Annual subscriptions to the Great Basin Naturalist are $25 for private individuals and $40 for institutions (outside the United States, $30 and $45, respectively), and $15 for student subscriptions. The price of single issues is $12. All back issues are in print and are available for sale. All matters pertaining to subscriptions, back issues, or other business should be directed to the Editor, Great Basin Naturalist, Brigham Young University, 290 Life Science Museum, Provo, Utah 84602. The Great Basin Naturalist Memoirs may be purchased from the same office at the rate indicated on the inside of the back cover of either journal. Scholarly Exchanges. Libraries or other organizations interested in obtaining either journal through a continuing exchange of scholarly publications should contact the Brigham Young University Exchange Librarian, Harold B. Lee Library, Provo, Utah 84602. Manuscripts. See Notice to Contributors on the inside back cover. 4-89 650 39866 ISSN 017-3614 TABLE OF CONTENTS Volume 49 Number 1—31 January 1989 On the distribution of Utah's hanging gardens. Stanley L. Welsh 1 Changes in mule deer size in Utah. Dennis D. Austin, Robert A. Riggs, Philip J. Unless, David L. Turner, and John F. Kimball 31 White-tailed prairie dog (Cynomys leucurus Merriam) diggings in western harvester ant, Pogonomi/rmex ocddentalis (Cresson), mounds. William H. Clark and Cynthia [. Clark . . 36 Amphibians of western Chihuahua. Wilmer W. Tanner 38 Observations on recruitment and ecology of razorbaek sucker: lower Colorado River, Arizona-California-Nevada. Paul C. Marsh and W. L. Minckley 71 Competition between adult and seedling shrubs of Ambrosia dumosQ in the Mojave Desert, Nevada. Richard Hunter 79 Aquatic insects in Montezuma Well, Arizona, USA: a travertine spring mound with high alkalinity and dissolved carbon dioxide. Dean W. Blinn and Milton W. Sanderson 85 A new Haliplus from Warm Springs, Nevada (Coleoptera: Haliplidae). Samuel A. Wells 89 On the genus Paracarinolidia (Cicadellidae: Coelidiinae: Teruliini). M. W. Nielsen 92 Two genera and two new species of Teruliine leafhoppers (Homoptera: Cicadellidae: Coelidiinae). M. W. Nielson 96 A new species oiAsclepias (Asclepiadaceae) from northwestern New Mexico. Kenneth D. Heil, J. Mark Porter, and Stanley L. Welsh 100 Effect of timing of grazing on soil-surface cryptogamic communities in a Great Basin low-shrub desert: a preliminary report, fames R. Marble and Kimball T. Harper '. ' 104 Size and overlap of Townsend ground squirrel home ranges. Nicholas C. Nydegger and Donald R. Johnson 108 Pisulithus tinctorius, a Gasteromycete, associated with Jeffrey and Sierra lodgepole pines on acid mine spoils in the Sierra Nevada. R. F. Walker Ill Spatial and temporal variability in perennial and annual vegetation at Chaco Canyon, New Mexico. Anne C. Cully and Jack F. Cully, Jr 113 Associations of small mammals occurring in a pluvial lake basin. Ruby Lake, Nevada. Mark A. Ports and Lois K. Ports 123 Utah chub (Gila atraria) from the latest Pleistocene Gilbert shoreline, west of Conine, Utah. Stuart Murehison 131 Mediation of nutrient cycling by arthropods in unmanaged and intensively managed brush habitats. T. A. Christiansen, J. A. Lockwood, and J. Powell 134 Locality, habitat, and elevation records for the desert shrew, Notiusorex crawfordi. Russell Davis and Ronnie Sidner 140 Number 2—30 April 1989 Snake Creek Burial Cave and a review of the Quaternary mustelids of the Great Basin. Emilee M. Mead and Jim I. Mead 143 Diatom flora of Mink Creek, Idaho, USA. Christopher T. Robinson and Samuel R. Rushforth 155 Responses of Utah deer hunters to a checking station questionnaire. Dennis D. Austin and Lucy Jordan 159 Nomenclatural changes and new species of Scolytidae (Coleoptera), Part IV. Stephen L. Wood 167 Species composition, emergence, and habitat preferences of Triehoptera of the Sagehen Creek basin, California, USA. Nancy A. Erman 186 Influence of substrate water content on neonate size in the prairie skink, Eumeces septentrionalis. Louis A. Somma 198 Review of selenium in soils, plants, and animals in Nevada. Stephen C. Poole, Verle R. Rohman, and James A. Young 201 Classification of the riparian vegetation of the montane and subalpine zones in western Colorado. William L. Baker 214 Winter habitats and foods of Blue Grouse in the Sheeprock Mountains, Utah. Peter J. Pekins, Frederick G. Lindzey, Jay A. Roberson, Gregory McDaniel, and Randy Berger 229 New species of Harasupia with a revised key to the species (Homoptera: Cicadellidae: Coelidiinae). M. W. Nielson 233 Quadrat and sample sizes for frequency sampling mountain meadow vegetation. Jeffrey C. Mosley, Stephen C. Bunting, and M. Hironaka 241 Foods and feeding periodicity of the White River springfish, Crenichthys baileyi. Gene R. Wilde 249 Habitat use and selection and home ranges of Merriam's Wild Turkey in Oregon. R. Scott Lutz and John A. Crawford 252 Gerridae (water striders) of Idaho (Heteroptera). R. C. Biggam and M. A. Brusven .... 259 Comparison of Sage and Sharp-tailed Grouse leks in south central Wyoming. James H. Klott and Frederick G. Lindzey 275 Notes on Hells Canyon birds. Daniel M. Taylor 279 Number 3—31 July 1989 Systematica and distribution of the winter stonefly genus Capnia (Plecoptera: Capni- idae) in North America. C. Riley Nelson and R. W. Baumann 289 Mass mortality of salamanders (Ambystoma tigrinum) by bacteria (Acinetobacter) in an oligotrophy seepage mountain lake. Kathleen Muriel Worthylake and Peter Hovingh 364 Comparison of regression methods for biomass estimation of sagebrush and bunch- grass. Robin J. Tausch 373 List of Montana Scolytidae (Coleoptera) and notes on new records. Sandra J. Gast, Malcolm M. Furniss, James B. Johnson, and Michael A. Ivie 381 First North American record of Cichlasoma managuense (Pisces: Cichlidae). Paul C. Marsh, Thomas A. Burke, Bruce D. DeMarais, and Michael E. Douglas 387 Co-occupancy of a den by a pair of Great Basin black bears. John M. Goodrich and San J. Stiver 390 High-elevation records for Neotoma cinerea in the White Mountains, California. Donald K. Grayson and Stephanie D. Livingston 392 Opportunistic foraging by the kangaroo rat (Dipodomys deserti) Stephens (Rodentia: Heteromyidae). Richard W. Rust 396 Additional new species of Teruliine leafhoppers with key to species (Cicadellidae: Coelidiinae: Teruliini). M. W. Nielson 398 Habitat use by breeding male sage grouse: a management approach. Kevin L. Ellis, Jimmie R. Parrish, Joseph R. Murphy, and Gary H. Richins 404 Physiographic characteristics of peregrine falcon nesting habitat along the Colorado River system in Utah. Brandon L. Grebence and Clayton M. White 408 Soil-site relationships of white locoweed on the Raft River Mountains. Michael H. Ralphs, Brock Benson, and J. Cameron Loerch 419 Role of post-Pleistocene dispersal in determining the modern distribution of Abert's squirrel. Russell Davis and David E. Brown 425 Influence of experimental habitat manipulation on a desert rodent population in southern Utah. Jiping Zou, Jerran T. Flinders, Hal L. Black, Steven G. Whisenant 435 Evaluation of wildlife response to a retained mine highwall in south central Wyoming. John P. Ward and Stanley H. Anderson 449 Effects of arthropods on root:shoot ratio and biomass production in undisturbed and modified mountain shrub habitats. Tim A. Christiansen, Jeffrey A. Lockwood. and JeffPowell 456 A new perennial species of Gilia (Polemoniaceae) from Utah. Frank J. Smith and Elizabeth C. Neese 461 Marmot scats supplement hay pile vegetation as food energy for pikas. James A. Gessa- man and Andrew G. Goliszek 466 Number 4— 31 October 1989 Soil characteristics of mountainous northeastern Nevada sagebrush community types. Mark E. Jensen 469 Hyobranchial apparatus of the Cryptobranchoidea (Amphibia). Douglas C. Cox and Wilmer W. Tanner 482 Impact of cattle on two isolated fish populations in Pahranagat Valley, Nevada. Frances R. Taylor, Leah P. Gillman, and John W. Pedretti 491 An extant, indigenous tortoise population in Baja California Sur, Mexico, with the description of a new Xerobates (Chelonii: Testudinidae). John R. Ottley and Victor M. Velazques Solis 496 Status of S pea stagnalis Cope (1975), Spea intermontanus Cope (1889), and systematic- review of Spea hammondii Band (1839) (Amphibia: Anura). Wilmer W. Tanner . . 503 Variations in Thamnophis elegans with descriptions of new subspecies. Wilmer W. Tanner and Charles H. Lowe 511 On the structure and function of white-tailed prairie dog burrows. James A. Burns, Dennis L. Flath, and Tim W. Clark 517 Bibliography and subject index of the prairie skink, Eumeces septentrionalis (Baird) (Sauria: Scincidae). Louis A. Somma and Philip A. Cochran 525 Orientation of zooplankton to the oxycline in Big Soda Lake, Nevada. Michael A. Bozek 535 Coexistence of two species of sucker, Catostomus, in Sagehen Creek, California, and notes on their status in the western Lahontan Basin. Lynn M. Decker 540 Diet similarity between elk and deer in Utah. Kerry J. Mower and H. Duane Smith . . . 552 Effect of indomethacin-treated wheat on a wild population of voles. Rodney R. Seeley and Timothy D. Reynolds 556 Litter decomposition by arthropods in undisturbed and intensively managed moun- tain brush habitats. Tim A. Christiansen, Jeffrey A. Lockwood, and JeffPowell. . . 562 Arthropod community dynamics in undisturbed and intensively managed mountain brush habitats. Tim A. Christiansen, Jeffrey A. Lockwood, and JeffPowell 570 Demographic characteristics of American marten populations in Jackson Hole, Wyo- ming. Tim W. Clark, Thomas M. Campbell III, and Tedd N. Hauptmann 587 Terrestrial vertebrates of Scotts Bluff National Monument, Nebraska. Mike K. Cox and William L. Franklin 597 Response of nesting waterfowl to flooding in Great Salt Lake wetlands. A. Lee Foote . . 614 Note on fungi in small mammals from the Nothofagus forest in Argentina. Javier G. Perez Calvo, Zane Maser, and Chris Maser 618 Note on mound architecture of the black-tailed prairie dog. Richard P. Cincotta 621 Lesser prairie-chicken nest site selection and vegetation characteristics in tebuthiuron- treated and untreated sand shinnery oak in Texas. D. A. Haukos and L. M. Smith 624 Mourning dove use of man-made ponds in a cold-desert ecosystem in Idaho. Frank P. Howe and Lester D. Flake 627 Radio transmitter attachment for Chukars. Bartel T. Slaugh, Jerran T. Flinders, Jay A. Roberson, M. Ray Olson, and N. Paul Johnston 632 Checklist of Recent Mollusca of Wyoming, USA. Dorothy E. Beetle 637 Horizontal and vertical diameter of burrows of five small mammal species in south- eastern Idaho. John W. Laundre 646 Ecology of Comandra umbellata in western Wyoming (Santalaceae). W. R. Zentz and W. R. Jacobi ' 650 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young University ISSN 0017-3614 Volume 49 31 January 1989 No. 1 ON THE DISTRIBUTION OF UTAH'S HANGING GARDENS Stanley L. Welsh1 Abstract. — This is a summary monograph of the hanging gardens as they occur in the Colorado River and Virgin River portions of the Colorado Plateau in Utah. Discussed in this paper are the hanging gardens, their geography, geomorphology, aspects of distribution and diversity, and principal vascular and algal plant species. Animal trapping studies and plant productivity aspects are reviewed. The sea of aridity that overlies southern Utah and vicinity is broken by seasonal influ- ences and by the dendritic trenches of the Colorado River and its tributaries. The effects of the river are restricted to its banks and adjacent alluvial terraces; the riparian vegeta- tion is generally both monotonous and pre- dictable. Away from the riverbanks aridity is the general rule. However, here and there on the canyon walls are moist places clothed in green. They are well-watered islands in an ocean of drought (Figs. 1 and 2). It is with these patches of greenery that this paper is involved. They must be placed within their setting in order to understand the contrast of their mesophytic vegetation with the xeric communities that surround them. The Colorado River system is entrenched into a great platform supported by a geological substructure more than a billion years old. Impressive as the inner gorges of this canyon are, the broader aspect of the system is evi- dent to the east or south of the Wasatch Plateau in central Utah. The canyon of the Colorado at that point is more than a hundred miles wide, having yielded to the processes of erosion hundreds of cubic miles of alluvium. Despite its huge size, the canyon is of rela- tively recent origin, geologically speaking (Hintze 1972). The geological strata are remarkably evi- dent in this arid setting, where vegetative cover is thin and where rate of soil develop- ment is exceeded by processes of erosion. No great bodies of contemporary alluvium serve to obscure the underlying geology as in the Great Basin to the west. The Colorado River and its tributaries have excavated the allu- vium almost as it has formed. The canyon is open to the south, and the products of erosion have been transported in the great river. Ped- iments of ancient erosional deposits persist for a while perched atop highlands between arms of modern drainages, but raw geological strata are exposed over huge areas of the basins of the Colorado. Reason for the sparse protective layer of plants and for the limited soil development are related to the general aridity of the region. The dryness is a function of both low precipi- tation and high evaporation. Eubank (1979) records the following long-time precipitation means (in inches, followed by centimeters in parentheses) for the following stations: Hanksville 5.19 (13.18), Green River 6.06 (15.39), St. George 8.78 (22.3), Moab 8.82 Life Science Museum and Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602. Great Basin Naturalist Vol. 49, No. 1 Fig. 1. Map of Utah showing distribution of hanging gardens discussed in this paper. (22.4), Price 9.88 (25.1), Blanding 13.21 (33.55), and Zion National Park 14.61 (37.11). Mean temperatures Fahrenheit (centigrade in parentheses) for those stations are: Zion National Park 61.2 (16.2), St. George 60.1 (15.6), Moab 55.0 (12.8), Hanksville 52.3 (11.3), Green River 52.5 (11.2), Blanding 49.4 (9.7), and Price 48.8 (9.3). Extreme tempera- tures are probably more important than means to the survival of plants. Summer tem- peratures greater than 100 degrees Fahren- heit (38 degrees centigrade) are common at all of the selected stations, and winter tempera- tures of below zero on that scale have been recorded at all stations. My first experience with this grandly arid country occurred almost four decades ago when I visited Glen Canyon and the townsite of Hite. I was traveling as a student in a class led by Professor Bertrand F. Harrison. No measurable rain had fallen for more than a year at the pioneer community, along the Col- orado River at the mouth of Trachyte Wash, that 17th of May 1950 when I visited there. Fig. 2. Ribbon Garden, Ribbon Canyon, Lake Powell, San Juan Co., Utah. Sparingly vegetated Navajo Sandstone is cut by Glen Canyon, Ribbon Canyon, and Cottonwood Canyon. January 1989 Welsh: Utah's Hanging Gardens Table 1. Geological strata and distributions of hanging gardens. Geologic age Strata Region Cretaceous Jurassic Triassic Permian Wahweap Straight Cliffs Morrison (Bluff Sst) Entrada (various members) Navajo Sandstone Navajo Sandstone Kaventa Moenave (Springdale Sst) Chinle (Shinarump Egl) White Rim Cedar Mesa Pennsylvanian Hermosa Kaiparowits Kaiparowits Bluff Arches Kaiparowits Lake Powell Canyonlands Lake Powell Zion Canyon Canyonlands Zion Canyon Lake Powell St. George Zion Canyon Zion Canvon Zion Canyon Canyonlands Cataract Canyon Natural Bridges Cataract Canvon Despite the aridity, there had been sufficient unmeasured water to allow for germination of some seeds, and a few diminutive plants of red brome had each matured a solitary seed, re- placing those from which they had germi- nated. And, plants of datura displayed their huge, sweetly scented white flowers, which contrasted with the red, barren background. Princes plume grew against a backdrop of purple siltstone, the difference in hue both pleasing and startling. Later on that same trip to the canyon coun- try of the Colorado, we reached Natural Bridges National Monument, where we camped. The following day we explored Arm- strong and White canyons, looked at the amazing bridges, and observed the small, ver- tical wet seeps, occupied by mesophytic plants. This was my first introduction to the peculiar vegetative assemblages known as hanging gardens. At the time their peculiarity was lost in the immense amount of informa- tion thrust upon a student in this remarkable land for the first time. The hanging gardens result from coinci- dence of water in perched bedding planes within sandstone strata intersected by the dendritic drainages of the Colorado River system (Fig. 1). The kind of garden develop- ment, whether alcove, terrace, or window- blind (Welsh and Toft 1981), is determined by the nature of the geological formation and the presence or absence of joint systems. Com- plexity of the plant community within a hang- ing garden is a function of quantity and quality of water, developmental aspects, and accessi- bility of plant species to it. Hanging gardens occur in sandstone forma- tions and sandstone members of several for- mations ranging in age from Pennsylvanian to Cretaceous. Massive sandstones seem to be best suited for alcove development coinciden- tal with garden formation, some better than others. The formations with greatest develop- ment are the Navajo and Entrada, both of them cross-bedded, massive formations com- posed of wind-blown sand and containing an- cient pond bottoms that serve as impervious bedding planes. The Wingate Formation is of similar composition but lacks significant hang- ing gardens. More thinly bedded sandstone formations tend not to form alcove gardens similar to those of the Navajo or Entrada. Exceptions occur, however. Main formations bearing hanging gardens are listed in Table 1. Less than massive, though not especially bedded strata such as the Kayenta at St. George, Springdale Sandstone at Zion, and the Bluff Sandstone at Bluff, are alcove form- ers. The base of the alcove is not in the sand- stone formation, however. Instead, the base is on the impervious formation beneath the sandstone. The Colorado River is entrenched into geo- logical strata that are displayed over vast re- gions in flat or only somewhat inclined posi- tions. The strata are those exposed in that more or less stable geological highlands east of the hingeline in Utah (Hintze 1972). The canyons dissected into those highlands dis- play vast sandstone surfaces along their walls. The sands of formations suitable for hanging garden development were deposited mainly on land, as dunes with interdune valleys. The interdune valleys were often the sites of lakes, whose bottoms were made impervious by ac- cumulations of dust and other fine particles. Thin layers of limestone are evident in many of the bedding planes. Turned to stone, the ancient lake and pond basins continue to exist within the strata. Water percolating through the porous rock encounters the ancient bed- ding planes, still impervious and capable of Great Basin Naturalist Vol. 49, No. 1 holding water. When filled to overflowing, these bedding planes carry the water down- ward to the next bedding plane beneath or to another impervious stratum at the base of the formation. Joint systems within the rock act as passage- ways for water, which follows the vertical gra- dient of the crack downward until it encoun- ters some obstacle to that flow. Where the joint systems are exposed along canyon walls, the water flows over the moist surfaces. Here gardens of the windowblind type are formed. Alcove gardens develop in massive sand- stones with minimal jointing. Terrace gardens result when water at the base of a stratum encounters an obstacle to its movement, flows laterally to the margin of the formation along a canyon wall, and tumbles over the terracelike margin of that impervious layer. Flow of water from the margin of the bed- ding plane varies in amount from that hardly sufficient to moisten the surface of the rock to substantial quantities that collect into streams below the gardens. Many of the canyons carved in sandstone in southern Utah feature crystal-clear steams that flow perennially. Whatever the type of garden development (alcove, terrace, or windowblind), no wet sur- face exists for long prior to invasion of plant propagules (Malanson 1980). Spores and dust- sized seeds are carried to the moist sites by air currents that sweep the canyons. Sticking to the moist surface, the propagules germinate to form prothallia seedlings. Algae, ferns, and seed plants are involved in community devel- opment on the wet sites. Hanging garden formation as a geomorpho- logical process has been discussed by Welsh and Toft (1981). The gardens are positionally unique. They tend to occur at all exposures of the canyon walls, but whatever the direction of exposure, they are shaded for much to most of each day. Indeed, some of the gardens never receive direct sunlight. A hygrother- mograph placed within the Step Garden al- cove along Glen Canyon recorded smooth rounded curves of temperature and humidity, as if the instrument had been placed within a house. Temperatures are moderated by the shade of canyon walls and frequently by an enclosing margin of trees and other vegeta- tion. Air movement is restricted in well- developed gardens also, but in some they are exposed directly to both intense light and winds. Thus, there is a diversity of gardens. They vary in size, aspect, exposure to the elements, water quantity and quality, num- ber of bedding planes, and amount of light received. Water quality, in some degree, controls the kinds of plants in hanging gardens. Quality of water is dictated by the nature of the forma- tions through which the water passes. Most gardens are the products of water of drinkable quality. However, water in some formations is saline and leaves a crust of various salts upon drying. In others the water is laden with cal- cium, which results in tufa deposits in the gardens. Generally, however, water from the gardens is potable. Hanging garden vegetation is frequently closely juxtaposed to that of riparian plant communities immediately down the drain- age. Several common components of the riparian communities occur in the gardens, but there are a series of species that are unique to this peculiar vegetative type. The unique species are more than mere extensions of the riparian vegetation. Many taxa of the gardens are widely distributed plants of di- verse habitats elsewhere; others are known only from this habitat. Some of the latter, the endemics, and the distributionally unique species have taxonomic relationships with species of wide distribution in North America (Welsh and Toft 1981). Nevertheless, the plants that occupy hanging gardens are oppor- tunists. The habitats are available to plants from specific areas, and the plants of adjacent or contemporaneously disjunct floristic re- gions are those that now occur in the gardens. Hanging garden algal floras in Utah have been studied by Clark (1972), Rushforth et al. (1976), and Johansen et al. (1983). Rushforth and Merkley (1988) have published a compre- hensive list by habitat of the algae of Utah. Wet walls (i.e., hanging gardens) are included by Rushforth and Merkley (1988) as one of nine habitat categories. Lists of algal species of hanging gardens cited below in Tables 2 and 3 are abstracted from the paper by Rushforth and Merkley (1988). The small number of species known cur- rently only from wet walls or hanging gardens might be simply an artifact of collection. They could yet be found elsewhere in subsequent collections. January 1989 Welsh: Utah's Hanging Gardens Table 2. Algal components of Utah's hanging gardens. Taxonomic group No. HG 1 other habitat 2 + other habitats Cyanophyta Cyanophyceae Chroeoccales Oseillatoriales Chlorophyta Chlorophyceae Tetrasporales Ulotrichales Trentepohliales Oedogoniales Cladophorales Chlorocoecales Zygnematales Chrysophyta 15 3 1 26 4 8 2 0 1 2 0 1 1 1 0 1 0 0 1 0 0 3 0 0 9 3 2 Cyanophyta Gloeocapsa nigrescens Naeg. Gloeothece palea (Kutz) Rabh. Gloeothece rupestris (Lyngb.) Born. Nostoc microscopicum C. A. Ag. Oscillatoria subbrevis Sehmidle f. minor Desik. Scytonema datum (Carm.) Borzi Stigonema mamillosum (Lyngb.) C. A. Ag. Chlorophyta Trentepohlia aurea (L.) Martius Cosmarium mineghinii Breb. var. continuum Rabh. Cosmarium undulatum Corda var. crenulatum Zygnema sterile Trans. Chrysophyta Caloneis alpestris (Grun.) CI. Cymbella incerta (Grun.) CI. var. naviculacea (Grun.) CI. Hantzschia amphioxys (Ehr.) Grun. var. linearis (O. Mull.) Cl.-Eul. Mastogloea grevillei W. Sm. Pinularia biceps Greg. var. minor (Peters.) Cl.-Eul. 11 14 Bacillariophyceae 144 Totals 204 5 16 20 34 119 154 Table 3. Hanging garden algae and distributions. Zion Glen Canyon Arches Hanging garden algae represent approxi- mately 10.7% of the 1,900 species known from Utah. The greater proportion of the species (75%) are evidently generalists, being re- ported from two or more habitats besides hanging gardens. The 20 species reported from only a single habitat other than the wet walls are from lakes and reservoirs (15), not from rivers and streams, which would be more likely due to juxtaposition of the walls to those features. The others are from rivers and streams (2), thermal springs (2), and soils (1). These data might be indicative of the unique- ness of the wet wall habitat, or they might be an artifact of collection or merely an indication of commonality between species of wet walls Great Basin Naturalist Vol. 49, No. 1 and the great preponderance of species that occur in the lakes and reservoirs. The current list of algae known only from wet walls in Utah and their geographic locali- ties are given in Table 3. Whether the apparent differences in distri- bution represent reality or merely lack of col- lection is not known. Certainly the differences noted are similar to those demonstrated con- clusively for vascular plant species. The Chyrsophyta is the largest single group of algae in the hanging gardens. The 144 spe- cies comprise 70.5% of the known algal flora. The group is represented in the gardens in greater proportion than would be expected based on the ratio of diatoms to other algae in Utah, where only slightly more than 50% of the algal flora is composed of diatoms. Evi- dently the wet walls present excellent habitats for diatom species. Johansen et al. (1983) note: In most cases we have studied, the moist wall is inhabitated primarily by mucilage secreting green and bluegreen algae. As such species become estab- lished an abundance of secondary species colonize the mucilage. These include green and bluegreen algae and numerous diatom species as well as occa- sional Euglenophyta and chrysophytes. Despite the designation as opportunists, not all plants that grow within this community type should be regarded as hanging garden species. Only those whose distribution is ex- clusively (or almost so) within the garden (the endemics), whose range is almost entirely from them (the elevational or spatial dis- juncts), or of unusual distribution should be so categorized. Hanging gardens are typical of the canyons of the Colorado, which occupy two main por- tions of Utah. They are best formed along Glen Canyon and northward to the vicinity of Moab and Arches National Park along the Col- orado and Green rivers (the Canyonlands), and along the Virgin River in southwestern Utah. The gardens of the two areas are of fundamentally different structure geomor- phologically. Those from along the Virgin River are best developed on the jointed walls of Zion Canyon. Alcove types along the Virgin do occur, however, especially in the Spring- dale member of the Moenave in Zion Canyon and vicinity, and in the Kayenta on the Red Hill in St. George. In southeastern Utah the gardens are mostly of the alcove type within poorly jointed rocks or where the jointing is not controlling the water source. Terrace gar- dens are present along the steplike margins of bedded sandstones such as the Wahweap and Straight Cliffs formations in western Kane County. Alcoves sometimes form in the less than massive White Rim (Cutler), Bluff (Mor- rison), and members of the Entrada forma- tions. The impervious layer associated with some garden formation is the immediately un- derlying stratum. The Kayenta is often the layer immediately below the Navajo Forma- tion. It acts to halt the flow of water down- ward, and, where intersected by a canyon, the water tumbles down its margin in a cascading, terrace garden. Above the Kayenta, huge al- coves sometimes are formed near the base of the Navajo proper (Fig. 15). Several bedding planes are present in some alcoves (Fig. 11). Alcoves within the Navajo Sandstone are frequently more than 100 m in height. In some of them, water drips from superposed bedding planes separated by sev- eral meters of sandstone. Each of the bedding planes supports one or more of the species typical of hanging garden habitats. The hanging garden habitat does not form a sharp boundary with the desertic vegetation externally. Some of the desertic species occur near or within the more mesic margins of the garden, and some of them are occasionally present in the gardens (e.g. , species of yucca). Typically, the gardens give way gradually to plants that are more drought tolerant. Plants of intermediate moisture tolerance include a series of grass species, especially those com- mon elsewhere in western North America on prairies and plains. The classic alcove type of hanging garden in the Canyonlands of southeastern Utah consists of an overhanging back wall, a vaulted face wall, a detrital slope, and a plunge basin. The back and face walls support clinging plants of maidenhair fern (Adiantum capillus- veneris L.), cave primrose (Primula specui- cola Rydb.), Eastwood monkey-flower (Mim- ulus eastwoodiae Rydb.), rock plant [Petro- phytum caespitosum (Nutt.) Rydb.], and several other species. Some of these species occur also on the detrital slope, but the wet, sandy detritis supports the Garber phase of the golden sedge (Carex aurea Nutt.), small-flowered columbine (Aquilegia mi- crantha Eastw.), Jones reedgrass (Calama- grostis scopidorum Jones), helleborine orchid January 1989 Welsh: Utah's Hanging Gardens Fig. 3. Late winter view of Weeping Rock, Zion Canyon, Washington Co., Utah. Note the striated tufa deposit to the right of the main alcove. Cardinal monkey-flower and golden columbine are components of the vascular flora. (Epipactis gigantea Dougl. ex Hook.), alcove orchid (Habenaria zothecina Higgins & Welsh), bundle panic (Panicum acuminatum Swartz), Rydberg thistle (Cirsium rydbergii Petrak), alcove death camas [Zigadenus vagi- nalis (Rydb.) Macbr.], and several other spe- cies also. A fringing margin of western redbud (Cercis occidentalis Torr. ex Gray), netleaf hackberry (Celtis reticulata Torr.), and Gam- bel and Eastwood oaks (Que reus gambelii Nutt. and Q. x eastwoodiae Rydb.) often occurs outward from the foot slope where the plants tend to conceal the alcove base. Toward the drier margins of the garden are grasses typical of the prairies and plains of the west- ern United States. Little bluestem [Schiza- chyrium scoparium (Michx.) Nash in Small] is a common component, occasionally growing with Indiangrass (Sorghastrum nutans Nash in Small), switchgrass (Panicum virgatum L.), bushy bluestem [Andropogon glomeratus (Walter) B.S.P.], and big bluestem (Andro- pogon gerardii Vit.). The gardens tend to form a kind of microcosm of the prairies and deciduous summer forests more typical of portions of North America eastward from the Colorado Plateau. Other species of the back and face wall of some of the alcoves are the alcove daisy (Erigeron zothecinus Welsh) and alcove rock- daisy (Perityle specuicola Welsh & Neese). Only the cave primrose, Eastwood monkey- flower, Rydberg thistle, small-flowered col- umbine, alcove orchid, alcove daisy, and al- cove rock-daisy are endemic to the hanging gardens of the Colorado. Both terrace and windowblind gardens exist along the canyons of the Colorado also. Their floristic composition is frequently simi- lar to that noted above. However, unless there is at least some alcove development, the typical garden species are lacking or occur in reduced numbers. Hanging gardens in Zion Canyon are often of the windowblind type, with a flat face wall and a vaulted, dry arch at the top (Figs. 3-5). Sometimes the wet wall is curving and with- out an apparent arch at the top. The joints are mostly evident as cracks that can be seen be- tween the dry capstone arch and the smooth Great Basin Naturalist Vol. 49, No. 1 Fig. 4. Late winter view of Upper Emerald Pool Gar- den, Zion Canyon, Washington Co., Utah. A snowbank persists in the foreground. Joints in the Kayenta Sand- stone control water flow. Fig. 5. Late winter view of Narrows Trail Garden, Zion Canyon, Washington Co., Utah. This garden sup- ports the Zion shooting-star, maidenhair fern, yellow columbine, and western columbine. face wall, but they are markedly apparent as vertical cracks in the rock faces of gardens such as those at Upper Emerald Pool. Alcove development does occur to some extent within the windowblind gardens (Fig. 6), and alcoves are present below the Springdale Sandstone member of the Moenave Forma- tion (Fig. 7). There are no classic alcove gar- dens in Zion Canyon. Lack of classic alcoves can be attributed to the control of downward movement of water along the joint systems in the Navajo Sandstone, which is sometimes transitional into the bedded Kayenta Forma- tion. Alcove development within the window- blind gardens is minimal, due in part to the peculiarity of climatic conditions within Zion Canyon. Narrow, deep, and shaded for much of each day, more so than the gardens of Canyonlands, the windowblind gardens of Zion Canyon have low winter temperatures that result in ice formation over the wet sur- faces of the gardens. Moderating weather re- sults in ice melt. Sheets of ice cascade from the cliff face, shearing plantlets clinging pre- cariously to the stone. The attachment fails prior to building up an accumulation of detri- tis and prior to creation of minor indentations in the rock surface. Alcove development is slowed. Trees adjacent to the wet walls of Zion Canyon are etiolated as a result of shading through much of each day. The long, slender branches of many trees are unable to support the weight of the crown and break under the stress. Proportionally, the trees are too tall for the thickness of the trunks. Even the alcove at Weeping Rock is not well developed (Figs. 3, 6). Ice forms on its flat upper surface also, during some winters at least. And, though water from its aquifer is abundant during late winter and spring (Fig. 6), the water flow diminishes during early summer, and the garden is often merely damp during late summer and autumn. Tufa de- posits in the Weeping Rock garden provide microhabitats where plants are protected from erosion by ice. Tiny caverns and depres- sions beneath overhanging tufa accumulations January 1989 5 mE&''lk\ Welsh: Utah's Hanging Gardens Fig. 6. Late winter view south from the alcove of Weeping Rock, Zion Canyon, Washington Co. , Utah. Water drips abundantly from the tufa deposit and sandstone walls early each year but dries as the season progresses. support miniature gardens. The tufa deposits also protect the alcoves from exfoliation, which is apparent in the north portion of Weeping Rock where tufa is absent. Clinging plant species in the Zion Canyon gardens tend to be few in number (Malanson 1980, 1982, Malanson and Kay 1982). Fewer still are coincidental species with the hanging gardens of the Colorado. Jones reedgrass (Calamagrostis scopulorum Jones), Garber sedge (Carex aurea Nutt.), and maidenhair fern (Adiantum capillus-veneris L.) are pres- ent in gardens of both places. Northern maid- enhair fern (A. pedatum L.) is present also in some gardens in Zion Canyon but is missing in all Colorado gardens except for a few near the head of the Escalante drainage. Mostly the representation within the Zion gardens is con- generic, not conspecific, with that of the Canyonlands. Columbine (Aquilegia) is pres- ent in both areas, but that of the Colorado is Aquilegia micrantha Eastw., while those of Zion Canyon are A. formosa Fisch. in DC. and A. chrysantha Gray and their hybrids. A peculiar phase (var. fosteri Welsh) of A. formosa simulates A. micrantha in being glandular overall, but the plant is obviously allied to A. formosa. Red-flowered species of monkey-flower oc- cur in gardens of both areas, but the species are different. That of the Canyonlands is Mim- ulus eastwoodiae Rydb., while that of Zion is M. cardinalis Dougl. ex Benth. (including M. verbenaceus Kearney & Peebles). Eastwood monkey-flower is evidently confined to the hanging garden habitats of the Colorado Plateau, while the cardinal monkey-flower is a disjunct garden plant only at the northern limits of its distribution in Zion Canyon. Oth- erwise the cardinal monkey-flower is a wide- spread species of moist sites of the Southwest. The two species likewise differ in flowering time. The plant of the Colorado gardens flow- ers from August to October or even to Novem- ber in some years. The Zion plants have an initial flush of flowering in May and June, with fewer flowers produced thereafter into the summer months. The cave primrose (Primula specuicola Rydb.) of the Canyonlands gardens has a con- familial representative in Zion, i.e., the beau- tiful, broad-leaved shooting-star, Dodecatheon 10 Great Basin Naturalist Vol. 49, No. 1 Fig. 7. Lower and Upper Emerald Pool gardens, Zion Canyon, Washington Co., Utah. The upper garden is in Kayenta Sandstone, the lower in the Springdale member of the Moenave Formation. pulchellum (Raf.) Merrill var. zionense (Eastw.) Welsh. While both the cave primrose and the Zion shooting-star begin growth and flower early, the cave primrose is the more precocious, flowering as early as late January in some years. The typical period of main flowering is March to May. The Zion shooting- star seldom flowers prior to early April, with greatest flowering occurring during May. The genus Primula per se is not known in Zion. The Zion variety of the pretty shooting-star extends into some minor gardens along the Colorado, especially in those of lower Last Chance Canyon east of Wahweap and in se- lected gardens as far north as the mouth of the Escalante along Glen Canyon. Zion Canyon has other species unique to Utah hanging gardens. American spikenard, Aralia racemosa L. ssp. bicrenata (Woot. & Standi.) Welsh & Atwood, is perhaps the most peculiar of Zion hanging gardens species. It is typically present on margins and lower shelves below the great wet walls. Occasion- ally it clings, attached in crevices, to the walls of the grottos associated with the gardens, and seldom the plants grow on sandy benchlands and terraces in the Narrows portion of Zion Canyon, removed from the gardens alto- gether. The spikenard occasionally grows to a height of almost 2 m and has ternate-pinnate leaves to almost 1 m in width. The Zion daisy, Erigeron sionis Cronq., is an endemic of moist sites in Zion Canyon. Habitats of the Zion daisy vary in size from minute areas, wet only in springtime by water percolating shal- lowly in sandstone, to the largest of the hang- ing gardens in the canyon. Growth of this attractive small plant with white flowers and lobed leaves is aided by production of stolons that bind the plant closely to the moist, sandy surface. Grasses typical of prairies and plains, such as occur within and on the fringes of gardens of the Colorado, seldom form such stands in Zion Canyon. Many of the same species occur in the vicinity, but in Zion they are typically riparian components. Hanging gardens occur on the red sand- stone cliffs immediately north of the business district of St. George, Utah. These gardens are exposed to direct sunlight through much of each day, except where the alcoves are sufficiently developed to provide shade in early morning and late afternoon. Maidenhair fern (Adiantum capillus-veneris L.) is a prin- cipal component of these gardens also, but the bright flowers of columbine, primrose, and shooting-star are missing. Instead, the blue flowers of Sisyrinchium demissum Greene grace these gardens. The thistle spe- cies, Cirsium virginensis Welsh, grows in them but is not confined to this habitat. Sol- idago spectabilis (D.C. Eaton) Gray, the Nevada goldenrod, is an opportunist in the St. George gardens. The gardens have served as dumping grounds for residents of the region and contain old refrigerators, tires, washing machines, and other refuse, a kind of permanent condemnation of the humanity of our time. The examples discussed here give an indi- cation of the importance of position and ecol- ogy in the determination of garden diversity and species composition. Not indicated is the variation from place to place within a major drainage system or from garden to adjacent garden at a given place. January 1989 Welsh: Utah's Hanging Gardens 11 Virgin River Gardens Hanging gardens are present at St. George, in Zion Canyon along the North Fork of the Virgin River, and along Parunuweap, a canyon cut by the main fork of the Virgin. The largest and best developed are those of Zion Canyon, described above. The Parunuweap gardens are mainly associated with the base of the Springdale Sandstone, and they are gen- erally small and lack the diversity of those in Zion Canyon. Eastward in Parunuweap the canyon is incised into the Navajo Sandstone, and gardens are larger and better vegetated. The Virgin thistle (Cirsium virginensis Welsh) was described from plants of the al- coves at the north margin of St. George. The species was first taken, evidently, by Charles Christopher Parry during his visit there in 1874 (Welsh 1988). Subsequently, the plant was collected by other botanists but remained unnamed until this decade (Welsh 1982). The thistle is evidently a riparian species, which reaches its northernmost distribution in these hanging gardens. The plant is known other- wise from moist habitats in adjacent Mohave County, Arizona, and Clark County, Nevada. Colorado River Canyons Gardens Glen Canyon There is a land that was That we who are can never see For it is drowned In crystal waters of a Stone-bound inland sea — Glen Canyon SLW, Goldfield, Nevada, 2 April 1982 Moist spots on the canyon walls immedi- ately downstream from Glen Canyon Dam support hanging garden species. These can be viewed by looking almost vertically into Glen Canyon from the visitor center of the recre- ation area. They are only an indication of the gardens to the east along Glen Canyon proper in Utah. Glen Canyon was named by John Wesley Powell, who entered it on 29 July 1869 (Pow- ell 1875). Powell (1875) states: On the walls, and back many miles into the country, a number of monument-shaped buttes are ob- served. So we have here a curious ensemble of wonderful features — carved walls, royal arches, glens, alcove gulches, mounds, and monuments. From which of these shall we select a name? We decided to call it Glen Canyon. On 3 August 1869, Powell (1875) gave the following description: Sometimes the rock are overhanging; in other curves curious narrow glens are found. Through these we climb by a rough stairway, perhaps several hundred feet, to where a spring bursts out from under an overhanging cliff and where cottonwoods and willows stand, while along the curves of the brooklet oaks grow and other rich vegetation is seen, in marked contrast to the general appearance of naked rock. We call these Oak Glens. Thus, Powell not only chose to name the canyon after the glens observed in its length, but he gave a general description of these unique botanical features. The oak glens are now known as hanging gardens. Glen Canyon, which begins at the conflu- ence of the Dirty Devil River and the Colo- rado, near the Hite Rridge, and terminates at Lee's Ferry in Arizona, is readily divisible into the following three main segments. Wahweap to Confluence of the San Juan Vegetation along the shores of Lake Powell from Wahweap eastward to Rock Creek and vicinity is composed mainly of species of shrubs, perched atop sands derived from the Navajo and Carmel formations. North of the lake the Entrada forms cliffs below the escarpments carved into the soft Tropic Shale member of the Mancos Shale. The Entrada changes appearance east from Wahweap; at Padre Bay it has become a candy-striped to red sandstone, instead of the chalky white, stained brown cliff former at Wahweap Bay. Small hanging gardens of the terrace type are in Crosby Canyon, a minor tributary of Warm Creek. Other similar terrace gardens are present in Last Chance Canyon, which intersects bedded portions of the Entrada. A small garden at Nipple Spring on Nipple Bench, north of Big Water (formerly Glen Canyon City), includes maidenhair fern in a tiny, well-watered alcove in the Straight Cliffs formation. Excess water supports Fremont cottonwood and other riparian plants (Fig. 8). Water quality in these formations is generally saline, and white encrustations of salts are commonplace. Zion shooting-star is a compo- nent of the gardens, which support large stands of little bluestem. Etiolated, sprawling plants of roundleaf buffaloberry (Shepherdia rotundifolia Parry) and skunkbush (Rhus aro- matica Ait.) sometimes persist in minor al- coves in this formation. These gardens have not been studied in detail. 12 Great Basin Naturalist Vol. 49, No. 1 ^Vf'4; -^Px^ ",; Fig. 8. Nipple Spring, Nipple Bench, Kane Co., Utah. Water from the Straight Cliffs Formation supports a small hanging garden of maidenhair ferns and gives moisture that allows growth of Fremont cottonwood. There are some small gardens in the red, pockmarked phase of the Entrada in the Rock Creek vicinity. Others in that formation were inundated by Lake Powell. The best devel- oped of the hanging gardens remaining in this sector of Glen Canyon are to the east of Dan- gling Rope, where the Navajo Formation reappears at the water's edge. Possibly the best examples are those in Driftwood Canyon, whose backdrop to the north is the great gray, water-stained cyclorama of cliffs below the summit of Fifty-Mile Mountain. On the east side of Driftwood Canyon, around the first meander bend north of its mouth, are the remains of a small but unique plant assem- blage known as Step Garden. For the sake of reference the hanging gardens studied previ- ously by this author have been given binomi- als, not unlike those used scientifically for plants. Step Garden received its name from the historic steps, now beneath the lake, carved by optimistic early prospectors in search of gold in ancient terrace gravels perched high above the present inner gorge. Though small, Step Garden is almost unique among hanging gardens in Glen Canyon in supporting a stand of saw-grass [Cladium californicum (Wats.) O'Neil in Tidestr. & Kittel]. This saw-grass, a close relative of the warm temperate and subtropi- cal C. mariscus R. Br. with which it is some- times united as a variety, grows here in huge clumps. The accumulated bases beneath the current seasons of growth are now more than 2 m thick. A sample of leaf bases from near the bottom of the accumulation was radiocarbon dated at more than 400 years in age. The saw-grass is unusual in Utah, reaching its northern limits here. The species grows also along Furnace Creek in Death Valley, Califor- nia. In the broad sense of C. mariscus, the species is known from California, Nevada, Arizona, Mexico, Central America, eastern North America, and the Old World (Munz 1970). In Step Garden the saw-grass occurs with western redbud, another species of broad distribution to the south and west of Glen Can von. January 1989 Welsh: Utah's Hanging Gardens 13 The alcove at Step Garden is very small, only about 3-5 m in height. It has an over- hanging upper ledge, whose underside is clothed with maidenhair fern and Eastwood monkey-flower. The face wall has more maid- enhair fern, rock plant, and helleborine or- chid. A small plunge basin, scoured by water that pours over the cliff margin from a drainage above, is filled with cool water that seeps from the sandstone of the alcove. The alcove was formed from interaction of plants and the wet surface at the margin of a small bedding plane almost at the base of the Navajo Sandstone. Such alcoves are a dominant fea- ture of this portion of Glen Canyon, and it is suggested that all alcoves along the canyon, whether now vegetated or not, are the result of previous hanging garden and wet bedding plane interaction. There is a developmental sequence in hanging garden and alcove development from a simple wet wall or wet spot on a wall, to a simple alcove without a plunge basin, to a classic alcove with a plunge basin (Welsh and Toft 1981). A final category of alcoves is that designated as decadent, those in which the alcove has become too deep and the roof has collapsed, sealing the moisture of the bedding plane or those in which the bedding plane has dried. Step Garden is a small classic alcove, shel- tered in front from the searing heat of the sun and from wind action by a cluster of redbud and the growth of saw-grass. The plunge basin at Step Garden is sheltered behind the wall of redbud. In some gardens the plunge basin is un vegetated, the falling water from seasonal storms or from melting snow scouring the basin into bedrock. However, in some classic- gardens the plunge basin is in accumulated sandy detritis from both the falling water and exfoliation of the alcove. In those situations the plunge basin will be surrounded by meso- phytes such as seep willow (Baccharis spp.), bushy bluestem, alcove death camas, and var- ious sedges. Pot sherds adjacent to Step Garden and other gardens with plunge basins suggest their importance as sources of water for pre- historic peoples. Pipelines and water cisterns of contemporary civilization are prominent features of hanging gardens in St. George and in Zion National Park. Water in the gardens is utilized by animals of many kinds (Welsh, Wood, and Raines 1975). Birds drink the water and roost and nest in the vegetation. And, since the water supply of the gardens is more or less indepen- dent of the climatic regime of the region, plant growth in them is independent of current an- nual rainfall. There is vegetation in abun- dance each year (Welsh and Wood 1975). Be- cause of the productivity of the gardens, several species of small desert mammals sur- vive in them during periods of less than ade- quate rainfall in the sea of aridity abounding beyond the gardens proper. Woodrats and deer mice especially live in the gardens dur- ing periods of climatic stress and move out- ward from them during seasons of adequate moisture and concurrent vegetative growth in the surrounding desertic lands (see discussion under Three Garden below). Canyon tree frogs (Hyla arenicolor Cope) and red-spotted toads (Bufo panctatus Baird & Girard) congregate in the plunge basins of the gardens and in canyons with permanent streams (Toft 1972). They begin to call as evening falls, often accompanied by the dying calls of a canyon wren, with its lilting song that begins high and falls gradually prior to ceas- ing. In spring and early summer croaking of the frogs and toads rises to an amazing cacaphony soon after nightfall, often magni- fied by the vaulted arch of a hanging garden alcove. Eastward from Driftwood Canyon, on the south side of the lake, is Forbidding Canyon, which contains Rainbow Bridge, sculpted from Navajo Sandstone and perched on shelf- rock of Kayenta Formation, an impervious sandstone. A stream channel bearing peren- nial water is entrenched into the Kayenta at Rainbow Bridge. The Kayenta supported sev- eral small hanging gardens prior to flooding by Lake Powell. Plants of Toft yucca occur on sand adjacent to the bridge, and western red- bud grows here and there against the cliffs, watered by runoff from the slickrock. Not far west from the confluence with the San Juan Arm, Hidden Passage Canyon en- ters Glen Canyon from the west. Its walls are vertical and the canyon is short, ending abruptly in a boxed end. Along the canyon, at the base of vertical cliffs atop rounded sand- stone slopes that drop to the water, are a 14 Great Basin Naturalist Vol. 49, No. 1 mggmamm Fig. 9. Plunge basin and lower bedding plane at Pool Garden, Reflection Canyon, Lake Powell, Kane Co., Utah. This garden now lies beneath Lake Powell. Sheathed death camas grew on the wet wall, and the plunge basin, more than 15 m in width, served as home for beavers. couple of clumps of saw-grass, a portion of what existed prior to the high water of Lake Powell. Across the lake from the mouth of Hidden Passage is Music Temple Canyon, or what is left of it. The canyon mouth plunged for more than 100 feet into a huge alcove or grotto, elongated oval in form. The walls were clothed in part with patches of maidenhair fern and monkey-flower. It was in this grotto that Powell (1875) and his party encamped. Powell noted: On entering, we find a little grove of box-elder and cottonwood trees, and turning to the right, we find ourselves in a vast chamber, carved out of the rock. At the upper end there is a clear, deep pool of water, bordered with verdure. Standing by the side of this, we can see the grove at the entrance. The chamber is more than 200 feet high, 500 feet long, and 200 feet wide. Through the ceiling, and on through the rocks for a thousand feet above, there is a narrow, winding skylight. ... It was doubtless made for an academy of music by its storm-born architect; so we name it Music Temple. To the north of Music Temple Canyon, across Lake Powell, is Reflection Canyon, certainly one of the most photogenic tribu- taries of Glen Canyon. Interlocking canyon spurs now plunge into the placid waters of the lake, giving reflections that deceive the mind — air passes into water without percep- tion of the difference in state of matter. Some of the meander bends contain remnants of hanging gardens. One such garden, named Reflection (Fig. 9), had an alcove more than 60 m in height and breadth. A plunge basin more than 15 m across served as home for a family of beavers. Alcove death camas, bearded blue- stem, and seep willow grew around the plunge basin and on the foot slope of this enormous garden. The wet wall of this and of most other gardens was covered by a mat of green and blue-green algae, often with globes of Nostoc staring eyelike from the glistening algal mat (Clark 1972, Rushforth et al. 1976). Now the lower portion of the garden is flooded, including the plunge basin, a part of Lake Powell. The beaver have moved up canyon in pursuit of cottonwood trees, their main source of food. January 1989 Welsh: Utah's Hanging Gardens 15 Fig. 10. Three Garden, ca 1.5 km north of the confluence of the San Juan and Glen Canyon arms of Lake Powell, San Juan Co., Utah. This superposed set of gardens was selected for study of species composition, cover, productivity, and rodent interaction. Lake Powell now covers the lower garden up to the vegetated stripes on the face wall. San Juan Confluence to Bullfrog North of the confluence of Glen Canyon with the San Juan, on the east side of the lake, is an area designated as Gardens Cove. Sev- eral small hanging gardens existed in this area, and a large one, designated as Three Garden, is present at the north end (Fig. 10). Three Garden was named for its three super- posed alcoves, each with hanging garden de- velopment. It was a classic set. The lowermost garden had a plunge basin, a lower shelf, a gently angled foot slope, a rounded, arching face wall, and an overhanging back wall, with all of the plant components noted previously. It became the type with which all other gar- dens were compared by me in later studies. The basal two-thirds of the lower garden is now drowned, but Middle and Upper gar- dens persist. The two upper gardens can be reached by judicious scrambling up slickrock. Middle Garden is dryish, with a pothole arch in its back wall. Upper Garden (Fig. 11) is huge, with a scoured shelf where a plunge basin might develop. Down the drainage toward the lip of the overhanging margin of Middle Garden are a series of swirl holes carved into the stone, one of which became the pothole arch. 16 Great Basin Naturalist Vol. 49, No. 1 Fig. 11. Upper Three Garden, with two upper bedding planes and a larger basal one, Navajo Sandstone, San Juan Co. , Utah. Water drips from the bedding planes at all seasons. At Three Garden we undertook investiga- tions of small mammal populations (Welsh and Toft 1972, Welsh, Wood, and Raines 1975) and studied the use of the gardens by rodents especially. Three groups of rodents were found in or near the garden habitats. They are the cricetids, heteromyids, and sciurids. Cricetid rodents include the native rats [Neotoma cinerea (Ord) and N. mexicana (Baird)] and deer mice [Peromyscus boijlei (Baird) and P. crinitus (Merriam)]; the het- eromyids are the kangaroo rats [Dipodomys ordii (Woodhouse)] and pocket mice [Per- ognathus apache (Merriam) and P. inter- medins (Merriam)]; and the sciurids are the antelope ground squirrel [Ammospermo- philus leucurus (Merriam)] and chipmunk [Eutamias quadrivittatus (Say)]. The cricetid and heteromyid rodents are nocturnal animals who forage at night and sleep in burrows during the daytime. The sciurids are diurnal. The trapping design was primarily for the nocturnal animals. Traps were opened and baited in early evening, and the traps were checked and the animals released prior to sunrise. Inclusion of the sciurids in the traps at all represented chance occurrences of late afternoon or early evening visitation. Three principal habitats occurred in the vicinity of the gardens, i.e., the gardens proper (HG), the immediately adjacent talus (TDS) slopes dominated by sparse cover of shrubs and grasses, and a semidesert shrub (SDS) on the more gently sloping, sandy ground away from the garden and talus slopes. The hanging garden habitat was tested in the summer of 1972 (Welsh and Toft 1972) to de- termine species presence and potential of movement from garden to garden in the su- perposed set. Animals moved from Middle to Lower Garden during the three-night test trapping. In 1973 and 1974 a larger trapping design was imposed on the HG, TDS, and SDS communities. Results of those studies indicated that there was a partitioning of habi- tat by species of the three rodent groups. Heteromyids avoided the gardens altogether (except for Perognathus intermedins, which lives almost exclusively in the adjacent TDS habitat). Sciurids visited the gardens proba- bly for food and occasionally for water, even January 1989 Welsh: Utah's Hanging Gardens 17 Fig. 12. Double Garden, west side of Glen Canyon, ea 1 km west-northwest of Three Garden, Kane Co., Utah. Bedding plane control of this linear garden is apparent. Hanging garden plants are restricted to the strike of the plane. though the visitation was not adequately tested in the trapping design. Cricetids lived in all of the habitats available but reacted dif- ferently on a species-by-species basis to each habitat. A trapability index was proposed as a device wherein the dynamics of rodent species could be partitioned within each of the habitat types. Neotoma mexicana and Peromyscus boylei showed a definite preference for the hanging garden habitat both in 1973 and 1974. Neotoma cinerea was indiscriminate with re- gard to habitat, but was not captured in the SDS during the dry year of 1974. Peromyscus crinitus was important in all habitat types but showed a definite preference for the SDS community. The heteromyids are habitat se- lective and evidently do not depend on the hanging gardens for either food or water. They are well adapted to the dry conditions in the desert plant communities external to the gardens. Although woodrats and deer mice used the gardens at all seasons, they moved from the surrounding dry habitats into the gardens during seasons of low total precipita- tion and corresponding low food production. Welsh and Wood (1975) conducted produc- tivity, cover, and composition studies on the plant community within lower Three Garden. Productivity of the gardens was high during both 1973 and 1974, even though 1974 was a dry year and productivity of the SDS commu- nity declined considerably during that year. Similar studies of animal communities and plant productivity should be conducted in other hanging garden sites in both the Colo- rado and Virgin basins. To the west, across Lake Powell from Three Garden, is Double Garden whose hanging garden vegetation is aligned along a bedding plane near the base of the Navajo Sandstone (Fig. 12). North of Hole-in-the-Rock, where the San Juan pioneers labored so diligently, on the east side of Glen Canyon, a canyon named Ribbon enters through sentinellike monoliths of Navajo Sandstone. Immediately within the mouth of the canyon, on the south side, is a huge, perpetually shaded alcove, bearing the monument-sized Ribbon Garden (Figs. 2, 13). Within the huge alcove is a smaller one, facing westerly. Both gardens support plant 18 Great Basin Naturalist Vol. 49, No. 1 •■ S$$^4$£, Fig. 13. Ribbon Garden, Ribbon Canyon, Lake Powell, San Juan Co., Utah. The detrital slope in the shadow is mainly on Kayenta Formation terraced margins. The slope supports a growth of New Mexico raspberry and Knowlton ironwood. A smaller alcove is present within the large alcove, which is shaded almost continuously. Its detrital slope is covered with Rydberg thistle, and small-flowered columbine clings to the wet wall. species unusual in the Utah flora. It was here that we first discovered the New Mexico rasp- berry (Rubus neomexicanus Gray), a plant with leaves more like a currant or gooseberry and with pure white, roselike flowers to 4 cm wide. The species is known in Utah only in hanging gardens along Glen and Cataract canyons, where shaded for most of each day. Growing with the raspberry is the Knowl- ton ironwood (Ostrya knowltonii Cov.), an- other rarity within the Utah flora. The small trees flower early in springtime, with stami- nate catkins borne pendulous from buds near branch ends. Pistillate catkins appear later, finally evident from the hoplike, inflated, papery bracts. The species is known in Utah only from along Glen Canyon and its tribu- taries, from the Needles section of Canyon- lands National Park, and from along the Colo- rado River near Moab. The small alcove on the east side of Ribbon Garden is almost perpetually in shade, and the foot slope is overgrown with a dense car- pet of Rydberg thistle (Cirsium rijdbergii Pe- trak), another plant confined to or near hang- ing gardens in the canyons of the Colorado. January 1989 Welsh: Utah's Hanging Gardens 19 The basal cluster of leaves can be up to 1 m across, much larger than any other native Utah thistle, but the flower heads are small, seldom more than 20 mm in length. Ribbon Canyon is only a few thousand feet in length, ending abruptly in boxed ends. There are numerous alcoves along its margin, some with well-developed plunge basins. Those on the south side of the canyon support the New Mexico raspberry; those on the north do not. Water from the gardens forms a stream that flows into the lake. Reduced evaporation results in maximum stream flow at night; some of the streams and portions of hanging gardens dry completely each day in summer. East of the mouth of Escalante Canyon, on shelfrock of Kayenta Sandstone rising gently from the water's edge on the north side of the canyon, are the Escalante gardens. There are eight alcoves carved into the base of the Navajo Formation, some of them with plunge basins. They contain the usual hanging garden species for Glen Canyon. Additionally, one of them is the type locality for the alcove daisy, Erigeron zothecina Welsh. Tall spires of the Toft yucca (Yucca toftiae Welsh) stand sen- tinellike along the shelfrock associated with the gardens. Bullfrog to Hite Near the lake end in Moki Canyon, in the first meander bend east of Halls Crossing, there is a large hanging garden along the south side of the canyon. The approach is through drowned Fremont cotton wood trees. The al- cove is classic with a plunge basin, but the foot slope is brush clad, with poison ivy (Toxico- dendron rydbergii Small) as a principal com- ponent. Poison ivy exists in many hanging gardens, but nowhere as abundantly as in this garden. Knowles Canyon has the remnants of mag- nificent hanging gardens on the south side. Possibly they had plunge basins in the past, but now the lake receives the water from pour points and from the gardens. New Mexico raspberry is included in the vegetation. Gar- den development in alcoves on the north side is not as great, but the gardens there support an abundance of grass, with little bluestem and Jones reedgrass being common. On the east side of the lake at Good Hope Bay there are springs surrounded by a pecu- liar phase of Gambel oak. The spring sites are not situated on exposed sandstone walls, but rather arise on the Chinle Formation, with water possibly developed from joint systems in the Wingate Sandstone to the east. The oak there has acorns much larger than the species elsewhere in Utah. It is called Quercus gam- belii Nutt. var. bonina Welsh, the Good Hope oak. Possibly this phase of Gambel oak has been derived through hybridization and back- crossing between Q. gambelii and Q. havardii Rydb., the shinnery oak. Potentially it could exist in hanging gardens in this portion of Glen Canyon, but further exploration is nec- essary. Ticaboo Canyon enters Good Hope Bay from the west. Water flows perennially down the canyon bottom for a short distance, and there is minor hanging garden development near its juncture with the lake. Here along the stream is the northernmost known locality for bushy bluestem and for the scarlet lobelia or cardinal flower (Lobelia cardinalis L.). The cardinal flower is known from hanging gar- dens in the vicinity of Gardens Cove near the San Juan confluence, and in moist situations in Zion Canyon also. Prior to the existence of Lake Powell, a peculiar species of aster was taken at Ticaboo. Aster spinosus Benth., the Mexican devilweed, was known from a garden in the mouth of Llewellyn Gulch, which was flooded during the high water of 1983. Other- wise the species exists as a plant of sandy and gravelly flood plains and bars north to near the confluence of the Colorado and Green rivers. Escalante Canyon — Waterpocket Fold Escalante Canyon was formed by entrench- ment of Escalante Creek in the synclinal flex- ure west of the Waterpocket Fold anticline, whose east side is steeply dipping and whose west side falls more gradually to the Es- calante. Two canyons near the south end of the fold drain into Glen Canyon proper. They are Bowns Canyon and Long Canyon. Bowns Canyon is of interest because of the numerous small alcoves, most of them with at least some hanging garden development. It is in this canyon that evidences of prehistoric animals, including mammoths, have been discovered amidst evidences of vegetation now confined to higher elevation, cooler portions of the state. The entrance to Long Canyon is blocked by a huge nickpoint, occupied by a moderately developed hanging garden. A 20 Great Basin Naturalist Vol. 49, No. 1 > Fig. 14. Cow Canyon Garden, at the end of the Left Fork of Cow Canyon, Waterpoeket Fold, Kane Co., Utah. Principal vegetation on the foot slope of the upper garden is Rydberg thistle. This is one of many alcove gardens at the base of the Navajo Formation in Cow Canyon. route to the west will lead into the canyon, which requires additional investigation. There are several canyons that drain west from the crest of the fold into Escalante Canyon. Explorer Canyon had a perennial stream and some garden development, but the most interesting of the Waterpoeket canyons is that named Cow. Indeed, it is per- haps the most intriguing of all canyons that drain water to Lake Powell. Cow Canyon is deeply entrenched through the Navajo Sand- stone into the Kayenta Formation. The slope of the canyon bottom approximates that of the dipping west slope of the Waterpoeket Fold. Alcoves line the canyon walls, with one or more in each meander bend. Typically the alcoves are perched atop the Kayenta plat- form. The gardens often are classic alcoves with developed plunge basins, but the species complement varies with each garden — there are no two gardens alike. The canyon has two main branches, and the left-hand branch forks again near its apex. Each of the forks is termi- nated by box ends that form hanging gardens of huge vertical relief. The left fork terminates in paired, owl-face-like alcoves filled with greenery (Fig. 14). The right fork terminates in an alcove of great height, possibly as much as 120 m. One alcove west of the forks of the left-hand fork of Cow Canyon is occupied by Knowlton ironwood almost to the exclusion of other woody vegetation. Other of the alcoves stud- ied lacked this plant. Evidences of past occu- pation by Indians are also present. Alcoves where the bedding planes had dried support small stone structures where grain and other items derived from primitive agriculture could be stored. Dwellings are also present. White Canyon White Canyon drains from the west margin of Elk Ridge and enters Glen Canyon south of the Hite Marina. The canyon proper is en- trenched through the Permian Cedar Mesa Formation, which consists of water-deposited sands alternating with fine-textured materi- als. The sands are sometimes massive, and the fine materials serve to halt the flow of perco- lating water. Small alcove gardens, mainly lacking plunge basins, are present in this for- mation. Some of the exposed aquifers are lin- eal, extending as strips along the exposed sandstone margin, only a few decimeters to a few meters in width. Maidenhair fern grows in these wet sites along with small-flowered columbine and the Kachina daisy (Erigeron kachinensis Welsh & Moore). Gardens occur here and there along much of the length of the canyon, but possibly the best of them occur within Natural Bridges National Monument, in both White and Armstrong canyons. The alcove death camas was initially collected by P. A. Rydberg and A. O. Garrett in 1911, probably from the wet spot in the canyon immediately south of Owachomo Natural Bridge. The Kachina daisy was taken from gardens on the east side of the Kachina Mean- der in 1963 by Welsh and Moore (1964). This plant was at first thought to be endemic to the gardens, but more recent collections show it to occur on sandy sites in ponderosa pine forests in the Abajo Mountains. It is known from hanging gardens in Dark Canyon and from along the Delores River in Colorado also. January 1989 Welsh: Utah's Hanging Gardens 21 Cataract Canyon At the Hite Marina a buff sandstone, the Cedar Mesa Formation, is intersected by Lake Powell. This very old sandstone is a portion of the Cutler Group of formations. Slickrock margins of the lower portion of Cataract Canyon are in this formation. The slickrock supports scattered specimens of Utah juniper, narrowleaf yucca, blackbrush, and many other shrubs and grasses. The lower end of Cataract Canyon is marked by cliffs, with shelves supporting desert shrubs such as single-leaf ash (Fraxinus anomala Torr. ex Wats.). Redbud grows along the shelfrock also, with the vicinity of Dark Canyon mark- ing its northern limits. Just north of Dark Canyon, on the east side of Cataract, is a poorly developed hanging garden. The gar- den faces north and is stained by carbona- ceous black water stripes. The foot of the gar- den supports some of the typical garden species noted earlier. This garden is the southernmost known locality for the alcove rock-daisy, Perityle specuicola Welsh & Neese. It is a peculiar, rushlike member of the sun- flower family with slender, drooping branch- lets and small heads of cream disk ilowers. The plant is known otherwise from gardens in the vicinity of Moab. There is another poorly developed hanging garden on the south side of the first bend west of the Gypsum Canyon reentry, east of Clear- water Canyon. The garden, named Ron's Gar- den, is peculiar in that it is formed in lime- stone and has a large central pillar of tufa (similar to the tufa deposits along the trail to Zion Narrows), derived from calcium carbon- ate in water that has percolated through the Hermosa Formation. Besides several of the usual hanging garden plants — such as maid- enhair fern, cave primrose, and the small- flowered columbine — the New Mexico rasp- berry is present. This is the northernmost location known for the raspberry. San Juan Arm At the juncture of the San Juan with Glen Canyon, the Navajo Formation rises above lake level. The canyon of the San Juan Arm of Lake Powell is entrenched in a series of mean- der bends, each different from the others and each with hanging gardens. On the west side of the first meander bend, almost straight through the rock less than 1 km from Three Garden, is a huge alcove, dry in its tremen- dous upper portion but wet along the terraced base formed from the Kayenta Sandstone. A cottonwood tree grows on the shelfrock, and a terraced type of hanging garden exists. Ryd- berg thistle, redbud, maidenhair fern, and alcove death camas are components of the vegetation. This is Death Camas Garden (Fig. 15). It is an excellent example of an alcove with upper bedding planes either dried or buried with detritus, but with the impervious Kayenta being wet from water percolating to its surface and flowing down its margin. Death Camas Garden demonstrates the dif- ference in species composition of closely spaced garden assemblages. It contains the alcove death camas, which did not occur in Three Garden, but was present in other gar- dens almost due west of Three Garden across Glen Canyon. The death camas occurred also in Reflection Garden. Other species show similar peculiarities of distribution. A small canyon enters the San Juan at the north end of the first meander curve. Several hanging gardens are present in its box end and margins, in alcoves perched atop the Kayenta high above the lake surface. Most of these gardens lack plunge basins, the water having scoured the rock surface smooth and in some instances having worn swirl holes in the stone. Nasja Canyon enters from the south in the beginning of the second meander bend. Its entry is blocked by a nickpoint in stone, whose alcove contains remnants of a de- pauperate hanging garden clad with small- flowered columbine, cave primrose, and a few other plant species. Eastward along the lake, Wilson Creek enters across from Trail Canyon. The mouth of Wilson Creek is also blocked by a nickpoint, which is sometimes above and sometimes be- low lake level. Perennial water flows from Wilson Creek. Above the nickpoint grows a stand of saw-grass, a third known locality, if it still survives following the high lake level of 1984 and 1985. Along the stream grow cotton- wood, seep willow, willow, and the beauti- fully plumed satintail grass (Imperata brevifo- lia Vasey), which is currently known from only this location in Utah. Wilson Creek is evidently that described by Platte D. Lyman of the Hole-in-the-Rock expedition (Miller 1959). Lyman and party 22 Great Basin Naturalist Vol. 49, No. 1 (t^^ > \ *■-*+? Fig. 15. Death Camas Garden, in the first meander bend east of the confluence of the San Juan and Glen Canyon arms of Lake Powell, San Juan Co., Utah. The alcove is more than 100 m wide and perhaps half as high. It supports a terrace-type garden on the bedded Kayenta Formation margin, principally Rydberg thistle, sheathed death camas, and maidenhair fern. Western redbud is prominent. followed it down to the San Juan River. The canyon was described as having a small stream and lush vegetation. Deep water-filled holes occurred here and there, and Lyman indi- cated that they caught several mud turtles as large as a man's hand. These have not been found in recent times, but an occasional turtle shell has been recorded for Glen Canyon. Doubtlessly, Lyman's observation of 1 De- cember 1879 is correct. The Great Bend of the San Juan is marked in its outer margin by huge alcoves with hang- ing gardens. The alcoves are some distance from the lake. Limited investigations yielded no new distributional information from these gardens. Bluff Pioneer plant taxonomist Alice Eastwood (1896) took plants from the moist garden-clad alcoves near Bluff, Utah, in the 1890s. She was probably the first botanist of consequence to visit and describe the unique hanging gar- den habitat. She (1896) stated that the habitat "is a boreal oasis in the midst of a Sonoran desert." Gregory (1938), in speaking of the canyons of the San Juan and of Butler Wash, noted that "the spring line at the base of the bare Bluff sandstone is marked by a bank of green vegetation formed by plant species that seem out of place in the present scheme of distribution." The types of Mitnulus east- woodiae Rydb., Primula specuicola Rydb., and Cirsium rydbergii Petrak (C. lactucinum Rydb.) were collected by Rydberg and Gar- rett during 1911 (Rydberg 1912, 1917, Petrak 1917). Aquilegia micrantha Eastw. was taken from the gardens by Alfred Wetherill in 1894 (Eastwood 1896). The Bluff Sandstone has been variously regarded, either as a member of the Entrada, a separate formation, or, more recently, as a basal member of the Morrison Formation. The stratum is not especially thick, and al- coves worn in its margin rest on the Wanakah Formation (previously regarded as Sum- merville). January 1989 Welsh: Utah's Hanging Gardens 23 The gardens are readily accessible in Cot- tonwood Canyon north of Bluff, or along the strike of the formation east or west of that canyon. The cliffs of the Bluff Sandstone are less imposing to the eastward, and the gar- dens are accessible at ground level. Indeed, some of the easternmost gardens are grazed by sheep and goats, who drink the water draining from them and that which accumu- lates in the plunge basins following storms. The cave primrose is most impressive when in full flower, and coincidence of flowering in some years with the Easter season has led to the local common name of Easter flower. The gardens and their flowers are easily seen from the main highway through Cottonwood Canyon. Moab and Vicinity Examination of the plant collection in the herbarium of Brigham Young University in 1960 demonstrated the existence of Ostrya knowltonii Cov., represented by several sheets of a single collection taken many years earlier by Walter P. Cottam, pioneer plant ecologist at the University of Utah and prior to that professor of botany at Brigham Young University. The locality for the plant was sim- ply "Moab." In a later discussion with Dr. Cottam I asked him from where the plant had been taken, and he replied: "I won't tell you! It is so much fun when you find it for yourself!" The collection came from a hanging garden near Moab, and I later rediscovered it there. There are distinctive classic alcove gardens west of Moab at The Portals, east of the Colo- rado Biver bridge north of Moab, and in the reentry canyons along the Colorado Biver. Negro Bill Canyon has several gardens rang- ing in size from a few meters square to huge classic alcoves with plunge basins. The re- entry canyon west of Negro Bill has a darkly shaded alcove and plunge basin at its boxed end. The garden is accessible only by consid- erable effort of clawing one's way through a thick growth of oak and scrambling over rock falls. Cave primrose, maidenhair fern, alcove death camas, and other species characterize these gardens. The alcove rock-daisy and al- cove bog-orchid were named and typified on plants taken from alcoves in the first meander bend east of the Colorado Biver bridge. The gardens nearest the road are heavily impacted by humans, and they have not survived as they were when first visited by botanists fol- lowing the turn of this century. Several intro- duced tree species have escaped and grow within or immediately in front of the gardens. Catalpa, Siberian elm, and tamarix are now growing in them. Befuse from campers and hikers clogs some of the gardens. The Delicate Arch section of Arches Na- tional Park displays many hanging gardens. A fault line north of the trail head to that arch trends generally east-west and the Entrada Formation has an offset of some meters. The wall of that offset is marked by alcoves that continue along the strike of the formation west to Fresh Water Canyon. The trail to Delicate Arch is on the downthrown side of the fault, and the gardens are readily visible along it. Dead Tree Garden is the larger of those dis- played (Figs. 16, 17, 18). It is approachable up a slickrock drainage. There is an upper alcove with a low, horizontal back wall and darkly shaded face wall. A dead juniper occupies a place on the lip of this upper alcove. Maiden- hair fern and Eastwood monkey-flower grow on the back wall. The lower alcove is not so deeply cut. Its slightly indented and curved wet wall is occupied by cave primrose, mag- nificent in early springtime when in full flower. Alcove death camas grows along the base of the face wall and sometimes clings to the surface of the wall itself. Plants of Garber sedge and helleborine orchid hug the wall base at the back of the detrital slope. The dryish margin is clothed with little bluestem. Island in the Sky portion of Canyonlands National Park is margined by cliffs of Wingate Sandstone. Only rarely do alcoves form in the Wingate, possibly because the capstone Kay- enta is so impermeable. Exceptions to hang- ing garden formation in the Wingate occur at the head of Trail Canyon off the north side of the island, where a huge alcove is present in the Wingate and a small saline seep hints at a garden there in the past, and at Natural- ists Cove south of the first switchback of the Schafer trail. There behind a grove of Douglas-fir and Bocky Mountain maple is a small garden supporting Fendler barberry (Berberis fendleri Gray), cave primrose, and Jones reedgrass. The elevation is about 1,770 m. Fendler barberry is known also from a dryish garden below The Neck. The White Bim formation is encountered along the edge of the Green Biver west of 24 Great Basin Naturalist Vol. 49, No. 1 Fig. 16. Dead Tree Garden, Entrada Formation, north of the trail to Delicate Arch, Arches National Park, Grand Co., Utah. This alcove garden is controlled by a series of bedding planes in the Entrada Formation, exposed along a fault line. The upper alcove is very deep and densely shaded. Maidenhair fern clings to its flat roof. Island in the Sky, north of Anderson Bottoms. It is best displayed downriver from Anderson Bottoms, where the formation gradually rises above the river as a prominent marginal feature. The gardens occur along its strike, with alcoves developed to the underlying im- pervious Cutler. The gardens have not been explored but are known to support cave prim- rose. More work is indicated. A poorly devel- oped garden at Anderson Bottoms was tenta- tively explored as a water source by a private owner prior to establishment of the Canyon- lands as a national park. Hanging Garden Species Hanging garden species show diverse affinities as was outlined by Welsh and Toft (1981). Some are boreal plants, whose phenol- ogy still reflects their northern habitat regi- men. Others are southwestern in their rela- tionships, and still others have affinities with plants to the south. The following list is not meant to be exhaustive of all species that oc- cur in the gardens. Especially listed are the plants that regularly occupy wet walls in the alcove, terrace, and windowblind gardens. Adiantum capillus-veneris L. Maidenhair fern This plant grows in practically all hanging gardens. It occurs in the salt-encrusted gar- dens in the Kayenta Formation in St. George, in the gardens of Zion, and in practically all of the gardens in the Colorado drainage. Else- where the species is widespread. In North America it occurs from British Columbia east to South Dakota, Missouri, Florida, and Tex- as. It grows in much of Europe, often on cal- careous tufa. The species is also present in the subtropics of both hemispheres. It fits the requirements for a hanging garden species both in being disjunct and in being the most consistent of the hanging garden wall plants. Adiantum pedatum L. Northern maidenhair fern Northern maidenhair fern is disjunct in the hanging gardens of Zion Canyon mainly, but it occurs in some gardens in the head of the Escalante drainage of the Colorado Plateau. January 1989 Welsh: Utah s Hanging Gardens 25 Fig 17. View east from upper alcove of Dead Tree Garden, showing clinging back and face wall plants of maidenhair fern and Eastwood monkev-flower. Its distribution even in those areas is spotty. It grows usually beneath boulders where it is consistently more shaded than the habitats occupied by A. capillus-veneris. In Utah this species is not always a hanging garden plant. It grows in mesic sites in the Wasatch Moun- tains. The species occurs from Alaska to the Atlantic and south to California, Oklahoma, and Georgia. It is also known from Asia. Andropogon glomeratus (Walter) B.S.P. Bushy bluestem Plunge basins and detrital slopes of alcove gardens in the Glen Canyon vicinity support bushy bluestem at the northern limits of its distribution in Utah. The plants occupy a posi- tion in the gardens somewhat similar to that of saw-grass. The two plants also grow together along Furnace Creek in Death Valley, Califor- nia. Bushy bluestem is a plant of warm tem- perate affinities through much of the southern and coastal eastern U.S. It also grows in Mex- ico. It is typical of a group of prairies and plains species that occur marginally to the gardens. Sorghastrum nutans (L.) Nash in Small, Panicum virgatum L. , Andropogon gerardii Vit., and Schizachyrium scoparium (Michx.) Nash in Small are other examples of such species that fringe many of the hanging gardens. Aquilegia chrysantha Gray Golden colum- bine Golden columbine is restricted to the Zion Canyon gardens in Utah. The first flowers are very large and attractive, with spurs 4-7 cm long. It hybridizes with A. formosa (see be- low) in the gardens, and the hybrids are con- spicuous and recognizable by their telltale reddish sepals and spurs. The species is known from Arizona, New Mexico, Colorado, and Mexico. Aquilegia formosa Fisch. in DC. Western columbine Zion Canyon materials of A. formosa are of two types. Plants of only moderately mesic crevices at 1,678 to 2,288 m elevation are glandular overall and belong to the var. fosteri Welsh. Seeds cascade into the canyons where some plants of that type occur in the hanging 26 Great Basin Naturalist Vol. 49, No. 1 Fig. 18. Reverse view of Figure 16, showing the cling- ing plants on the bedding plane, whose erosion has re- sulted in the dark upper alcove. Etiolated plants of Rhus aromatica Ait. sprawl along the edge of the alcove. gardens. There they coexist with formosa plants that are glandular only in the inflores- cence and with A. chrysantha. The formosa type of plants that lack glandular foliage is thought to represent hybrids between var. fosteri and A. chrysantha. Their lack of glan- dularity on the foliage, flowers that average larger, and more mesic habitat requirements (than var. fosteri) suggest such an origin. Plants of the formosa type are known from gardens only in Zion. Elsewhere in Utah A. formosa grows in seeps and along streams. The species is known from Alaska and Yukon south to Baja California, Nevada, and Mon- tana. Aquilegia micrantha Eastw. Alcove colum- bine Alcove columbine is an endemic of hanging gardens of the canyons of the Colorado. It occurs in most of the gardens of that area. The small, white to cream-colored (less commonly pinkish to pale blue) flowers and overall glan- ularity are diagnostic for the species. Alcove columbine is not known from the Virgin River drainage. The species is known only from the Colorado Plateau of Arizona, Colorado, and Utah. Aralia racemosa L. American spikenard This plant forms clumps that reach a height of more than 2 m and have leaves to almost 1 m wide. The umbels of small white flowers are conspicuous in early summer and are fol- lowed by rounded clusters of fruits. The plants occur primarily on the floors of grottos, where they are shaded almost constantly. Oc- casional plants do occur on the garden walls and in the riparian communities, especially in the Narrows. The plant is known in Utah only from Zion Canyon. The phase of the species occurring in Utah is ssp. bicrenata (Woot. & Standi.) Welsh 6c Atwood. The species occurs from southeastern Canada and northeastern U.S. southward to Arizona and New Mexico. Calamagrostis scopulorum Jones Jones reed- grass This species was named and described on the basis of plant specimens taken from Zion Canyon by Marcus E. Jones in 1894. The species is elevationally disjunct in the hanging gardens of Utah. Its principal distribution oth- erwise is in boreal and alpine sites. The spe- cies occurs from Montana to Arizona and New Mexico. It is the main grass in the gardens in Zion Canyon and is often a main component of gardens along the canyons of the Colorado in eastern Utah. Carex aurea Nutt. Garber sedge; Golden sedge The common sedge of hanging gardens in both Zion and Glen canyons vicinities differs in mainly intangible features from that typical of the species as it occurs in montane and alpine sites. It has been referred to as C. garberi Fern. It is usually a larger plant and has greenish rather than golden fruits of its alpine counterpart. However, it seems best to regard it as little more than an ecotype of the more broadly ranging C. aurea. The species is present in most of the Colorado gardens and some in Zion. It is widely distributed in North America and has been considered by some authors as a portion of the C. bicolor All. circumboreal complex. Carex curatorum Stacey Canyonlands sedge This primocarex is easily recognized by its solitary, unisexual spikes borne on separate January 1989 Welsh: Utah's Hanging Gardens 27 plants. The species is elevationally disjunct in the gardens of the Colorado canyons from the confluence of the San Juan with Glen Canyon to the head of the Escalante drainage. Occa- sional specimens have been taken at high ele- vations as far north as central Utah. The spe- cies has recently been detected in Zion Canyon. Otherwise the species is known only from Arizona. Its affinities are with the boreal C. scirpoidea Michx., with which it has been combined at varietal rank. Cercis occidentals Torr. ex Gray Western redbud This beautiful plant produces masses of red- pink flowers in early springtime. It is a regu- lar occupant of hanging gardens in Glen Canyon, where it also occurs as a crevice plant not associated with hanging gardens. The spe- cies is present as far north as the mouth of Dark Canyon, a tributary of Cataract Canyon, where it grows on sandstone outcrops on the shaded side of the canyon. In the Virgin River drainage the species does not occur in hang- ing gardens. Rather, it occupies moist north- facing sites in sandstone along the Santa Clara River and shaded, dryish alcoves in limestone in the Beaver Dam Mountains. In hanging gardens the plants sometimes grow attached to the face wall, but more often they grow on the moist ledges at the base of the detrital slope, where they are typically associated with Celtis reticulata Torr. , Quercus gambelii Nutt, and Q. x eastwoodiae Rydb. Affinities of this redbud lie with Cercis canadensis L. , a species of eastern American summer decidu- ous forest affinities, with which it is some- times combined at varietal rank. Cirsium rydbergii Petrak Rydberg thistle The Rydberg thistle is an endemic of hang- ing gardens in the canyons of the Colorado. Rarely it occurs outside the gardens along drainages. It is a perennial, with huge basal rosettes sometimes more than 1 m across. The flower heads are small, however. It does not occur in Virgin River drainage gardens and has no counterpart. Cirsium virginensis Welsh grows in the gardens in St. George but is principally a riparian species. The same is true for C. calcareum (Jones) Petrak, which grows in some of the hanging gardens in the Colorado Plateau proper. Cladium californicum (Wats.) O'Neill in Tid- estr. Saw-grass This plant reached its northern distribution in hanging gardens along Glen Canyon. It was collected historically at the mouth of Kane Creek, now under several hundred feet of water held back by Glen Canyon Dam. Subse- quent collections were taken at Step Garden in Driftwood Canyon, Cladium Garden in Hidden Passage Canyon, and Wilson Creek along the San Juan Arm. The high water of 1983 and 1984 lapped at the margin of the great clump of this species at Step Garden and apparently drowned a portion of the plants at the mouth of Wilson Creek. Cladium Garden has long since been inundated. A second small patch of saw-grass is present at the foot of one of the vertical walls in Hidden Passage Canyon, however. The plant is precariously situated in the flora of the region. Elsewhere it is known from along Furnace Creek in Death Valley, California, and when consid- ered as a portion of the broadly distributed C. mariscus, its distribution is across much of the southern U.S. and southward. Dodecatheon pulchellum (Raf.) Merrill Pretty shooting-star The Zion Canyon phase [var. zionense (Eastw.) Welsh] of the shooting-star is also known from hanging gardens in Little Valley tributary of Last Chance Creek, north of Lake Powell, in the Escalante drainage, and along Lake Powell as far north as the confluence of the Escalante. Nowhere does it grow so pro- fusely as in the shaded gardens of Zion Canyon. The species occurs in much of Utah and is widely distributed in western North America and disjunctly in the eastern U.S. Epipactis gigantea Dougl. ex Hook. Helle- borine This plant is mainly palustrine in its distri- bution, occurring around seeps and springs and along streams in much of western North America from the Canadian provinces south- ward. It is a common component of the alcove hanging gardens in southeastern Utah. The purplish to greenish purple or purple-brown flowers are displayed on plants that are usual at the juncture of detrital slopes and adjacent face walls in many gardens. It grows there with Garber sedge, bundle panic, and other garden species. The plant is not especially noted as a garden species in Zion Canyon, however. It is in the canyon but occupies marshy sites in The Narrows, where garden 28 Great Basin Naturalist Vol. 49, No. 1 development is not especially prominent. A single clump of the plant is present in a dryish garden associated with maidenhair fern along the trail to Refrigerator Canyon in Zion Canyon. Erigeron kachensis Welsh & Moore Kachina daisy This pretty dwarf daisy produces stolons that allow it to grow on wet walls and moist sand in hanging gardens in Natural Bridges National Monument and Dark Canyon, Utah, and along the Dolores River in Colorado. The plants also occur as mere mesophytes in other plant communities besides hanging gardens at higher elevations in the Abajo Mountains. It is a Colorado Plateau endemic. Erigeron sionis Cronq. Zion daisy The Zion daisy is a stoloniferous meso- phyte. The stolons tend to bind the plant in colonies on moist sandstone. The nature of the stone in Zion Canyon allows water to perco- late in shallow stone lenses, only to come to the surface along minor defiles. Some of the moist lenses are hand-sized and still support this daisy. The plants are provided with water for some time following snow melt in spring- time or following rains at any season. Hanging gardens also serve as habitat for this plant, which is especially common along minor ter- race margins and on rock faces in the grotto floors. Occasionally the plants grow along drainages below the gardens, and rarely the plants are found growing well on apparently dry sand. The plant is endemic to Zion Canyon. Erigeron zothecinus Welsh Alcove daisy Alcove daisy is known only from a suite of hanging gardens (the North Escalante Gar- dens) on the north side of Glen Canyon imme- diately east of the confluence of the Escalante Canyon. It is confined to the wet walls and foot slopes of small gardens near the base of the Navajo Sandstone or top of the Kayenta Formation. The eight main alcoves in the Navajo to the north apparently do not support this plant. Habenaria zothecina Higgins & Welsh Alcove bog-orchid Long thought to be a phase off/, sparsiflora Wats., this species with long-spurred flowers was named from the hanging gardens of the first meander bend east of the Colorado bridge north of Moab, Grand County, Utah. The species occurs in hanging gardens from Arch Canyon, west of Comb Reef, north to Arches National Park. It is endemic to the hanging gardens. Lobelia cardinalis L. Scarlet lobelia This third of a triumvirate of plants with scarlet or brilliant red flowers is less likely to occur within the garden flora than adjacent to it. The species is rather common on the wet walls of Zion Canyon, however. The other two, Mimulus cardinalis Dougl. ex Benth. and Mimulus eastwoodiae Rydb., are charac- teristic of the gardens proper, even though they also occur on the downslope portions of the garden and less commonly along the drainages below them also. Those downslope and down drainage localities are the common sites occupied by scarlet lobelia in the Colo- rado canyons gardens especially. The species is present in Zion Canyon and along Glen Canyon as far north as the mouth of Ticaboo Canyon. It is widely distributed in the eastern United States, Mexico, and Central America. Mimulus cardinalis Dougl. ex Benth. Cardi- nal monkey-flower This plant with brightly colored, large flow- ers reaches the northern limits of its distribu- tion in the hanging gardens of Zion Canyon. Its entire distribution otherwise is in New Mexico, Arizona, Nevada, California, Ore- gon, and Mexico (including Baja California). Utah materials have sometimes been treated as a separate species, M. verbenaceus Kear- ney & Peebles, but the ranges, when treated separately, overlap and the characters used as diagnostic fail individually and in tandem. The plants are common in the gardens of Zion where they occur both as wall and floor plants in the grotto and windowblind gardens. Mimulus eastwoodiae Rydb. Eastwood monkey- flower Back and face walls are the common sites of placement for this species in the alcove gardens of the canyons of the Colorado. The species is in most of the gardens where it shares the walls with cave primrose. It is a smaller plant in all respects than is the cardi- nal monkey-flower. Flowers are borne in late summer and autumn in this species. Its coun- terpart in Zion Canyon is mainly a spring- flowering plant, but flowering continues through the summer. The species is endemic to the hanging gardens. January 1989 Welsh: Utah's Hanging Gardens 29 Panicum acuminatum Swartz Bundle panic Gardens of the Colorado canyons typically support this plant, growing intermixed with Garber sedge on the detrital slopes of alcoves. It is less common in Zion Canyon. This is a Dicanthelium panicum, which has broad basal leaves. Flowers of two types are pro- duced, open pollinated flowers at the summit of the plant and flowers that remain in the enclosing sheath that are self-pollinated. The species has passed under several names. It is a widespread plant in the U.S., portions of Canada, and Mexico. Perityle specuicola Welsh & Neese Alcove rock-daisy Type locality for this species is the alcove immediately adjacent to the highway on the first meander bend east of the Colorado River bridge north of Moab. The plant grows there, hanging from crevices in the back wall of the alcove. It simulates Stephanomeria tenuifolia (Torr.) Hall with which it grows in both habit and position. The plants also grow in the hang- ing garden on the west side of the same mean- der bend, but there on the dry edge of the garden the plants stand erect. A second popu- lation is known from a small garden in Cata- ract Canyon north of its confluence with Dark Canyon. The plant is not otherwise known. Petrophytum caespitosum (Nutt.) Rydb. Rock spiraea Rock spiraea grows in hanging gardens both in Zion and along the canyons of the Colorado. It does well in some of the alcove gardens where the substrate is wet, even growing as- sociated with the algal mat on the walls, but is common in gardens that are dry. In Zion Canyon it is mainly a plant of garden margins. Often the plants occur as mats conforming to the contour of the surface, but sometimes the plant swings away from the wall and hangs pendulously from the overhanging back of al- cove gardens. Spikes of creamy flowers are borne in late summer and autumn. The spe- cies grows elsewhere on limestone and ig- neous outcrops, typically at middle eleva- tions, from Washington to South Dakota and south to California, Arizona, New Mexico, and Texas. Primula specuicola Rydb. Cave primrose; Easter flower The cave primrose was evidently first col- lected by Alice Eastwood in 1895 in the alcove gardens at Bluff, Utah. She mistakenly identi- fied the plants with an Old World member of the genus but recognized the unique nature of the gardens and the misplaced distributions. The species might be regarded as the botani- cal motif of Canyonlands hanging gardens, but that role is shared practically always with Eastwood monkey-flower and small-flowered columbine. The flowers vary in color from pink (the usual color) to lavender, maroon, and white. Rarely the inflorescences are su- perposed as in some cultivated species, and rarely too the flowers are double. The plant is a hanging garden endemic. Smilacina stellata (L.) Desf. Solomonseal Widespread in much of North America from the subarctic southward, this species is a common mesophyte. It grows in moist sites throughout Utah, not surprisingly in hanging gardens. The hanging gardens are elevation- ally placed below the general distribution of the species, and thus the species seems to be a somewhat regular occupant. Solomonseal oc- curs in gardens in Zion and Glen canyons, mainly on the foot slope or downward along the drainages. Solidago sparsiflora Gray Alcove goldenrod Gardens of the Canyonlands region of east- ern Utah are practically always partially clothed with alcove goldenrod. It is a plant of face walls and detrital slopes. Stream courses below hanging gardens and sometimes the gardens themselves also support specimens of S. canadensis L. Intermediates are known be- tween these species, and some garden plants are difficult to assign to one or the other of them. The plant is less often a component of Zion Canyon gardens. Nevada goldenrod, S. spectabilis (D.C. Eaton) Gray, grows in the hanging gardens in St. George. Elsewhere it is a palustrine or riparian species. Alcove goldenrod is a common species, especially at elevations far above that typical for hanging gardens, in the western U.S., from Wyoming and South Dakota south to Arizona and Ne- vada. Zigadenus vaginatus (Rydb. ) Macbr. Sheathed death camas This death camas occurs sporadically in the hanging gardens of the canyons of the Colo- rado. Pool Garden in Reflection Canyon, Double Gardens on the west side of Glen Canyon, Three Garden, and Death Camas 30 Great Basin Naturalist Vol. 49, No. 1 Garden in the first meander bend of the San Juan east of the confluence with Glen Canyon were essentially aligned and distant directly not more than 3-4 km. Sheathed death camas was in Pool Garden, Double Garden, and Death Camas Garden, but not in Three Gar- den. This hit-and-miss distribution was char- acteristic of the large alcove gardens along Glen and San Juan arms. The plants occur in hanging gardens north to Arches and in seeps evidently north to Dinosaur National Monu- ment. Relationships of this species lie to the south with Z. volcanicus Benth. of Central America. The type ofAnticlea vaginata Rydb. (Rydberg 1912) was taken at Natural Bridges National Monument, probably in the seep be- neath Owachomo Bridge. Literature Cited Clark, R. L. 1972. A study of the algal flora of selected hanging gardens of Glen Canyon, Utah. Unpub- lished manuscript. Brigham Young University Herbarium. 90 pp. + IX plates. Eastwood, A. 1896. Report of a collection of plants from San Juan County, in southeastern Utah. Proc. California Acad. Sci. 116:271-329. Eubank, M 1979. Utah weather. Horizon Publishers, Salt Lake City, Utah. 284 pp. Gregory, H. E. 1938. The San Juan country: a geographic and geologic reconnaissance of southeastern LItah. U.S. Geol. Surv. Prof. Paper 188. 123pp. Harrison, BF, S.L.Welsh, and G. Moore 1964. Plants of Arches National Monument. Brigham Young Univ. Sci. Bull., Biol. Ser. 5(1): 1-23. Hintze, L F. 1972. Geologic history of Utah. Brigham Young Univ. Geology Studies20(3): 1-181. JOHANSEN, J R., S R RUSHFORTH, R ORBENDORFER, N. Fungladda, and J. A. Grimes. 1983. The algal flora of selected wet walls in Zion National Park, Utah, USA. Nova Hedwigia38: 765-808. Malanson, G. P 1980. Habitat and plant distributions in hanging gardens of the Narrows, Zion National Park, Utah. Great Basin Nat. 40: 178-182. 1982. The assembly of hanging gardens: effects of age, area, and location. Amer. Nat. 119: 145-150. Malanson, G P., and J Kay 1980. Flood frequency and the assemblage of dispersal types in hanging gar- dens of the Narrows, Zion National Park, Utah. Great Basin Nat. 40: 365-371. Miller. D E 1966. Hole-in-the-Rock. University of Utah Press, Salt Lake City. 229 pp. Petrak, F. 1917. Die nordamerikanischen Arten der Gattung Cirsium. Beih. Bot. Centralbl. 35(2): 223-567. Powell, J W 1895. Canyons of the Colorado. Chau- tauqua-Century Press, Meadville, Pennsylvania. 400 pp. Rushforth, S. R., and G. S. Merkley. 1988. Comprehen- sive list by habitat of the algae of Utah, USA. Great Basin Nat. 48: 154-179. Rushforth, S. R , L. L. St.Clair, T. A. Leslie, K H Thorne, and D C Anderson 1976. The algal flora of two hanging gardens in southeastern Utah. Nova Hedwigia 27: 231-323. Rydberg, P. A 1912. Studies on the Rocky Mountain flora— XXVI. Bull. Torrey Bot. Club 39: 99-111. 1917. Flora of the Rocky Mountains. Published by the author, New York. 1, 143 pp. Toft, C. A 1972. Keys to the lizards (Sauria), frogs, and toads of the Kaiparowits Basin of southern Utah and northern Arizona. In-house document. Brigham Young University Herbarium. 11 pp. Welsh, S. L , and G. Moore. 1968. Plants of Natural Bridges National Monument. Proc. Utah Acad. Sci., Arts, and Letters 45: 220-248. Welsh, S L , and C A Toft. 1972. Site thirty-three. In: Environmental studies of the Navajo and Kai- parowits generating stations. Annual report 1972. Center for Health and Environmental Studies, Brigham Young University. 183 pp. Welsh, S L , and C. A. Toft 1981. Biotic communities of hanging gardens in southeastern Utah. Nation- al Geographic Society Research Reports 13: 663-681. Welsh, S L , and B W, Wood. 1975. Structure and pro- ductivity of a selected hanging garden in Glen Canyon, San Juan County, Utah. Unpublished manuscript. Brigham Young University Herbar- ium. 23 pp. Welsh, S. L , B W. Wood, and J A Raines 1975. Three Garden animal trapping — 1973-1974. Unpub- lished manuscript. Brigham Young University Herbarium. 46 pp. CHANGES IN MULE DEER SIZE IN UTAH1 Dennis D. Austin2, Robert A. Riggs2, Philip J. Urness2, David L. Turner3, and John F. Kimball4 Abstract. — Trends in age-speeifie, eviscerated carcass weights were determined for hunter-harvested yearling and two-year-old buck mule deer. Carcass weights declined over an 11-year period from two areas of similar management, but with independently collected data sets. Carcass weights also declined between the opening and second weekends of the hunt. Management implications are discussed. Mule deer (Odocoileus hemionus) bucks, especially large, mature animals, have sport hunting, economic, and intrinsic values (Wennergren 1968). However, perceptions of a quality hunt or buck vary considerably among hunters as evidenced by the various types of hunts established by state wildlife agencies in response to hunter input. As hunt- ing intensity has augmented and increased the impacts on wildlife populations, and as human population growth has usurped range areas traditionally used by wildlife, game managers have been increasingly pressed to maintain quality programs. Consequently, in Utah either-sex hunting regulations during the 1960s were replaced by buck-only restric- tions in the early 1970s to compensate for increasing hunter numbers using a limited resource. Antler restriction and limited-entry hunts have become increasingly common in the 1980s, with motivation for more restric- tive regulations coming from hunters and bi- ologists interested in quality hunting in terms of maintaining high numbers of mature, har- vestable bucks, and restricting hunter num- bers. This paper examines long-term trends of age-specific changes in the size of hunter- harvested mule deer in Utah. Methods Study areas. — The Daniels Canyon check- ing station located in north central Utah stopped about 17,000 hunters per year (1975-85), with about 70% of the hunters returning from the Current Creek and Avin- taquin deer units. The Blacksmith Fork sta- tion in northern Utah checked about 2,700 hunters (1973-83) from a portion of the Cache deer unit, mostly within the Blacksmith Fork drainage. Between 1973 and 1985 both areas had 11-day buck-only hunts, except in 1973 when the area served by the Daniels Canyon station held a three-day either-sex hunt fol- lowed by eight days of buck-only hunting. All deer hunts began the Saturday closest to 20 October. Data COLLECTION. — Checking station data were collected and reviewed for the two ar- eas. Eviscerated carcass weight, age, and antler tine data were collected at checking stations during most years. Data were col- lected during the first and/or second week- ends of the hunt. Deer were weighed to the nearest .5 kg and field aged as IV3, 2V3, or 373+ years by tooth eruption and wear methodology (Robinette et al. 1957). All antler tines exceeding 2.5 cm (Robinette et al. 1977), but excluding brow tines, were counted on intact antler pairs only at Daniels Canyon. Data analyses. — The factors of year, age class, and weekend were initially used for ana- lyzing the carcass weight data (1975-85) from Daniels Canyon. However, because there were many missing three-way cells, and there appeared to be a significant difference be- tween weekends, the data were divided into four sets for analysis: (1) first week, age 1, (2) first week, age 2, (3) second week, age 1, and (4) second week, age 2. Least squares pro- cedures and a two-factor linear model were 'This paper is, in part, a contribution of Utah State Division of Wildlife Resources, Federal Aid Project W-105-R. department of Range Science. Utah State University, Logan, Utah 84322. department of Mathematics and Statistics, Utah State University, Logan, Utah 84322. 4Utah Division of Wildlife Resources, 515 East 5300 South. Ogden, Utah 84405. 31 32 Great Basin Naturalist Vol. 49, No. 1 Table 1. Deer carcass weights and tine counts from Daniels Canyon and Blacksmith Fork checking stations, Utah (sample sizes in parentheses). First weekend Second weekend IV3 Age (in years) Weight1 2V3 Weight lVa 2Va 2V.3 Year Weight Weight Tines"3 Tines4 Daniels Canvon 67-68 44.9 (19) 57.2(14) — — — — 75 45.1 (85) 58.3(29) 43.7 (42) 56.1(10) 16.5(127) 17.9(39) 76 44.0(140) 58.5(38) 43.1 (95) 55.5 (24) — — 77 43.8 (38) 59.5 (9) — — — — 78 42.5 (46) 52.7 (7) — — — — 79 — 41.9 (50) 55.7(27) — — 80 — 41.0 (52) 50.6 (2) 44.2 (52) 50.0 (2) 81 — 40.6(104) 51.8 (5) 26.0(104) 20.0 (5) 82 — 42.1 (82) 53.4 (7) 18.3 (82) 57.1 (7) 85 41.1 (59) 52.3 (6) 40.6 (53) 51.9(10) 33.0(273) 27.2(33) 87 41.1 (51) 54. 1 (53) — — 26.8(1627) 30.2(69) Blacksmith Fork 66 47.2 (33) 63.9(12) 73 50.1 (22) 64.8(11) 79 44.2 (44) 58.9(14) 80 43.0 (45) 60.6 (6) 81 41.9 (40) No Data 83 41.6 (43) 56.7 (6) 86 44.4 (52) No Data 87 42.8(102) 54.5(32) Eviscerated carcass weight (kg). A tine is defined as a minimum length of 2.5 cm excluding brow tines. Percentage of yearling deer with two tines on both antlers combined (spikes). Percentage of 2'/3-year-old bucks with four or fewer tines on both antlers combined. used. Deer aged 37.3+ years were not consid- ered in this manuscript because of the poten- tial variation of actual age within the class. The data from Blacksmith Fork (1973-83) were consistently collected on the first weekend. These data were independently analyzed us- ing the same statistical test as a check to the conclusions of the Daniels Canyon analysis. Antler data from Daniels Canyon were ana- lyzed using chi-square comparisons of tine counts across years. Because the weekend ef- fect was not significant, the data were pooled. Because of the important implications de- rived from the data sets, additional, supple- mentary data were obtained. Two sets of carcass weight data, collected before 1970, but under either-sex hunting regulations, were compared using the 95% confidence in- tervals to the endpoints of the regression equations. Three sets of recent data were col- lected after 1985 for data trend verification and were similarly compared. Besults Carcass weights from Daniels Canyon de- creased (P = .02) between the first and second weekends. Yearling bucks decreased a mean 2.1% in weight, and two-year-old bucks de- clined 3.3% between weekends. Over an 11-year period, trends in carcass weight have been negative at Daniels Canyon (Fig. 1). Between 1975 and 1985 the weight of yearling bucks decreased 8.9% (P = .03) and 7. 1% (P = .02) for the first and second week- ends, respectively. Weight in two-year-old bucks decreased 10.1% (P = .12) and 7.5% (P = .03) for weekends one and two, respec- tively. Besults of the Blacksmith Fork analysis during a comparable period (1973-83) showed a similar decrease in carcass weight. Weight in yearlings declined 17.0% (P = .005) and 12.5% (P = .04) in two-year-old bucks. Tine counts of yearling bucks checked at Daniels Canyon (Table 1) were greater in 1975 than during 1980-85 (P == .01). In 1975, 16.5% of the harvested yearling bucks were spikes, whereas the mean percentage of spikes harvested 1980-85 was 30.4%, with three of four years being significantly different January 1989 Austin etal.: Utah Mule Deer 33 U<0 60n CT> J* £ 50- 40 y=59.38-0.68x r2=0.46 p<0.12 y=44.87-0.37x r2=0.84 p<0.02 1974 1976 1978 1980 1982 1984 1986 Year Age 1 Age 2 'a> (1b) 60-i 50- 40 1974 y=56.74-0.43x r2=0.58 p<0.03 y=43.55-0.30x r2=0.60 p<0.03 I976 1978 1980 Year 1982 1984 1986 70 -i S 60- > 50 -I 40 (1c) y=65.44-0.81x r2=0.88 p<0.04 y=50.39-0.88x r2=0.94 p<0.005 1972 1974 1976 1978 1980 1982 1984 Year Fig. 1. Decline of mean eviscerated weight of hunter-harvested buck deer aged 1 lh and 21k years during the first (la) and second (lb) weekends of the regular Utah deer hunt from Daniels Canyon checking station and during the first weekend from Blacksmith Fork (lc). 34 Great Basin Naturalist Vol. 49, No. 1 from 1975 (P == .05). However, significant differences were also found among years from 1980 to 1985 (P = .01). Tine counts of two- year-old bucks at Daniels Canyon in 1975 did not differ from combined years 1980-85 (P = .21) or among years 1980-85 (P = .26). How- ever, 17.9% of bucks in 1975 had four or fewer antler tines, whereas in 1980-85, the mean was 31.9%. Data collected at Daniels Canyon 1967-68 were combined because of small sample sizes (Table 1). The mean weights for yearling and two-year-old bucks (1967-68) were not differ- ent from the 1975 predicted weights. Yearling weights were different from the 1985 value, but two-year-old buck weights were not dif- ferent. Both yearling and two-year-old buck weights from Daniels Canyon (1987) were not different from the 1985 predicted weights, but were different from those of 1975. Weight data from Blacksmith Fork (1966) were not different from the 1973 weights, but differed from the 1983 weights. The 1986 yearling weights differed from both the 1973 and 1983 predicted weights. Weights for both yearlings and two-year-old bucks in 1987 did not differ from those of 1983, but did differ from the 1973 weights. Discussion The decrease in carcass weight between the first and second weekends of the hunt indi- cates the need for consistent timing of data collection. The data also suggest that physical condition indices (e.g., Austin 1984) collected during the first weekend of the hunt would not be representative of the herd at the hunt's end. A probable cause of the weight loss between weekends is hunter harassment, although other factors, including rutting activity and hunter selection, may also be important. Nonetheless, where overwinter survival is questionable, this degree of weight loss may be important, particularly its impacts on does and fawns. Beduction in carcass weight and the cor- responding, although weaker, reduction in number of antler tines suggest that age- specific deer size has declined in Utah. In our study, and others, weight appears to be the more sensitive index (Kie et al. 1983, Williams and Harmel 1984). Besults from Daniels Canyon and Blacksmith Fork show a gradual decline in buck size from the same deer popu- lations over more than a decade. Both areas were under similar management and showed about the same loss in size of deer. Data collected since 1985 at both Daniels Canyon and Blacksmith Fork supported the findings of reduction in body size, in that mean weights and numbers of antler tines remained low. Data collected previous to 1970 suggest age-specific deer size probably did not decrease before the early 1970s. A probable consequence of size decline in younger age classes is a parallel reduction in age-specific size of mature bucks. Williams and Harmel (1984) reported changes in num- ber of antler points and live-body weights of 60 pen-reared white-tailed buck deer, fed 16% protein diet ad libitum, during ages IV2, 2V2, and 3V2 years. They found that the num- ber of antler points and the weights at ages 2V2 and 3V2 years were directly correlated with the number of antler points and weights at- tained by the same deer at younger ages. The corollary is that weight and antler size of older-aged bucks are also related to yearling characteristics. Consequently, the probabil- ity of older bucks being large trophy has likely also declined in Utah. Begulations of Utah's deer hunt changed from either-sex to buck-only hunts in the 1970s, with 1973-74 being transition years. Because data collected prior to 1970 sug- gested no age-specific changes in size, atten- tion should be given the potential effects of changing regulations and their effect on buck populations and quality of animals. However, partitioning the important factors potentially responsible for the observed decline was not possible. Nonetheless, several factors should be considered. First, phenotypic changes in deer populations due to hunter selectivity for larger bucks may have occurred. Scribner et al. (1984) demonstrated through modeling that selective removal of spike white-tailed bucks will gradually lower the incidence of spikes in the buck population; conversely, se- lective removal of nonspike yearlings would increase the incidence of spikes. Second, with increasingly wide buck-to-doe ratios and the lowering of the mean age of the buck popula- tion, both of which result from intensive buck- only hunting, a delay in the mean breeding date causing a similar delay in the fawning date may have occurred. In support, Beimers January 1989 Austin et al.: Utah Mule Deer 35 (1983) showed a significant relationship be- tween a delay in the date of calving and re- duced dressed weight of females 2+ years in wild reindeer (Rangifer). Third, a density-de- pendent response to buck-only hunts may have occurred. Buck-only regulations may have allowed population density of females and fawns to increase and, consequently, nutrition limited phenotypic expression of genetic potential. In support, the modest rebound in weight from Blacksmith Fork in 1986 followed a marked population decline due to previously harsh winters (1983-85). McCullough (1979) discussed population growth within finite available resources and demonstrated a decrease in recruitment as carrying capacity was approached. A similar decrease apparently occurs in animal size, as observed for yeld hinds (Clutton-Brock and Albon 1983). Kie et al. (1983) reported re- duced body weights and number of antler tines with increased density of white-tailed deer. Finally, climate has been shown to af- fect fluctuations in deer size on a yearly basis (Bobinette et al. 1977), and, consequently, long-term weather trends might also be involved. Additional research is needed to identify the specific factors involved, as well as management alternatives to address the problem. In summary, eviscerated carcass weight and number of antler tines for yearling and two-year-old buck deer were shown to de- crease from 1973 to 1985 in two areas of Utah under buck-only hunting regulations. Al- though the reduction in size corresponded with changes to buck-only hunting, the effects of individual factors could not be partitioned. Literature Cited Austin. D D 1984. Fat depth at the xiphoid process — a rapid index to deer condition. Great Basin Nat. 44: 178-181. Clutton-Brock, T H . and S D Albon 1983. Climatic variation and body weight of red deer. J. Wildl. Manage. 47: 1197-1201. Coles, F. H., and J. C Pedersen 1970. Big game range inventory. Utah Div. Wildl. Res. Publ. No. 70-1. Jense. G. K. 1985. Utah big game annual report. Utah Div. Wildl. Res. Publ. No. 85-1. Kie, J G ., M White, and D L Drawe. 1983. Condition parameters of white-tailed deer in Texas. J. Wildl. Manage. 47: 583-594. McCullough, D K 1979. The George Reserve deer herd. University of Michigan Press, Ann Arbor. Olson, B C , and J R. Logan. 1974. Big game range inventory. Utah Div. Wildl. Res. Publ. No. 74-11. Reimers, E 1983. Growth rate and body size differences in Rangifer, a studv of causes and effects. Rangifer 3:3-15. Robinette. W L . N. V. Hancock, and D A. Jones. 1977. The Oak Creek mule deer herd in Utah. Utah Div. Wildl. Res. Publ. No. 77-15. Robinette, W. L., D A Jones, G Rogers, and J. S. Gash- wiler. 1957. Notes on tooth development and wear for Rocky Mountain mule deer. J. Wildl. Manage. 21: 134-153. Scribner. K T . M H. Smith, and P. E Johns 1984. Age, condition and genetic effects on incidence of spike bucks. Proc. Ann. Conf. Southeast Assoc. Fish and Wildl. Agencies 38: 23-32. Wennergren, E B 1968. What is the value of Utah's deer hunting resource? Utah Sci. 29: 55-59. Williams, J D . and D F Harmel. 1984. Selection for antler points and body weight in white-tailed deer. Proc. Ann. Southeast Assoc. Fish and Wildl. Agencies 38: 43-50. WHITE-TAILED PRAIRIE DOG (CYNOMYS LEUCURUS MERRIAM) DIGGINGS IN WESTERN HARVESTER ANT, POGONOMYRMEX OCCIDENTAEIS (CRESSON), MOUNDS William H. Clark1 and Cynthia J. Clark1 Abstract. — We report observations of the white-tailed prairie dog, Cynomys leucurus Merriam, digging and burrowing into mounds of the western harvester ant, Pogonomyrmex occidentalis (Cresson), in Emery County, Utah. On 16 July 1987 we observed evidence of white-tailed prairie dog (Cynomys leucurus Merriam) digging into the mounds of Pogono- myrmex occidentalis (Cresson) near Welling- ton, Emery County, Utah. We observed two white-tailed prairie dogs sitting on the edge of a P. occidentalis mound at 1630 hr. Investiga- tion of the ant nest revealed an active prairie dog entrance at the base of the mound and 12 large exploratory holes dug into the mound (Fig. 1). The entrance (11 cm diameter) was actively used, with numerous tracks, scats, and bits of vegetation adjacent to it. The bur- row angled downward to the southeast and went directly under the ant mound and pre- sumably through the center of the ant nest. The exploratory holes dug into the ant nest averaged 10 cm diameter and 15 cm deep. Of 20 ant mounds surveyed in the immediate area, 55% had the large exploratory holes; *«v **'« I Fig. 1. Mound of Pogonomyrmex occidentalis showing prairie dog burrow entrance and exploratory diggings. Prairie dog droppings can be seen in the lower portion of the photo. Scale: the field book is 11 x 19 cm. 'Museum of Natural History, College of Idaho, Caldwell, Idaho 83605. 36 January 1989 Clark, Clark: White-Tailed Prairie Dog 37 however, holes were not found in the adjacent ant clearings or nearby vegetated areas. Thus, the mounds appeared to be selected as ex- ploratory digging sites. Effects of the digging on the ants are not known. Clark and Comanor (1973, Occ. Pap. Biol. Soc. NV 34: 1-6) reported heteromyid rodent digging activity into mounds of P. occidentalis in Nevada and Utah apparently to obtain seeds stored by the ants. Allred (1982, Great Basin Nat. 42: 415-511) made many collec- tions from P. occidentalis nests in Utah but reported only one mound with two rodent burrows. Little else has been reported in the literature concerning the use of these ant mounds by small mammals. Acknowledgments Mary and Karen Clark assisted with the field observations. Eric Yensen reviewed the manuscript and offered valuable suggestions. Voucher specimens of the ants (WHC #8187) were deposited in the Orma J. Smith Museum of Natural Historv, College of Idaho, Caldwell (CIDA). AMPHIBIANS OF WESTERN CHIHUAHUA Wilmer W. Tanner ABSTRACT. — This third report on the herpetofauna of Chihuahua deals exclusively with amphibians. The first plethodonid salamander is reported, the species Ambystoma rosaceum is discussed in greater detail than before, and two subspecies are recognized. Spea is elevated from subgeneric to generic rank, and S. stagnalis Cope is removed from synonymy and is recognized as a subspecies of hammondii. The species listed include the following: 2 salamanders and 19 anurans (1 Scaphiopus, 2 Spea, 9 Bufo, 1 Elentherodacttjlus, 2 Hyla, 3 Rana, and 1 Microhyla). Reference is made to various habitats that are associated with elevations arising from lower deserts and extending into the western mountains. The role played by the dry and wet annual cycles is also noted. This is the third report on the herpetology of western Chihuahua (Tanner 1985 [1986], 1987). It deals only with the amphibians. As indicated in the preceding reports, the state of Chihuahua is a large area and includes deserts on the east, steppe foothills and valleys in the central part (north to south), and the north end of the Sierra Madre Occidental in the west. This diverse geographical terrain provides numerous and multiform types of habitat. One wishing to gain a general understand- ing of the terrain of southern and northwest- ern Chihuahua would benefit by reading the account by Goldman (1951), which deals with the explorations of Nelson and Goldman dur- ing the late 1890s. Although their activities did not include exploration of the entire state, they did include considerable travel in north- ern Chihuahua extending into the deserts, mountains, and valleys and to the northeast and west of Casas Grandes. The trip by Gold- man from Parral to Batopilas is replete with descriptions of this extensive area. The cen- tral area west from Ciudad Chihuahua to Madera and southwest to the extensive areas north and west of Creel were not explored. Although these exploratory field trips pro- vided valuable information concerning the general nature of the terrain, faunas, and floras, only a relatively few herpetological specimens were collected and these were de- posited in the U.S. National Museum (Smith- sonian Institution). Thus, it has remained for others to explore and report the rich herpeto- logical faunas of this state. References to other important studies are cited in the first of this series (Tanner 1985 [1986]). One cannot tra- verse this area without becoming enamored with its rugged diversity and beauty. There are few areas where one can stand in a conifer- ous forest on the rim of a mighty canyon and observe an entirely different biome approxi- mately a mile below, where wild figs and man- goes grow along a river, which is in turn sur- rounded by an invading thorn forest from the bench lands of Sinaloa (Fig. 1). The climate in Chihuahua can be character- ized by its cool to cold winters (November into March), dry to very dry springs (March through June), and moderately to heavily rainy summers (July into September), al- though the amounts will vary greatly from location to location and from year to year. The fall months (mid-September and October) are delightful, with warm days and cool nights. As indicated below, the climate and terrain com- bine to provide numerous habitats and a mod- erately rich amphibian fauna. During the dry season (March through June) few amphibians are seen and those only along permanent streams and springs. Much of central Chihuahua is at an elevation be- tween 5,000 and 6,000 feet; and, therefore, it is cool to cold until May with occasional frost during April. It is not until the summer rains come, usually from July to the first part of September, that an abundance of amphibians is seen. A trip from western Durango to Ciu- dad Chihuahua during a rainstorm corre- sponded with the emergence of large num- bers of frogs and toads from areas that had 'Life Science Museum, Brigham Young University, Provo, Utah 84602. 38 January 1989 Tanner: Chihuahua Amphibians 39 Fig. 1. View from the west rim of the Barranca de Urique east of Cerocahui, Chihuahua, Mexico. been very dry a few days before, but teemed with life soon after the heavy rain. The following is from mv field notes taken on 24 July 1958: Near El Salto, Durango, about 4:00 p.m. it began to rain. We decided to return to Chihuahua City and drove through the rain until 2:00 a.m. About 18.5 miles north of Durango City (and for the next 50 miles) we encountered large numbers of amphibians on the road, in the pools along the road and calling from the entire area, in concert, but with several distinct choruses. The following were collected: Bufo d. insidior, Bufo cogna- tus, Hyla arenicolor, and Scaphiopus hammondi. They were so numerous on the road that we could not drive without hearing a constant "pop" as their inflated bod- ies were crushed. We outran that storm but were in it again at Ciudad Chihuahua on the afternoon of 28 July. The following is recorded in my field notes: Packed and left Chihuahua City about 4:00 p.m. in a light rain. The rain continued and intensified during our trip to Colonia Juarez. The amphibians were seen on the road for most of the trip. In some of the valleys between Sueco and Casas Grandes, we encountered large numbers of the same species seen in Durango. The phenomenon of amphibian emergence after heavy late spring or summer storms oc- curs throughout many areas in North Amer- ica. Inasmuch as I have seen the same phe- nomenon in Kansas (summer 1948), I would classify Chihuahua as a southern extension of the areas to the north and not an exception. A similar emergence occurred in the mountains at Chuhuichupa (3 July 1958) and was re- peated again at Cerocahui (17 July 1958). In both cases Bufo microscaphus and Hyla ex- imia wrightorum were abundant. Careful collecting was done around most habitats, such as small mountain streams, springs, seeps at the base of cliffs, and, in general, areas that serve as salamander habi- tats in southern Mexico. We found only one species of salamander, Ambystoma rosaceum, and concluded that the radiation of pletho- dontids had not extended north of Nayarit. The discovery of Pseudoeurycea belli in eastern Sonora (Lowe et al. 1968), and more recently in adjoining west central Chihuahua, revives the belief that representatives of the Mexican plethodontids may indeed occur in 40 Great Basin Naturalist Vol. 49, No. 1 select habitats along the western front of the Sierra Madre Occidental. The western moun- tains, which extend north from central Mex- ico, have served as dispersion pathways for representatives of major groups to reach north into Sonora and Chihuahua, as may be seen in such genera as Eleutherodactylus (tara- humaraensis), Eumeces (brevirostris), and Thamnophis (melanogaster), to mention only a few. Not only have these mountains pro- vided a series of habitats conducive to disper- sion, but they have also provided places of refuge from desiccation since the last pluvial period in an area now nearly surrounded by deserts. We did not collect in all areas of southwest- ern Chihuahua. Those areas near the barran- cas in western Chihuahua were not exten- sively worked. Thus, much must be done in these areas before we can present complete distributions of the species listed in this report. Localities we visited are listed on the map published in the first report on the snake fauna. The map has been revised for reprint- ing in this report. It should be noted that most records are in areas outside of towns and cit- ies. The map provides general locations and should aid in the identification of most locali- ties cited in the text (Fig. 2). Since our first trip into Chihuahua in the spring of 1956, considerable change has oc- curred in some of the waterways. This has undoubtedly affected habitat areas along streams in the central area. Dams have been built along some streams, particularly in the fertile valleys of central Chihuahua. This oc- curred basically in the valleys directly east of the higher mountains to the west and has essentially eliminated some of the streams that originally flowed north into the catch basins of northern Chihuahua. Such streams as the Santa Carmen, Santa Maria, and Rio de Casas Grandes provide little, if any, stream- flow north of such towns as Rancho Flores Magon, Galeana, and Colonia Dublan. The diversion of streamflow has also eliminated most of the wetlands that were once a part of the areas adjacent to the streams. This situation was also referred to by Conant (1974 [1977]: 485-89). To compound the difficulties, bullfrogs (Rana catesbeiana) have been introduced in some of the areas, thus interjecting a serious predator to other amphibians that may have originally inhab- ited ponds and streams in these valleys. Man's agrarian activities in these valleys and others to the south may actually provide additional habitat in some areas to partially compensate for the diversions that their activities have wrought on the northern parts of these valleys. It should be noted that precipitation occurs in Chihuahua primarily during July, August, and early September, during which time heavy rains may occur as thunderstorms rather than as general rains that cover all or the major part of the state. These thunder- storms present a hit-and-miss pattern. Thus, the rainy season presents a climate character- istic of steppe deserts. It is during this time that the amphibians are most active, not only breeding but also, during the evening hours, feeding in open areas not always associated directly with streams or ponds. As noted above, the dry season in the spring and early summer provides little opportunity for most anurans to be active. Also, the evening tem- peratures may fall sharply in September or early October, thus terminating the period of activity as well as the feeding season at ap- proximately three months for most species. In the species accounts additional ecological information is presented. Because of the varied geographical terrain of the area studied, the amphibian fauna of Chihuahua may be placed into three rather distinct biotic groups as follows: 1. Central Chihuahua. This area includes the valleys, low hills, and mountain ranges lying between Highway 45 (Ciudad Juarez to Parral) and west to the Sierra Madre. This central part of the state is divided into two major drainage systems. The uplift of the Sierra Del Nido and the accompanying lower ranges extending west from Del Nido to the vicinity of Cuauhtemoc and Ciudad Guerrero serve to divide the northern system (Rio de Carmen, Rio Santa Maria, and Rio de Casas Grandes) from the more extensive Rio Con- chos to the south and east. Extending into this central area are lower desert valleys and ranges of eastern Chihuahua (Conant 1974 [1977]). It should be noted that the Rio Pa- pigochic (a tributary of the Rio Yaqui) enters west central Chihuahua through a low area in the Sierra Madre and drains some of the high valleys. The area extending north from January 1989 Tanner: Chihuahua Amphibians 41 MAP OF CHIHUAHUA MIL Fig. 2. Revised map of Chihuahua. Minaca to Yepomera and Madera is a part of the Rio Papigochic drainage system. Within this central area are found a series of species closely associated with the desert habitats to the north and east. The northern part of this area is basically a northwestern extension of the Chihuahuan Desert. It is our observation that the major populations of am- phibians in this central area are found in the valleys, which are dry except during the rainy season but then, with the onset of heavy sum- mer rains, literally erupt into breeding popu- lations. This is in contrast to the few indi- viduals seen along the small or intermittent streams during the dry season. Those species most commonly seen in the central area are the following: Scaphiopus coachi Baird, Bufo debilis insidior Girard, and Bufo woodhousei Girard. 42 Great Basin Naturalist Vol. 49, No. 1 Some species have ranges that extend from the mountains into the lower river valleys and basins; Spea hammondii Baird, Bufo punc- tatus Baird & Girard, Hyla arenicolor Cope, and Rana pipiens Schreber are found, but our records indicate that they are less abundant in these lower valleys than are populations of these species in the higher valleys and west- ern mountains. Becords for Rana pipiens for 1931 are from Colonia Dublan (BYU 301 and 3657). We found them approximately 14 miles south of Casas Grandes. Bullfrogs were abun- dant at the river and springs near the old Bancho San Diego, downstream from Colonia Juarez. Apparently, pipiens had been greatly reduced or eliminated from most of the Bio de Casas Grandes by bullfrogs and water diver- sion. This may be true for other species that were present in the recent past. 2. Sierra Madre. The second group, and perhaps the largest series, of species in Chi- huahua occurs in the Sierra Madre. Mountain streams, meadows, and springs provide a vari- ety of habitats that are less frequently found in the central areas of the state or are unavailable or unusable in the rapid streams of the deep barrancas to the west. The following are listed: Ambystoma rosaceum Taylor, Spea hammondii Cope, Bufo woodhousei Girard, Bufo simus Schmidt, Bufo punctatus Baird & Girard, Eleuthrodactijlus tarahumaraensis Taylor, Hyla arenicolor Cope, Rana pipiens Schreber, and Rana tarahumarae Boulenger. 3. Western barrancas. In the western bar- rancas, particularly at Urique, two species, Bufo marinus and Bufo mazatlanensis , were abundant in the streets of the town and along the river in the evenings (14 July 1958). Other species occurring in northern Sinaloa and southern Sonora may enter the mouth of the barrancas in extreme southwestern Chi- huahua. We collected one specimen of Bufo punctatus at Urique and a small series at Ce- rocahui and Cuitaco, an indication that this species has also entered the lower basins of the Bio Oteros. A single representative of Pachymedusa dacnicolor Cope was examined by Peter Warren on 4 January 1980, approxi- mately 5 km south of the junction of the Bio Chico and Bio Yaqui. This is about 50 km by air northeast of Ciudad Obregon, Sonora. Its presence in the Bio Yaqui suggests its pres- ence in western Chihuahua. From the few records available it is obvious that the areas along the state borders between Sonora, Sinaloa, and Chihuahua are not well known. Thus, the distributions of many amphibian and reptile species are not as yet defined. Species Accounts Family Ambystomatidae Genus Ambystoma Tschudi Ambystoma rosaceum Taylor Ambystoma rosaceum Taylor, 1941, Copeia 1941: 143-144, Figs. 1A, IB. Type locality: Majorachic, Chihuahua, Mexico; Anderson 1978: 206. 1-. 2. In small side streams of Rio Bavispe, below Tres Rios near Chihuahua-Sonora border, 4 larvae (BYU 13727-30). Approximately 15 mi W Pacheco (Rio Gavilan), 1 (BYU 16877 small adult S-V 50.8 mm). 18-20 mi from Colonia Juarez up Tinaja Canyon, 14 (BYU 14417-30, larvae). Nortena, 5 mi NW Chuhuichupa, 35 (BYU 14432-66, larvae with one transforming-14442). 1 mi S Chuhuichupa, 8 (BYU 13942-44, 14414, 4 recently transformed adults ranging from 44.0 to 69.5 mm in S-V length; 14488-91, larvae, 30.0-42.5 in S-V length). Black Canyon approx 8 mi W Chuhuichupa, 3 (BYU 14256-58, larvae). 10 mi NW San Juanito, 2 (BYU 14483, 15684, larvae). 8 mi SW Maguarichic, 11 (BYU 16879-89, larvae). 25 mi SE Creel, 34 (BYU 17110-28, 22771-22794, all larvae, 32.0-42 mm S-V length). 8 mi N Basihuari, 12 (BYU 22646-57, 1 adult 51.0 mm S-V and larvae 34-47 mm S-V length). 15.5 mi N Basiberachic, 11 (BYU 22747-57, lar- vae). 9 mi SE Colonia Garcia, 1 (MVZ 20681, adult). Del Nido, Arroyo Mesteno, 19 (MVZ 70554-59, adults; 68714-20, 65939-41, 70547, 70552 lar- vae). Rio Milpillas, 1/2 mi S Milpillas, 1 (UAZ 45882, small adult, S-V 57 mm). Arroyo la Cienega Prieta, approx 1.5 mi N Las Chinacas (N of Milpillas), 1 (UAZ 46562, larvae S-V 55 mm). 14 mi SE Madera (Hwy 16), 1 (UAZ 34641, adult male). 4.6 mi SE Madera (Hwy 16), 1 (UAZ 34642, adult male). Yepomera, 2 (UAZ 34777-8, adults). 2 mi S Yepomera, 1 (UAZ 34779, adult). 6.8 mi N W Yepomera, 1 (UAZ 34780, adult). Mpio, Guadalupe y Caluo, E side of Cerro Mohi- nora, 3,750 ft, (UTEP9307, 1 adult, 73.5 mm S-V, and one lot of 20 larvae). 8 mi E Guadalupe de Los Reyes (29. 1 mi E Cosala, Sinaloa), 1 (UAZ 46090, this locality is approx 75 mi N W of El Salto, Durango, larva-S-V 65.2 mm). January 1989 Tanner: Chihuahua Amphibians 43 Neviera, 4 mi W La Ciudad, Durango (MVZ 47288-9, 57274, 57276-57291, 58648-63, 65890, 22 adults and 14 larvae). Rio Hondo near Los Adjuntas, Durango, 1 (MVZ 57305, larvae). 11 mi W El Salto, Durango, 1 (USNM 154571, adult). El Salto, Durango, 1 (USNM 123581, larva— type of Ambystoma rosaceum nigrum Shannon). SW of Ciudad Durango, 1 (CAS 87350, larva). Specimens collected in Durango are indi- cated; all other collection citations are from the state of Chihuahua. This salamander is present in nearly all small streams, springs, and ponds throughout the mountains of west- ern Chihuahua. We did not find it in the larger streams and rivers where fish occurred. They are more commonly observed in habitats 6,000-8,000 feet in elevation, although some descend into the canyons and high valleys such as the Tinaja and Yepomera, to at least 6,000 feet. We observed the life history from eggs to transforming larvae and two breeding adults. Although some aspects are not yet clearly understood, the life history data avail- able from our studies are reported in the fol- lowing sections. Eggs and Larvae. — Eggs were first ob- served on 2 April 1963 in large potholes in Tinaja Canyon approximately 18 miles north- west of Colonia Juarez. The potholes were fed by seeps along the floor of the canyon and were filled with large, embedded boulders. Eggs were found in two clear ponds, both 2-3 feet deep. In my field notes, the following is recorded: In two potholes there were eggs, in one they were attached between two large rocks (boulders), some were hatching, some out, others not. There were 20-25 eggs and larvae. In the second pool, eggs were laid on tree roots that were extending out and down into the pool. The eggs were attached in clusters of 2-7. On the opposite side of this pool, on the lower edge of a large boulder, 30 eggs were securely attached to the surface of the rock. None of these were hatching. In this pool, a large cluster of frog eggs (Rana?) was found. Because the boulders were large and embed- ded, we could not investigate for additional eggs or adults that may have been beneath (Fig. 3). On 3 April 1963, at a small spring, we found eggs and two adults. This spring and small stream is called the Turkey Tank and is a small tributary of the Rio Juarez. The locale is about 15 miles southwest of Colonia Juarez and a short distance off the road to Pacheco. Only potholes were filled with water, with the small stream filtering through sand and gravel leav- ing little exposed water. Here, as at Tinaja, only potholes provided one to two feet of clear water, and in two of these, eggs were attached to rocks, pine needles, and other debris. Here also the eggs were in clusters of 3-4 to as many as 10-12. Some were also in strings of 3-6 eggs. Eggs were scattered over the rocks in an irregular pattern but with each cluster se- curely attached. Eggs and/or large adults were not found during our field studies from mid-May to late July and August. During this time only larvae and small transforming adults were seen. An- derson (1961) states that "mating probably oc- curs soon after the onset of the summer rainy season, which would be any time from mid- June to mid-July in Durango and Chihuahua." I have no data for Durango, but in Chihuahua most eggs are laid during April or perhaps during early May in the higher elevations (Webb and Baker 1984). This is not to say that some eggs may not be laid during the first heavy rains of July. In our study of Am- bystoma tigrinum nebulosum in Utah (Tanner et al. 1971) the breeding season occurred in May as soon as the melting snow filled the ponds. However, after a heavy July rain a few adults were seen in the lake, and by August we noted that there were large and small lar- vae. We suspect that this may occur in rosaceum; however, the observations of An- derson (1961) at Yaguirachic and those we have made only indicate the possibility of such an occurrence. In Chihuahua, March through June is the dry season. During this period large larvae would have difficulty securing food in the lim- ited aquatic habitats, but small larvae are able to establish themselves so that by the onset of the rainy season they are greatly benefited by the increased feeding areas provided by the rains. Unfortunately, I do not have precise data on the rate of growth by larvae. What is known is that eggs are laid much earlier than July and that small larvae 20-30 mm snout- to- vent length are found in May (Tinaja and below Tres Rios). By mid-July and into August, some larvae are 45-55 mm and some are transforming into small adults with a snout-to-vent length ranging from 44 to 54 mm 44 Great Basin Naturalist Vol. 49, No. 1 Fig. 3. An example of potholes along the streambed in upper Tinaja Canyon. This photo was taken 28 May 1956. (Fig. 4). Our data do not indicate that larvae winter over to the next July as was suggested by Anderson (1961) and Anderson and Webb (1978). All data available to us strongly indi- cate that the majority of populations of rosaceum in Chihuahua grow from larvae spawned in April and transform into small adults before the onset of winter. Our obser- vations of populations in the higher habitats suggest that the observations at Yaguirachic by Anderson were not at variance with ours at Chuhuichupa and that the larger larvae were probably spawned in April or early May. We did not find large larvae in any habitats (Tinaja or Tres Rios) during April or May. I believe that rosaceum has adapted to an early breed- ing season even though the aquatic habitat in most areas consists of seeps and small streams and is thus limited in space until the rains arrive. Furthermore, small larvae would be at a great disadvantage during the rainy season when the heavy runoff usually occurs. There- fore, an early breeding season provides lim- ited but adequate water in the habitat for small larvae. By July larger larvae can cope with the increase of streamflow and, as noted above, a greater feeding area and perhaps a greater abundance of food. July appears to be a period of rapid growth, resulting in larvae that reach their full growth and transform -=S3 fV:,- v ^aaas***- Fig. 4. Selected Ambystoma r. rosaceum larvae from the mountains of western Chihuahua, eastern Sonora, and Sinaloa: A, Chuhuichupa, July 1958, S-V 30-40 mm. (Figs. 4B through 4D continued on facing page.) January 1989 Tanner: Chihuahua Amphibians 45 \ \ Fig. 4B continued. Chuhuichupa, 25 August 1957, S-V 45-50 mm (dorsal fin and gills reduced). **4* Fig. 4C, D continued. C, 25. 15 mi (by road) N Yecora, Sonora, UAZ 45879, S-V 58.2 mm; D, 48.2 mi NE Mocorito along road to Surutato, Sinaloa, UAZ 46091, S-V 54.8. during August or at the close of the rainy season in September or October (Fig. 4). Webb and Baker (1984) found eggs and small larvae in late May. These were in per- manent streams and consisted of small and large larvae (at an elevation above 10,000 ft, locality 3, northeast side of Cerro Mohinora, ca 3,750 m). As noted above, elevation may delay egg laying until May and also increase the time needed for larvae to transform. 46 Great Basin Naturalist Vol. 49, No. 1 Fig. 5. Recently metamorphosed adults of Ambystoma r. rosaceum collected 1 mi S of Chuhuichupa, Chihuahua, 25 August 1957. BYU 13943, 13944, and 13942; S-V 45.2, 44.7, and 51.8, right to left. Neoteny. — We observed no neoteny in the rosaceum of Chihuahua between 1956 and 1972. We saw no large larvae during April (1963 and 1985) and May (1956) in habitats where we found large numbers of them in July and August. There were perhaps two basic reasons for this: It is (1) too dry to provide a habitat that would permit survival or (2) too cold to allow wintering over. Under these conditions, adults survive by burrowing, whereas neotenes cannot. Our data strongly suggest a life cycle of early egg laying (April and perhaps May) with little breeding at later dates (July), a rapid growth of larvae, particu- larly after the onset of summer rains, and a period of metamorphosis from mid-July into October with little or no carryover of larvae in most, if not all, habitats even in the higher mountains of western Chihuahua. That breeding larvae occur in Durango and Sinaloa is not questioned, but we did not find them in western Chihuahua and can find no data in the reports of Anderson (1961) or Anderson and Webb (1978) to support their occurrence. Webb and Baker (1984) report large larvae in late May. This indicates a wintering over of larvae but does not confirm larval reproduc- tion since adults were also present. Adults. — Two mature adults and 10 trans- forming or recently transformed adults were collected (Fig. 5). The latter ranged in size from 44 to 53 mm S-V, one adult 69.5 mm (Fig. 8). During late July and August numer- ous larvae were observed with reduced gills and the dorsal fin greatly reduced or absent from most of the body. Time limits precluded our gathering data on the length of time in- volved in the process of metamorphosis. Indi- viduals with only stumps of gills were seen along the edge of the water or out on the bank. Apparently, rosaceum has adapted to take advantage of the rainy season. The larvae, having reached the size and age for metamor- phosis, are ready to leave the aquatic habitat while the terrain is moist. Thus, the new ter- restrial habitat is less stressful, having a moist ground litter and softer soil for burrowing. Two breeding adults were collected at the Turkey Tanks (3 April 1963). In our attempt to return them alive they died and spoiled while January 1989 Tanner: Chihuahua Amphibians 47 Fig. 6. Adult Ambystoma r. rosaceum taken at the Turkey Tanks SW Colonia Juarez, 3 April 1963. in transit. The pair was marked with yellowish cream spots and blotches on a dark greenish ground color. Their general appearance was reminiscent of Ambystoma tigrinum tigrinum except that the spots were more uniformly round and small (Fig. 6). The smaller and perhaps younger adults still retained some features of the larval pattern (Fig. 5). One transforming specimen was returned alive to the laboratory where it slowly developed the yellowish spots similar to the breeding adults from the Turkey Tanks (Fig. 7). The labora- tory specimens readily accepted earthworms. The series of specimens from Guadalupe y Calvo (adult, UTEP 9307 and lot 9308, 20 specimens) is important in that all specimens are from an area in southern Chihuahua and near Durango. Furthermore, the series con- sists of individuals ranging in size from small larvae, less than 20 mm in total length, to a mature larva, 51.3 mm S-V with gills and dor- sal fin reduced. There is also a recently meta- morphosed adult 53.0 mm S-V. The series (collected 23 May 1982) is re- markable in that it contains all stages in the life cycle. Collections made in northern Chi- huahua (25-31 May 1956) at Rio Bavispe and seen in Tinaja Canyon were of a nearly uni- form size (30-40 mm). I have observed little variation in the color pattern of the adults and larvae from Chihuahua. The degree of pig- mentation may vary, some darker than others, but the pattern is essentially the same in all. However, in widely separated populations there may be some variation in the life cycles. Systematics. — Dunn (1940) included six adults and four larvae from Chihuahua (no localities or collections are listed) under the subspecies Ambystoma tigrinum velasci I ETHIC 1 £t\ huh 3! l 4 5 ' 61 7. 8. 9 10 11 ) V Fig. 7. Small adult collected (19 July, 25 mi S Creel) while still with gill buds and larvae color pattern. Returned to the laboratory where it developed the adult color pattern. Photo taken mid-September 1960. 48 Great Basin Naturalist Vol. 49, No. 1 Duges. He described them as being "marked with round yellow spots and a high gill raker count" (9-15 gill rakers on the anterior face of the third arch). In his closing remarks Dunn (1940) states, "Some specimens from Arizona and New Mexico with circular yellow spots cannot be assigned." He further states that "this form has rather large larvae with 19-20 gill rakers on the anterior face of the third arch, thus differ- ing from A. t. velasci. " The specimens in question were from Prescott, Arizona (2); Rio Mimbres near Deming (1), Ft. Wingate (9), Pescao (3), and Nutria (2), all from New Mex- ico. The location of these specimens was not indicated. It is obvious that he examined more material from Chihuahua and areas immedi- ately to the north than had been done in previ- ous studies. Taylor (1941) described A. rosaceum (based on larval characters) and made no reference to the study by Dunn (1940). Shannon (1951) described the subspecies of rosaceum, r. nigrum from Salto, Durango, and r. sono- raensis from 32 miles south of the Arizona border, Sonora. Both Shannon subspecies were based on larval specimens. Anderson (1961) reviewed the life history and systematics of A. rosaceum, concluding that the two Shannon subspecies were not valid. His reasoning was an extension of Dunns (1940) concern that larval characters were not reliable in a final determination of Ambijstoma taxa. Apparently, Anderson and Webb (1978) were still of the same opinion. I agree with the conclusion reached by An- derson (1961) that Ambijstoma fluvinatum Taylor is a color phase of rosaceum. The type of fluvinatum as figured by Taylor appears to be an individual in an early state of metamor- phosis, with the dorsal fin greatly reduced. I have seen one of the larvae in the type series of A. r. sonoraensis Shannon (USNM 17255) and have compared it to specimens from the Rio Bavispe drainage. In most characters there are only slight variations. However, the color pattern is of a more diffuse lateral pat- tern and there are fewer gill rakers (20) than in the Bavispe series (22—26). Shannon (1951) described the color pattern as a "dark brown, ground color extends to edge of, but not or barely onto venter — other- wise pattern similar to that of A. r. nigrum. " We have noted significant differences in color patterns between the above description and specimens collected during our field trips into the Rio Bavispe drainage system (below Tres Rios, Chuhuichupa, and Black Canyon). We, therefore, assumed that Ambijstoma rosace- um had developed a series of subspecies be- cause of isolation in the major river systems. It is reasonable to suggest that the Rio Bavispe, a tributary of the Rio Yaqui, may contain popu- lations as distinct as those in northern Sonora (sonoraensis), the Rio El Fuerte (rosaceum), or the Durango subspecies (nigrum). We did describe and differentiate the larvae of the Bavispe area but hesitated to complete the report until a series of adults from the river systems are available. The teeth of six larvae were examined. The larval specimen (BYU 13727, S-V44.7) is representative of the series and has the following: pterygoid 11, vomerine 20, premaxillary 13-14, maxillary 12-13, dentary 28, splenial 35, max-premax 35-27, dent-splenial 62. Van Devender (1973) and Van Devender and Lowe (1977) reported the occurrence of Ambijstoma tigrinum and Ambystoma rosace- um in the environs of Yepomera, Chihuahua. The adults were identified as tigrinum and the series of larvae taken from the streams and springs as rosaceum. This cells into question the presence of two ambystomid species (lar- vae of one and adults of the other) in the same habitat. During a recent meeting with Dr. Charles Lowe, we examined the series of adults from the area in and near Madera and Yepomera, Chihuahua, which were reported by Van De- vender (1973) and Van Devender and Lowe (1977) as Ambijstoma tigrinum. We concluded that this series represented only adults of Am- bijstoma rosaceum. There is reason to believe that the occurrence of adults with bright yellowish spots did not correspond to the de- scription of A. rosaceum as described by An- derson (1961). Furthermore, Dunn (1940) had referred spotted specimens from Chihuahua as representatives of A. tigrinum. Most, if not all, of the mature adults seen by Anderson (1961) were the dark subspecies from Durango, with most of his Chihuahua speci- mens being larval or small, recently meta- morphosed adults. Thus, the finding of large greenish adults with bright, round, yellow spots was not expected to be Ambijstoma rosaceum. Futhermore, Lowe (1964) and January 1989 Tanner: Chihuahua Amphibians 49 Fig. 8. Ambystoma r. rosaceum collected 1 mi N Chuhuichupa, Chihuahua, 2 July 1958. This was the largest recently transformed specimen collected, S-V 69.5, BYU 14414. Webb and Roveche 1971 listed only Am- bystoma tigrinum for Arizona and New Mexico. The biology of Ambystoma rosaceum as stated above has a direct relationship to its systematics. It appears that eggs are laid in April and/or May. Larvae grow rapidly, with some reaching 45-55 mm in S-V length by late July and others by the end of the rainy season in September. Larvae begin to trans- form in July, retaining for a time the larval color pattern, and slowly develop a uniform darker color before the yellow spots appear that are characteristic of mature adults (see Figs. 5, 7). These data are for Chihuahua populations taken from areas west of Colonia Juarez and south to Maguarichic and areas near and southeast of Creel. Throughout this mountainous area, little variation occurred between the larval populations. Transforming specimens from Chuhuichupa and south of Creel developed similarly from mature larvae to young adults, with the mature color pattern of yellow spots appearing at 55-60 mm in S-V length. Specimens collected at Chuhuichupa on 25 and 26 August 1957 included four recently transformed adults, 44-53 mm S-V. On 2 July 1958 we collected one specimen, 69.5 mm S-V, that still had three gill buds on each side; it was the largest one we collected or saw that had larval characters (BYU 14414, Fig. 8). These specimens were taken from and around the large springs south of town. All other lar- vae or transforming adults collected or seen were smaller (45-55 mm in S-V length). A collection from 8 miles north of Basihuari (southeast of Carmen Bridge in Rio Urique drainage) consisted of 12 specimens collected 4 October 1964 and included one small adult (51.0 mm S-V), one with reduced gills and dorsal fin greatly reduced (47 mm S-V), and a series of 10 larvae ranging in size from 34 to 45 mm. All were spotted and with a yellowish ground color. One can only speculate as to the age of these larvae. Other data suggest that eggs for these were laid in May. However, it is obvious that the larger individuals were trans- forming and that this would continue until the aquatic habitat was reduced or disappeared after the rains stopped and the soils dried. A small adult from 1/2 mile south of Mil- pillas (UAZ 45882) in west central Chihua- hua near the Sonora border and from the Rio 50 Great Basin Naturalist Vol. 49, No. 1 Fig. 9. Adults of Ambystoma r. rosaceum Taylor: A, male, UAZ 34642, collected 1 July 1971, 4.6 mi SE Madera on Hwy 16, Chihuahua, S-V length 89.0 mm; B, female, UAS 3477, collected at Yepomera, Chihuahua, 12 June 1972, S-V 75.5 mm. Milpillas drainage has the round, light spots but with a dark ground color. It is possible that adults from Sonora may be spotted, but with a melanistic ground color, in contrast to the Chihuahua populations and the nearly spot- less populations in Durango. Although additional data from larger series of adults would be helpful, there are at pres- ent adequate materials available, both larvae and adults, from Chihuahua and Durango to recognize the following subspecies: Ambystoma rosaceum rosaceum Taylor Tarahumara Salamander Ambystoma rosaceum Taylor, 1939(1938). Univ. Kansas Sci. Bull. 25: 385-405. This subspecies is distinguished by bright yellowish spots on an olive green ground color in adults (Figs. 6, 9). Larvae have dark spots and reticulations on a yellowish ground color that may be tinged with a pinkish color and with a lateral, irregular, yellowish stripe in the lateral line area. It is presently known from the high valleys and mountains of western Chihuahua, moun- tains of eastern Sonora, and the drainage of the Rio Sinaloa in northeastern Sinaloa and southern Chihuahua. The type locality is Mo- jarachic, Chihuahua. This distribution does not include northern Sonora north and west of the Rio Yaqui drainage. Ambystoma rosaceum nigrum Shannon Durango Salamander Ambystoma rosaceum nigrum Shannon, 1951. Proc. U.S. Nat. Mus. 101(3284): 465-484. This subspecies is distinguished by a dark ground color without the bright yellow spots. Any spots are faint and in most individuals are indistinguishable in adults. Larvae have small, irregular, light spots and reticulations on a dark ground color; the central lateral light stripe may or may not be present (Fig. 10). It is presently known from the mountains of west central Durango and east central Sina- loa. The type locality is El Salto, Durango. Remarks. — At present we do not have all the data that would detail the complete life January 1989 Tanner: Chihuahua Amphibians 51 Fig. 10. Ambystoma r. nigrum Shannon: A, holotype (larva), USNM 123581, El Salto, Durango, S-V 59.0 mm. Fig. 10. Ambystoma r. nigrum Shannon: B, adults MVZ 57285, S-V 78.0, and 57279, S-V 72.8 collected at Neviero, 4 mi W La Ciudad, Durango. Fig. 10. Ambystoma r. nigrum Shannon: C, adult USNM 154571, collected 11 mi W El Salto, Durango. 52 Great Basin Naturalist Vol. 49, No. 1 history of rosaceum from egg to mature adult. Such a study presumably would provide de- tails of the size of larvae from hatchling to one transforming, rate of larval growth, age and size of larvae at metamorphosis, and size of breeding adults in the subspecies. Only parts of this sequence are now fully understood. In spite of life history gaps that must yet be filled, data now available do provide a more com- plete understanding than has been available. It is obvious that elevation plays an impor- tant role in determining the time of egg lay- ing. In Sinaloa (Sierra Surutatas, 48 airline km northeast of Guamuchil, about 3,500 ft) An- derson and Webb (1978) report that eggs were laid in February. We found eggs in Chihuahua in April at about 6,000 ft; Webb and Baker (1984) found them in May on the northeast side of Cerro Mohinora at about 10,000 ft. Whether eggs are laid in late June or July at the onset of summer rains as suggested by Anderson (1961) has not as yet been observed in Chihuahua populations. Our data suggest that this does not often occur for the following reasons: (a) competition with the larger larvae and life in the increased streamflow, at times in flooding channels, would make survival dif- ficult in many habitats; (b) egg laying that late in the year may not provide enough time for growth and metamorphosis before the water in most habitats is reduced or disappears soon after the close of the rainy season. However, as noted above, other ambystomatid species are known to be stimulated to reproduce late in the season after heavy summer storms, and, though this has as yet not been observed in the rosaceum of Chihuahua, it may occur. The suggestion of Dunn (1940) and Ander- son (1961) that larval characteristics are too variable to be useful in systematics may have merit. However, I find that the larvae of Am- bijstonia rosaceum do have a very distinctive color pattern. Both Taylor (1941) and Shannon (1951) used larval characters in establishing and differentiating taxa. Variability should not be a justifiable criterion for eliminating the usefulness of characters that may appear in the sequence of a species' life history. The external anatomy of Ambystoma rosaceum as it appears in the text was pre- pared before studies dealing with elec- trophoretic data were reviewed. It is of inter- est to note that both studies (Shaffer 1983 and Jones et al. 1988) arrive at essentially the same conclusions, namely that rosaceum is a spe- cies distinct from tigrinum. Shaffer also con- cluded that rosaceum is a polytypic species with A. r. rosaceum in the north (Chihuahua) and A. r. nigrum in the south (Durango). Family Plethodontidae Genus Pseudoeurijcea Taylor Pseudoeurycea belli sierraoccidentalis Lowe, Jones, & Wright Pine-Oak Plethodon Pseudoeurijcea belli sierraoccidentalis Lowe, lones, & Wright, 1968, Contributions in Science, Los An- geles County Museum 140: 1-11. Type locality 21 km WSW Yecora, Sonora. 6 km WNW Ocampo (on road to baseball field El Aguila) (Lowe, Van Devender, and Holm, in press). Four adults were observed on 24 June 1987 by field parties on a trip organized by Paul S. Martin of the University of Arizona. The spec- imens were emerging from holes in rotted tree roots in the bank of a road cut. Above the road cut is a pine-oak woodland with volcanic boulders and deep leaf litter. The elevation is approximately 1,830 m. Pseudoeurycea b. sierraoccidentalis is a black salamander with dark red spots on its upper surfaces. Ground color above is a uni- form black throughout with little variation be- tween the dorsal and ventral areas. The num- ber of spots for three of the Ocampo animals was 0, 11, and 13, based on field notes and color slides; field photo specimen vouchers are UAZ 47824-25 PSV. The samples include the first spotless individual for the species. The area of Ocampo, Chihuahua, and the type locality southwest of Yecora, Sonora, are in the headwaters of the Bio Mayo. These uplands range above 6,000 ft and are near, if not a part of, the divide between the Bio Mayo and Bio Yaqui drainage systems. Further- more, the drainage basins to the east (ca 50 km) include the headwaters of streams flowing north into the Bio Yaqui or south into the Bio Fuerte (Oteros). This suggests a much larger distribution than is now known for both east- ern Sonora and west central Chihuahua. The type locality and habitat of the type series is in an east-west oriented canyon (bar- ranca) with the upper end extending into an. area in or near Chihuahua (Lowe et al. 1968). Thus, the western slopes of the mountains extending from about 5,000 feet in Sonora to at least 6,000-7,000 ft in west central Chi- huahua are apparently the present habitat. January 1989 Tanner: Chihuahua Amphibians 53 Family Pelobatidae Within this family are genera from a wide distribution that includes representatives from Europe to the South Pacific. In North America only the genus Scaphiopus is presently recognized by most authors (Tanner 1939, Zwiefel 1956, Kluge 1966, Estes 1970, Brown 1976). Cope (1889) placed the Ameri- can spadefoot toads in the family Scaphiopodi- dae and divided the species into two genera as follows: Scaphiopus (holbrookii and couchii), Spea (hammondii with three subspecies, ham- mondii, bombifrons, and intermontana, and multiplicata as a species). The genus Scaphiopus has a range from the eastern United States west to the Great Plains (Dakotas south to Texas). The excep- tion is couchii, with a range west from south- ern Texas through the low desert valleys of southern New Mexico, southern Arizona, and into southwestern California between Nee- dles and Vidal Junction (Tinkham 1962). The range also includes northern Mexico from Tamaulipas to Baja California. The genus Spea occurs in the western United States from the Great Plains west to California and south throughout western and southern Mexico. At present four species are listed for this genus (bombifrons, intermon- tana, hammondii, and multiplicatus) (Brown 1976). There is, however, some question as to whether all are valid as species. This is dis- cussed further below with a brief review of the history of the genus Spea. The genus Spea was established by Cope (1875) to include the spe- cies stagnalis from northwestern New Mex- ico. In this report Cope also proposed the family Scaphiopodidae and included the spe- cies bombifrons in the genus Spea. In 1889 he listed Spea hammondii and included bomb- ifrons and intermontana as hammondii sub- species; species status was retained for multi- plicata. Cope (1889) gave the following description for the genus Spea: "Cranial derm free from cranium; the latter generally with a fronto- parietal fontanelle; vomerine teeth present; toes webbed; cuneiform process large." Bec- ognizing that skull modifications had occurred in the genus, he stated: "In one of the subspe- cies of S. hammondii the ossification of the cranium has progressed so far as to close the frontoparietal fontanelle, but not so as to pen- etrate the cranial integument." Based on our present knowledge of the skull characters in Spea we can assume that Cope may have been referring to intermontana. Although Cope did not figure the skulls of the species he included in Spea, his descriptions strongly indicate that he recognized the uniqueness of the cranial characters, particularly the presence of a fron- toparietal fontanelle. The latter is not present in the genus Scaphiopus. Tanner (1939) examined specimens of the subgenus Scaphiopus (holbrookii, Massachu- setts to Florida; hurterii, Benton and Lytle, Texas; couchii, San Pedro, Baja California; Waco, Fairbanks, and San Antonio, Texas; San Xavier Mission, Arizona) and figured a representative of each species (Fig. 11). The subgenus Spea was examined (bombifrons: Goodnight, Texas; Lexington, Oklahoma; Elk- hart, Kansas; intemwntanus: Carbon County, Garfield County, Kane County, Juab County, Uintah County, Washington County, Utah; hammondii: Cochise County, Arizona; San Jacinto, Biverside County, and San Diego, California; Ojos Negros and Punta Banda, Baja California) and skulls representing each taxon were figured (Fig. 11 reproduced from Tanner 1939, Plate I). He regarded each taxon as a species, in contrast to Cope (1889), who assigned them subspecies status. Schmidt (1953) accepted bombifrons and hammondii as species but placed intermontana as a sub- species of hammondii. The dorsal skull figures of Tanner (1939) clearly illustrate the differences between Scaphiopus and Spea. It is difficult to assume that such structures should be considered as belonging within a single generic group. In Scaphiopus there is a bony plate extend- ing from the parietal area to the nasal. It cov- ers the interorbital area and is suggestive of a primitive dermal plate. Whether it is a re- tained primitive dermal bone or one more recently developed, it is a unique, but uni- formly simple, covering for the interorbital area. The interorbital skeletal structures in Spea are not readily comparable to those of Scaphiopus. Zwiefel (1956) retained Spea as a subgenus within Scaphiopus and recognized Pelobati- dae rather than Scaphiopodidae. His figures of the skulls vary little from those of Tanner (1939), and the relationship of the skin (derm) to the cranium was not considered. The study by Zwiefel (1956) provided an abundance of 54 Great Basin Naturalist Vol. 49, No. 1 STUDY OT GENUS SCAPH/OPUS VA3C0 M. TANNER FIG. | SCAPH/OPUS H0LBR0OKU \ FIGr. 8 FI&. 7 SCAPH/OPUS HAM/10ND// F\Cj. 5 SCAPH/OPUS COUCH// FIG. JO FIG-. 12 FlOr. <5 SCAPH/OPUS /MTCPrtOMTAA/US FIG. II SCAPH/OPUS QOrtBIFRONS Plate I Fig. 11. Representative skulls of the genera Spea and Scaphiopus as prepared by V. M. Tanner (1939). January 1989 Tanner: Chihuahua Amphibians 55 data; my only concern is that cranial charac- ters were not adequately weighted when generic values were considered. It is not only skull characters that vary, but also external characters such as the shape of the spade, smaller body size (S-V length), color pattern, and perhaps area of distribu- tion. It appears that Spea is a western and southern group, whereas Scaphiopus is an eastern genus, in which couchi has more re- cently extended its range west to overlap that of S. bombifrons in the Great Plains and ham- mondii in Texas and west to California, includ- ing the adjoining border states of Mexico. Cope (1889) in his key separated Scaphio- pus from Spea on the character of derm in- volved in the cephalic ossification, which in Spea is "distinct from cranium, which is usu- ally only ossified superiorly in two supercil- iary bars." Restated, the interorbital area is completely ossified in Scaphiopus (Fig. 11), whereas in Spea only bars of bone and an interorbital fontanelle provide little surface for the derm to be ossified with or to the cranium. In recent studies other distinguishing char- acters have been established. Kluge (1966) summarized, in his Table 5, 17 diagnostic- characters that distinguish the genus Spea from the genus Scaphiopus. To retain Spea as a subgenus in the genus Scaphiopus does not seem to be justified. I am persuaded to recog- nize Spea (as did Cope originally and Tihen [I960]) as a full genus based primarily on the distinct differences in the skull characters. Other characters as indicated by Kluge (1966) do not distract from such a taxonomic change. Key to the Genera 1. Frontoparietal and nasal bones broad and com- plete, no interorbital fontanelle; derm ossified and adhering to cranium in the interorbital area; large, S-V length of adults 50+ mm; area be- tween orbits wide, 5-7 mm; spade sickle- shaped and long Scaphiopus Holbrook — Frontoparietal area with narrow bars between the orbits, separated by an interorbital fon- tanelle in hammondii and bombifrons but thin bone may extend between the interorbitals in intermontanas- smaller, S-V of adults 50 mm or less; area between orbits narrow, 4-5 mm; spade not sickle-shaped Spea Cope Genus Scaphiopus Holbrook Scaphiopus couchii Baird Desert Spadefoot Scaphiopus couchii Baird, 1854, Proc. Acad. Nat. Sci. Philadelphia 7: 62. Scaphiopus couchii: Wasserman, 1970, Cat. Amer. Amph. and Rept. 85: 1. Near Cd. Chihuahua, 5 (BYU 10440-44). Colonia Dublan, 5 (BYU 2142-46i, 2771). 13 mi E Rancho Flores Magon, 4 (BYU 13967-70). Colonia Juarez, 2 (BYU 14522, 15321). Southern edge of Cd. Chihuahua, 4 (BYU 10424-27). 2 mi SE Colonia Juarez, 6 (BYU 13446-48, 15452-53, 15586, 15839). Along road (Hwy 10) 15-30 mi SE Nuevo Casas Grandes, 33 (BYU 14075-14107). 12 mi S Samalayuca, 1 (UAZ 7671). 5 mi SE Galeana, 9 (UAZ 36428-36). 1.6 mi N Galeana, 7 (UAZ 34457-63). 7. 1 mi N Cd. Chihuahua, 1 (UAZ 34458). 1. 1 mi SW Nuevo Casas Grandes, 1 (UAZ 34456). 1.9 mi S Buenaventura, 2 (UAZ 36427, 36431). 5 mi N Cerro Campana, 17 (MVZ 68776-80, 70608-620). 7 mi N 3 mi E Cerro Campana, 5 (MVZ 70603-7). Ojo de Laguna 1, (MVZ 12778). 29 mi W Gallego, 7 (MVZ 70600-2, 72776-7, 72779-80). In the late afternoon of August 1957, Dr. Gerald Robison and I left Ciudad Chihuahua for Colonia Juarez. We intended to do night collecting along the road. Soon after dark a light rain began, and by the time we reached Sueco we were in heavy rain. We saw a few anurans on the road and noted an increase as we drove toward Rancho Magon. The heavy rain continued, and the numbers of anurans on the road increased. By the time we reached Galeana and crossed the Rio Santa Maria, the desert was literally alive with frogs and toads. What had been a few days before a dry, barren landscape was suddenly an expanse teeming with thousands of croaking, hopping creatures that had literally erupted from the earth. In- deed, this was a "show time" not often seen, and, at this time, Scaphiopus was a primary participant (Fig. 15). As we stood on the road and witnessed this assemblage of energetic creatures, we real- ized that only yesterday they were in a burrow and had perhaps been there for months or even a year awaiting this, their day, to fulfill their biological role in life's program. What a remarkable phenomenon. Genus Spea Cope A reexamination of the skulls as figured by Tanner (1939) and Zwiefel (1956) and the 56 Great Basin Naturalist Vol. 49, No. 1 preparation of others from localities in the range of S. hammondii indicate that the latter is a polytypic species with at least three sub- species. The following key, which uses skull characters to separate the species in the genus Spea, will serve to identify the species. Key to the Species of the Genus Spea 1. Frontoparietal bones narrow, ridgelike, unmod- ified and separated by a large fontanelle hammondii — Frontoparietal bones modified by a boss or in- creased bony tissue reducing or nearly eliminat- ing the frontoparietal fontanelle 2 2(1). An enlarged boss near anterior ends of fronto- parietals, producing an external swelling be- tween eyes; fontanelle small and posterior to bosses; skin smooth with few tubercles; body length of adults 45-55 mm bombifrons — Frontoparietals enlarged ridges but without a distinct boss, fontanelle eliminated or greatly reduced by thin bone between ridges of fronto- parietals; skin more rugose; body length of adults 50-60 mm intermontana Spea bo?nbifrons Cope Plains Spadefoot Scaphiopus bombifrons Cope, 1863, Proc. Acad. Nat. Sci. Philadelphia 15: 53. Spea hammondii bombifrons: Cope, 1886, J. Acad. Philadelphia 2, 6: 81. Spea bombifrons: Cope, 1889. U.S. Nat. Mus. Bull. 34: 5-525. Colonia Dublan, 1 (BYU 415). Outskirts Cd. Chihuahua, 5 (BYU 10440-44). 13 mi E Rancho Flores Magon, 5 (BYU 13962-66). Neither Cope (1889), Kellogg (1932), nor Smith and Taylor (1948) listed this species for Mexico. Conant (1975) and Stebbins (1985) listed the range as extending from south cen- tral Canada through the plains of central U.S. and into northern Chihuahua. Firschein (1950) and Shannon (1953, 1957) listed speci- mens for Samalayuca and Cd. Chihuahua. The existing records place this species in the lower valleys of north and central Chi- huahua and not in the higher valleys such as Babicora or Madera just east of the Sierra Madre. In the higher valleys (7,000-9,000 ft) of the mountains only rnultiplicata was found. Spea hajynnondii Baird Western Spadefoot Scaphiopus hammondii Baird, 1857 (1859). Explorations and Surveys for a Railroad Route from the Missis- sippi River to the Pacific Ocean 10, Pt. 4, No. 4: 12. Spea hammondii: Cope, 1889. U.S. Nat. Mus. Bull. 34: 5-525. There has been much uncertainty and doubt concerning the proper systematics for populations inhabiting a wide area in south- western United States and Mexico. Brown (1976) included all populations of Scaphiopus hammondii that were east and south of Cali- fornia in a single species, Scaphiopus multipli- catus. His careful research of two populations (southern California and southeastern Ari- zona) added greatly to our understanding of life-history variables in these two segments of the S. hammondii complex. There are, how- ever, two basic areas that are not yet fully researched. Apparently, there is little or no difference in the skeletal characters, and the two populations are interfertile with no appar- ent postulating isolation (Brown 1976: 2). If we accept the skull characters as being the basic characteristic of the species in the genus Spea (as has been done for bombifrons and intermontana), then the open, unmodi- fied frontoparietal fontanelle (Fig. 11 from Tanner 1939, his Fig. 7) is the basic character for the species S. hammondii. It is obvious that isolation and habitat modification were brought about by the desiccation after the Pleistocene period (Morafka 1988). External characteristics of the habitat were modified, of course, and then these variations directly affected the life-history characters of the populations in a wide area originally occupied by this species. It seems reasonable to believe that in this widespread hammondii complex of popula- tions there is a series of geographical subspe- cies that are related through cranial characters but show subspecific variations in external characters and life-history variables. Those who have examined the skulls (Tanner 1939, Zweifel 1956) have found no significant differ- ences. The various populations show variation in the external anatomy (size S-V, skin tex- ture, nature of mating calls, and adaptation to changes in the aquatic environment; Bragg 1945, Brown 1976). The dorsal view of skulls from numerous localities indicates rather uniform skull char- acters for populations ranging over a wide area in southwestern United States and northwest- ern Mexico (Figs. 11, 12, 14). The following key is an attempt to identify the subspecies that may presently be recog- nized. It should be noted that far more re- search dealing with all internal and external Fig. 12. Dorsal view of the skulls of Spea hammondii representing the following populations: A, Brewster County, Texas (BYU 2767); B, 0.5 mi W Kirtland, San Juan County, New Mexico (UNM 47614); C, Chirieahua Mt, Cochise County, Arizona (BYU 8932); D, Navajoland, Many Farms, Apache County, Arizona (LACM 127298); E, 11.1 mi NW Yepomera, Chihuahua (UAZ 34818, see Fig. 14A); F, 3.8 mi SE Yepomera, Chihuahua (UAZ 34814, see Fig. 14B); G, 18.8 mi N Cd. Durango(BYU 15526); H, El Rosario, Baja California, (BYU 34551); I, San Diego, California (BYU 2141). Note: These drawings are designed to show the size, shape, and position of the frontoparietal fontanelle to the dorsal skull structure in the species Spea hammondii Baird. These figures were prepared by Kaye H. Thorne. characters must be done before a complete understanding of S. hainmondii is obtained, particularly for the Mexican populations. Key to the Subspecies of Spea hammondii Baird 1. Skin of upper parts unusually warty, dorsal tu- bercles large and numerous; heel usually reach- ing to tympanum (Fig. 13-A) h. multiplicata — Skin smooth or with fewer low tubercles; often a dorsal pattern of lighter stripes or spots; heel usu- ally not reaching tympanum (Fig. 13-B) 2 2(1). Adults large, 50-55 mm S-V; interorbital fontanelle large and usually extending medially beyond the posterior margins of the orbits .... h. hammondii — Adults smaller, 45-50 mm S-V; interorbital fontanelle large, but not extending posterior to the orbits hammondii stagnalis Remarks. — Differences in mating calls and adaptation to seasonal precipitation (March in California and July in Arizona) cannot be used as key characters for preserved museum spec- imens. Such life-history characters are impor- tant in demonstrating differences brought about by isolation and changes in the environ- ment that have resulted from desiccation in an area of a formerly more widespread spe- cies. The data presented by Brown (1976) are important in that the biological evolutionary 58 Great Basin Naturalist Vol. 49, No. 1 Fig. 13. Scaphiopus hatnmondii: A, Spea h. multiplicatus , UAZ 34818, 11. 1 mi NW Yepomera, Chihuahua, on Road 16, ca 7,000 ft, in pine forest; B, Spea h. hammondii, UAZ 34814, 3.8 mi SE Yepomera, Chihuahua, on Road 16, ca 6,200 ft, in grassland; a distance of 15 mi and an elevation differential of 800 ft provide an environmental change that ecologically separates these populations. divergence is established and provides popu- lation data that cannot be ascertained by a study of the external or internal anatomy. There are, however, few divergences in the crania of these widespread hammondii popu- lations. They are apparently interfertile, although further research is needed for the extensive Mexican populations. In view of the similarity in cranial char- acters and the fact that the widespread ham- mondii populations are interfertile, I am persuaded to consider hammondii a polytypic species, consisting of at least three subspe- cies: h. hammondii in California and Baja Cali- fornia; h. stagnalis in Arizona, New Mexico, southwestern Texas, and the valleys on each side of the Sierra Madre in Chihuahua, Durango, and Sonora. In the higher mountain valleys extending from western Chihuahua south into central and southern Mexico is h. multiplicata. Spea hammondii stagnalis Cope Scaphiopus hammondii Baird, 1859, Expl. Surv., R.R. January 1989 Tanner: Chihuahua Amphibians 59 Miss. Pacific, 10, Pt. 4, No. 412. Fort Reading, California (part). Spea stagnalis Cope, 1875, in Yarrow, U.S. Geol. Surv. W. 100th Meridian 5: 525. Northwestern New Mexico. Scaphiopus hammondi hammondi Schmidt, 1953, Check- list North American Amphibians and Reptiles. Amer. Soc. Ichth. and Herpet. 6th ed., p. 59 (part). Scaphiopus multiplicatus: Brown 1976. Los Angeles County Nat. Hist. Mus. 286: 1-15 (part). 12 mi SE Babicora (approx halfway between Babi- cora and Gomez Farias), 14 (BYU 14453-67, 15571-80). 10 mi W San Francisco del Oro, 1 (BYU 15677). Chuhuichupa, 1 (BYU 15481). 0.3-18.3 mi SE Madera (along Hwy 16), 15 (UAZ 34649-51, 34656-61, 34663-64, 34666-68, 35040). Yepomera 14 (UAZ 34652-55, 34804-13). 3.8 mi SE Yepomera, 1 (UAZ 34814). 6.2-6.6 mi NW Yepomera, 4 (UAZ 34815-17, 35041). 11.1 mi NW Yepomera, 1 (UAZ 34818). 2 mi S Santa Clara, 24 (MVZ 70622-70645). 1 mi S 1/2 mi E Santa Clara, 1 (MVZ 72790). 5 mi N Cerro Campana, 1 (MVZ 72791). Ojo de Laguna, 1 (MVZ 72792). Arroyo Mesteno (Sierra del Nido), 3 (MVZ 72787-9). Specimens collected in the higher val- leys (Babicora, Madera, and Yepomera) have fewer and lower tubercles and thus a smoother skin (Fig. 13). In other characteris- tics they do not vary greatly from specimens taken in the mountains. The relationships be- tween those populations previously referred to as hammondii and multiplicatus show a close relationship but do show differences that suggest, at least in Chihuahua populations, that distinctions between mountain popula- tions (7,000-9,000 ft) and those at lower ele- vations can be made on the basis of skin tex- ture, if not on other characteristics (Fig. 13). Remarks. — In 1875 Cope described as a new species Spea stagnalis. The type locality is northwestern New Mexico on the Eocene Plateau. This population has been considered a part of the subspecies occurring in the lower southern valleys extending from southwest Texas west through southern New Mexico, Arizona, and the adjoining Mexican states of Chihuahua and Sonora. The question arises as to whether S. stag- nalis belongs to the southern populations or if indeed those populations in northern New Fig. 14. Spea stagnalis Cope: cotype USNM 25335, northwest New Mexico, dorsal view of skull. Mexico belong to the northern species S. intermontana that have been listed as occur- ring in the Upper Colorado Basin and the Great Basin. An examination of the cotype USNM 25335 (that was figured by Cope 1875 in Yarrow, Plate 25, Figs. 6-8) determined that the cotypes had the same cranial charac- ters as other populations in the Spea ham- mondii complex (Fig. 14). By recognizing the California populations as Spea h. hammondii, the eastern populations (Arizona, New Mex- ico, Texas, and those in the adjoining states of Mexico) must now be recognized as Spea h. stagnalis Cope. The distribution of the species of Spea is uncertain in the headwaters of the Rio Grande and the Upper Colorado Basin. It appears that bombifrons extends well into the Upper Basin. A further study including populations in the adjoining areas north of h. stagnalis may provide a clarification of distribution as well as any further subspeciation that might have occurred in this rugged area consisting of mountains and plateaus dissected by deep canyons and isolated by deserts. To complete such a study is beyond the scope of this report. This report on the species S. hammondii is preliminary not only to an extensive study of this widespread species but also to a study of the entire genus Spea. Spea hammondii multiplicata Cope Mexican Spadefoot Scaphiopus multiplicatus Cope, 1863, Proc. Acad. Nat. Sci. Philadelphia 15: 52. Scaphiopus hammondii multiplicatus: Kellogg, 1932, U.S. Nat. Mus. Bull. 160:22-24. Scaphiopus nmltiplicatus. Smith and Taylor, 1948, U.S. Nat. Mus. Bull. 194: 36. Chihuichupa, 35 (BYU 14388-403, 15391-409). Cerocuhui, 2 (BYU 15502-3). 2 mi S Creel (along road to La Bufa), 6 (BYU 15598, 17801; UAZ 31220-23). 3.2 mi E Agostadero de Aguire, 1 (UAZ 46815). 60 Great Basin Naturalist Vol. 49, No. 1 Fig. 15. Scaphiopus couchi: BYU 15839, collected 2 mi SE Colonia Juarez, 30 July 1958, S-V 69.5. The distribution of multiplicata was consid- ered by Smith and Taylor (1948) to be on "the plateau of central Mexico from Durango and Zacatecas southward." Conant (1975) did not recognize multiplicata and extended the range of hammondii to occupy not only south- western United States but also the area previ- ously designated for multiplicata by Smith and Taylor (1948). Stebbins (1985) apparently accepted the conclusions of Brown (1976) and replaced hammondii with multiplicata for es- sentially the same area indicated by Conant. The relationship of these two taxa is close and thus warrants additional study. Until ad- ditional investigation establishes the degree of relationship, I will not attempt to further define the taxonomic status of the Mexican populations of S. h. multiplicata. Family Bufonidae Genus Bufo Laurenti The state of Chihuahua contains nine spe- cies of Bufo. Some have entered the state from the north and east as a part of the desert habitat (cognatus, debilis, punctatus, specio- sus, and woodhousei). Others occur in the mountains of the south (simus), the northwest (microscaphus), and the west (marinus and mazatlanensis). Thus, this large state, with its eastern deserts, central foothills and low ranges, and western mountains and deep western can- yons, provides a variety of habitats singular to Chihuahua. Within this diverse area is per- haps the largest assemblage of bufonid toads to be found in Mexico. In this report no attempt is made to cite all the synonyms for these bufonid species. These can be found in the reports of Cope (1889), Kellogg (1932), or Smith and Taylor (1948). Bufo cognatus Say Great Plains Toad Bufo cognatus Say, 1823, Long's Exped. Rocky Mts., Vol. 2: 190. Footnote: Arkansas River, Prowers County, Colorado. Bufo dipternus Cope, 1879, Amer. Nat. 13. 437. Fort Benton, Choteau County, Montana. Southern outskirts of Cd. Chihuahua, 4 (BYU 10457-60). 13 mi E Rancho Flores Magon, 7 (BYU 13955-61). Approx 15 mi SE Nuevo Casas Grandes, 2 (BYU 14109, 15838). 25 mi SE Las Varas (Mennonite Village), 3 (BYU 15433-35). Vallev Road approx halfway between Babicora and Gomez Farias, 9 (BYU 15564-67, 15757-61). Hidalgo del Parral, 6 adults and 21 untagged small ones (BYU 15558-63). 13.2 mi S Nuevo Casas Grandes, 2 (UAZ 34444-5). 4.9 mi N El Arco (36.5 mi S El Sueco), 2 (UAZ 34386, 34397). 2.0 mi N Villa Ahumada, 5 (UAZ 36281-2). 4.7 mi S Galeana, 2 (UAZ 36281-2). 1.0 mi NW Temosaehic, 1 (UAZ 34832). 28 mi N Cd. Chihuahua, 1 (UAZ 34206). 3.7 mi S Buenaventura, 1 (UAZ 36280). 28 mi S Cd. Juarez, 1 (UAZ 34449). 5 mi N Cerro Campana, 6 (MVZ 70678-83). Ojo de Laguna (25 mi S Gallego), 11 (MVZ 72764-80, 72784-7). 2 mi S Santa Clara, 4 (MVZ 70676-7, 72762-3). Smith and Taylor (1948) list the following locali- ties: Rio Santa Maria near Progreso, near Villa Ahumada, 3 mi E Carmen and Colonia Juarez. The above records indicate that the distri- bution of this species is in central Chihuahua January 1989 Tanner: Chihuahua Amphibians 61 primarily between the western mountains and eastern desert. Smith, Williams, and Moll (1963) report two specimens from Cuchillo Parada taken after a heavy rain. This locality is near the Rio Conchos in northeastern Chi- huahua. Bufo debilis insidior (Girard) Western Green Toad Bufo debilis Girard, 1854. Proc. Acad. Nat. Sci. Philadel- phia 7: 87. Bufo debilis insidior: Smith, 1950, Misc. Publ. Univ. Kansas, Mus. Nat. Hist. 1: 75. 13 mi E Rancho Flores Magon, 7 (BYU 13948-54). 15—30 mi SE Nuevo Casas Grandes along Hwv 10, 6 (BYU 14108, 14110-14). 27. 6 mi N Villa Ahumada (Hwy 45), 3 (UTEP 2059, UAZ 34830-1). 8.4 mi N Villa Ahumada, 1 (UAZ 11471). 3.9 mi N Villa Ahumada, 1 (UAZ 34472). 31.0 mi N Villa Ahumada, 1 (UAZ 34473). 2.0 mi N Villa Ahumada, 2 (UAZ .34828-9). 11.8 mi N El Arco Iris (31.3 mi S Sueco), 1 (UAZ 34471). 5 mi N Cerro Campana, 1 (MVZ 70647). 7 mi N 3 mi E Cerro Campana, 9 (MVZ 70648-56). Ojo de Laguna, 10 (MVZ 72752-61). During and immediately after heavy sum- mer rains, this species may appear in great numbers along roads in Durango and Chi- huahua. We collected 34 specimens 18.8 mi north of Durango City after a heavy evening rain on 24 July 1958 and could have collected a bushel from one roadside pond. This species was unusually abundant, and we found ourselves in a community of toads and frogs the likes of which I had not seen before, but which I experienced again on the night of 28 July on our trip from Ciudad Chi- huahua to Colonia Juarez. This species was seen along the road (Hwy. 45 and 10), at times in great numbers, but was particularly numer- ous in Chihuahua from Buenaventura to near Nuevo Casas Grandes. I have followed Smith (1950), Schmidt (1953), and Conant (1975) in placing insidior as a subspecies of debilis. Determining whether there are two species, one east (de- bilis) and one west (insidior) as originally de- scribed by Girard (1854) and accepted by Smith and Taylor (1948), or whether this widespread group is indeed a series of subspe- cies must await an intensive study of the en- tire assemblage of available specimens. At present there is an abundance of material to be studied. This includes specimens from Zacatecas north through Durango, Chi- huahua, Sonora, and into Arizona and New Mexico. The eastern segment extends from Tamaulipas north into Texas. A careful com- parison of these populations may lead to an understanding of this interesting group of toads. Kellogg (1932) lists two specimens (USNM 2622) from Chihuahua and notes the designa- tion in the museum catalogue as the cotypes of Bufo insidior. No specific locality is listed in the catalogue, only Chihuahua, Mexico. Bufo marinus Linnaeus Giant Toad Rana marina Linnaeus, 1758, Svstema Naturae, Ed. 10, 1:211. Bufo horribilis Wiegmann, 1833, Isis von Oken 26: 654-655. Bufo marinus Kellogg, 1932, U.S. Nat. Mus. Bull. 160- 31-21, pp. 53-57, Fig. 11. Easteal 1986, p. 395.1-4. Urique, 10 (BYU 14355-64). We arrived at Urique on the evening of 14 July and were soon greeted by large toads in the streets and along the nearby river. A moderate rain shower had occurred in the afternoon. Thus, it was warm and humid and we were soon aware that on such evenings toads and fruit bats were active. In the dim light of a nearby dwelling and from a radius of a few yards, 12 large toads were counted in less than one minute. As I moved about, their numbers did not seem to diminish. The next morning the toads were gone and were re- placed by curious youngsters. Thus began three days in the old mining town of Urique. Bufo punctatus Baird & Girard Red-Spotted Toad Bufo punctatus Baird & Girard, 1852, Proc. Acad. Nat. Sci. Philadelphia 6: 173. 36 mi S Ciudad Juarez, 1 (BYU 15207). Ciudad Chihuahua, 1 (BYU 14252). Cerocahui, 4 (BYU 14342, 14367, 14576-7). Cuitaco, 3 (BYU 14527-29). Urique, 1 (BYU 14341). Crossing at Los Chales, 5 (BYU 15704-8). 35 mi SW Nuevo Casas Grandes, 1 (BYU 15454). Colonia Juarez, 3 (BYU 17045-47). Sierra del Nido Complex, 4.7 mi (by road) W Encinillas, 1 (UTEP 546). 7.5 mi (by road) ESE Buenaventura, 1 (UTEP 547). 62 Great Basin Naturalist Vol. 49, No. 1 NW La Junta at junction of road to Ciudad Guer- rero, 1 (UTEP 558). 6.5 mi (by road) NE Ciudad Guerrero, 1 (UTEP 559). Cd. Chihuahua, 1 (UAZ 11296). Coyame, 5 (UAZ 34965-69). 8.3 mi W Covame (hwy to Aldama), 1 (UAZ 34971). 11.2 mi W Coyame, 7 (UAZ 34972-78). 20.6 mi W Coyame, 1 (UAZ 34979). 15.0 mi S Nuevo Casas Grandes, 1 (UAZ 34470). 6.3 mi S Buenaventura, 4 (UAZ 36276-79). 6.5 mi NW Yepomera, 1 (UAZ 34826). 4.3 mi NW Yepomera, 1 (UAZ 34827). 2.7 mi S Milpillas (road to San Antonio, Sonora), 1 (UAZ). 24.6 mi S Ahumada, 7 (MVZ 52167-73). Hidalgo del Parral, 1 (MVZ 58736). 5 mi N Cerro Campana, 18 (MVZ 68764-75, 70662-67). 2 mi S Santa Clara, 5 (MVZ 70657-61). This species is seemingly widespread in Chihuahua, based on the above collection records. It should be noted here that much time was spent during the dry season when this species may not have been abroad. Our records place them in the desert valleys and foothill area usually between 5,000 and 7,000 ft in suitable habitats in central and western Chihuahua. We did not find them at the higher elevations. Smith et al. (1963) report two specimens from eastern Chihuahua, one from near Bene- ficio and one from near Alamo. Both were near the Bio Conchos. Bufo speciosus Girard Texas Toad Bafo compactilis Wiegmann, 1833, Isis, p. 661 (part). Bufo speciosus Girard, 1854, Proc. Acad. Nat. Sci. Philadelphia 7: 85-87. Bufo compactilis speciosus: Smith, 1947, Herpetologica 4(1): 7-13. Bufo speciosus: Conant, 1975, Field Guide: p. 313, map 268; Stebbins 1985, Field Guide: pp. 74-75, map 36. 1.7 mi NW Matachic, 1 (UAZ 34387). 2.5 mi NW Matachic, 1 (UAZ 34823). 2.6 mi SE Matachic, 2 (UAZ 34388-89). The record for Guadalupe y Calvo may be in question. Webb and Baker (1984) collected B. microscaphus in this general area, and it is suspected that the Kellogg (1932) citation (USNM 47244) may be a confusion of specio- sus (compactilis) with microscaphus. The lat- ter is commonly found in the mountains, whereas speciosus is in the grassland plain habitat at lower elevations. Valleys in the gen- eral area of Matachic, Yepomera, and be- tween Babicora and Gomez Farias are suit- able habitats. Kellogg (1932) also listed Bufo compactilis for Colonia Garcia, and Meadow Valley, where Bufo microscaphus occurs. Stebbins (1985) did not recognize the records of Van De vender and Lowe (1977) for Matachic and confined the distribution of spe- ciosus to a narrow edge of eastern Chihuahua. Conant (1975) extended the range into central Chihuahua, approaching the area of Mata- chic, but gave no locality records. Originally, the valley between Gomez Farias and Babicora, to the northwest, was a grassy plain that served as a large cattle ranch. This may have been the case for other nearby valleys such as Matachic and Madera. In their original state these valleys seemingly served as havens for many species since they were at elevations above the hot, dry deserts and yet not in the mountains. On examination of the species listed in the report of Van Devender and Lowe (1977) one is impressed with the importance of this general habitat as a refugium as the deserts slowly claimed pre- vious habitable areas after the last pluvial period. Bufo microscaphus mexicanus Brocchi Western Madre Toad Bufo mexicanus Brocchi, 1879, Bull. Soc. Philom., Ser. 1, 3: 23-24. Bufo microscaphus Cope, 1867, Proc. Acad. Nat. Sci. Philadelphia 18: 301. Bufo woodhousii microscaphus Shannon, 1949, Bull Chicago Acad. Sci. 8(15): 301-12. Bufo microscaphus mexicanus Webb, 1972, Herpetolog- ica 28(1): 1-6. Near Colonia Juarez, 2 (BYU 13516-17). Chuhuichupa environs, 16 (BYU 13974, 14129- 31, 14234-37, 14239-42, 15390, 15485). Hope Valley at junction of Rio Juarez, 1 (BYU 14238). Mouth of Tinaja Canvon near Colonia Juarez, 1 (BYU 15581). 12 mi SE Babicora, 1 (BYU 15763). Black Canyon, approx 8 mi W Chuhuichupa, 3 (BYU 14253-55). 26 mi W San Juanito, on road to Maguarichic, 1 (BYU 16957). 11 mi W San Juanito, 1 (BYU 16878). San Juanito, 2 (BYU 15768, 17036). 2 mi SE Creel, 5 (BYU 15638-40, UAZ 37365-6). January 1989 Tanner: Chihuahua Amphibians 63 25.5 mi S Creel, 2 (BYU 17048, 17050). Cerocahui, 7 (BYU 14543-48, 14578). Cuiteco, 3 (BYU 14294, 15507, 15781). 1 mi W La Laja, 1 (BYU 15872). Maguarichic, 2 (BYU 16936-37). Along Rio Urique near Carmen Bridge, 9 (BYU 22703-22711). Rio San Miguel at San Juan de Dios, 1 (BYU 22623). Rio Gavilan, 1 mi SE Gavilanisto, 1 (UAZ 8960). Basaseachic Parque Nacional, 2 (UAZ 47005, 47240). Yaguirachic (130 mi W Cd. Chihuahua), 15 (MVZ 65982-93, 65996-8). This is a widespread species in the western mountains of Chihuahua. It descends into the mouths of some canyons but has not been found in the desert valleys. The distribution of mexicanus, at least in Chihuahua, is appar- ently restricted to elevations at or above 6,000 ft. This is in contrast to microscaphus, which is found along desert streams in the south- western United States (northwestern Ari- zona, and adjoining California, Nevada, and Utah) at much lower elevations. At none of the collecting localities were large numbers seen. The nine specimens taken near the Carmen Bridge were adults and juveniles. Since microscaphus Cope (1867) and mexi- canus Brocchii (1879) were described, various taxa have been used to include representa- tives of this species. Thus, for the past hun- dred years, populations of this species have been assigned as a part of Bufo compactilis, Bufo columbiensis , Bufo woodhousei, or as subspecies within one of the above or with a reference to such subspecies as californicus or speciosus. The close resemblance of microsca- phus to other Bufo species in the southwest- ern United States and northwestern Mexico contributed to the taxonomic confusion. Shannon (1949) recognized the distinctness of microscaphus, separated it from com- pactilis, but retained it as a subspecies of Bufo woodhousei. The detailed study by Webb (1972) has presumably extricated microsca- phus from a confused and uncertain past and described it as a unique species, sharing simi- lar characters with other sympatric species but with morphological characteristics such as uniqueness of foot tubercles, size, shape, and position of parotoid gland, warts on back low, rounded and mostly smooth, and color pat- tern of light brownish spots on the body and without spotting on the venter. The nomenclatural review and description of microscaphus and its subspecies mexicanus by Webb (1972) has not only clarified the posi- tion of microscaphus and its subspecies but has also improved our understanding of its distribution in Mexico. Bufo woodhousei australis Shannon & Lowe Southwestern Woodhouse Toad Bufo woodhousei Girard, 1854, Proe. Acad. Nat. Sci. Philadelphia 7: 86. Bufo frontosus Cope, 1866, Proc. Acad. Nat. Sci. Philadelphia 18: 301. Bufo woodhousei woodhousei Smith, 1934, Amer. Midi. Nat. 15: 449-57. Bufo woodhousei australis Shannon & Lowe, 1955, Her- petologica 11: 185. Garcia, 1 (BYU 160). Colonia Juarez, 2 (BYU 13581-2). Colonia Duhlan, 2 (BYU 13971, 15455). Rio Bavispe, below Tres Rios, 2 (BYU 13458 and 13497). Nuevo Casas Grandes, 1 (BYU 15455). Yepomera, 1 (UTEP 2052). 2 mi N Yepomera, 1 (UAZ 34442). 6.3 mi NW Cd. Guerrero, 1 (UAZ 34443). Yepomera, 12 (UAZ 34372-79, 34340-1, 34381, 34824). 6.9 mi NW Yepomera, 1 (UAZ 34645). 5 km N Yepomera, 1 (UAZ 36285). 2 mi SE Matachic, 2 (UAZ 12742, 34384-5). Matachic, 1 (UAZ 34342). 4.5 mi SE Temosachic, 1 (UAZ 34382). 50 mi SE Galeana (Hwy 10), 1 (UAZ 34669). 2 mi N Janos, 3 (UAZ 34956, 36283-4). 6.5 mi N Nuevo Casas Grandes, (Hwv 10) 1 (UAZ 36286). 12.7 mi NW Gran Morales, 4 (MVZ 52179-82). Rio San Pedro, Meoqui, 2 (MVZ 52183-4). Minaca (Rio Papigochic), 5 (MVZ 52184, 58727- 30). Ojo de Laguna, 1 (MVZ 75873). 2 mi S Santa Clara, 8 (MVZ 70668-72, 70674-5, 72751). The above records place this species in or near the Sierra Madre and north of the Rio Papigochic. Conant 1974 (1977) extends its range southeast along the Rio Grande and south along the Rio Couchos into south cen- tral Chihuahua. Bufo woodhousei has been collected in streamside habitats and in the higher valleys near the western mountains where permanent streams are fed by springs. We did not find them in the mountains of western Chihuahua. 64 Great Basin Naturalist Vol. 49, No. 1 Fig. 16 Bufo simus Schmidt: BYU 17134, collected 25.5 mi S Creel, Chihuahua, 18 July 1960. Bufo simus Schmidt Bufo simus Schmidt, 1858, Denkschr, Akad. Wiss. Wien, math-nat. Class. 14: 254-55). Bufo intermedins Gunther, 1858 (1859), Catalogue of the Batrachia Salientia in the collection of the British Museum, p. 140, pi. 9, Fig. FF. Cerocahui, 1 (BYU 14542). 25.5 mi S Creel, 5 (BYU 17049, 17134-37). SW Chihuahua, Rio San Miguel, 1 (BYU 39373). We found this species only in the Rio El Fuerte basin and near streams during the rainy season. Smith and Taylor (1948) listed this species for Chihuahua but gave no locali- ties. Webb and Baker (1984) did not report it for the Cerro Mohinora region during the dry season of late May (21-30). Its occurrence in adjoining Durango and Sinaloa suggests a dis- tribution in the mountains of southern Chi- huahua, at least southwest and southeast of Creel (Fig. 16). Bufo mazatlanensis Taylor Mazatlan Toad Bufo mazatlanensis Taylor, 1939 (1940), Univ. Kansas Sci. Bull. 26: 492-494, p. 53, Fig. 1. Urique, 14 (BYU 14343-54, 15556-7). 3 mi NN W Moris on Rio Santa Maria (ca 800 m), Rio Mayo drainage, 1 (UAZ). This species was abundant along the Rio Urique and in the streets of Urique on the evening of 14 July 1958. We were at Urique for three nights and two and one-half days. Between rain showers we collected speci- mens and tried to sample as many habitats as possible. It was at Urique that we first wit- nessed the emergence of multitudes of Bufo at dusk each evening. While we were at Urique, the weather was hot and humid. Mornings were clear, but light rain showers occurred in the afternoons, contributing to the humidity and providing a proper environment for the toads each evening. Family Leptodactylidae Genus Eleutherodactylus Dumeril & Bibron Eleutherodactylus tarahuniaraensis Taylor Tarahumara Barking Frog Eleutherodactylus tarahumaraensisTayXor , 1940, Copeia 1940: 250-253. Eleutherodactylus augusti: Bogert and Oliver, 1945, Bull. Amer. Mus. Nat. Hist. 83: 405-6. Eleutherodactylus tarahuniaraensis. Zweifel, 1956, Amer. Mus. Novitates. No. 1813: 28-33. 2 mi E Cerocahui, 1 (BYU 14385). Maguarichic, 1 (BYU 16926). 6 km WNW Ocampo, 3 (UAZ 24741-2 and 47237). The type (EHT-HMS 13008) collected at Mojara- chic by Irving Knobloch. 7 mi SW El Vergel (Lagunita), 1 (MVZ 58797). Yaguirachic, 2 (MVZ 65974-5). The specimen taken at Cerocahui was un- der a rock in a moist area. It was spotted green on the back, had larger feet and eyes than other frogs of similar size, and had a snout- vent length of 23.5 mm. The specimen from Maguarichic was taken two years later (1960) at the same time, 13 July, and is 30 mm in S-V (Fig. 17). Both specimens are juvenile, and in neither locality could we find other speci- mens. Both were collected at a moist hillside habitat with permanent water at least 1/2 mile away. It should be noted that during the rainy season seeps occur in hillside depressions. The figure by Zweifel (1956) of a specimen from southern Chihuahua has essentially the same color pattern as the juveniles listed above. Spotting is more diffused or less pat- terned than appears to be the case for other January 1989 Tanner: Chihuahua Amphibians 65 50 cm total length [TL]) are presently common in Lake Mohave, Arizona-Nevada, and larvae of the species (<15 mm TL) are seasonally abundant. Size- groups between the larval and adult life stages are, however, essentially absent from col- lections, despite intense sampling. Adults comprised an average of —25% of total fishes caught in annual trammel net samples be- tween 1975 and 1988 (Minckley 1983, Minck- ley and Marsh, unpublished data). Larvae occupy the littoral zone of Lake Mohave (Bozek et al. 1984, Marsh and Langhorst 1988), where 10-100 or more can be attracted to a strong light in a few minutes at night (Langhorst and Marsh 1986). They rarely occur in open water of the reservoir; Langhorst and Marsh (1986) captured only a single specimen in 22 tow-net hours in 1985, although larvae were at the same time com- mon in near-shore habitats. Four juvenile specimens (three preserved, 33 to 54 mm TL; ASU 11567 and 11568), collected by AZGFD personnel in July 1987 (T. Liles, AZGFD, personal communication), are the only natu- rally spawned juveniles recorded from Lake Mohave since the 1950s. In marked contrast, except for a small resi- dent population in Senator Wash Reservoir, California (Medel-Ulmer 1980, Ulmer 1987), there are confirmed records since 1962 for only 42 adult razorback suckers from the en- tire lower Colorado River mainstream and as- sociated habitats downstream from Davis Dam (Fig. 1), despite intensive fisheries sur- veys in that area. Sixteen adults were from Lake Havasu proper: five averaging 56.9 cm TL were caught in 1962; four >50 cm TL were observed in 1975; three averaging 65.4 cm were electrofished in 1976; one (56.9 cm) was collected from the Bill Williams Arm of the reservoir in 1979; two (unmeasured) were caught by anglers in 1984 (Ulmer and Ander- son 1985); and a single fish 50.6 cm TL was gill-netted in 1986 (M. Giusti, CADFG, personal communication). Riverine reaches yielded 23 individuals: 12, all >50 cm TL, but unmeasured, were taken by various means from Blythe, California, downstream to Impe- rial Reservoir from 1969 to 1985 (Ulmer and Anderson 1985); nine others, mostly >60 cm TL, were angled, electrofished, trammel- netted, or observed in the Needles-Topock Gorge reach from 1972 through 1985 (Minck- ley 1983, Ulmer and Anderson 1985); and two, 57.2 and 61.0 cm TL, were trammel- netted from Laughlin Lagoon, Nevada, an ar- tificial backwater about 8 km below Davis Dam in 1986 (M. Burrell, NVDOW, personal communication). An additional three adults, —50 cm TL, were caught from the Central Arizona Project (CAP) Granite Reef Aqueduct in October 1986 (USBR 1986), which began withdrawing water in 1983 from the Bill Williams Arm of Lake Havasu. Two of the last were 25+ years of age, as determined by otolith analysis (original data; following methodology of McCarthy and Minckley 1987), and thus originated from Lake Havasu. There are no indications that adult razorback suckers in the lowermost Colorado River are occurring less frequently in the 1980s than in the 1960s, which is likely a reflection of low adult mortality and individual longevity (to at least 44 years; McCarthy and Minckley 1987). Larval razorback suckers are as rare as adults in the lower Colorado River down- stream from Lake Mohave. None was found in shoreline surveys with bright light at night in Lake Havasu in 1988. Razorback suckers ac- counted for only 0.56% of 6,617 larval speci- mens caught in tow-net samples in Lake Havasu and upstream riverine reaches in 1985 and 1986 (Marsh and Papoulias, in press). Eight individuals were taken in 1985 and 29 in 1986. Although catch per unit effort varied between years and among stations, similar abundances were indicated in riverine and reservoir habitats, and no areas of larval con- centration were evident (Marsh and Papou- lias, in press). Two larval razorback suckers, — 15 mm TL, were also identified among 5,036 larval specimens from the CAP canal in 1987 (G. Mueller, USBR, personal commu- nication). Twenty-four juvenile razorback suckers, — 15 to 37. 1 cm TL, fish of sizes not otherwise known from Lake Mohave or elsewhere in the January 1989 Marsh, Minckley: Razorback: Sucker 73 \NV V-ML-J- 1 CokDraOo/ + ca \, r Needles *\ ARIZONA Phoenix Sa " 1 N. Biyine */ OMR/"^--— \ 7* Vuma \ £ — C"-— . ._! \ \ MEXICO \ Legend O Larva O Juvenile • Adult — •— Major Canals Central Arizona Project Canal Fig. 1. Sketch map of the lower Colorado River, western North America, showing place names referred to in text and capture locations since 1962 for larval, juvenile, and adult razorback suckers. 74 Great Basin Naturalist Vol. 49, No. 1 Colorado River basin and thus the major sub- jects of this report, have been caught be- tween 1974 and 1988 downstream from Lake Havasu. All but one were from the extensive system of artificial waterways that have been constructed for irrigated agriculture. In Arizona, canals near Parker produced two fish (32.3 and 37.1 cm TL) in 1980, two specimens (each 30 cm) in 1981, and four aver- aging 35.2 cm in 1986 (Minckley 1983, Ulmer and Anderson 1985, S. Yess, USFWS, per- sonal communication). Intake of water for the Parker area canal system is at Headgate Rock Dam, about 23 km downstream from Parker Dam. In California, one specimen (23.4 cm) was angled from a canal east of Palo Verde in 1983, which obtains its water from Palo Verde Diversion Dam, 21 km upstream from Blythe. Farther south, in Imperial Valley, the Coa- chella Canal (Fig. 1), its laterals, and its equal- izing reservoirs produced three fish (average 34.8 cm TL) in 1984, and four others averag- ing 35.5 cm in April 1985 (Ulmer and Ander- son 1985). Five young (—15 cm TL) were taken in 1973 and 1974 from the East Highline canal and adjacent ponds at Niland (St. Amant et al. 1974), a single fish measuring 22.5 cm fork length (FL x 1.085 ± 0.021 = 24.4 cm TL; unpublished data) was taken from the canal at Niland in 1974 (Ulmer and Anderson 1985), and another —30.5 cm long was cap- tured there from a canal-fed pond in Decem- ber 1985 (E. Milstead, Niland, California, personal communication). Intake of water for the Imperial Valley is mostly through the Ail- American Canal, which originates near Impe- rial Dam (Fig. 1). The only small fish from the mainstem Colorado was a 35. 1-cm individual captured 16 km downstream from Parker, Ari- zona, between Headgate Rock and Palo Verde Diversion dams, in summer 1987 (Langhorst 1988). That specimen was two years of age by otolith examination. Assuming all these juvenile fish exhibited growth rates similar to those from Lake Mo- have, hatchery ponds (McCarthy and Minck- ley 1987), and a variety of other waters where reintroduced populations have been studied (Marsh, in press, Marsh and Minckley, un- published data), none was more than five years old. Only the eight fish captured in 1983 or later (three fish in 1984 and five others in 1985) from canals and other waters confluent with the Ail-American Canal (Fig. 1) could have been derived via West Pond from artifi- cially propagated stocks (see below); all others were wild fish. Reintroductions, 1980-88 The first razorback sucker reintroduction to the lower Colorado River area was in 1980. It consisted of 17 hatchery-produced adults (average 32.5 cm TL; 1974 year class, Toney 1974) and 3 Lake Mohave adults (average 56.6 cm TL, ages unknown) into the isolated West Pond, Imperial County, California (Fig. 1; W. Loudermilk, CADFG, personal commu- nication). An unknown number of progeny of Senator Wash Reservoir fish, artificially prop- agated and reared by CADFG personnel, also were stocked in West Pond between 1981 and 1983 (L. Ulmer, CADFG, personal communi- cation). In November 1983, 457 razorback sucker juveniles (average 95 mm TL) from Lake Mohave broodstock also were stocked into an artificial rearing enclosure constructed in West Pond by USBR; samples of those fish averaged 115 mm (N = 5) in December 1983 and 156 mm (N = 5) in January 1984 (Ulmer, personal communication). West Pond and the enclosure were not again monitored until 1988, when no razorback suckers were en- countered. Because water from West Pond is pumped into the Ail-American Canal, these stocked fish could have contributed to the eight post- 1983 juvenile occurrences downstream in the confluent Coachella and East Highline canals or their adjacent ponds and reservoirs. Fur- thermore, progeny of Senator Wash Reservoir adults, artificially propagated in spring 1983, were also reared in aquaria in Blythe, Califor- nia, and later transferred for grow-out in local ponds. A total of 57 survivors (average 28.5 cm TL for 39 measured) was stocked into the Colorado River mainstem near Blythe in April 1985 (Ulmer, personal communication). A dozen others (unmeasured) from the same group were stocked into an isolated pond on federal lands in February 1986 (Ulmer, per- sonal communication). These last two stock- ings could not have contributed to subsequent captures from the areas of Parker, Arizona, or Palo Verde, California, because they were downstream from barriers created by Head- gate Rock and Palo Verde Diversion dams (Fig. 1); however, fish could have made their way downstream to Imperial Valley. January 1989 Marsh, Minckley: Razorback Sucker 75 The first major reintroduction of razorback suckers in the mainstream was in March 1986, when nearly 1.4 million larvae (10-18 mm TL) were released by CADFG and USFWS at various localities along the Colorado River from Devil's Elbow and Blankenship Bend (in the Topock Gorge area upstream from Lake Havasu, Fig. 1), downstream to Imperial Na- tional Wildlife Refuge near Yuma, Arizona. AZGFD and USFWS placed an additional 70,000 juveniles (—5.1 cm TL) into the Colo- rado River near Parker, Arizona, in May 1986, and CADFG stocked 4,163 juveniles (-20 cm TL) in the same area in October-November 1986. Since then, more than a million addi- tional larvae and juveniles have been stocked downstream from Parker Dam (Langhorst 1988). These last stockings were all conducted later than the collections of all but one (Niland, December 1985) of the juveniles tabulated above, and of larvae reported by Marsh and Papoulias (in press). Captures, 1987-88 Excluding the 1987 collection (Langhorst 1988) of a two-year-old individual in the Colo- rado River mainstream upstream from Palo Verde Diversion Dam (thus wild-hatched), a total of 41 juvenile razorback suckers was cap- tured from canals downstream from Parker, Arizona, on the east (Arizona) side of the Colo- rado River in 1987 and 1988 (S. Yess, US- FWS, personal communication). Thirty-eight fish caught in 1987 averaged 28.8 cm, and three taken in 1988 averaged 45.1 cm TL. Unfortunately, none from the first group was aged, but based on mean size they could have been one-year-old fish and therefore origi- nated, at least in part, from the 1986 stock- ings. None could have been derived from ear- lier reintroductions, all of which were placed downstream from Headgate Rock Dam (Fig. 1). Fish of the second (1988) group had otolith ages of three, four, and seven years, having hatched, respectively, in spring 1981, 1984, and 1985. These were naturally produced wild fish, since dates of hatching do not corre- spond with those of any reintroductions in areas from which they could have moved to the collection sites. Discussion and Summary Captures between 1974 and 1988 of at least 19 young, wild-hatched razorback suckers in the lowermost Colorado River system down- stream from Lake Mohave provide convincing evidence of potential recruitment to that population. Numbers recruited nonetheless appear insufficient to maintain a population of adults, since fish of reproductive size are ex- ceedingly rare and scattered in distribution (42 adult individuals recorded in the period 1962-1988). Further, artificial canals where most young fish were recorded may act not only as a refuge for early development but as death traps later, during annual dewatering for maintenance of the irrigation system. Be- cause of this, potential recruits may ultimately be lost to the population. Waterways of Colorado River irrigation sys- tems consist of two major components, canals and drains (or wasteways). Canals vary down- ward from maximum flows of 400 m /sec. Water is withdrawn by gravity at diversion structures (e.g., Headgate Rock, Palo Verde, Imperial, and Laguna diversions) or through pumps (CAP and Colorado River Aqueduct intake facilities; USBR 1980; Fig. 1). Small laterals, which deliver water to agricultural fields or other points of use, are the least permanent, carrying water for only a few days or hours per month. Most canal habitats from which razorback suckers have been taken are of intermediate sizes that are dewatered at least annually for cleaning and repairs. Some of the largest canals may not be dewatered for periods of years. Periodic cleaning and repair of canals is typically in the irrigation off-season, usually December or January. Fishes are decimated by dewatering and mechanical cleaning, and few survive (Marsh and Minckley 1982). How- ever, razorback suckers spawn early, in late January through March, and larvae are thus available (generally from February through April; Marsh and Langhorst 1988) to colonize canals as they are placed back in service. De- pleted populations of potential predators en- hance larval survival, and razorback sucker growth rates (to 25+ cm in six months; unpub- lished data) are such that they rapidly grow out of predation range of small, abundant, nonnative predators (e.g., green sunfish, Lep- omis cyanellus) and attain capabilities suffi- cient to avoid larger species (largemouth bass, Micropterus salmoides, and ictalurid catfishes, especially flathead catfish, Pylo- dictis olivaris). Further, annual drainage of 76 Great Basin Naturalist Vol. 49, No. 1 canals makes young razorback suckers suscep- tible to collectors. Now that the species is known from such places, biologists are alert for their occurrence. Interest among biolo- gists along the lower river and extensive infor- mation exchange stimulated by active reintro- duction programs have contributed to and increased the probability that razorback suck- ers will be noted. Drains transport excess water used for leaching of salts from agricultural lands back to the river. They are fed by over-surface flow during irrigation cycles and subsurface perco- lation the rest of the time, which results in slow-moving, enriched aquatic habitats that are often densely vegetated by algae and macrophytes and may be characterized by chemical and thermal extremes (Minckley 1979). Drains are far more permanent than canals. Large drains rarely dry and are only sporadically disturbed by cleaning and maintenance operations in ways, such as dredging, that do not involve dewatering. There are, however, no known records for razorback suckers from drains. Origins of Recruits Larvae are most likely passively entrained into canals. Currents at intakes are substan- tial, and larval razorback suckers tend to be near shorelines, at least in reservoirs (Lang- horst and Marsh 1986) and also in the only historic collection of aggregated larvae and juveniles of the species recorded from the mainstream Colorado River in 1950 (R. R. Miller, in Sigler and Miller 1963). Drifting catostomid larvae (Catostomus insignis, Pan- tosteus clarki) in the Gila River, New Mexico, were concentrated by a factor of 6.5 near banks compared with samples in midstream (Bestgen et al. 1987). Reintroduced juvenile razorback suckers also show a marked procliv- ity to move downstream (Brooks 1985). Such behavior would obviously enhance the proba- bility of encountering a withdrawal point. The absence of razorback suckers in drains may result from a lack of sampling. As noted above, these habitats are far more permanent than canals. In addition, they are more com- plex, and thus exceedingly difficult to sample. Drains may also suffer seasonal chemical and physical extremes that are lethal to fishes. They nonetheless often support substantial populations of nonnative species (Minckley 1979, Matter et al. 1986), including cen- trarchids and ictalurids that are demonstrated predators on young razorback suckers (Os- mundson 1987, Marsh and Brooks, in press). Drains furthermore flow into the river, and larval or juvenile razorback suckers may ei- ther not actively ascend against current or may be blocked from ascent by structures designed to prevent headward erosion. Young razorback suckers in the lower Colo- rado River system may have hatched in a number of places. Downstream, the Senator Wash Reservoir population occupies a small (190 ha) pump-storage impoundment, where they behave similarly to fish in Lake Mohave. A small number of adults (estimated popula- tion 54 ± 22 individuals [95% confidence limits] in 1980-81, which averaged —60 cm TL in 1973-74 [10 fish]) spawn and produce larvae each year, which then disappear before achieving juvenile size (Ulmer and Anderson 1985). Some of these could conceivably pass through penstocks of the reservoir and enter downstream intakes that lead to Imperial Val- ley canals. Fish appearing in the Parker, Ari- zona, and Blythe, California, areas could simi- larly originate from reproduction in Lake Havasu (Marsh and Papoulias, in press), pass through epilimnetic penstocks of Parker Dam, and then be diverted into canals at Headgate Rock Dam or Palo Verde Irrigation Diversion, respectively. Other spawning areas are un- known but certainly may exist (Loudermilk 1985). Occurrence in the Parker, Arizona, area in 1987 of 38 juveniles of a size attributable to the 1986 reintroductions, and three in 1988 that were wild fish, underlines a number of needs and factors to consider. First, it is imperative that reintroduced fish be marked, by fin re- moval or with oxy tetracycline, for example, so they may be certainly and readily discrimi- nated from naturally produced individuals. Only in this way can the relative contributions of natural and reintroduction recruitment be evaluated. Second, assuming that some or all fish caught in 1987 were reintroduced, stocked and naturally produced larvae and juveniles must behave similarly, since they both appear to have passed from the river into canals. This provides information that razor- back sucker larvae and/or juveniles drift or move downstream after hatching (or introduc- tion), and likely did so in the natural state. January 1989 Marsh, Minckley: Razorback Sucker 77 Last, survival to the juvenile stage in preda- tor-poor habitat of canals further strengthens the hypothesis (Minckley 1983, Marsh and Langhorst 1988) that attributes lack of recruit- ment by this imperiled species to direct pre- dation by introduced, nonnative piscivores. Literature Cited Bestcen, K R . D L Propst, and C. W Painter 1987. Transport ecology of larval fishes in the Gila River, New Mexico. Proc. Desert Fish. Counc. 17(1985): 175. Bozek. M A . L J. Paulson, and J. E. Deacon. 1984. Factors affecting reproductive success of bonytail chubs and razorback suckers in Lake Mohave. Final report, U.S. Bur. Rec. Contr. 14-16-0002- 81-251, Boulder City, Nevada. 136 pp. Brooks. J E 1985. Annual reintroduction and monitor- ing report for razorback sucker Xyrauchen texanus in the Gila River basin, Arizona, 1985. U.S. Fish Wildl. Serv., Albuquerque, New Mexico. 23 pp. DOUGLAS, P A 1952. Notes on the spawning of the hump- back sucker, Xyrauchen texanus (Abbott). Califor- nia Fish Game 38: 149-155. Johnson, J. E 1985. Reintroducing the native razor- back sucker. Proc. Desert Fish. Counc. 13(1981): 73-79. Jonez. A , J Hemphill, R D Beland, G. Duncan, and R. A. Wagner 1951. Fisheries report on the lower Colorado River. Arizona Game Fish Comra., Phoenix. 40 pp. Jonez, A , and R C Sumner 1954. Lakes Mead and Mohave investigations: a comparative study of an established reservoir as related to a newly created impoundment. Nevada Fish Game Comm., Reno. 186 pp. Langhorst, D R. 1988. A monitoring study of razorback sucker reintroduced into the lower Colorado River. Final report, California Dept. Fish Game Contr. C-1888, Long Beach. 76 pp. Langhorst, D R., and P. C. Marsh 1986. Early life his- tory of razorback sucker in Lake Mohave. Final report, U.S. Bur. Rec. Contr. 5-PG-30-06440, Boulder City, Nevada. 36 pp. Lanigan, S H. and H M Tyus 1988. Population size and status of the razorback sucker in the Green River basin, Utah and Colorado. N. Amer. J. Fish. Manag. Loudermilk, W E 1985. Aspects of razorback sucker (Xyrauchen texanus, Abbott) [sic] life history which help explain their decline. Proc. Desert Fish. Counc. 13(1981): 67-72. Loudermilk, W. E., and L. C. Ulmer. 1985. A fishery inventory with emphasis on razorback sucker (Xyrauchen texanus) status in the lower Colorado River. California Dept. Fish Game, Reg. 5 Inf. Bull. 0012-9-1985. 255 pp. Marsh, P. C In press. Native fishes at Buenos Aires National Wildlife Refuge and Arizona State Uni- versity Research Park, Arizona: opportunities for management, research, and public education on endangered species. Proc. Desert Fish. Counc. 19(1987). Marsh, P C, and J. E. Brooks In press. Predation by ictalurid catfishes as a deterrent to re-estab- lishment of hatchery-reared razorback suckers. Southwest. Nat. Marsh. P C , and D. R Langhorst 1988. Feeding and fate of wild larval razorback sucker. Env. Biol. Fish. 21:59-67. Marsh. P C . and W. L Minckley 1982. Fishes of the Phoenix metropolitan area, central Arizona. N. Amer. J. Fish. Manag. 2: 395-402. 1987. Aquatic resources of the Yuma Division, lower Colorado River. Final report, phase II, U.S. Bur. Rec. Contr. 2-07-30-X0214, Boulder City, Nevada. 300 pp. Marsh, P C ., and D Papoulias. In press. Ichthyo- plankton of Lake Havasu, a Colorado River im- poundment, Arizona-California. California Fish Game. Matter. W J J C Tash.andC D Ziebell 1986. Evalu- ation of the commercial potential of tilapia popula- tions to support a commercial fishery in the lower Colorado River area. Final report, Arizona Game Fish Dept. Contr. 650018, Phoenix. 10 pp. McAda, C W , and R S Wydoski 1980. The razorback sucker, Xyrauclien texanus, in the upper Colorado River basin, 1974-1976. U.S. Fish Wildl. Serv., Tech. Pap. 99: 1-15. McCarthy. M S . and W L Minckley 1987. Age esti- mation for razorback sucker (Pisces: Catostomi- dae) from Lake Mohave, Arizona and Nevada. J. Arizona-Nevada Acad. Sci. 21: 87-97. Medel-Ulmer. L 1980. Movement and reproduction of the razorback sucker (Xyrauchen texanus) inhabit- ing Senator Wash Reservoir, Imperial County, California. Proc. Desert Fish. Counc. 12(1980): 106. Minckley, W L 1979. Aquatic habitats and fishes of the lower Colorado River, southwestern United States. Final report, U.S. Bur. Rec. Contr. 14-06- 300-2529, Boulder City, Nevada. 478 pp. 1983. Status of the razorback sucker, Xyrauchen texanus (Abbott), in the lower Colorado River. Southwest. Nat. 28: 165-187. Osmundson, D B 1987. Growth and survival of Colo- rado squawfish (Ptychocheilus lucius) stocked in riverside ponds, with reference to largemouth bass (Micropterus salmoides) predation. Unpub- lished dissertation, Utah State University, Logan. 179 pp. Sigler. W E , and R R. Miller 1963. Fishes of Utah. Utah St. Dept. Fish Game, Salt Lake City. 203 pp. St Amant, J A , R Hulquist, C. Marshall, and A Pick- ard. 1974. Fisheries section including information on fishery resources of the Coachella Canal study area. Pages 64-88 in Inventory of the fish and wildlife resources, recreational consumptive use, and habitat in and adjacent to the upper 49 miles and ponded areas of the Coachella Canal. Final report, U.S. Bur. Rec. Contr. 14-06-300-2555, Boulder City, Nevada. Toney, D. P. 1974. Observations on the propagation and rearing of two endangered fish species in a hatch- ery environment. Proc. Ann. Conf. West. Assoc. State Game Fish Comm. 54: 252-259. TYUS, H M 1987. Distribution, reproduction, and habi- tat use of the razorback sucker in the Green River, 78 Great Basin Naturalist Vol. 49, No. 1 Utah, 1979-1986. Trans. Amer. Fish. Soc. 116: 111-116. Ulmer, L. C 1987. Management plan for the razorback sucker (Xyrauchen texanus) in California. Proc. Desert Fish. Counc. 17(1985): 183. Ulmer, L. C, and K R Anderson 1985. Management plan for the razorback sucker (Xyrauchen texanus ) in California. California Dept. Fish Game, Reg. 5 Inf. Bull. 0013-10-1985. 26 pp. U.S. Bureau of Reclamation 1980. Statistical com- pilation of major pumping plants and carriage facilities (canals, pipelines, and tunnels) on Water and Power Resources Service projects. U.S. Bur. Rec. (Wat. Pow. Resour. Serv.), Denver, Colo- rado. 29 pp. 1986. Central Arizona Project Granite Reef Aque- duct fishery investigations progress report. U.S. Bur. Rec. , Boulder City, Nevada. 27 pp. COMPETITION BETWEEN ADULT AND SEEDLING SHRUBS OF AMBROSIA DUMOSA IN THE MOJAVE DESERT, NEVADA Richard Hunter Abstract — Seeds of the perennial shrub Ambrosia dumosa germinated in abundance following 11 days of rain during August 1983 at a study site in the northern Mojave Desert. Seedling establishment, growth, and reproduction were observed in natural vegetation and in an area that had been previously cleared of vegetation. For 5,527 A. dumosa seedlings, percent survival in April 1986 averaged 3% in the undisturbed vegetation and 58% in the denuded area. Seedlings occupying the cleared area had grown to sizes up to 0. 1 m by October 1984; some produced flowers and fruit in the spring of 1985. Surviving seedlings in the undisturbed vegetation were all smaller than 0.001 m3 and did not reproduce. These pronounced differences in growth, survival, and reproduction associated with the presence or absence of adult shrubs demonstrated an intense competition that is incompatible with indications of mild competition from nearest-neighbor analyses. I therefore hypothesize that competition for water occurred, not by competition for water in two dimensions but by rapid use of a common resource, as if several people were drinking with straws from a common cup. This temporal mechanism would strongly favor adults over seedlings. Attempts to detect the occurrence of "competition" in desert vegetation have de- pended largely on analyses of spacing and the distribution of shrub populations (reviewed recently by Fowler 1986, Ismail and Babikir 1986). Regular spacing has been taken to im- ply that competition has occurred in the past. However, neither pattern nor spacing with respect to neighbors has produced unequivo- cal evidence that competitive interactions have major effects in desert plant communi- ties (Wallace and Romney 1972, Yeaton and Cody 1976, Ebert and McMaster 1981, Fonteyn and Mahall 1981, Phillips and Mac- Mahon 1981, Wright 1982, Schlesinger and Jones 1984). Some workers have presented direct experimental evidence that competi- tion for soil moisture occurred between desert shrubs by demonstrating that removal of shrubs resulted in statistically significant in- creases in the water potentials of remaining plants (Ehleringer 1984, Fonteyn and Mahall 1978, Robberecht et al. 1983). Though com- petition in deserts has thus been acceptably "proven," it would appear from these studies to be a minor process in shrub population dynamics producing subtle effects subject to considerable debate. To assess the role of adult-seedling compe- tition on seeding survival, growth, and repro- duction, I compared seedling establishment in undisturbed vegetation with that in an adja- cent area denuded of shrubs. Heavy rains in August 1983 over an area including several denuded areas resulted in germination of nu- merous seeds of the shrub Ambrosia dumosa, initiating the study. Study Site The focus of the study was a 43 X 1,000-m area in Jackass Flats, Nevada (36°42'N, 116°24'W, 1,100 m altitude), cleared of vege- tation by surface blading in 1979 (Major W. Jacobs, personal communication) for a pur- pose unrelated to this study. Prior to the A. dii7nosa germination, the denuded area supported a sparse population of the peren- nial grass Oryzopsis hymenoides. The abun- dant germination of A. dumosa appeared to be a highly localized event, extending approxi- mately 1.3 km E, 3.3 km N, and 3-5 km S and W of the study site. The rain that caused germination fell almost daily 9-19 August 1983, totaling 99 mm at the nearest NOAA weather station 11 km away. That rainfall amount in such a short period was unprece- dented in 23 years of record. Soils in west Jackass Flats are sandy to a depth of at least 1 m (Romney et al. 1973, site 72). Natural plant cover consists largely of the shrubs Larrea tridentata and A. du- mosa, together with sparse individuals of Ceratoides lanata, Acamptopappus shockleyi, and O. hymenoides. 'Laboratory of Biomedical and Environmental Sciences, University of California, Los Angeles. Present address: Box 495, Mercury, Nevada 89023. 79 80 Great Basin Naturalist Vol. 49, No. 1 Methods In January 1984 a 100-m steel tape was laid at right angles to the long axis of the denuded area extending 25—32 m into undisturbed habitat on either side of the denuded plot. All seedlings were counted (but not measured) in a 1-m swath on one side of the tape. In addition, measurements were made of the heights, maximum widths, and the perpen- dicular widths of all mature shrubs (> 10 cm in any dimension) and the grass O. hyrnenoides occurring in a 2-m swath centered on the tape. (The greater width was chosen to in- crease the number of individuals sampled for the sparser O. hyrnenoides and adults.) At subsequent censuses in October 1984, June 1985, and April 1986 all plants, seedling and adult, were measured and their reproductive states recorded. Seedlings that occurred in clusters were measured individually. The area covered by each shrub was calculated as an ellipse with the two radii equal to half of the two measured widths. Total cover was cor- rected for overlapping canopies. Shrub vol- ume was estimated as the volume of an ellipti- cal cylinder the height of the shrub. Results In January 1984 seedling density of Am- brosia dumosa ranged from 0 to 535/m2. In the undisturbed vegetated area the seedlings were quite uniform in size and appearance. Individuals ranged up to 3 cm tall and lacked branches and expanding leaves. In the dis- turbed area they were somewhat larger and vegetative. They bore short branches and healthy green leaves and appeared to be ac- tively growing. All but three new A. dumosa on the denuded area were less than 10 cm in any dimension. Remains of cotyledons, some still green, were apparent. There were 5,527 A. dumosa seedlings in the first census. At the same time there were 15 Larrea tridentata and 3 Acamptopappus shockleyi seedlings. Ten of the L. tridentata and all A. shockleyi seedlings occurred in the undisturbed area; none of these survived to June 1985. Three of the 5 L. tridentata that germinated on the denuded area survived. Along the 2-m-wide transect there occurred 43 Oryzopsis hyrnenoides seedlings in the undisturbed area, 17 of which survived to June 1985, and 9 in the denuded area, 5 of 40 40 60 LOCATION, melers Fig. 1. Population characteristics of a cohort of Am- brosia dumosa seedlings spanning a 43-ni-wide denuded area: A, density of seedlings in January and October 1984 (three meter moving average); B, percent survival in June 1985 (three meter moving average); C, average aboveground volumes; D, cover by seedlings and mature plants (Larrea tridentata, A. dumosa, and Oryzopsis hyrnenoides). which survived. Cover by adult perennials in the undisturbed sections was by A. dumosa (10.5%), L. tridentata (7.1%), and other spe- cies (0.3%). Ambrosia dumosa seedlings were more nu- merous in quadrats in the natural vegetation area than in quadrats located on the denuded aStea (Fig. 1A). They also were more dense on January 1989 Hunter: Ambrosia Dumosa Competition 81 Table 1. Logarithmic growth rates (k = ln(V2/Vl)/dT) and volume-doubling times (days) for Ambrosia dumosa seedlings and adults between September 1983 and October 1984. Location (m) Habitat n k ± sem Doubling time 0-25 Control 21 3.5 ± 0.2 72 26-30 West edge, scraped 5 7.5 ± 0.2 34 29-63 Central scraped 27 8.6 ± 0.2 29 64-68 East edge, scraped 3 7.6 ± 0.6 33 69-99 Control 23 3.2 ± 0.3 79 0-99 Control adults 28 0.6 ± 0.1 420 the edges than in the center of the denuded area. Many occurred near or in obstructions, such as around the bases of O. hymenoides clumps, around standing dead wood, in shal- low depressions where litter and seeds col- lected (Reichman 1984), or in small sand de- posits in the lee of dead twigs lying on the surface. Percent survival in June 1985 is plotted in Figure IB. Survival near the center of the disturbed area reached 100% in some quadrats and averaged 65%, whereas in the natural vegetation it averaged only 5% (X2 = 1461, 1 d.f., p < .0005). Survival on the dis- turbed area appeared to taper off near the edges (Fig. IB). A decrease in survival be- tween 48 and 55 m was due to infrequent vehicle traffic (over which I had no control) crossing the transect. (Survival exceeding 100% in Figure IB arose from small counting errors at the initial census due to clustering of seedlings and slight variations in reposition- ing the steel tape.) By October 1984 there was a striking differ- ence in seedling size between the disturbed and undisturbed plots (Fig. 1C). Within 14 months of germination, individual plants in the center of the disturbed area had reached nearly 0. 1 m in canopy volume, while none in the natural vegetation exceeded even 0.001 m . Consequently, by October 1984 average cover on the disturbed area (26 ± 3%) ex- ceeded cover in the natural vegetation of the undisturbed sections of the transect (18.6%) (Fig. ID). In June 1985 and March 1986 cover was lower than in October 1984. Logarithmic growth rates (Erickson 1976) were calculated for the period from germina- tion to October 1974 (Table 1). They demon- strated what is visually apparent in Figure 1C: the growth of surviving seedlings was much slower in the control areas than on either the edges or center of the denuded strip. Proportionally, growth of surviving seed- lings in the control areas was more rapid than that of adults, but in absolute volume growth, adults (January-October 1984; 134 m3/ha) far surpassed seedlings (germination-October 1984; 2.6 m7ha). Adult survival was 100% between January and October 1984, com- pared to 5% for seedlings. In June 1985 there were 207 seedlings lo- cated between 31 and 62 m along the transect (the central denuded section), of which 31 had produced flowers and/or fruit. The mean vol- ume of the reproductive plants was .031 m3 (s.e. = .004), and the smallest was 0.0016 m3. Slightly more than 50% (24 of 47) of the seedlings between 40 and 55 m flowered in 1985. All other seedlings along the entire transect remained nonreproductive. Discussion Dominant shrub species in the Mojave Desert are long-lived (Johnson et al. 1975, Hunter et al. 1980, Vasek 1980), and turnover in established populations is correspondingly slow (Shreve and Hinckeley 1937, Beatley 1980, Hunter et al. 1980,' Goldberg and Turner 1986). Storm-initiated germination and subsequent high seedling mortality are typical (Went and Westergaard 1953, Sheps 1973, Friedman and Orshan 1975, Ackerman 1979, Ebert and McMasters 1981), though the densities I observed were extreme. The excel- lent survival I found on the denuded area has not been previously reported. Denuded areas in the Mojave Desert nor- mally require many years to regain even a semblance of original cover and diversity (Vasek et al. 1975, Romney et al. 1980, Wallace et al. 1980, Webb and Wilshire 1980). The rapid restoration of cover by a dominant species is evidently unusual but concurs with observations of Shreve (1942). 82 Great Basin Naturalist Vol. 49, No. 1 There were bare patches in the vegetated area several meters in diameter associated with slight improvements in survival (6 and 98 m; Fig. IB); yet growth was not improved in those patches (Fig. 1C). In contrast, at the edge of the denuded area growth increased dramatically within 1 m (Fig. 1C). In other words, there was relative uniformity of growth and survival within the two areas and a sharp divergence at the boundaries. In order to explain the uniformity, I postulate that soil moisture levels in each of the two areas were relatively uniform, and that the differences in moisture content were due to plant transpira- tion in the vegetated area. Thus, the appar- ently patchy shrubs resulted in uniformly dry soil. I therefore propose, as a generalization, that in deserts, though shrub aboveground biomass is relatively patchy, soil moisture is relatively uniform. This could arise through rapid equilibration of soil water pools, de- pending on relatively rapid hydraulic conduc- tivity through soils and roots (Bichards and Caldwell 1987). It could also arise through uniform withdrawal of soil moisture by shrubs, but that mechanism is contradicted for wet soil by data of Cable (1977) and Fer- nandez and Caldwell (1975) on the spatial as- pects of moisture withdrawal and root growth. Probably, as soil goes from wet (near 0 MPa) to dry (near 5.0 MPa), the mechanism producing uniformly available soil moisture would switch from equilibration by water movement to relatively uniform withdrawal by shrubs. Nevertheless, whatever the mechanism, this hypothesis relates significantly to current re- search on desert plant ecology, which at- tempts to infer competition by spatial analysis of aboveground biomass. Went (1973), observing density-dependent growth of desert ephemerals, proposed that plants share rather than compete for re- sources, each individual growing in propor- tion to its share of the available resources. A less teleological view is that they had equal capabilities for resource utilization, thus mak- ing competition ineffective rather than ab- sent. This is reasonable for desert ephemer- als, which all start at a given time as seeds and then germinate and grow rapidly to maturity. My hypothesis is similar, except that it in- volves both adults and seedlings. It appears that though they shared a common pool of water, seedlings had a much greater probabil- ity of drought-induced mortality. They had less biomass to "store" water, smaller root systems, and were less self-shading. Their small size was, therefore, a probable cause of the differential mortality (see e.g. Cook 1979, Paine 1976, and Sebens 1982 on size-related mortality). The result was that adults utilized the majority of the resource, and, although seedlings grew proportionately faster than adults while water was available, they failed to become established because of their small size. I offer no plausible alternative hypothesis to explain the large differences in growth and survival between seedling populations on and off the denuded site. Browsing damage was apparent on a few A. dutnosa seedlings and on many O. hymenoides , but it was almost totally restricted to the denuded sections (Hunter 1987). The possibility of competition between seedlings did occur, but only on the denuded area, where seedling cover averaged 26 ± 3% (s.e.) in October 1984. In the vegetated areas, seedling cover was then only 0.7 ± 0.1%, while adult cover was 17.9% (Fig. ID). To suggest that fertility or soil compaction might cause such major differences would be spe- cious. Fertilizers have had little effect in the Mojave Desert without added water (Bomney et al. 1978, Lajtha and Schlesinger 1986). There was no evidence of soil disturbance other than a slight compaction in the denuded area. Allelopathic interactions have been sug- gested for both A. dumosa (Muller 1953) and L. tridentata (Knipe and Herbel 1966), with some positive effects seen in the lab. But a study by Wallace and Bomney (1972) showed positive association of A. dumosa with 17 spe- cies, L. tridentata with 12 species, and nega- tive association with only 2 and 1 species, respectively. That, together with the im- proved growth of annuals under A. dumosa canopies (Muller 1953), implies that allelopa- thy is not significant in the field. An analogy to my hypothesis is several peo- ple drinking a single soda, each with his own straw. The soda disappears at a rate propor- tional to the number of drinkers and their drinking rates, but independent of the dis- tances between them. Similarly, my hypothe- sis suggests the evidence from spatial analy- ses for competition among desert shrubs is incomplete. Of what importance is distance January 1989 Hunter: Ambrosia Dumosa Competition 83 between neighbors if root systems overlap, if soil water flows at significant rates, if a neigh- bor is orders of magnitude smaller, or has a different phenological pattern? I would argue that distance to the nearest neighbor is of minor importance. Of more relevance would be the sum of biomasses of neighboring plants, as in studies by Fowler (1984) and McAuliffe (1984), or modelled by Weiner and Conte (1981). But even that addition would ignore competitive aspects that are tempo- rally determined or that are related to vertical rather than horizontal spacing, and would therefore be incomplete. There is some circumstantial evidence for my hypothesis. The improved survival of seedlings under adult canopies (Friedman and Orshan 1975), the high percentage sur- vival of individuals in clusters of seedlings (this paper, Ebert and McMasters 1981), the finding of only contagious distributions in the Qatari Desert (Ismail and Babikir 1986), and the weak correlation between distance to neighbor or neighbor's size and mortality (Yeaton and Cody 1976, Howe and Wright 1986) all suggest distance between neighbors is of only slight competitive importance in undisturbed desert. I do not mean to suggest that competition is totally unrelated to horizontal spacing. There have been limited successes at inferring com- petitive interactions using analyses of spacing (Fowler 1986). The findings that larger (older?) shrubs are regularly spaced, while small ones are random (Phillips and MacMahon 1981, Cody 1986), suggest small effects operating over long periods of time. They could be ef- fects of spatial patterns of moisture use by shrubs and of differences between species in phenologies and root distribution patterns. I suggest these are all small variations on a nearly uniform background of intense compe- tition for water that is largely independent of plant spacing. Acknowledgments I am grateful for the help and suggestions of E. M. Bomney, P. D. Greger, B. B. Vance, and J. B. McAuliffe. Work was conducted under contract DE-AMO#-76-SF00012 be- tween the United States Department of Energy and the University of California, with supplemental support from the Nevada Applied Ecology Group, US DOE, Nevada Operations Office. Literature Cited Ackerman, T. L 1979. Germination and survival of perennial plant species in the Mojave Desert. Southwest. Nat. 24: 399-408. Beatley, J. C 1980. Fluctuations and stability in climax shrub and woodland vegetation of the Mojave, Great Basin and transition deserts of southern Nevada. Israel J. Bot. 28: 149-168. Cable, D R 1977. Seasonal use of water by mature velvet mesquite. J. Range Manage. 30(1): 4-11. Cody. M L 1986. Spacing patterns in Mojave Des- ert plant communities; near neighbor analyses. J. Arid Environ. 11: 199-217. Cook. R E 1979. Patterns of juvenile mortality and re- cruitment in plants. Pages 207-231 in O. Solbrig, S. Jain, B. Johnson, and P. H. Raven, eds., Topics in plant population biology. Columbia University Press, New York. Ebert. T. A., and G. S. McMaster 1981. Regular pattern of desert shrubs: a sampling artifact? J. Ecol. 69: 559-564. Ehleringer, J R 1984. Intraspecific competitive effects on water relations, growth, and reproduction in Enceliafarinosa. Oecologia63: 153-158. Erickson, R O 1976. Modelling of plant growth. Ann. Rev. Plant Physiol. 27: 407-434. Fernandez. O A , and M M Caldwell 1975. Phenol- ogy and dynamics of root growth of three cool semi-desert shrubs under field conditions. J. Ecol. 63: 703-714. Fonteyn. P J . and B E Mahall 1978. Competition among desert perennials. Nature 275: 544-545. 1981. An experimental analysis of structure in a desert plant community. J. Ecol. 69: 883-896. Fowler, N L 1984. The role of germination date, spatial arrangement, and neighborhood effects in com- petitive interactions in Linum. J. Ecol. 72: 307-318. Fowler, N 1986. The role of competition in plant com- munities in arid and semiarid regions. Ann. Rev. Ecol. Syst. 17:89-110. Friedman. J., and G Orshan 1975. The distribution, emergence and survival of seedlings of Artemisia herba-alba Asso in the Negev Desert of Israel in relation to distance from the adult plants. J. Ecol. 63: 627-632. Goldberg, D E , and R M. Turner. 1986. Vegetation change and plant demography in permanent plots in the Sonoran Desert. Ecology 67: 695-712. Howe, H F , and S. J Wright 1986. Spatial pattern and mortality in the desert mallow (Sphaeralcea am- bigua). Nat. Geo. Res. 2: 492-499. Hunter, R. B 1987. Jackrabbit-shrub interactions in the Mojave Desert. Pages 88-92 in F. D. Provenza, J. T. Flinders, and E. D. McArthur, Proceed- ings— Symposium on Plant-Herbivore Interac- tions. USDA, Forest Service, Intermountain Res. Sta. Rept. INT-222. Ogden, Utah. 179 pp. 84 Great Basin Naturalist Vol. 49, No. 1 Hunter, R B , E M Romney, A. Wallace, and J E. Kinnear. 1980. Residual effects of supplemental moisture on the plant populations of plots in the northern Mojave Desert. Great Basin Nat. Mem. 4: 24-27. Ismail, A. M. A., and A. A. A Babikir 1986. The contro- versy over distribution of desert plants and the pattern of perennial shrubs in the eastern part of the Arabian Desert. J. Arid Environ. 10: 29-38. Johnson, H. B., F. C. Vasek, andT. Yonkers. 1975. Pro- ductivity, diversity and stability relationships in Mojave Desert roadside vegetation. Bull. Torrev Bot. Club. 102:106-115. Knipe, D , and C. H. Herbel. 1966. Germination and growth of some semidesert grassland species treated with aqueous extract from creosote bush. Ecology 47: 775-781. Lajtha, K, andW H Schlesinger 1986. Plant response to variations in nitrogen availability in a desert shrubland community. Biogeochemistry 2: 29-38. McAuliffe, J R 1984. Sahuaro-nurse tree associations in the Sonaran Desert: competitive effects of sa- huaros. Oeeologia 64: 319-321. Muller. C. H. 1953. The association of desert annuals with shrubs. Amer. J. Bot. 40: 53-60. Paine, R. T. 1976. Size-limited predation: an observa- tional and experimental approach with the Myti- lus-Pisaster interaction. Ecology 57: 858-873. Phillips. D. L, and J. A. MacMahon. 1981. Competition and spacing patterns in desert shrubs. J. Ecol. 69: 97-115. Reichman, O. J. 1984. Spatial and temporal variation of seed distributions in Sonoran Desert soils. J. Bio- geography 11: 1-11. Richards, J H , and M. M. Caldwell. 1987. Hydraulic lift — substantial nocturnal water transport be- tween soil layers by Artemisia tridentata roots. Oeeologia 73: 486-490. ROBBERECHT, R . B. E. MAHALL, AND P. S NOBEL. 1983. Experimental removal of intraspecific competi- tors— effects on water relations and productivity of a desert bunchgrass, Hilaria rigicla. Oeeologia 60: 21-24. Romney, E M , V. Q Hale. A. Wallace, O R Lunt, J D Childress, H Kaaz, G V. Alexander, J E Kin- near, andT L. Ackerman. 1973. Some character- istics of soil and perennial vegetation in Northern Mojave Desert areas of the Nevada Test Site. NTIS, U.S. Dept of Commerce, Springfield, Vir- ginia. 340 pp. Romney, E. M , V. Q. Hale, A Wallace, and R. B Hunter. 1978. Plant response to nitrogen fertil- ization in the Northern Mojave Desert and its relationship to water manipulation. Pages 232- 243 in N. E. West and J. Skujins, eds.. Nitrogen in desert ecosystems. Dowden, Hutchinson and Ross, Inc., Stroudsburg, Pennsylvania. 307 pp. 1980. The pulse hypothesis in the establishment of Artemisia seedlings at Pahute Mesa, Nevada. Great Basin Nat. Mem. 4: 28-30. Schlesinger, W. H., andC. S Jones. 1984. The compar- ative importance of overland runoff and mean an- nual rainfall to shrub communities of the Mojave Desert. Bot. Gaz. 145: 116-124. Sebens, K P. 1982. Competition for space: growth rate, reproductive output, and escape in size. Amer. Nat. 120: 189-197. Sheps, L O 1973. Survival of Larrea tridentata S. & M. seedlings in Death Vallev National Monument, California. Israel J. Bot. 22: 8-17. Shreve, F. 1942. Desert vegetation of North America. Bot. Rev. 8: 195-246. Shreve, F . and A. L. Hinckeley 1937. Thirty years of change in desert vegetation. Ecology 18: 463-478. Vasek, F. C. 1980. Creosote bush: long-lived clones in the Mojave Desert. Amer. J. Bot. 67: 246-255. Vasek, F. C , H B Johnson, and G D. Brum 1975. Effects of power transmission lines on vegetation of the Mojave Desert. Madrono 23: 114-130. Wallace, A., and E M Romney. 1972. A study of a measure of species association between pairs of perennial plants in desert hardpan soil. Pages 205-229 in Radioecology and ecophysiology of desert plants at the Nevada Test Site. NTIS, U.S. Dept. of Commerce, Springfield, Virginia. Wallace, A , E M Romney, and R B Hunter 1980. The challenge of a desert: revegetation of dis- turbed desert lands. Great Basin Nat. Mem. 4: 216-225. Webb, R H . and H G Wilshire 1980. Recovery of soils and vegetation in a Mojave Desert ghost town. J. Arid Environ. 3: 291-303. Weiner, J , AND P T Conte 1981. Dispersal and neigh- borhood efforts in an annual plant competition model. Ecol. Model. 13: 131-147. Went, F W 1973. Competition among plants. Proc. Nat. Acad. Sci. USA 70: 585-590. Went. F. W , and M. Westergaard. 1953. Ecology of desert plants. III. Development of plants in the Death Valley National Monument, California. Ecology 30: 26-38. Wright, S J 1982. Competition, differential mortality, and their effect on the spatial pattern of a desert perennial, Eriogonum inflatum Torr. & Frem. Oeeologia 54: 266-269. Yeaton. R I . and M. L. Cody 1976. Competition and spacing in plant communities: the Northern Mo- jave Desert. J. Ecol. 64: 689-696. AQUATIC INSECTS IN MONTEZUMA WELL, ARIZONA, USA: A TRAVERTINE SPRING MOUND WITH HIGH ALKALINITY AND DISSOLVED CARBON DIOXIDE Dean W. Blinn1 and Milton W. Sanderson1 Abstract. — An annotated list of aquatic insects from the high carbonate system of Montezuma Well, Arizona, USA, is presented for collections taken during 1976-1986. Fifty-seven taxa in 16 families are reported, including new distribution records for Arizona (Anacaena signaticollis, Laccobius ellipticus, and Crenitulus sp. [nr. debilis]) and the USA (Enochrus sharpi). Larval stages for Trichoptera, Lepidoptera, Megaloptera, Neuroptera, Chironomidae, and Anisoptera were absent even though the habitat lacks fish, and water temperature, dissolved oxygen, available food, and substrata appear adequate in Montezuma Well. The potential importance of alkalinity in restricting these insect groups is discussed. Previous collections from the near-constant environment of Montezuma Well, Arizona, have yielded several endemic species of plant and animal taxa (Polhemus 1976, Cole and Watkins 1977, Czarnecki and Blinn 1979, Landye 1981, Davies et al. 1985). Therefore, we believe that a thorough survey of the aquatic insects occupying this high carbonate system was warranted. To date, there have been only a few published reports on the aquatic insects in Montezuma Well. Cole (1965) prepared a list of aquatic insect species in the Well, and Polhemus (1976) described a new heteropteran species (Ranatra mon- tezuma) in the Well; Blinn et al. (1982) dis- cussed the nocturnal planktonic behavior of the endemic Ranatra species. Recently, Pol- hemus and Sanderson (1987) reported Mi- crovelia rasilis from the Well, which was a new distribution record for the USA. Montezuma Well is an active, collapsed travertine spring mound in the upper Sonoran Desert grassland of Arizona that has consider- ably higher concentrations of dissolved CO, (>550 mg I"1) and alkalinity (>600 mg 1 T CaC03) than other aquatic habitats in the re- gion. The system encloses an area of 0.76 ha and has a mean depth of 6. 7 m (Cole and Barry 1973). The littoral zone supports a dense stand of Potamogeton illinoensis Morong; a precipi- tous drop beyond the submerged vegetation delineates a well-defined pelagic region (0.33 ha, maximum depth 17 m). Warm (24 C) water enters through three or four artesian spring vents at the bottom of the pelagic zone and exits (4, 163 1 min ) through a side wall cavern in the travertine deposits (Cole and Barry 1973). The water level remains constant throughout the year, and water temperature never varies more than ± 4.0 C, with an annual mean water temperature of 21.1 C (Boucher et al. 1984). Concentrations of dis- solved oxygen are greater than 6.0 mg 1 ' in the littoral vegetation over a diel period. There are no fish in Montezuma Well, appar- ently due to high concentrations of dissolved CO,' (Cole 1983). The high alkalinity of the artesian water maintains a constant pH (6.5, s.d. ± 0.02) and a moderately high specific conductance (925-1,394 u,S cm"2, 25 C). Ad- ditional physical-chemical information on the Well may be found in Cole and Barry (1973) and Boucher et al. (1984). Methods Seasonal collections were initiated during 1976 and continued through 1986. Samples were taken during both day and night with net tows, grabs, and bottom dredges. The abun- dance of dominant insect taxa was estimated from harvests of macrophytes taken from cir- cular quadrats (573 cm ) or vertical 1-m tows with a plankton net (153 |xm mesh). Ultra- violet light traps were also employed dur- ing night periods to obtain flying adults. Assistance with species identification and/or verification was provided by the following Department of Biological Sciences, Northern Arizona University. Flagstaff, Arizona 86011 85 86 Great Basin Naturalist Vol. 49, No. 1 specialists: F. N. Young (Dytiscidae), H. P. Brown (Elmidae), W. U. Brigham (Halipli- dae), P. J. Spangler (Hydrophilidae), W. W. Wirth and VV. L. Grogan, Jr. (Ceratopogo- nidae), B. V. Peterson (Diptera), J. Polhemus (Veliidae), D. Bloodgood (Ephemeropta), and D. G. Huggins (Odonata). Besults and Discussion Fifty-seven taxa of aquatic insects in 16 families and 5 orders were collected in Mon- tezuma Well during this study (Table 1). Seven additional taxa have been reported by Cole (1965). We assumed the presence of larval stages indicated that insects reproduced in the Well. Our collections reported three new records of Coleoptera for the state of Arizona, includ- ing Anacaena signaticollis, Crenitidus sp. (nr. debilis), and Laccobius ellipticus, and the first records of Enochrus sharpi in the United States. In addition, undescribed species of adult Brachypogon, Ceratoculicoides, Dasij- helea, and Trichomyia were collected (per- sonal communication, W. L. Grogan and B. V. Peterson) in light traps near the shore of Montezuma Well. The biting midge family, Ceratopogonidae, had the greatest diversity of species with 30 taxa, while the Hydrophilidae (Coleoptera) were the next most diverse group (7 taxa plus Hydrophilus reported by Cole 1965). The soft bottom muds and abundant plant material in Montezuma Well may have provided a variety of suitable substrata for members of the biting midge family. Although not represented by many individual taxa, the Nepidae (Ranatra montezuma ), Coenagrionidae (Telebasis salna ), and Mesoveliidae (Mesovelia mulsanti) were numerically the most abundant aquatic in- sects in Montezuma Well. Collections in the top 20 cm of the vegetation yielded 87 nymphs m~2 (s.e. ± 21) of R. montezuma and 57.4 nymphs m 2 (s.e. ± 19.8) of Mesovelia mul- santi during the summer, while integrated collections in a 1-m water column in the vege- tation yielded 2,119 nymphs m3 (s.e. ± 372) of Telebasis salna from June through August. The absence of major aquatic insect groups in Montezuma Well may be of more ecological interest than the actual occurrence of insect species reported in Table 1. There were no larval representatives of Trichoptera, Lepi- doptera, Megaloptera, Neuroptera, Plecop- tera, and Anisoptera (Odonata) in our collec- tions from Montezuma Well. In addition to these groups, Chironomidae larvae were also absent, but Cole (1965) reported the occur- rence of Pentaneura sp. Furthermore, Cal- libaetis (Ephemeroptera) larvae were present in only one collection during July 1986. The absence of Plecoptera is not surprising because of their strict requirements for clean, cool running waters (Harper and Stewart 1984); however, the absence of Trichoptera, Lepidoptera, Megaloptera, and Anisoptera and the limited collections of Chironomidae and Ephemeroptera are notable since larval stages for these aquatic insects are commonly associated with hydrophytes and/or soft sedi- ments within lakes and ponds (Brigham et al. 1982, Merritt and Cummins 1984). The lack of predaceous fish, the abundance of potential prey (Blinn et al. 1987), the near-constant annual water temperature (21.1 ± 4 C), the adequate dissolved oxygen concentrations (6-14 mg l1), and the extensive vegetative refuge along the shore of Montezuma Well would appear to provide a suitable habitat for all of these insect groups. Adults of Anisoptera have been observed flying over Montezuma Well and ovipositing, but no nymphs have been collected. The absence of Trichoptera, Lepidoptera, Megaloptera, Neuroptera, and Anisoptera, as well as the infrequent occurrence of Chirono- midae and Ephemeroptera (Callibaetis) in Montezuma Well, suggests that high concen- trations of dissolved C02 (550 mg 1" ) and/or alkalinity (600 mg l"1 CaC03) restrict hatching and/or larval development by individuals in these aquatic insect groups, because other physicochemical conditions in the Well ap- pear to be suitable for occupation. This agrees with the findings of Winget and Mangum (1979), who reported a significant negative correlation between alkalinity and number of macroinvertebrate taxa in aquatic eco- systems. One hypothesis for the restriction of these insect groups is that the high carbonate alka- linity in Montezuma Well may interfere with cutaneous respiratory activities of the imma- ture stages by forming deposits of CaC03 on the body surface due to a shift in the carbonate equilibrium (Cole 1983). These deposits could greatly reduce the cutaneous surface area January 1989 BLINN, SANDERSON: MONTEZUMA WELL INSECTS 87 Table 1. Annotated list of aquatic insects from Mon- tezuma Well, Arizona, USA: * indicates new Arizona record; ** indicates new USA record; *** indicates record reported only by G. A. Cole (1965). All identifica- tions for Ceratopogonidae were made from adults, and larval stages were assumed to be present in Montezuma Well. COLEOPTERA Dytiscidae Celina occidentals ? (Young) Cybister ellipticus LeConte Desmopachria (Pachriodesma) mexicana Sharp Neoclypeodytes cinctellus (LeConte) Thermonectes marmoratns (Hope) Elmidae Microcylloepus similis (Horn) Haliplidae Peltodytes dispersus Roberts Peltodytes simplex (LeConte) Hydraenidae Ochthebius puncticollis LeConte Hydrophilidae *Anacaena signaticollis (Fall) Chaetarthria sp. *Crenitulus sp. (nr. debilis Sharp) **Enochrus sharpi Gundersen ***Hydrophilus sp. *Laccobius ellipticus (LeConte) Tropisternus columbianus Brown Hydroscaphidae ***Hydroscapha natans LeConte Ephemeroptera Baetidae Callibaetis sp. Diptera Ceratopogonidae Alluaudomyia needhami Thomsen Atrichopogon occidentalis Wirth Atrichopogon transversus Wirth Atrichopogon sp. Bessia sandersoni Grogan & Wirth Brachypogon sp. Ceratoculicoides sp. Culicoides brookmani Wirth Culicoides butleri Wirth & Hubert Culicoides haematopatus Malloch Culicoides salihi Khalaf Culicoides variipennis (Coquillett) Dasyhelea ancora (Coquillett) Dasyhelea cincta group (Coquillett) Table 1 continued. Dasyhelea fasciigera Kieffer Dasyhelea grisea group (Coquillett) Dasyhelea messersmithi Waugh & Wirth Dasyhelea mutabilis (Coquillett) Dasyhelea pollinosa Wirth Dasyhelea pritchardi Wirth Dasyhelea sp. Forcipomyia brevipennis (Macquart) Leptoconops (L.) torrens (Townsend) Palpomyia occidentalis Grogan & Wirth Parabezzia biennis (Coquillett) Parabezzia sp. Phaenobezzia fulvithorax (Malloch) Stilobezzia antennalis (Coquillett) Stilobezzia fuscula Wirth Stilobezzia pruinosa Wirth Chironomidae ***Pentaneura sp. Culicidae *** Anopheles freeborni Aitken Psychodidae Trichomyia sp. Stratiomyiidae ***Auparyphus sp. ***Odontomyia sp. Heteroptera Belostomatidae ***Abedus breviceps Stal Belostoma bakeri Montandon Corixidae Cenocorixa wileyae (Hungerford) Hydrometridae Hydrometra aemula Drake Mesoveliidae Mesovelia mulsanti White Naucoridae Ambrysus woodburyi Usinger Nepidae Banatra montezuina Polhemus ***Banatra quadridentata Stal Veliidae Microvelia mulsanti White Microvelia rasilis Drake Odonata Coenagrionidae Enallagma civile (Hagen) Telebasis salna (Hagen) available for gas exchange. It has been re- ported that many aquatic insect larvae, partic- ularly the trichopterans, rely exclusively on the rich tracheal network located just beneath the thin cuticle for gas exchange (Wiggins 1977). Those insect larvae that rely almost exclusively on cutaneous respiration but lack supplemental tracheal gills or the ability to transport air stores may be unable to meet the metabolic demands for oxygen in the high Great Basin Naturalist Vol. 49, No. 1 carbonate waters of Montezuma Well. It has also been proposed that the tracheal gills are highly susceptible to damage by environmen- tal extremes since they serve as active uptake sites (Eriksen et al. 1984). The constant low pH (6.5) may also restrict the above described insect groups from Montezuma Well. Acknowledgments This study was supported in part by funds provided to DWB from the Whitehall Foun- dation, Inc., and an Organized Research Grant from Northern Arizona University. We thank Clay Runck for assistance with the col- lection and quantification of dominant insect species and the personnel at Montezuma Well, especially Jack Beckman and Jim Cole- man, for their excellent cooperation during this study. Literature Cited Blinn, D W . R W Davies, and B Dehdashti. 1987. Specialized pelagic feeding by Erpobdella mon- tezuma (Hirudinea). Holarctic Ecology 10: 235-240. Blinn, D. W., C. Pinney, and M. W Sanderson 1982. Nocturnal planktonic behavior of Ranatra mon- tezuma Polhemus (Nepidae: Hemiptera) in Mon- tezuma Well, Arizona. J. Kansas Ent. Soc. 55: 481-484. Boucher, P., D W Blinn, and D B Johnson 1984. Phytoplankton ecology in an unusually stable environment (Montezuma Well, Arizona, USA). Hydrobiologia 119: 149-160. Bricham, A R , W U Brigham, and A Gnilka, eds. 1982. Aquatic insects and oligochaetes of North and South Carolina. Midwest Aquatic Enter- prises, Mahomet, Illinois. COLE, G A 1965. Final report to Montezuma Castle National Monument of investigations of Mon- tezuma Well. Report presented to National Park Service. 86 pp. 1983. Textbook of limnology. 3d ed. C. V. Mosby Co., St. Louis, Missouri. 401 pp. Cole, G A , and W T Barry 1973. Montezuma Well, Arizona, as a habitat. J. Arizona Acad. Sci. 8: 7-13. Cole, G A , and R. L. Watkins 1977. Hyalella mon- tezuma, a new species (Crustacea: Amphipoda) from Montezuma Well, Arizona. Hydrobiologia 52: 175-184. Czarnecki, D B , and D W Blinn 1979. Observations on southwestern diatoms. 2. Caloneis latiuscula var. reimeri n. var., Cyclotella pseudostelligera f. parva n. f. and Gomphonema montezumense n. sp., new taxa from Montezuma Well National Monument. Trans. Amer. Microsc. Soc. 98: 110-114. Davies, R W , R N. Sinchal, and D W Blinn 1985. Erpobdella montezuma, a new species of fresh- water leech from North America (Arizona, USA). Canadian J. Zool. 63: 965-969. Eriksen, C H , V H Resh, S S Balling, and G A Lam- berti Aquatic insect respiration. Pages 27-37 in R. W. Merritt and K. W. Cummins, eds., An introduction to the aquatic insects of North Amer- ica. 2d ed. Kendall/Hunt Publ. Co. 722 pp. Harper. P P., and K W Stewart 1984. Plecoptera. Pages 182-230 in R. W. Merritt and K. W. Cum- mins, eds., An introduction to the aquatic insects of North America. 2d ed. Kendall/Hunt Publ. Co. 722 pp. Landye, J J 1981. Current status of endangered, threat- ened, and/or rare mollusks of New Mexico and Arizona. U.S. Fish and Wildl. Serv., Office of Rare and Endangered Species, Albuquerque, New Mexico. 35 pp. Merritt, R W , and K W Cummins, eds 1984. An intro- duction to the aquatic insects of North America. 2d ed. Kendall/Hunt Publ. Co. 722 pp. Polhemus, J. 1976. Notes on North American Nepidae (Hemiptera: Heteroptera). Pan-Pacif. Ent. 52: 204-208. Polhemus, J., and M. W. Sanderson. 1987. Microvelia rasilis Drake in Arizona: a species new to the United States (Heteroptera: Veliidae). Great Basin Nat. 47: 660. Wiggins, G. B. 1977. Larvae of the North American cad- disfly genera. University of Toronto Press, Toronto, Ontario. 401 pp. WiNGET, R. N., AND F. A. Mangum 1979. Biotic condition index: integrated biological, physical, and chemi- cal stream parameters for management. U.S. De- partment of Agriculture, Forest Service, Inter- mountain Region, Ogden, Utah. 51 pp. A NEW HALIPLUS FROM WARM SPRINGS, NEVADA (COLEOPTERA: HALIPLIDAE) Samuel A. Wells1 Abstract. — Haliplus eremicus is described as a new species of the crawling water beetle family Haliplidae. A brief description of the species, illustrations, and diagnosis that compares it with Haliplus mimeticus Matheson are provided. In the early fall of 1984, 11 specimens of Haliplus eremicus were collected from a small pond near the Warm Springs recreation area in Nevada. The place is named for the springs that feed a small stream and ponds. Haliplus eremicus was collected in a cool pond not far from the stream. Haliplus eremicus, n. sp. Male. — Rody broadly oval. Length 3.10- 3.60 mm long, 1.70-1.85 times as long as wide. Head vitelline. Area between eyes 42-46% width of head. Pronotum vitelline to yellow, 1.75-1.95 times wider than long, 1.55-1.70 times wider at base than at apex, punctate throughout except for a thin, glabrous area slightly basad of center. Prosternum marginate at sides, slightly constricted between procoxae, ante- rior margin straight. Elytra cinnamomeus with areas of yellow maculae as follows: along lateral margin with expanded area beginning at basal third and extending distally beyond middle, and mesad to stria 5; macula on disc beginning near elytral base between striae 1 and 3 and ex- tending distally to area one-third the length of elytra from base; macula between striae 2 and 4 joining or adjacent to discal macula and ex- tending distally to area three-fifths length of elytral from base; small macula 0.50 mm long between striae 4 and 5 near base. In one speci- men the entire disc is yellow. Left paramere pale yellow, apex straight and more heavily sclerotized, tuft of long hair one-third length of paramere and one-third distance from apex and base. Aedeagus regu- larly curved above and below. Female. — Similar to male. Diagnosis. — Haliplus eremicus (Fig. 1) appears to be most closely related to H. mimeticus Matheson (Fig. 2), which was de- scribed from one female collected on the Pacific Slope. Haliplus mimeticus is unicolor- ous, whereas H. eremicus has light maculae as described above. The anterior margin of the prosternum is sinuate with the median area modestly recurved in H. mimeticus (Fig. 4) and straight in H. eremicus (Fig. 3). The left paramere in H. eremicus (Fig. 6) is more heav- ily sclerotized at the tip and with the tuft of hair arising more than 0. 12 mm from the tip; on H. mimeticus the tuft of hair arises 0. 10 mm or less from the tip (Fig. 8). Type material. — The male holotype, fe- male allotype, and nine paratypes were collected from Warm Springs in Clark County Nevada, 28-IX-1984, by R. W. Raumann, E. Nutall, and myself; one paratype was col- lected from Tucson, Pima County, Arizona, 25-111-1925, by Rryant, and one was collected from Lakeside, Navajo County, Arizona, 22- VIII-1952, by R. Malkin. The holotype and allotype are in the U.S. National Museum, four paratypes are in the California Academy of Science, one paratype is at Cornell University, two paratypes are at Rrigham Young University, and the remain- der are in my collection. 'Monte L. Bean Life Science Museum, Brigham Young University. Provo, Utah 84602. 89 90 Great Basin Naturalist Vol. 49, No. 1 Figs. 1-8. Haliplus spp.: 1, H. eremicus; 2, H. mimeticus; 3, prosternum of H. eremicus; 4, prosternum of H. mimeticus; 5, aedeagus of//, eremicus; 6, left paramere of//, eremicus; 7, aedeagus of//, viimeticus; 8, left paramere of//, mimeticus. January 1989 Wells: A New Nevada Haliplus 91 Acknowledgments Literature Cited I thank Dr. P. J. Spangler for examining the new species and offering suggestions and Sun Yung Kim for Figures 1-2. Matheson, R 1912. The Haliplidae of North America, north of Mexico. J. New York Ent. Soc. 20: 156-191. Usinger, R L 1963. Aquatic insects of California. University of California Press, Rerkeley and Los Angeles. Wallis. J R 1932. Revision of the North American species (north of Mexico), of the genus Haliplus, Latreille. Trans. Royal Canadian Inst. 19: 1-76. ON THE GENUS PARACARINOLIDIA (CICADELLIDAE: COELIDIINAE: TERULIINI) M. W. Nielson1 ABSTRACT. — Two new species, Paracarinolidia longiseta and P. glabra from Brazil and French Guiana, respectively, are described and illustrated. A revised key to males of five known species is also presented. The genus is now known to occur in Brazil, Ecuador, Peru, Colombia, and French Guiana. Paracarinolidia Nielson is a small Neo- tropical teruliine genus that occurs in a rather broad region from Peru, Ecuador, and Colombia on the west to French Guiana and Brazil on the east. With the addition of two new species described in this paper, there are now five known. Three species occur exclu- sively in Brazil. One of the new species is from French Guiana, the other from Brazil. Members of the genus are small and slen- der, with dark brown to black forewings punc- tuated with numerous small to large, pale ochraceous markings. The narrow, produced head and carinate lateral margins of the crown are distinctive. These characters together with the long, very slender styles separate the group from its nearest relative, Carinolidia Nielson. Key to Males of Paracarinolidia 1. Aedeagus with numerous short to long setae on shaft 2 — Aedeagus without setae or with few very short setae on shaft 4 2(1). Aedeagus with several setae restricted to middle of shaft 3 — Aedeagus with several setae in apical 1/4 of shaft and with a single, very long, subterminal seta (Figs. 3, 4) longiseta, n. sp. 3(2). Aedeagus with deep longitudinal cleft medially, setae uniformly short (Nielson 1979, Figs. 87, 88) amabilis (Linnavuori) — Aedeagus without such cleft, two setae moder- ately long, the remainder uniformly short (Niel- son 1979, Figs. 79, 80) differta Nielson 4(1). Aedeagus with few very short setae medially, apex of shaft narrowed, without teeth (Nielson 1979, Figs. 73, 74) guttulata (Stal) — Aedeagus without setae, apex of shaft enlarged, with teeth on anterior margin (Figs. 11, 12) ... glabra, n. sp. Paracarinolidia longiseta, n. sp. Figs. 1-7 Length. — Male 7.20 mm, female 7.70 mm. General color dark brown to black with small to large, ivory or pale ochraceous mark- ings on forewings, larger markings at apex of clavus, along costa, and near apex of fore- wings; small yellowish markings on disk of crown; clypeus and eyes dark brown to black; genae, lorae, and clypellus yellowish. Head much narrower than pronotum; crown narrow, produced distally beyond anterior margin of eyes, lateral margins dis- tinctly carinate; ocelli near anterior margin of crown; eyes large, nearly globular; pronotum short, median length less than median length of crown; scutellum moderately large, median length greater than median length of prono- tum; forewings elongate, venation typical; clypeus long and narrow, with prominent me- dian longitudinal carina; clypellus narrow, lat- eral margins nearly parallel. Male. — Pygofer with long, narrow, cau- dodorsal lobe and very short, caudoventral lobe (Fig. 1); aedeagus asymmetrical, tubelike in ventral view, constricted subapically and curved dorsally at apex, with several short setae subapically and one long, subterminal seta extending basally in lateral view (Figs. 3, 4), gonopore near middle of shaft on ventral surface; style long and slender in distal 2/3, enlarged at basal 1/3, tapered distally (Figs. 5, 6); plate long and moderately broad medially with few short distal setae (Fig. 2). Female. — Seventh sternum with caudal margin broadly bilobed (Fig. 7). Holotype (male). — Brazil: Rondonia, Vil- hena, 3. VIII. 1983, Norman Penny (INPA). Monte L. Bean Life Science Museum, Brigham Young University, Provo, Utah 84602. 92 January 1989 Nielson: Neotropical Leafhoppers 93 Figs. 1-7. Paracarinolidia longiseta, n. sp. : 1, male pygofer, lateral view; 2, plate, ventral view; 3, aedeagus, lateral view; 4, aedeagus, ventral view; 5, left style, dorsal view; 6, left style, lateral view; 7, female seventh sternum, ventral view. Allotype (female), same data as holotype ex- cept 29. VII. 1983 (INPA). Remarks. — This species is nearest to dif- ferta Nielson and can be distinguished by a row of short setae on the apical 1/4 of the aedeagal shaft with a long, subterminal seta. Paracarinolidia glabra, n. sp. Figs. 8-15 Length. — Male 6.90 mm. General color as in longiseta; costal spots on forewing smaller, spot in 5th apical cell much reduced; markings on crown and face similar to longiseta. Similar in size and male genitalia to guttulata . Head much narrower than pronotum (Fig. 8); crown very narrow, produced distally beyond anterior margin of eyes, width much narrower than width of eyes, lateral margins distinctly carinate; eyes very large, nearly globular; pronotum short, median length shorter than median length of crown; scutel- lum moderately large, median length greater than median length of pronotum; forewing elongate, venation typical (right forewing missing in type); clypeus long and narrow, with prominent median longitudinal carina; clypellus narrow, lateral margins expanded distally (Fig. 9). Male. — Pygofer as in longiseta (Fig. 10); aedeagus asymmetrical, tubular throughout 94 Great Basin Naturalist Vol. 49, No. 1 Figs. 8-15. Paracarinolidia glabra, n. sp. : 8, head, pronotum, and scutellum, dorsal view; 9. same, ventral view; 10, malepygofer, lateral view; 11, aedeagus, lateral view; 12, aedeagus, ventral view; 13, right style, ventral view; 14, right style, lateral view; 15, plate, ventral view. shaft, slightly sinuate, without setae on shaft, apex curved dorsally and toothed on anterior margin (Figs. 11, 12), gonopore near middle of shaft; style very narrow in distal half (Figs. 13, 14); plate long and moderately broad, with several setae distally (Fig. 15). Female. — Unknown. Holotype (male). — French Guiana (Cay- enne): Oyac-Conti-Cacao-Bief, — .IX-X. 1914, R. Benoist(MNHN). Remarks. — Paracarinolidia glabra is simi- lar in male genital characters to P. guttulata but can be separated by the lack of aedeagal setae and by the shorter style. The internal male genital structures (aedeagus, connec- tive, and styles) of the holotype specimen were apparently lost after they were illus- trated. The abdomen, pygofer, and plates remain in the attached microvial. Acknowledgments I thank Dr. Norman Penny, formerly of Instituto Nacional de Pesquisas da Amazonia (INPA), now with California Academy of Sci- ence, San Francisco, and Dr. Michel Bou- lard, Museum National d'Histoire Naturelle, Paris (MNHN), for the loan of specimens de- scribed in this paper. I appreciate the fine January 1989 Nielson: Neotropical Leafhoppers 95 illustrations prepared by Mrs. Jeanette Price; Literature Cited and for reviewing the manuscript, I express , _r . . Nielson, M.W. 1979. A revision of the subfamily Coelidi- my gratitude to Dr. Paul Freytag, University inae (Homoptera: Cicadellidae). III. Tribe Teruli- of Kentucky, Lexington. ini. Pacific Insects Monograph 35. 329 pp. TWO NEW GENERA AND TWO NEW SPECIES OFTERULIINE LEAFHOPPERS (HOMOPTERA: CICADELLIDAE: COELIDIINAE) M. W. Nielson1 Abstract. — Two new genera and two new species of leafhoppers in the tribe Teruliini are described and illustrated. New genera include: Perspinolidia, type-species Perspinolidia peruviensis, n. sp., and Brevicapitorus, type-species Brevicapitorus elongatus, n. sp. Both genera are monobasic and occur in the Neotropical region. Two genera and their attendant species described in this paper represent a continuum of new taxa of a leafhopper group whose pres- ence is rare in the Neotropical region. Only a few specimens are usually collected at any given time, but they add generic diversity and taxonomic composition to the tribe Teruliini (Nielson 1979, 1983a, 1983b). There are now 47 genera assigned to the tribe, 23 of which are monobasic. The genus Perspinolidia represents an anomaly in characterization of the tribe. The median longitudinal clypeal carina, which separates the tribes Teruliini and Coelidiini, is incomplete in Perspinolidia and does not reach the anterior margin from its origin at the transclypeal suture. This deficiency may de- note variability in the length of the character or an evolutionary state of development in which the carina is being added to or deleted from the clypeus. In either case, the genus is more closely related to members of Teruliini than to members of any other tribe in the subfamily Coelidiinae. Perspinolidia, n. gen. Type-species. — Perspinolidia peruviensis, n. sp. Medium-sized, robust species. Similar in general habitus to Articoelidia Nielson but with distinctive male genitalia. General color dark brown with suffused light brown on cla- vus extending to apex of forewing. Head narrower than pronotum; crown broad, width greater than width of eyes; eyes large, semiglobular; pronotum and scutellum large; forewings with 5 apical cells, 3 ante- apical cells present, outer one closed; clypeus long and broad, with incomplete median lon- gitudinal carina, originating at the transclyp- eal suture but not reaching anterior margin; hind femoral setal arrangement 2-1-2+1. Male genitalia partly asymmetrical; pygofer with small caudodorsal lobe; aedeagus asym- metrical, long, somewhat tubular with tuft of setae distad of middle near large gonopore on lateral margin; connective Y-shaped with short stem; style large, very broad in lateral view; plate long, broad subapically, seta- ceous. The genus keys to Articoelidia in couplet 21 in Nielson (1979), but it can be distinguished by the lack of prominent spines on the pygofer and segment 10, the presence of the broad style, and the subdistal tuft of setae on the aedeagus. Perspinolidia peruviensis, n. sp. Figs. 1-8 Length. — Male 8.70 mm. General color dark brown. Crown light tan- nish; pronotum black except for tan anterolat- eral margins; scutellum black; forewings dark brown to black except for light brown from base of clavus to apex of wing, costal area light brown; face tan. Similar in general habitus to species of Articoelidia, but with distinctive male genitalia. Head narrower than pronotum (Fig. 1); crown produced and rounded anteriorly, broad, width much greater than width of eyes; eyes large, semiglobular; pronotum moder- ately long, median length about as long as median length of crown; scutellum large, median length greater than median length of pronotum; forewing and venation typical; 'Monte L. Bean Life Science Museum, Brigham Young University, Provo. Utah 84602. 96 January 1989 Nielson: Neotropical Leafhoppers 97 Figs. 1-8. Perspinolidia peruviensis, n. sp.: 1, head, pronotum, and scutellum, dorsal view; 2, face, ventral view; 3, male pygofer, lateral view; 4, aedeagus, dorsal view; 5, aedeagus, lateral view; 6, style, dorsal view; 7, style, lateral view; 8, plate, ventral view. clypeus long and broad, somewhat tumid, with incomplete median longitudinal carina arising at base of transclypeal suture and extending anteriorly to about 2/3 length of clypeus (Fig. 2). Male. — Pygofer with small caudodorsal lobe directed dorsally (Fig. 3); aedeagus asymmetrical, long, somewhat tubular, sinu- ate in lateral view and tapered toward apex, with tuft of dense, stout setae on lateral mar- gin just basad of gonopore (Figs. 4, 5), gono- pore large, subdistal on lateral margin; style large, compressed laterally and very broad in lateral view, dentate apically on dorsal margin (Figs. 6, 7); plate moderately long, lateral margins expanded before apex (Fig. 8). Female. — Unknown. Holotype (male). — Peru: Rio Santiago, 30.IX.1924. H. Bassler, F-6137, Ace. 33591 (AMNH). Remarks. — Perspinolidia peruviensis is the only known species in the genus, and it can be distinguished from members of Articoelidia by the incomplete median clypeal carina, by the lack of spines on the pygofer and segment 10, and by the presence of a subdistal tuft of setae on the aedeagus. Brevicapitorus, n. gen. Type-species. — Brevicapitorus elongatus, n. sp. Medium-sized, robust species. Similar in general habitus to large species of Docalidia Nielson, but with distinctive male genitalia. General color black. Head distinctly narrower than pronotum; crown short, broad, about as wide as eyes; eyes large, semiglobular; pronotum and scu- tellum large; forewings with 5 apical cells, 3 anteapical cells present, outer one closed; 98 Great Basin Naturalist Vol. 49, No. 1 Figs. 9-16. Brevicapitorus elongatus, n. sp. : 9, head, pronotum, and scutellum, dorsal view; 10, face, ventral view; 11, male pygofer, lateral view; 12, aedeagus, lateral view; 13, aedeagus, ventral view; 14, style, dorsal view; 15, style, lateral view; 16, plate, ventral view. clypeus long and narrow, with complete me- dian longitudinal carina; hind femoral setal arrangement 2+2+1. Male genitalia partly asymmetrical; pygo- fer with large caudoventral lobe; aedeagus asymmetrical, long, somewhat tubular with single row of stout setae along middle, gono- pore subbasal; connective Y-shaped with short stem; style very long; plate long, nar- row. The genus keys near Terulia Stal in couplet 5 of Nielson (1979) and can be separated by a row of stout setae along the middle of the aedeagus and by the very short, rounded head. Brevicapitorus elongatus, n. sp. Figs. 9-16 Length.— Male 9.70-9.90 mm. General color black. Crown and eyes light tannish brown, disk blackish in basal half; pronotum and scutellum black; forewings dark brown to black, veins black; face black. Similar in general habitus to species of Docalidia but with distinctive male genitalia. Head much narrower than pronotum (Fig. 9); crown short, rounded anteriorly, disk broad, width nearly equal to width of eyes; eyes large, semiglobular; pronotum large, median length much greater than median length of crown; scutellum large, median length greater than median length of prono- tum; forewing broad, apex acutely angled, venation typical; clypeus long and narrow, with prominent median longitudinal carina (Fig. 10); clypellus long, lateral margins ex- panded distally. Male. — Pygofer with broad caudoventral lobe (Fig. 11); aedeagus asymmetrical, long, somewhat tubular in dorsal view, slightly curved in lateral view and expanded between January 1989 Nielson: Neotropical Leafhoppers 99 apex and gonopore, with single row of stout setae along middle of shaft, setae directed more or less laterally (Figs. 12, 13), gonopore subbasal on lateral surface; style very long, nearly as long as aedeagus, curved in dorsal and lateral views, tapered distally (Figs. 14, 15); plate long and narrow, sparselv setose (Fig. 16). Female. — Unknown. Holotype (male). — Brazil: Sinop, Matto Grosso, — .X.1975, M. Alvarenga (UFP). Paratype, one male? (abdomen missing), same data as holotype (author's collection). Remarks. — This species is the only known representative of the genus, and it can be separated from members of the genus Do- calidia by the single row of stout spines on the middle of the aedeagal shaft. Acknowledgments The leafhopper specimens were kindly furnished by Dr. Jerome Rozen, American Museum of Natural History, New York (AMNH), and Dr. Keti Maria Rocha Zanol, Universidade Federal do Parana, Curitiba, Parana (UFP). The excellent illustrations were prepared by Jean Stanger, and the manuscript was reviewed by Ray Gill, Cali- fornia Department of Food and Agriculture, Sacramento. To all of these colleagues I extend my sincere gratitude for their help. Literature Cited Nielson. M. W 1979. A revision of the subfamily Coelidi- inae (Homoptera: Cicadellidae). III. Tribe Teruli- ini. Pacific Insects Monograph 35. 329 pp. 1983a. New genera in the tribe Teruliini with descriptions of new species (Homoptera: Cicadell- idae: Coelidiinae). J. Kansas Entomol. Soc. 56(4): 560-570. 1983b. New Neotropical species of teruliine leaf- hoppers (Cicadellidae: Coelidiinae: Teruliini). J. Kansas Entomol. Soc. 56(3): 365-370. A NEW SPECIES OF ASCLEPIAS (ASCLEPIADACEAE) FROM NORTHWESTERN NEW MEXICO Kenneth D. Heil1, J. Mark Porter2, and Stanley L. Welsh3 Abstract. — Asclepias sanjuanensis Heil, Porter, & Welsh, a new species from the pinyon-juniper woodlands of the San Juan River Valley, San Juan County, New Mexico, is described and illustrated. The species appears to be local and rare. Similar to A. ruthiae Maguire, it is distinguished by the greater number of flowers per inflorescence, the larger number of stems, and the pubescence characters. A new species of Asclepias was discovered among specimens taken recently from the San Juan Valley of northwestern New Mexico. The rare new entity apparently falls into the subgenus Asclepiodella (Small) Woodson according to Woodson (1954). The subgenus is represented by A. ruthiae Maguire in Maguire & Woodson, A. eastivoodiana Rarneby, A. cutleri Woodson, A. cordifolia (Benth.) Jepson, A. brachystephana Torr., and A. uncialis Greene, all occurring in west- ern North America. Additional species, A. cinerea Walt, and A. feaiji Chapm. ex Gray, occur in the eastern United States. The spe- cies described herein was discovered by Barbara Jenkins, Londa Smith, and Marc Werthington while performing a floristic study of the Fred Edwards Wilderness Walk at San Juan College, Farmington, New Mex- ico, during the spring of 1988. Following the initial discoveries, a survey of surrounding areas was conducted to ascertain the distribu- tion of the taxon. The plant is described as follows. Asclepias sanjuanensis Heil, Porter, & Welsh, sp. nov. Asclepiate ruthia Maguire affinis sed in floribus et caulibus plus numerosis et pubes- centis differt. Herbaceous perennial; stems prostrate to ascending, 4-8.3 cm long, glabrous below, becoming minutely tomentulose above, branched below ground, with 2-7 stems from a woody taproot; lowermost leaves scalelike; leaves (1) 2-4 cm long, 0.4-2.5 cm broad, oblong-lanceolate, narrowly acute, approxi- mate to opposite, petiolate, the petiole 0.2- 0.5 cm long, white tomentulose on leaf mar- gins and midrib of abaxial leaf surface only; inflorescence terminal, rarely axillary, sessile, sparsely pilosulose, umbelliform cyme, with 4-15 flowers; pedicels 1.2-2.8 cm long; flow- ers small; calyx lobes lanceolate, 1.8-3.2 mm long, reflexed; corolla reflexed-rotate, pale violet, the lobes 3.5-6 mm long; column 0.4-0.7 mm high, ca 1.3-3 mm thick, reddish green; hoods 1.5-2.5 mm long, saccate, trun- cate, reddish violet with cream to yellowish margins, glabrous, the marginal auricles more or less erect, lanceolate; horn ca 2.3 mm long, included to barely exserted from the hood, attached near the middle and erect from it; anther head 1.9-3 mm high, ca 1.3-2 mm in diameter, the wings narrow; pollinarum ca 0.23 mm long, the corpusculum ca 0.08 mm long, ca 0.04 mm wide, the translator arm ca 0.07 mm long, the pollinia 0.15 mm long; follicle 3.5-6.5 mm long, 1.1-1.6 mm wide, puberulent, smooth, erect on a reflexed pedicel; seeds ca 1 mm long. Type. — USA: New Mexico, San Juan County, Farmington, along the Fred Ed- wards Wilderness Walk on the campus of San Juan College; T30N, R13W, S35, NW 1/4, 22 May 1988, K. D. Heil 4338 (Holotype BRY; Isotypes ARIZ, MO, NY, NMC, SJNM). Asclepias sanjuanensis occurs on sandy or sandy loam soils, usually in disturbed sites, i.e., erosion channels, trails (human or animal), and two-track roadways. The popula- tions are known from 1,524 to 1,676 m on San Juan College, Farmington, New Mexico 87401. "Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721. Life Science Museum and Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602. 100 January 1989 Heiletal.: ANevvAsclepias 101 Fig. 1. Asclepias sanjuanensis: A, habit; B, follicle; C, flower; D, leaf; E, enlargement of cilliate leaf margin. slopes and floodplains of the San Juan River Valley. They occur in pinyon-juniper woodlands (with Finns edtdis, Junipems osteosperma, Que reus gambelii, Purshia tri- dentata, Mirabilis multiflora, Eriogonum microthecum, and Penstemon ophianthus. Subsequent to the initial discovery, two addi- tional collections were made at Farmington and near Bloomington, New Mexico. Only solitary plants were found at both of those locations. Preliminary field investigations in- dicate that the taxon is a local endemic in a region approximately 33 km in length along the San Juan River. The plants flowered first between early and late April in 1988 and continued to flower into May. Fruit matured by mid-June. The sessile or essentially sessile, erect, open hoods, which possess a deeply saccate basal attachment, are characters that associate A. sanjuanensis with members of subgenus Asclepiodella. The species is a near congener of both A. ruthiae and A. eastwoodiana, with which it can be easily confused. However, it is 102 Great Basin Naturalist Vol. 49, No. 1 Table 1. Morphological and ecological comparison of Asclepias sanjuanensis, A. ruthiae, A. eastwoodiana , and A. cutleri. Character sanjuanensis ruthiae eastwoodiana cutleri Number of flowers 4-15 (mean = 10) 2-6 3-5(14) 3-5(7) Number of branches 2-7 1(2) 1(3) 1(2) Herbage pubescence sparsely puberulent densely puberulent sparsely puberulent appressed puberulent Hood pubescence glabrous puberulent or glabrous glabrous glabrous Auricles of hood erect obscure, not erect erect erect Leaf shape lanceolate to broadly lanceolate broadly ovate to broadly lanceolate lanceolate to broadly ovate filiform Leaf pubescence margins and veins dense on entire surface margins and veins entire surface Habit prostrate to ascending prostrate prostrate to ascending erect- ascending Habitat disturbed undisturbed to disturbed disturbed disturbed Locale NW New Mexico SE Utah, N Arizona Central Nevada SE Utah, N Arizona allopatric with both of those taxa. The close relationship between A. ruthiae and A. eastwoodiana is not disputed (Cronquist et al. 1984); in fact, these taxa were at one time considered conspecific (Woodson 1954). Asclepias sanjuanensis fits well into this species group based on morphological and ecological characteristics (Table 1). It differs from A. ruthiae by the greater number of flowers per inflorescence, the larger number of branches from the summit of the root crown, the leaf pubescence characters, the leaf shape, and the erect auricles. From A. eastwoodiana it differs in flower number (but not as greatly as from A. ruthiae) and in the number of branches from the root crown. It appears that A. sanjuanensis is most like the remote A. eastwoodiana, which occurs in central Nevada (Barneby 1945, Cronquist et al. 1984). Between these two species is the much more widespread A. ruthiae, which occurs in southeastern Utah and north central Arizona (Welsh et al. 1987, Cronquist et al. 1984). It seems probable that both A. san- juanensis and A. eastwoodiana diverged from an ancestral series now called A. ruth- iae. Possibly the similarities between the two geographical isolates are the result of parallel evolution. Asclepias cutleri Woodson is quite distinct morphologically from the aforementioned western members of this complex (e.g., appressed pubescence and filiform leaves). Asclepias brachystephana, located from Trans- Pecos Texas to southern Arizona and south to central Mexico, differs in its linear lanceolate leaves, its sparse, more or less appressed pubescence of the leaves, and its ascending to erect habit. Asclepias sanjuanensis is, however, strik- ingly similar to these and other members of the subgenus in its ecological adaptations. As- clepias cutleri is adapted to unstabilized sand and is conspicuously absent from undisturbed localities. Likewise, A. eastwoodiana, A. cor- difolia, and A. brachystephana are adapted to sites that undergo continuous erosion or pro- longed disturbance, usually sandy or alkaline clay sites. We reject the proposition that these are pioneer species of succession but, rather, believe they are species highly adapted to sites that undergo continual disturbance, e.g., erosion channels, wash slopes, and dunes. As- clepias eastwoodiana is associated with highly erodable alkaline clay hills. Asclepias cordi- folia grows on gravel hills or talus slopes. Asclepias ruthiae, while found occasionally on January 1989 Heiletal.: ANewAsclefias 103 undisturbed habitats, grows often in areas of shifting sands and in gullies and other erosion channels (Cronquist et al. 1984). Until a detailed study of Asclepias is un- dertaken, the relationships of this complex to other subgenera will remain unclear. Barne- by (1945) suggested that A. eastwoodiana is closely allied to A. involucrata Engelm. ex Torr., considered by Woodson (1954) to be within subgenus Asclepias, series Macro- tides. We do not consider A. sanjuanensis to be closely allied to A. involucrata. Acknowledgments Much credit is due the systematic botany students Barbara Jenkins, Londa Smith, and Marc Werthington for the discovery of this plant. We thank Fred Edwards for searching for populations in the San Juan College area and acknowledge Kaye Hugie Thome for the illustrations. Literature Cited Barneby. R C 1945. A new species of Asclepias from Nevada. Leafi. W. Bot. 4: 210-211. Cronquist. A , A Holmgren, N Holmgren. J Reveal, and P Holmgren 1984. Intermountain flora 4: 35-50. Maguire. B , andR Woodson 1941. Two new asclepiads from the southwestern United States. Ann. Mis- souri Bot. Gard. 28: 245-248. Welsh. S L . N D Atwood, S Goodrich, and L C HlGGINS 1987. A Utah flora. Great Basin Nat. Mem. 9: 51-55. Woodson, R 1954. The North American species of Ascle- pias L. Ann. Missouri Bot. Gard. 41: 1-211. EFFECT OF TIMING OF GRAZING ON SOIL-SURFACE CRYPTOGAMIC COMMUNITIES IN A GREAT BASIN LOW-SHRUB DESERT: A PRELIMINARY REPORT James R. Marble1 and Kimball T. Harper Abstract. — Cover and species richness of vascular and cryptogamic components of the plant community were inventoried in experimental grazing paddocks at the USDA/FS Desert Range Experimental Station, Millard County, Utah. The grazing treatments considered have been applied continuously for over 50 years. The effects of heavy (ca 17 sheep days/acre) grazing treatment applied in two different seasons (early winter versus a split between early and late winter) differed significantly between seasons. Cryptogamic cover and cryptogamic species richness both showed larger decreases under early-late as opposed to early winter only grazing. Vascular plant cover (relative to controls) was also reduced by early-late winter grazing, but not to a significant degree. Late season grazing, likewise, had no significant effect on number of vascular species per transect. The importance of cryptogamic plants (algae, mosses, and lichens) as soil stabilizers on Great Basin rangelands has only recently been appreciated by range managers. Man- agement strategies that retain the positive ef- fects of these nonseed plants, without elimi- nating grazing animals from the land, have not yet been established. It is our purpose here to report the effects of different seasons of sheep grazing on cryptogamic plant covers at the Desert Range Experimental Station in west central Utah. Communities of nonvascular cryptogamic plants grow on or just below the soil surface. The organisms that coexist in such communi- ties include species of fungi, algae, lichens, and mosses. When well developed, the crusts play an important role in soil stabilization in deserts (Fletcher and Martin 1948, Kleiner and Harper 1972, 1977, Loope and Gifford 1972, Anderson et al., Recovery, 1982, An- derson et al., Factors, 1982). The effects of cryptogamic crusts on infiltration and sedi- ment release were studied by Loope and Gif- ford (1972). More recent studies have demon- strated the importance of cryptogamic crust in reducing soil erosion and increasing water in- filtration. Lusby et al. (1963, 1971) and Lusby (1979) studied the effect of grazing and sedi- ment yield on western Colorado watersheds. Although Lusby and collaborators did not record cryptogamic cover directly, their re- sults are congruent with the hypothesis that biological crusts are important in the soil sta- bilization process at that site. They found that altering the season of grazing from winter long (15 Oct. -15 May) to early winter only (15 Oct. -15 Feb.) without a change in grazing intensity reduced water runoff by 23% and sediment release by 31%. Lusby et al. (1963, 1971) and Lusby (1979) could not demonstrate a significant change in vascular plant cover associated with the large changes in runoff and sediment release. Harper and St. Clair (1985) found that physical disruption of cryptogamic crusts (as might be caused by trampling by herds of hooved animals) increased the aver- age loss of water as runoff by 51% and in- creased soil loss by 686%. Complete removal of the crust resulted in the loss of 92% more water than was lost from control plots and 1,441% more soil loss. The analysis of species composition of cryptogamic covers has been facilitated by the work of Anderson and Rushforth (1976). They provided photos and/or line drawings of the major algal, moss, and lichen compo- nents of cryptogamic crust communities in the Intermountain West. The role of cryptogamic crusts in nitrogen fixation has been studied by Rychert and Skujins (1974). The impact of 40 years of uncontrolled grazing on a soil cryp- togam community in Navajo National Monu- ment, Arizona, was studied by Brotherson et al. (1983). They found considerable damage to the cryptogamic community and less stable 'Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602. 104 January 1989 Marble, Harper: Cryptogamic Communities 105 soil conditions where grazing had been per- mitted. They recommended that future range research should include analysis of factors that favor persistence of cryptogamic crusts and the impact of established grazing systems on those microplant assemblages. Site Description The study area is located at the Desert Range Experimental Station in Millard County, southwestern Utah. Since the Forest Service, U.S. Department of Agriculture, established the Desert Experimental Range (DER) in 1934, the vegetational conse- quences of three intensities of grazing (light, moderate, and heavy) have been tested in all possible combinations with three different grazing seasons (early, mid-, and late winter). Each combination of intensity and time was replicated twice. Each treatment was applied to a separate 240-acre pasture. Each pasture contains two exclosures that initially had plant cover comparable to that on the remainder of the pasture. The 1-acre (0.40-ha) exclosures have been ungrazed throughout the period of study. Methods Within each of two grazed pastures (one heavily grazed in early winter and the other grazed at the same intensity but with half of the use applied in early winter and the re- mainder in late winter), four sets of paired transects were read. The heavy grazing treat- ment consisted of 17 sheep days of use per acre. Each grazed transect was paired with a transect approximately 16 m away and within a control exclosure. Each transect consisted of cover estimates at 10 subsample points. Cover readings were made using a nested frequency quadrat with eight points per subsample. The 0.25-m frequency quadrat was subdivided into three smaller quadrats nested within; the sizes of the four nested quadrats were 0.25, 0. 125, 0.0625, and 0.0025 m2. Cover percent- ages were based on the number of points in- tercepting each cover type. Percent cover for any cover type was computed by dividing the number of points intercepting that type by all points read along the transect (always 80) and multiplying by 100. The effect of each grazing treatment was analyzed in terms of cover and species rich- ness (species per transect) for the follow- ing cover parameters: absolute cryptogamic, absolute vascular, and percent of total con- tributed by cryptogams (cryptogamic/total). The mean value for each class along each transect was used as a single datum to avoid pseudoreplication. Data were compared us- ing the paired t-test (Zar 1974). Results The early-late winter grazing treatment showed significant differences between con- trol and treatment plots in respect to both cryptogamic cover and number of cryp- togamic species as a percentage of total spe- cies (Tables 1, 2). The average number of cryptogamic species per transect and the average percent of total species contributed by cryptogamic species were significantly re- duced relative to controls under early-late winter grazing (Table 2). Neither vascular plant cover nor the number of vascular species differed significantly between control and treatment plots. Early winter grazing treat- ment showed no significant differences for any variable. Discussion and Conclusions A significant reduction in cryptogamic cover is associated with late winter grazing as indicated by readings of absolute cryptogamic cover and cryptogamic cover as a percent of total living cover. Late winter grazing, how- ever, did not significantly reduce vascular cover. The early winter grazing treatment, in contrast, showed no significant difference in any cover type between treatment and control plots. Species richness comparisons show the same pattern of response to grazing season as cover comparisons. Early-late winter grazing significantly reduced both absolute cryp- togamic species richness and cryptogamic species as a percentage of total species. Early winter grazing, on the other hand, did not significantly reduce either absolute number of cryptogamic species or cryptogamic species as a percentage of all species. The number of vascular species was not significantly reduced under either grazing treatment. 106 Great Basin Naturalist Vol. 49, No. 1 Table 1. Cover values for cryptogamic and vascular plant species on grazed and ungrazed transects. Grazed transects are used at the same intensity (heavy) but in different combinations of seasons. Asterisked t-values de- note significant differences between treatments and con- trols. Table 2. Species richness values for cryptogamic and vascular plant species on grazed and ungrazed transects. Grazed transects are used at the same intensity (heavy) but in different combinations of seasons. Asterisked t-val- ues denote significant differences between treatments and controls. Percent cover No. of species/transect Cryptogamic species as % Cryptogamic species as % Crypto- Crypto- Treatment gamic Vascular of total Treatment gamic Vascular of total Early winter Early winter Grazed Grazed transect transect 1 0.0 7.5 0.0 1 1 13 7.1 2 0.0 18.8 0.0 2 1 12 7.7 3 1.3 27.5 4.5 3 4 10 28.6 4 0.0 27.5 0.0 4 2 8 20.0 Averages 0.3 20.3 1.1 Averages 3.8 10.8 15.9 Control Control transect transect 1 1.3 21.1 5.8 1 4 10 28.6 2 2.5 1.3 65.8 2 4 14 22.2 3 0.0 23.8 0.0 3 4 12 25.0 4 1.3 16.3 7.4 4 4 11 26.7 Averages 1.3 15.6 19.8 Averages 4 11.8 25.6 (t-values) 1.192 0.694 1.457 t-values 2.333 0.739 1.460 Early-late winter Early-late winter Grazed Grazed transect transect 1 0.0 15.0 0.0 1 2 11 15.4 2 0.0 11.3 0.0 2 2 11 15.4 3 0.0 15.0 0.0 3 2 10 16.7 4 0.0 20.0 0.0 4 1 8 11.1 Averages 0.0 15.3 0.0 Averages 1.8 10 14.7 Control Control transect transect 1 2.5 26.3 8.7 1 4 13 23.5 2 3.8 15.0 20.2 2 4 10 28.6 3 7.5 21.3 26.0 3 5 8 38.5 4 5.0 20.0 20.0 4 6 13 31.6 Averages 4.7 20.6 18.7 Averages 4.8 11 30.6 t-values 4.392* 2.251 5. 153* t-values 4.245* 0.632 4.940* t = 3. 182 t =3.182 0.05, (2), 3 0.05,(2), 3 These results suggest that cryptogamic cov- ers are less damaged by early winter grazing than by grazing at the same intensities but in late winter. Freedom from grazing in the late winter and spring while soil moisture is likely to be adequate to permit some regrowth of cryptogams may result in enough surface sta- bilization to significantly reduce runoff and sediment losses due to torrential summer rains. It also appears that avoidance of late winter grazing use permits a small increase in cryptogamic species richness. The results suggest that desert ranges in areas where late winter and early summer rainfall is low and/or torrential when it does occur will suffer depletion of cryptogamic covers when heavy grazing is permitted to continue into the late winter. Lusby (1979) has demonstrated that in western Colorado early winter grazing results in less runoff and January 1989 Marble, Harper.: Cryptogamic Communities 107 erosion than winter-long (to 15 May) grazing at the same intensity of use. Thus, avoidance of late winter use of desert ranges on the Colorado Plateau and in the Great Basin may reduce runoff and sedimentation downstream and prolong the useful life of large reservoirs, their hydroelectric plants, and other associ- ated values. Cryptogamic crusts have the po- tential of slowing soil erosion by both wind and water, enhancing infiltration of precipita- tion, and stimulating vascular plant growth through improved soil water and available nitrogen relations. Literature Cited Anderson, D. C, K T. Harper, and R C Holmgren 1982. Factors influencing development of cryp- togamic soil crusts in Utah deserts. Journal of Range Management 35: 18-185. Anderson, D C , K. T. Harper, and S R. Rushforth 1982. Recovery of cryptogamic crust from grazing on Utah winter ranges. Journal of Range Manage- ment 35: 355-359. Anderson, D. C, and S R Rushforth 1976. The cryp- togamic flora of desert soil crusts in Utah deserts. Nova Hedwigia 28: 691-729. Brotherson. J D , S. R. Rushforth, and J R Johansen 1983. Effects of long-term grazing on cryptogam crust cover in Navajo National Monument, Ari- zona. Journal of Range Management 36: 579-581. Fletcher, J. E , and W. P. Martin. 1948. Some effects of algae and molds in the rain-crust of desert soils. Ecology 29: 95-100. Harper, K T., and L. L St Clair 1985. Cryptogamic soil crusts on arid and semiarid rangelands in Utah: effects on seedling establishment and soil stability. Final report on BLM contract No. BLM AA 851- CTI-48. Kleiner, E F , and K T Harper 1972. Environment and community organization in grasslands of Canyonlands National Park. Ecology 53: 229-309. 1977. Soil properties in relation to cryptogamic ground cover in Canyonlands National Park. Jour- nal of Range Management 30: 202-205. Loope, W. L. , and G. F. Gifford. 1972. Influence of a soil microfloral crust on select properties of soils under pinyon-juniper in southeastern Utah. Journal of Soil and Water Conservation 27: 164-167. Lusby, G. C 1979. Effects of grazing on runoff and sedi- ment yield from desert rangeland at Badger Wash in western Colorado, 1953-73. Geological Survey Water-Supplv Paper 1532-1. Lusby, G C .. V H Reid. andO D Knipe. 1971. Effects of grazing on the hydrology and biology of the Bad- ger Wash Basin in western Colorado, 1953-66. Geological Survey Water-Supply Paper 1532-D. Lusby. G C , G T Turner, J R Thompson, and V. H. Reid 1963. Hydrologic and biotic characteristics of grazed and ungrazed watersheds of the Badger Wash Basin in western Colorado, 1953-58. Geo- logical Survey Water-Supply Paper 1532-B. 73 pp. Rychert, R C , and J Skujins 1974. Nitrogen fixation by blue-green algae-lichen crusts in the Great Basin Desert. Soil Science Society of America Proceed- ings 38: 768-771. Zar. J H 1974. Biostatistical analysis. Prentice-Hall, Inc., Englewood Cliffs, California. 620pp. SIZE AND OVERLAP OF TOWNSEND GROUND SQUIRREL HOME RANGES Nicholas C. Nydegger1 and Donald R. Johnson Abstract. — We evaluated movement distance (an index of home range size) based on capture histories of 32 postbreeding Townsend ground squirrels (Spermophilus townsendii) on a 15 X 15 trap grid in southwestern Idaho. Capture frequencies and movement distances of adult males were significantly greater than those of other sex/age groups. Members of the same sex/age group were rarely captured at the same grid location, evidence of mutual avoidance within sex/age groups. These results are compared with those for other species of ground squirrels. Ground squirrels offer favorable opportuni- ties to investigate space use because of their abundance, relative ease of capture, and diel pattern of activity. Conversely, some popula- tions demonstrate temporal differences in movement patterns related to sexual activity, food availability, density, dispersal, and inter- specific competition. Thus, comparisons be- tween congeners or even intraspecific popula- tions require caution. In this study we evaluate indices of home range size of Town- send ground squirrels in the Birds of Prey Area in southwestern Idaho during the post- breeding season when home ranges were as- sumed to be stabilized. Methods In 1983 we established a 280 x 280-m grid of 225 Pymatuning live traps (Tryon and Snyder 1973) with 20-m spacing in a stand of big sagebrush (Artemisia tridentata) and winterfat (Cerratoides lanata). The grid in- corporated Site 5 of Smith and Johnson (1985). Traps were baited with apple and opened dur- ing daylight hours 1-3 days per week between 15 February and 30 April, a total of 19 days. Captured animals were toe-clipped, weighed, and immediately released. Because trapping had been conducted at this site annually since 1975, the ages of residents first marked as juveniles were known. We classified un- marked squirrels when first captured as juve- niles, yearlings, adults, or unknown (at least one year of age) using the criteria of Smith and Johnson (1985). Capture records of juveniles (young of the year) and those in which > 50% of the captures occurred on the grid edge were eliminated from the analysis. We also eliminated capture histories that did not con- tain at least four grid locations, assuming that these inadequately sampled space use. Five movements (involving three animals) > 200 m were thought to be exploratory and were ignored in our calculations. Analysis of movement distances assumes independence of capture locations (Swihart and Slade 1985). Serial dependence results in biased estimates of movement distance de- pending upon the pattern of captures. We used the microversion of the HOMERANGE computer program (Samuel et al. 1985) to cal- culate T2/R2, a test for independence in cap- ture locations. Because 21 of 32 capture his- tories lacked independence in locations, we evaluated capture data of sex/age groups rather than individuals to minimize the effects of autocorrelation. There was no significant difference (P < . 13) in successive movement distances recorded during the same day and those separated by more than 24 hrs. Thus, serial dependence was due to the pattern of spacial use rather than insufficient time be- tween successive captures (Swihart and Slade 1987). We calculated (1) DS, the mean distance between successive captures, as d/irij where dj is the cumulative linear distance between successive capture locations and m, is the number of movements for squirrels of sex/age group i, and, (2) AD (Koeppl et al. 1977), the mean distance between all capture locations as d;/[2 n; (n, — 1)/2] where d{ is the cumulative linear distance between all capture locations Department of Biological Sciences, University of Idaho, Moscow, Idaho 83843. 108 January 1989 NYDEGGER, JOHNSON: SQUIRREL HOME RANGES 109 Table 1. Mean capture frequency and movement distance (± SE) in m ofTownsend ground squirrels: DS, mean distance between successive captures, and AD, mean distance between all capture locations. Sample size in parenthe- ses. * = significant difference (P < .05) from other sex/age groups. ** = significant difference between sexes. Sex/age No. captures DS AD Ad males 19.1 ± 2.0(191)* 28.6 ± 1.3(181)* 38.7 ± 2.9(327)* Yr males 8.1 ± 1.0 (58) 26.9 ± 4.4 (51) 31.4 ± 3.4 (54) Ad females 9.5 ± 2.7 (66) 23.0 ± 2.4 (58) 35.4 ±2.7 (57) Yr females 11.0 ± 1.9 (76) 25.3 ± 2.0 (70) 35.5 ±3.2 (96) All males 14.6 ± 1.8(249)** 28.2 ± 1.4(232) 37.7 ± 0.9(381)** All females 10.3 ± 1.7(142) 24.3 ± 1.5(128) 35.4 ± 1.5(153) All animals 12.2 ± 1.2(391) 26.8 ± 3.8(360) 37.0 ± 3.0(534) Table 2. Overlap (%) of capture locations among sex/age groups. * = significant difference (P < .03) from cells involving unlike sex/age groups. Traps capturir >g Also capturing Ad males Y r males Ad females Yr females Ad males 8* 23 17 22 Yr males 50 3* 13 20 Ad females 45 16 6* 26 Yr females 57 17 23 11* and ns is the number of locations of squirrels within a sex/age group i. Results Of the 146 squirrels captured, the capture histories of 32 animals one or more years of age (17 males and 15 females) met the criteria established for spacial use analysis. The mean capture frequency of adult males was significantly greater (F = 2.6; P < .05) than that of other sex/age groups (Table 1). Both measures of movement distance were significantly greater for adult males than other sex/age groups (DS: F = 3.59, P < .02; AD: F = 5.83, P < .001). Of these three vari- ables, capture frequency and AD were signifi- cantly different (F = 5.4, P < .03 and F = 6.2, P < .02, respectively) between the sexes (Table 1). The mean distance between centers of activity of 11 adults (7 males, 4 females) in successive years (1982 and 1983) was 52 ± 11 m (range 12 to 131 m), similar to that of adult Columbian ground squirrels (Spermophilus columbianus) (Murie and Harris 1984). Home range overlap was measured as the percentage of capture stations at which mem- bers of the same or different sex/age groups were also taken. Overlap was significantly less frequent among members of the same sex/age group than among different groups (F = 6.7, P < .03, Table 2). Differences in the mean and total number of capture locations among sex/ age groups produced asymmetries in home range overlaps. For example, yearling males occurred at 23% of the stations at which adult males were captured, but adult males oc- curred at 50% of the stations at which yearling males were captured (Table 2). Discussion Home range sizes of postbreeding ground squirrels may or may not differ between the sexes (Owings et al. 1977, Michener 1979). In our study home ranges of adult males were significantly larger than those of other sex/age groups (Table 1). Postbreeding males were also more mobile and prone to capture. It is likely that the difference in home range size between the sexes is increased during the breeding season when adult males search for females in estrus (Murie and Harris 1978, Michener 1979). Spacing in ground squirrel populations re- flects social status and in some cases kinship (Michener 1979). In our study overlap in space use by members of the same sex/age group was rare in comparison with that by members of different groups (Table 2). Squir- rels clearly avoided trap locations used by individuals of the same sex and age. Using direct observation, Michener (1979) found 110 Great Basin Naturalist Vol. 49, No. 1 that adult Richardson ground squirrels (Sper- mophilus richardsoni ) of the same sex main- tained greater distances when simultaneously active aboveground than distances between their centers of activity, another example of mutual avoidance. However, there was exten- sive overlap in the home ranges of both adult females and adult males. We found that over- lap in space use by yearling and adult females was no more frequent than that of other sex/ age combinations (Table 2). Thus, we are un- certain if female offspring are likely to take up residency near their mothers as observed in Richardson ground squirrels (Michener 1979). In summary, the home range characteris- tics of postbreeding Townsend ground squir- rels are similar to those of certain other spe- cies of the genus Spermophilus. Adult males occupy larger home ranges than those of other sex/age groups. There is significantly less overlap in space use by members of the same sex/age group than those of different groups. Acknowledgments We thank Karen Steenhof and an anony- mous reviewer for comments that signifi- cantly improved the manuscript. Tim Rey- nolds and his students at Boise State University provided field assistance. Fred Leban provided guidance regarding data anal- ysis. We thank Michael Kochert, research leader, Snake River Birds of Prey Area, for encouragement and cooperation. This work was part of the USDI, Bureau of Land Man- agement, Snake River Birds of Prey Research Project, funded under contract 52500-CT5- 1002 to the University of Idaho. Literature Cited Koeppl. J W , N. A. Slade. and R. S Hoffmann 1977. Distance between observations as an index of average home range size. Amer. Midi. Nat. 98: 476-482. Michener, G. R. 1979. Spatial relationships and social organization of adult Richardson's ground squir- rels. Canadian J. Zool. 57: 125-139. Murie, J O , and M A Harris 1978. Territoriality and dominance in male Columbian ground squirrels (Spermophilus columbianus). Canadian J. Zool. 56: 2402-2412. 1984. The history of individuals in a population of Columbian ground squirrels: source, settlement, and site attachment. Pages 353-373 in J. O. Murie and G. R. Michener, eds., The biology of ground- dwelling sciurids. University of Nebraska Press, Lincoln. Owings, D H . M Borchert, and R Virginia 1977. The behaviour of California ground squirrels. Animal Behav. 25: 221-230. Samuel, M . et al. 1985. User's manual for program HOMERANGE. Forestry, Wildlife and Range Expt. Sta. , University of Idaho, Tech. Rept. 15. 70 pp. Smith, G W , and D R Johnson 1985. Demography of a Townsend ground squirrel population in south- western Idaho. Ecology 66: 171-178. Swihart, R K . and N. A. Slade. 1985. Testing for inde- pendence of observations in animal movements. Ecology 66: 1176-1184. 1987. A test for independence of movements as shown by live trapping. Amer. Midi. Nat. 117: 204-207. Tryon, C A , and D P Snyder. 1973. Biology of the eastern chipmunk, Tamias striatus: life tables, age distributions, and trends in population numbers. J. Mammal. 54: 145-168. PISOLITHUS TINCTORIUS , A GASTEROMYCETE, ASSOCIATED WITH JEFFREY AND SIERRA LODGEPOLE PINE ON ACID MINE SPOILS IN THE SIERRA NEVADA R. F. Walker1 Abstract. — Basidiocarps o{ Pisolithns tinctorius , a gasteromycetous fungus adapted to harsh sites, were observed in association with Jeffrey and Sierra lodgepole pine on acid mine spoils in northeastern California. Subterranean mycelial strands were traced from these basidiocarps to the root systems of the two pine species, which had ectomycorrhizae characteristic of those formed by this fungus in symbiotic relationships with conifer hosts. The ectomycorrhizal fungus Pisolithus tinc- torius (Pers.) Coker & Couch occurs essen- tially worldwide in temperate, subtropical, and tropical zones and in symbiotic associa- tion with a wide variety of conifer and hard- wood hosts (Marx 1977). In the United States its basidiocarps have most frequently been observed on harsh sites in the East, South, and Midwest associated with various pine spe- cies (Lampky and Peterson 1963, Schramm 1966, Hile and Hennen 1969, Lampky and Lampky 1973, Marx 1975). Prompted by these reports of the ability of this Gas- teromycete to flourish on infertile and often toxic substrates, researchers have inoculated several southern pine species in forest nurs- eries with this mycobiont (Marx et al. 1984). Their efforts have resulted in substantial im- provement in seedling performance upon out- planting on a variety of adverse sites (Berry and Marx 1978, Marx and Artman 1979, Walker et al. 1985). Current research efforts include identification of new tree hosts for inoculation trials and development of more effective inocula incorporating locally adapted P. tinctorius isolates. Numerous basidiocarps of P. tinctorius were observed on spoils of the Leviathan Mine in Alpine County, California (38°42'30"N, 119°39'15"W), in mid-September 1988. This open-pit sulfur mine consists of approximately 100 ha at an elevation of 2,200 m and receives an average annual precipitation of about 50 cm, primarily as snowfall. Early attempts to revegetate the mine, inactive since 1962, failed. More recent efforts using woody and herbaceous species were somewhat success- ful, and limited natural vegetation, comprised largely of woody species, has become reestab- lished near adjoining undisturbed forest and woodland. Nevertheless, vegetative cover is sparse over most of the site. A comprehensive examination of soil chemical properties by Butterfield and Tueller (1980) revealed that most of these spoils have a pH of 4.0 to 4.5, a deficiency of plant-available N, and poten- tially phytotoxic concentrations of Al. A majority of the P . tinctorius basidiocarps observed were in close proximity to seedlings and saplings of Jeffrey pine (Pinus jeffreyi Grev. & Balf), resulting from both earlier plantings and natural colonization of the mine spoils. Fewer basidiocarps were found near naturally invading seedlings and saplings of Sierra lodgepole pine (Pinus contorta var. murrayana [Grev. & Balf] Engelm.). Califor- nia white fir (Abies concolor var. lowiana [Gord.] Lemm.) and occasional singleleaf pinyon (Pinus monophylla Torr. & Frem.), Utah juniper (Juniperus osteosperma [Torr.] Little), and quaking aspen (Populus tremu- loides Michx.) were also found growing in the mine, but no basidiocarps were observed in the immediate vicinity of any of these four species. These basidiocarps, dark yellow to brown in color, matched the description of Coker and Couch (1928) for P. tinctorius, which is very distinctive due to the presence of peridioles in the upper portion of the gleba. Specimens examined on this site included those of the stipitate, substipitate, and sessile forms, 'Department of Range, Wildlife and Forestry, University of Nevada, Reno. Nevada 89512. Ill 112 Great Basin Naturalist Vol. 49, No. 1 which varied in size from 8 to 21 em in height and from 3.5 to 8 cm in diameter. As many as five basidiocarps were observed encircling solitary Jeffrey pines, but no more than three were observed around isolated lodgepole pines. Clusters of either species tended to be accompanied by large numbers of basidio- carps, the total of which numbered several hundred over the entire site. Removal of the soil from around individual basidiocarps exposed mycelial strands that were traced up to 1.5 m through spoil materi- als to root systems of both Jeffrey and lodge- pole pine. With gold-yellow pigmentation, these mycelial strands compared favorably with the description of P. tinctorias rhi- zomorphs provided by Schramm (1966) and were joined at the root systems of both tree species with ectomycorrhizae closely resem- bling those Marx and Bryan (1975) identified as formed by this mycobiont. Subsequently, it is reasonable to conclude that the ectomycor- rhizal root systems of Jeffrey and Sierra lodge- pole pine on this site resulted from a symbi- otic association with P . tinctorius . Acknowledgments This paper contains results of the Nevada Agricultural Experiment Station Research Project 619 funded by the Mclntire-Stennis Cooperative Forestry Research Program. The author is indebted to P. M. Murphy of the Division of Forestry, Nevada Department of Conservation and Natural Resources, and to D. C. Prusso of the Department of Biology, University of Nevada, Reno, for their invalu- able assistance. Literature Cited Berry, C R . and D H. Marx. 1978. Effects of Pisolithus tinctorius ectomycorrhizae on growth of loblolly and Virginia pines in the Tennessee Copper Basin. USDA For. Serv. Res. Note SE-264. 6 pp. Butterfield. R I., and P T. Tueller. 1980. Revegeta- tion potential of acid mine wastes in northeastern California. Reclam. Rev. 3: 21-31. Coker. W C , and J N Couch 1928. The Gastero- mycetes of the eastern United States and Canada. University of North Carolina Press, Chapel Hill. 201 pp. Hile, N ., and J F Hennen. 1969. In vitro culture of Pisolithus tinctorius mycelium. Mycologia 61: 195-198. Lampky, S. A , and J. R Lampky. 1973. Pisolithus in cen- tral Florida. Mycologia 65: 1210-1212. Lampky, J. R , and J E Peterson. 1963. Pisolithus tincto- rius associated with pines in Missouri. Mycologia 55: 675-678. Marx, D H 1975. Mycorrhizae and establishment of trees on strip-mined land. Ohio J. Sci. 75: 288-297. 1977. Tree host range and world distribution of the ectomycorrhizal fungus Pisolithus tinctorius. Canadian J. Microbiol. 23: 217-223. Marx, D H , and J. D. Artman. 1979. Pisolithus tincto- rius ectomycorrhizae improve survival and growth of pine seedlings on acid coal spoils in Kentucky and Virginia. Reclam. Rev. 2: 23-31. Marx. D. H , and W C Bryan. 1975. Growth and ecto- mycorrhizal development of loblolly pine seed- lings in fumigated soil infested with the fungal symbiont Pisolithus tinctorius. For. Sci. 21: 245-254. Marx, D H , C E Cordell, D S Kenney, J G Mexal, J D Artman, J W Riffle, and R. J. Molina 1984. Commercial vegetative inoculum of Piso- lithus tinctorius and inoculation techniques for development of ectomycorrhizae on bare-root tree seedlings. For. Sci. Monogr. 25. 101 pp. Schramm, J. R. 1966. Plant colonization studies on black wastes from anthracite mining in Pennsylvania. Trans. Amer. Philos. Soc. 56: 1-194. Walker, R F , D C West, S B McLauchlin, and C C Amundsen. 1985. The performance of loblolly, Virginia, and shortleaf pine on a reclaimed sur- face mine as affected by Pisolithus tinctorius ecto- mycorrhizae and fertilization. Pages 410-416 in E. Shoulders, ed., Proc. Third Bienn. South. Sil- via Res. Conf. , November 1984, Atlanta, Geor- gia. USDA For. Serv. Gen. Tech. Rept. SO-54. 589 pp. SPATIAL AND TEMPORAL VARIABILITY IN PERENNIAL AND ANNUAL VEGETATION AT CHACO CANYON, NEW MEXICO Anne C. Cully1 2 and Jack F. Cully, Jr. ' 3 Abstract. — Annual plant populations in northwestern New Mexico were found to be spatially and temporally highly variable. During favorable years annual plant species have patterns of dominance and diversity that are different from those of perennial species. Measurement of perennial plant diversity in plant communities is a poor predictor of productivity. Both perennial and annual components of plant communities should be considered in measurements of diversity and productivity. The relationship between environmental diversity and animal populations is of interest to ecologists in their attempts to understand the factors that control the composition of bi- otic communities (Pielou 1974). Whitaker (1972) and Pielou (1974) discuss two different types of diversity. Alpha diversity is applied to small, homogeneous areas such as local plant communities or habitat types, defined by more or less natural boundaries. Beta diver- sity is the measurement of differences be- tween these small, homogeneous areas within a geographic region. Perennial plant species are often used in descriptive and quantitative studies of plant communities that include measurements of dominance and diversity. While perennial plants are the relatively more stable part of the plant community, annual plant species are a conspicuous component during years when environmental conditions are right for germi- nation, growth, and reproduction. The fluctu- ation of annual plant populations has been studied in several arid and semiarid locations in the southwestern United States (Beatley 1969, Juhren et al. 1956, Patten 1978, Tevis 1958). In desert and semiarid ecosystems, an- nual plant production is characterized by a fascinating cycle of years of low germination interspersed with years of superabundant ger- mination and reproduction. This production may be a significant resource for consumer populations, but it may also be overlooked by researchers because of the high degree of vari- ability from year to year. In this study we report on work done at Chaco Canyon, in northwestern New Mexico. Our work included measurements of peren- nial species to describe habitat types and their alpha diversity. In addition, we documented one superabundant year of annual plant pro- duction, followed by several consecutive years of little or no annual production. In northwestern New Mexico fluctuation in an- nual plant populations has been described (Jones 1972, Potter 1974, Scott 1980), but quantitative data have not been reported. Study Areas Chaco Canyon is in the central San Juan Basin in northwestern New Mexico. The San Juan Basin is an extension of the Colorado River drainage, although the Chihuahuan Desert to the south influences the climate and the composition of the vegetation. The soils are generally derived from shale and sand- stone. The vegetation is dominated by mem- bers of the Asteraceae and Chenopodiaceae, including sagebrush (Artemisia), rabbitbrush (Chrysothamnus), and various saltbushes (Atriplex) (Donart et al. 1978, Shreve 1942). As in the Great Basin to the west, winter precipitation is a significant component of the total in northwestern New Mexico, but the summer monsoonal rains from the south provide a large proportion of the rainfall during the growing season (Tuan et al. 1973). A summary of weather data at Chaco Can- yon over a 20-year period, 1957-1977 (Cully Biology Department. University of New Mexico, Albuquerque, New Mexico 87131. Present address: New Mexico Energy, Minerals and Natural Resources Department, Forestry Division, Villagra Building, Santa Fe, New Mexico 87503. Present address: Department of Biological Sciences, University of Notre Dame, Notre Dame, Indiana. 113 114 Great Basin Naturalist Vol. 49, No. 1 MAJOR HABITAT TYPES | Bench Flood Plain Pinyon - Juniper Wash 1 Shrub Grassland Fig. 1. Map of Chaco Canyon showing the locations of the study sites and the distribution of vegetation types. 1984b), showed that annual precipitation averaged 20.6 cm. The heaviest precipitation occurred during July through September, with each of these months averaging >3 cm. The balance of the rainfall was rather evenly distributed throughout the remainder of the year. The San Juan Basin can have two periods of plant productivity, one in spring if there is normal or above normal precipitation, and another more predictable period during late summer and fall in response to the monsoonal rains of late summer. The study sites for this report are five areas used by J. Cully (1984a) in his study of the bird and rodent communities of the San Juan Basin. They were selected from Kelley and Potter's (1974) vegetation map of Chaco Canyon. The principal criteria for the choices were that each area included habitat sufficient to contain a transect 120 m wide by 1.6 km long, that the areas were different from one another, and that in combination they repre- sented the major habitats at Chaco Canyon. The sites at Pueblo Alto and at the mouth of Werito's Bincon were added for this study to gather additional data on annual plant produc- tivity. Upland Areas The bench. — The bench lies within the Hilaria-Bouteloua-Atriplex vegetation type (Kelley and Potter 1974). The bench (Fig. 1) is elevated above the floodplain of the Chaco Wash and is bounded on the north and south by sandstone cliffs. Soils are thin, and there is a great deal of exposed bedrock. Pinyonjuniper. — The woodlands of the higher elevations of Chacra Mesa are domi- nated by one-seed juniper (Janiperus mono- sperma). Pinyon pine (Pinus edulis) is sub- dominant. Pueblo Alto. — Pueblo Alto is located in a shrub grassland, the Atriplex-Oryzopsis- Sporobolus vegetation type (Kelley and Pot- ter 1974). The site is located on the mesa north of the canyon proper. January 1989 Cully, Cully: Chaco Canyon Vegetation 115 Floodplain Areas Casa Chiquita and Pueblo Bonito. — Two study sites lie within the Atriplex-Sarcobatus vegetation type in the floodplain of the wash, which is dominated in the western portion of Chaco Canyon by four-wing saltbush, shad- scale, and black greasewood (Kelley and Pot- ter 1974). The sites are bordered on the south by the wash and on the north by sandstone cliffs and the bench habitat. The wash. — The wash is the erosion chan- nel of the Chaco River. It bisects the flood- plain through the length of the canyon in the park. Riparian, woody vegetation is character- istic of the wash habitat; the inner channel is dominated by rabbitbrush (Chrysothamnus nauseosus), sandbar willow (Salix exigua), and black greasewood (Sarcobatus vermicula- tus). Cottonwood (Pupulus fremontii), four- wing saltbush (Atriplex canescens), and tama- risk (Tamarix spp.) also occur. Weritos Rincon. — This study site is at the mouth of a large rincon, or side canyon, south- west of the main canyon. According to Kelley and Potter (1974), the vegetation is similar to that on the mesa tops surrounding the Pueblo Alto study area. Methods Perennial plant cover was measured using the line intercept method (Canfield 1941). Cover was measured on twenty-five 10-m lines stretched in alternate directions at 62-m intervals along a 1.6-km transect in each habi- tat. This yielded a total of 250 m sampled in each habitat. Each perennial plant species in- tercepted was measured to the nearest cm. The bench, Casa Chiquita, and the wash were sampled during April 1979. Pueblo Bonito and the pinyon-juniper sites were sampled during May 1981. To identify the species that most easily dis- tinguish habitats, we applied discriminant function analysis to the cover data from each 10-m line segment in each habitat. The habi- tats were the discriminating variables, and the 29 species of plants that were encountered on three or more (of a total of 125 segments) were the predictor variables. This analysis was done on a microcomputer using SPSSPC + Advanced Statistics (Norusis 1986). The cover data at each transect were ana- lyzed for species richness, or S (number of species) = H0; two indices of species diversity, 1/2 Pi = H, (Simpson 1949) and exp (— Epj In Pi) == H2 (Shannon and Weaver 1949); and evenness, or J, where J = H2/H0 (Peet 1974, Pielou 1974). Hill (1973) suggests that even- ness as measured above is subject to change with change in sample size, and that the ratio H,/H0 is a better ratio to describe evenness. This is partly because H2 always lies between H„ and H,. Since J is still common in the literature, we present both measures of even- ness. Annual plant densities were sampled at the bench, Casa Chiquita, Pueblo Bonito, and Pueblo Alto in June 1979, 1980, and 1981, and at Weritos Rincon during June 1979. At each area data were taken from twenty 1/2-m2 quadrats. These were placed at 10-m intervals along 100-m tapes laid at random in a 1-ha area along the transect used to measure cover at Casa Chiquita, Pueblo Bonito, and the bench. Cover was not measured at Pueblo Alto or at the mouth of Weritos Rincon; a transect was laid out in the same way at these two sites as at the previous sites to gather the annual data. The number of annual plants within each quadrat was counted by species. Then all veg- etation was picked and saved in plastic bags to determine the aboveground biomass produc- tion. The plant material was air-dried at room temperature for a minimum of two months. Each sample was weighed to the nearest 0. 1 g. Subsequently, the seeds were removed from foliage, stems, and flower parts in a seed sepa- rator and weighed separately. Results Perennial Plant Species Dominance In the upland areas the dominant species on the bench were Mormon tea (Ephedra viridis), Bigelow's sagebrush (Artemisia bi- glovii), wild buckwheat (Eriogonum spp.), Greene's rabbitbrush (Chrysothamnus greenei), galleta grass (Hilariajamesii), and Indian rice- grass (Oryzopsis hymenoides) (Table 1). The pinyon-juniper habitat on Chacra Mesa was dominated by Bigelow's sagebrush, mountain mahogany (Cercocarpus montanus), three- leaf sumac (Rhus trilobata ), pinyon pine, one- seed juniper, and galleta grass. On the flood- plain the dominant species at Casa Chiquita were broadscale (Atriplex obovata), black greasewood, torrey seepweed (Sueda torrey- 116 Great Basin Naturalist Vol. 49, No. 1 Table 1. Cover and diversity at five habitats at Chaco Canyon. The values are the number of cm intercepted for each species at twenty-five 10-m lines in each habitat. H„ = number of species; H, = 1/P, ; H2 = -2 P(LnP,; J = H2/H0; evenness = Hj/H,,. Casa Pueblo Pinyon- Species Bench Chiquita Bonito Wash juniper Agropyron spp. 0 0 0 71 0 Aristida spp. 0 0 0 0 17 Artemisia bigelovii 1195 0 0 0 665 Artemisia dracuncttloides 0 0 0 190 0 Artemisia tridentata 0 0 0 573 51 Astragalus spp. 0 0 0 0 34 Atriplex canescens 64 187 3829 0 0 Atriplex obovata 0 1808 0 0 0 Bouteloua gracilis 0 0 0 0 80 Cercocarpus montanus 0 0 0 0 376 Chrysopsis villosus 27 0 0 0 0 Chrysothamnus greenei 200 0 0 0 0 Chrysothamnus nauseosus 0 0 0 2953 100 Chrysothamnus pulchellus 107 0 0 0 0 Ephedra viridis 680 0 0 0 23 Eriogonum spp. 516 0 0 0 11 Eurotia lanata 49 0 0 0 83 Fallugia paradoxa 0 0 0 0 95 Gutierrezia sarothrae 10 0 21 0 100 Hilaria jamesii 327 0 325 0 181 Juniperus monosperma 0 0 0 0 2117 Lycium pallidum 0 0 44 0 29 Muhlenbergia pungens 92 0 0 0 0 Opuntia spp. 6 0 0 0 18 Oryzopsis hymenoides 324 0 4 80 10 Other 24 0 0 151 23 Pinus edulis 0 0 0 0 317 Populus fremontii 0 0 0 200 0 Rhus trilobata 79 0 0 0 217 Salix spp. 0 0 0 1772 0 Sarcobatus venniculatus 0 1215 2260 0 0 Sporobolus airoides 0 765 131 309 0 Sporobolus spp. 101 0 0 0 0 Sueda torreyana 0 286 202 0 0 Tamarix pentandra 0 0 0 2197 0 Yucca spp. 51 0 0 0 35 Total cover 3952 4261 6816 8496 4582 % cover 16% 17% 27% 34% 18% H„ 17 5 8 10 21 H, 6.30 3.33 2.33 4.19 3.96 H2 2.18 1.35 1.07 1.68 1.97 J 0.13 0.27 0.13 0.17 0.09 Evenness 0.37 0.67 0.21 0.41 0.19 ana), and alkali sacaton (Sporobolus airoides). Pueblo Bonito was similar to Casa Chiquita, except that broadscale was missing and four- wing saltbush was much more important than at Casa Chiquita. Black greasewood and seep- weed were also important at Pueblo Bonito. Galleta grass was absent at Casa Chiquita but contributed 2% cover at Pueblo Bonito. The wash was dominated by rabbitbrush, tama- risk, sandbar willow, and big sagebrush. Shrubs and forbs were the primary compo- nents of the vegetation at all study sites sam- pled for perennial vegetation. In the discriminant function analysis there were 12 species important in discriminating between the habitats (Table 2). The list in- cludes many of the species listed above as dominants plus Artemisia dracunculoides. Although galleta, wild buckwheat, mountain mahogany, three-leaf sumac, Torrey seep- weed, big sagebrush, and sandbar willow were all subdominant in one or more habitats, January 1989 Cully, Cully: Chaco Canyon Vegetation 117 Table 2. List of species used by the discriminant func- tion analysis and the univariate F value in a oneway analysis of variance of each species for differences be- tween habitats. All values are significant at P < .001. Species F to enter 1 Atriplex obovata 27.04 2 Atriplex canescens 16.75 3 Artemisia bigelovii 15.68 4 Chrysothamnus nauseosus 14.92 5 Tainarix pentandra 8.23 6 Oryzopsis hymenoides 8.06 7 Sarcobatus vermiculatus 7.57 8 Chrysothamnus greenii 7.47 9 Ephedra viridis 7.43 10 Sporobolus airoides 7.42 11 Juniperus monosperma 6.89 12 Artemisia dracunculoides 6.66 they were not used by the DFA. This may be partly due to the high probability level re- quired for inclusion with 27 variables (P < .002; .05/27). Figure 2 shows the relationships of the observations from each habitat on canonical variables (discriminant function axes) 1 and 2. Casa Chiquita and Pueblo Bonito are similar to each other. The bench and pinyon-juniper habitats are also very sim- ilar; in fact, they are almost completely over- lapping on canonical axes 1 and 2. (They are separated on axis 4 where juniper is an impor- tant variable.) The wash habitat was distinct from the others with a small area of overlap on canonical axes 1 and 2. One value of the DFA is the ability to show which habitats are most similar to each other and, by a jackknife procedure, to show how accurately the habitats can be distinguished on the basis of the predictor variables. The jackknife procedure takes each of the observa- tions used to derive the discriminant func- tions and tests, a posteriori, the accuracy with which the cases are attributed to the correct groups; it is a test of the accuracy of the dis- criminant functions to discriminate between groups. At all habitats the DFA correctly assigned 80% or more of the samples. At the bench two samples were misclassified to pinyon-juniper, and at the pinyon-juniper two were mis- classified to the bench. These two habitats are broadly overlapping on DF axes 1 and 2 C -1.8-| o O -2.7- — I- 1 .8 -.6 0 .6 1.8 Canonical Variable 1 Fig. 2. Relationship of each habitat based on discriminant function analysis of perennial vegetation on the first two discriminant function axes (canonical variables), which account for 65% of the vegetative variance. A = bench, B = Casa Chiquita, C = Pueblo Bonito, D = wash, and E = pinyon-juniper. 118 Great Basin Naturalist Vol. 49, No. 1 Table 3. Results of the discriminant function analysis bootstrap analysis to determine the accuracy of classifying cover samples to their correct habitats. See text for explanation. % Num ber of ( :'ases classified into group Casa Pueblo Pinyon- Group correct Bench Chiquita Bonito Wash juniper Bench 92 23 0 0 0 2 Casa Chiquita 80 0 20 4 0 1 Pueblo Bonito 88 0 1 22 0 2 Wash 80 0 0 2 20 3 Pinyon-juniper 92 2 0 0 0 23 Total 25 21 28 20 31 (Fig. 2). They are also similar in species com- position, species richness, and cover (Table 1). Casa Chiquita and Pueblo Bonito were classified by Kelley and Potter (1974) as belonging to the same habitat type; thus, it is not surprising that four samples from Casa Chiquita were misclassified to Pueblo Bonito and one from Pueblo Bonito was misclassified to Casa Chiquita. The perennial vegetation at Pueblo Bonito was more variable than that at Casa Chiquita, and two samples were misclas- sified to the pinyon-juniper, which shared four species with Pueblo Bonito. The wash was the most variable habitat on DF axes 1 and 2 and had three samples misclassified to pinyon-juniper and two to Pueblo Bonito. Perennial Plant Species Diversity and Cover The bench had the second highest species richness and the highest diversity according to the two diversity indices (Table 1). It also had the lowest cover of the five habitats. The pinyon-juniper habitat was similar to the bench in its high species diversity, particu- larly richness, and low cover values. On the floodplain, Casa Chiquita had considerably lower diversity than the bench, but cover that was similar in value. Pueblo Bonito had simi- lar diversity to Casa Chiquita, but much higher cover. The wash had similar cover, but greater diversity. Annual Plant Species Density and Diversity At all five areas sampled for annual plant densities the total densities were considerably higher in 1979 than in the following two years (Table 4). Annual species richness was also higher during 1979, with most species occur- ring only in that year. The densities varied considerably from site to site, the floodplain sites producing higher annual densities than the upland locations. The annual species composition at Werito's Rincon was different from the other sites (Table 4, Fig. 3). In 1979 pinnate tansy-mus- tard (Descurainia pinnata) was the dominant annual in terms of density at all sampling loca- tions except Werito's Rincon. There, stickleaf (Mentzelia spp.) was the most abundant. Biomass Measures At the floodplain sites of Casa Chiquita and Pueblo Bonito the high annual plant densities were accompanied by high biomass in 1979 (Table 5, Fig. 4). At Werito's Rincon, in spite of densities similar to those at Casa Chiquita, the total biomass and seed biomass were much lower in 1979. The differences between these study areas in biomass may have been related to the differences in species composi- tion of the annual populations at the two sites or to local differences in soil conditions and water availability. Overall, the flood plain sites were more productive than the upland sites during 1979. Biomass was drastically re- duced at Casa Chiquita and Pueblo Bonito in 1980. However, at the upland sites of Pueblo Alto and the bench, total plant biomass and seed biomass were higher than in 1979. The biomass figures include grasses, a component of the perennial vegetation whose growth and reproduction may not be affected by winter- spring moisture until later in the same year or the following year. Biomass was low at all sites in 1981. Discussion Because perennial plant species are rela- tively stable components of the plant commu- nity, measurements of their characteristics January 1989 Cully, Cully: Chaco Canyon Vegetation 119 Table 4. Anni jal plan t dens ities at Chaco Canyon (no. /m2). Casa Chiquita 1979 1980 19S1 Pue bio Bonito Bench Pueblo Alto Werito s Bincon 1979 1980 1981 1979 1980 1981 1979 1980 1981 1979 Astragalus 0 0 1.6 0 0 0 2.7 2.7 1.6 11.9 0 0 4.7 Atriplex 0 0.1 0 5.7 0 0 0 0 0 0 0.2 0 0 Chenopodium 0.2 0 0.1 1.4 0 0 0 0 0 0.2 0 0 0 Cryptantha 23.3 0 0 15.8 0 0.2 0.2 0.1 0 18.6 0 0 1.5 Descurainia 144.3 0 0 45.1 0 0 7.1 0 0 41.1 0 0 29.1 Ipomopsis 0.4 0 0 0 0 0 0 0 0 0 0 0 0.6 Lappula 1.0 0 0 3.7 0.2 0 0.4 0 0 0.2 0 0 0.1 Mentzelia 0 0.3 0 0 0 0 0 0 0 0.5 1.5 0 96.1 Phacelia 0 0 0 0.2 0 0 1.4 0 0 3.7 0 0 0 Plant ago 0 0 0 0 0 0 0 0 0 3.5 0 0.1 49.5 Purtidaca 0 0 0 1.4 0 0 0 0 0 0 0 0 0 Salsola 5.4 0 0 2.7 2.1 5.9 0 0 0 0.3 0 0 3.9 Senecio 0.4 0 0 0 0 0 0 0 0 2.3 0 0 16.3 Solanaceae 0.7 0 0 0.4 0 0 0 0 0 0 0 0 0 Sphaeralcea 0 0 0 0 0 0 0 0 0 0.2 0 0 8.1 Stephanomeria 0 0 0 0 0 0 0 0 0 0 0 0 2.3 Streptanthella 0 0 0 0 0 0 3.2 0 0 0.7 0 0 0 Townsendia 37.2 0 0 6.2 0 0 0 2.1 0 1.4 0.6 0 0 Other* 0 0 0 0.7 0.4 0.2 0.7 0 0.1 0.7 0.1 0 0.1 Unknown 7.1 0.3 0.1 1 0 0 1.2 0 0 0.9 0 0 10.6 Number sp. (richness) 10 3 3 12 3 3 8 3 2 15 4 1 13 Total density 220 0.7 1.8 84.3 2.7 6.3 16.9 4.9 1.7 86.2 2.4 0.1 222.9 *Other includes all identified species that never reached a density ol 0.5 per m . provide information that can be used to distin- guish one habitat type from another. The re- sults of this study indicate that the wash, the floodplain, the bench, and the mesa tops are distinct habitat types. Individual species dis- tribution may overlap habitat types, but each type is distinguished either by the presence of species unique to that habitat or by the greater dominance of particular species over others. One of the most conspicuous characteristics of annual plant productivity is the variability from year to year. Juhren et al. (1956) found that different constellations of annual species germinate under specific combinations of temperature and rainfall at Joshua Tree Na- tional Monument in southern California. During some years only a few individuals occurred. Likewise, in Nevada, Beatley (1969) observed that biomass of annual species fluctuated both spatially and temporally, de- pending on local conditions. At Chaco Canyon annual spring plant den- sities fluctuated drastically from year to year during the period 1979-1981. We believe that the dramatic abundance of annuals in 1979 was due to late winter and spring precipitation that fell at Chaco Canyon in 1978-1979. Dur- ing December and January mean monthly temperatures were low (Fig. 5), resulting in low water loss from soil and low evapotranspi- ration rates. These climatic conditions were favorable for the germination and growth of annuals to reproductive maturity. Conditions were particularly favorable for pinnate tansy- mustard, which, we believe (based on densi- ties), accounts for most of the increase in total biomass and seed biomass on the study plots. However, perennial grasses also contributed to the total. Different climatic conditions pro- duce different dominant species in the annual plant populations from year to year, and there are years when annual species are rare or absent. Another characteristic of annual pro- duction is spatial variability. Within a limited geographic area such as Chaco Canyon, the abundance and species composition of annual plant populations differ from habitat to habitat during the same growing season. Patten (1978) noted similar differences in productiv- ity in annual plants in different microhabitats within a Sonoran Desert shrub community. During the winter and spring of 1978-1979 the climatic conditions were probably the overriding factors in the abundance of annuals throughout Chaco Canyon. The differences 120 Great Basin Naturalist Vol. 49, No. 1 ll □ Chiquita Bonil ll CZZ1 Casa Pueblo Bench Pueblo Chiquita Bonito Alto I. I Casa Pueblo Bench Pueblo 1981 Chiquita Bonito Alto Fig. 3. Total plant and seed biomass at five sites at Chaco Canyon during three years. in the species composition and the produc- tivity between habitat types were probably a result of local characteristics of soil and physiography. Summary We measured attributes of both perennial and annual plant species in four habitat types at Chaco Canyon. These habitat types were found to be distinct, based on perennial spe- cies representation and/or relative dominance of perennial species. Diversity based on Table 5. Vegetation biomass of annual plants and grasses at five locations (g/m2). Total plant biomass (5m ) Location Year Mean SE Mean SE Casa Chiquita 1979* 98.8 12.68 21.8 2.52 1980 28.6 6.93 4.6 1.12 1981 3.8 0.97 0 0 Pueblo Bonito 1979 126.8 9.65 20.56 1.82 1980 24.26 5.02 2.9 0.55 1981 26.18 4.26 0 0 Bench 1979 5.52 0.88 0.74 0.12 1980 13.46 2.12 1.98 0.32 1981 5.08 0.74 0 0 Pueblo Alto 1979 32.50 2.27 5.44 0.53 1980 57.46 4.51 8.84 0.75 1981 15.94 1.29 4.72 0.45 Werito's Rincon 1979 42.34 2.91 4.72 0.45 *We collected biomass samples from only 10 jlots at Casa Chiquitt in 1979. R^ Descuraini a pinnata □ * antago purshii Casa Pueblo Bench Pueblo Werito's Chiquita Bonito Alto Rincon Fig. 4. Total density and density of dominant annual species from five sites at Chaco Canyon in 1979. perennial species was higher on the bench and in the wash, and the lowest by all mea- sures of diversity at Pueblo Bonito. January 1989 Cully, Cully: Chaco Canyon Vegetation 121 a average monthly temperatures, 1978-1979 mean monthly precipitation, 1951-1974 monthly precipitation, 1980-1981 no temperature data E 60 E A S 0 N D 1978 1 979 E 60 E — 5 0 c o Fig. 5. Monthly precipitation during the study and 20-year mean temperature and precipitation at Chaco Canyon. At Chaco Canyon annual plant species are an important component of the plant commu- nity even though their appearance is highly variable both spatially and temporally. An- nual plant species occurred in all habitat types in varying densities and profoundly affected the aboveground biomass from place to place and from year to year. Annual plant species density, diversity, and biomass were high on the floodplain sites during 1979 but low dur- ing the other two years. Annual plants on the bench had the lowest production measured during 1979, when the floodplain sites were producing at their peak. Annual plant dens- ity at Pueblo Alto and Werito's Rincon was also high during 1979; however, the greatest 122 Great Basin Naturalist Vol. 49, No. 1 biomass at Pueblo Alto was measured in 1980, during the year following the wet winter- spring of 1978-1979. This apparently reflects the development of grasses that responded to the wet period more slowly than annual plants. Thus, while the floodplain habitats were the most productive during exception- ally wet years, the upland habitats appeared to be more productive during drier years or during years when precipitation arrived dur- ing another season. Measurements of perennial plant diversity are poor predictors of productivity. Those habitats with the highest perennial alpha di- versity may be the poorest in terms of plant productivity. The bench, characterized by high perennial diversity, was low in annual diversity and productivity during 1979. On the other hand, the Pueblo Bonito study area was characterized by low perennial plant di- versity, but annual plant species were diverse and productivity was high during the favor- able year of 1979. Our study indicates that during favorable years, annual plant species have patterns of dominance and diversity that are separate from those of perennial species. Both perennial and annual components should be considered in measurements of diversity and productivity within plant com- munities. Acknowledgments The Division of Cultural Besearch, Na- tional Park Service, Santa Fe, New Mexico, provided funding for the fieldwork portion of this project. Glen Lanier helped gather the annual plant productivity information during 1979. Mollie Toll and Karen Clary helped during 1980 and 1981. Jerry Livingston of the Division of Cultural Besearch, National Park Service, provided the illustrations. Literature Cited Beatley, J. C. 1969. Biomass of desert winter annual plant populations in southern Nevada. Oikos 20: 261-273. Canfield, R 1941. Application of the line interception method in sampling range vegetation. J. Forest. 39: 388-394. Cully, J. F , Jr. 1984a. A multivariate analysis of niche relationships within a nocturnal rodent commu- nity in northwestern New Mexico. Unpublished dissertation, University of New Mexico, Albu- querque. 1984b. Diversity, stability and the deermouse: implications for the vegetative diversity model. In W. D. Judge and J. D. Schelberg, eds. Recent research on Chaco prehistory. Reports of the Chaco Center No. 8, National Park Service, Santa Fe, New Mexico. Donart, G B., D. D Sylvester, andW. C. Hickey. 1978. Potential natural vegetation. New Mexico. United States Department of Agriculture, Soil Conserva- tion Service, Albuquerque, New Mexico. Hill, M. O 1973. Diversity and evenness: a unifying notation and its consequences. Ecology 54(2): 427-432. Jones, K. 1972. Ecology of Chaco Canyon. Unpublished manuscript on file. Division of Cultural Research, National Park Service, Santa Fe, New Mexico. Juhren, M , F. W. Went, and E. Phillips. 1956. Ecology of desert plants. IV. Combined field and labora- tory work on germination of annuals in the Joshua Tree National Monument, California. Ecology 37(2): 318-330. Kelley, E , andL D Potter 1974. Preliminary vegeta- tion type map of Chaco Canyon. Unpublished manuscript on file, Division of Cultural Research, National Park Service, Santa Fe, New Mexico. Norusis, M. J. 1986. Advanced statistics SPSSPC+ for the IBM PC/XT/ AT. SPSS Inc., Chicago. Patten, D. T. 1978. Productivity and production effi- ciency of an Upper Sonoran Desert ephemeral community. Amer. J. Bot. 65(8): 891-895. Peet, R K 1974. The measurement of species diversity. Ann. Rev. Ecol. Syst. 5: 285-308. Pielou, E. C. 1974. Population and community ecology: principles and methods. Gordon and Breach, Sci- ence Publishers, Inc. , New York. Potter, L D 1974. Ecological survey of Chaco Canyon. Unpublished manuscript on file. Division of Cul- tural Research, National Park Service, Santa Fe, New Mexico. Scott, N. J , Jr., and J M Duke. 1980. The ecology of seeds and their predators on three arid sites in Arizona and New Mexico: phenology of the peren- nial plants. National Fish and Wildlife Laboratory, Museum of Southwestern Biology, University of New Mexico, Albuquerque. Shannon, C. E , and W Weaver 1949. The mathemati- cal theory of communication. University of Illinois Press, Urbana. Shreve, F 1942. The desert vegetation of North Amer- ica. Bot. Rev. 8(4): 195-246. Simpson, E. H. 1949. Measurement of diversity. Nature 163: 688. Tevis. L, Jr 1958. A population of desert ephemerals germinated by less than one inch of rain. Ecology 39(4): 688-695. Whitaker, R H 1972. Evolution and measurement of species diversity. Taxon 21: 213-237. ASSOCIATIONS OF SMALL MAMMALS OCCURRING IN A PLUVIAL LAKE BASIN, RUBY LAKE, NEVADA Mark A. Ports1 and Lois K. Ports1 Abstract. — Ruby Lake is a highly mesic and vegetationally diverse pluvial lake basin of east central Nevada. Small mammal associations were examined in six plant communities at Ruby Lake using transects of live traps. Small mammal activity was recorded for these six habitats plus an additional three other specialized habitats. A total of 11 species of small mammals were trapped from the six habitat types; from the entire study area 26 species were trapped or observed. Two greasewood shrub habitats and a shadscale-spiny hopsage habitat held the highest number of trapped species, 6, 5, and 7, respectively. The mesic haymeadow and spring habitats, as well as the big sagebrush-antelope bitterbrush habitat held 4 trapped species each. Peromyscus maniculatus and Perognathus parvus made up 76% of the total captures and were found in all habitat types except marshlands. Eutamius minimus was found in four of the six habitat types, while Dipodomys ordii, Dipodomys microps, Perognathus parvus, and Microtus montanus were limited to specialized habitats. Mesic adapted, wetland species such as Mustela vison. Ondatra zibethicus, and Sorex vagrans possibly dispersed into Ruby Valley from the northeastern drainages and valleys during the late Pleistocene or Holocene. Analyses of small mammal communities in the Great Basin have added much to our knowledge of their ecology and biogeographi- cal distributions (Hall 1946, Borrell and Ellis 1934, O'Farrell 1974). The majority of these studies have been concentrated in the more mesic, isolated mountain ranges in plant com- munities above 2,100 m (Brown 1978). Such research has shown that apparent relictal pop- ulations occur isolated on various mountain ranges, analogous to populations on oceanic islands, with little chance of restoring such gene pools by immigration (Brown 1978). It has been only in recent years that the shrub steppe desert of the valley floors and pluvial lake basins has received attention. Only minor differences in species composi- tion and abundances of small mammals have been shown to occur between the eastern and western halves of the Great Basin. The west- ern and southern deserts are characterized by a transitional zone of desert shrubs such as Artemisia, Sarcobatus, Atriplex, Larrea, and Chenopodium. The small mammal communi- ties of the Mojave and southern Great Basin deserts are dominated by such heteromyids as Dipodomys merriami, D. microps, Perogna- thus longimembris , and P. formosus. As one moves north and east, the plant communities lose their Mojave Desert affinities, and the small mammal composition of lower elevation communities is dominated by two species, Peromyscus maniculatus and Perognathus parvus (Durrant 1953, O'Farrell and Clark 1986, Jorgensen and Hayward 1965). The 53 defined pluvial lake basins of central and eastern Nevada provide an environment for several plant communities that have been progressively adapting to a drier, cooler cli- mate and more alkaline soils since the end of the Pleistocene (Mifflin and Wheat 1976). Ly- ing within the rain shadow of the Sierra Ne- vada and in the rain shadow of an adjacent mountain range, the pluvial basins today are characterized by low atmospheric precipita- tion, no external runoff, and limited inflow from a yearly snowmelt that varies from year to year (Mifflin and Wheat 1976). Evaporation of once extensive lake waters left behind high concentrations of saline deposits and thus ex- posed new habitat and soils for the rapidly evolving phraeophytic shrubs of the genera Sarcobatus and Atriplex (Young et al. 1986). While several studies have described the small mammal communities of valley floors dominated by Artemisia tridentata (O'Farrell 1974), few have concentrated on the phraeo- phytic plant communities around pluvial lake basins (Jorgensen and Hayward 1965, Young and Evans 1974). Durrant (1952) and Hall (1946) list several species collected from vari- ous localities representing pluvial lake basins in the Great Basin. 'Northern Nevada Community College, Department of Life Science, 901 Elm St., Elko, Nevada 89801. 123 124 Great Basin Naturalist Vol. 49, No. 1 Our purpose here is to describe associations among small mammals within six plant com- munities surrounding the perennial Lake Ruby, the southern remnant of pluvial Lake Franklin, in northeastern Nevada. We will also compare these small mammal associa- tions with small mammal communities exist- ing in other pluvial lake basins within the Great Basin and possible routes of late Pleis- tocene or Holocene dispersal. Methods and Study Area This study was conducted in Ruby Valley, southern Elko County, on the southeast- ern flank of the Ruby Mountains, Nevada, latitude 40°07'30" and longitude 115°30'00". During the late Pleistocene, Ruby Valley, ap- proximately 3,662 sq km, held a large pluvial lake covering some 628 sq km and had esti- mated depths of 60 m (Mifflin and Wheat 1979). Climatic changes over the last 500,000 years caused this large body of water to recede to its present-day level at an elevation of 1,860 m, resulting in a large seasonal lake and playa near the center of the valley. Today this rem- nant is called Franklin Lake (USFWS 1987). The northern half of the valley probably had some inflow of water from pluvial Clover Lake to the northeast (Mifflin and Wheat 1979). Today this end of the valley consists of big sagebrush and rabbitbrush (Chrysothamnus spp.) uplands, and lower-elevation hay mead- ows and greasewood (Sarcobatus sp.) shrub- lands. The southern end of the valley is under the jurisdiction of the Fish and Wildlife Ser- vice, U.S. Department of the Interior. In this locality over 200 perennial carbonic rock springs, coming from the eastern flank of the Ruby Mountains, provide a relatively constant source of water for the 15,053-ha Ruby Lake National Wildlife Refuge (USFWS 1987). The perennial springs provide a water source that sustains some unique and diversi- fied plant communities that are very mesic when compared to other pluvial lakes of the region (Mifflin and Wheat 1979). Delineation of six habitat types was deter- mined by the dominant plant species and comparison with plant communities de- scribed for other pluvial lake basins in Nevada (Young etal. 1986). Of the 15,053 ha of refuge, approximately 32% consists of a series of man- aged wetlands, dikes, and collection ditches used for both fishing and waterfowl produc- tion (USFWS 1987). This habitat was not quantitatively sampled due to its aquatic na- ture; however, observations of mammal use and mammal sign were recorded. All plant names were based on Cronquist et al. (1977) and Johnson et al. (1981). The following habi- tat types were all sampled for small mammals using quantitative techniques. Habitat 1. — A mixed plant community of black greasewood (Sarcobatus vermiculatus), big sagebrush (Artemisia tridentata), and rubber rabbitbrush (Chrysothamnus nause- osus) occupies approximately 10% of the refuge in a belt extending from the southern tip, proceeding north on the east side of the refuge, and terminating near the north end. The understory plants of this habitat are sparse but diverse, including Sandberg's bluegrass (Poa sandbergii) and long-leaved phlox (Phlox longifolia). A subunit of this habitat included a series of stabilized sand dunes with black greasewood clones on the top of the dune and big sagebrush around the perimeter. Other plants characteristic of the dunes include Indian ricegrass (Oryzopsis hy- menoides), needle and thread grass (Stipa co- mata), and Hooker's evening primrose (Oenothera hookeri). The stabilized dunes make up only 1% of the refuge area, while on other pluvial lake basins in Nevada they are a much more prominent land form (Young et al. 1986). Habitat 2. — As one moves closer to the center of the lake from Habitat 1, the soils become noticeably more alkaline and the plant community changes. The dominant shrubs here include black greasewood, shad- scale (Atriplex confertifolia), alkali rabbit- brush (Chrysothamnus albidus), and rubber rabbitbrush. Understory plants are dense but low in diversity. These include salt grass (Dis- tichlis spicata), Great Basin wildrye (Elymus cinerus), alkali bulrush (Scirpus paludosus), and western seepweed (Suaeda occidentalis), all of which are salt-tolerant species (Young et al. 1986). Habitat 3. — Due to the concentration of springs on the west side of the refuge, the majority of the mesic hay meadows and sea- sonally wet wire rush meadows are found di- rectly below the slopes of the Ruby Moun- tains. This habitat is extremely dense and January 1989 Ports, Ports: Small Mammals 125 varies in height from 10 cm to 1 m. The domi- nant vegetation in these meadows consists of Baltic wirerush (J uncus balticus), sedges (Carex spp.), bulrushes (Scirpus spp.), and grasses of the genera Festuca, Hordeum, and Agrostis. Seasonal grazing and mowing for hay is done in these habitats after waterfowl nesting season is over in July (USFWS 1987). This habitat makes up 10% of the total refuge area. Haritat 4. — The 200 springs and their drainages make up only 4% of the refuge area but are an extremely important habitat for many mammals. Not only is there a dense growth of J uncus and Scirpus in such areas but also many mesic shrubs, such as Scouler's willow (Salix scoueriana), Wood's rose (Rosa woodsii), and golden currant (Ribes aureum). The forb and grass species around these springs are very diverse and are usually part of an ecotone area bordered by big sagebrush or hay meadows. Haritat 5. — Alluvial fans coining from the eastern slopes of the Ruby Mountains at ele- vations between 1,800 and 2, 100 m constitute 22% of the refuge. The co-dominant shrubs of the community include big sagebrush, an- telope bitterbrush (Purshia tridentata), west- ern serviceberry (Amelanchier alnifolia), and green rabbitbrush (Chrysothamnus viscidi- florus). The introduced cheatgrass Bromus tectorium, found in many of the high-valley shrub communities, is the dominant grass along with such bunchgrasses as bottlebrush squirreltail (Sitanion hystrix) and Indian rice- grass. A very diverse forb component is found here with such dominants as long-leaved phlox, Aster spp., scarlet gilia (Ipomopsis spp.), and western yarrow {Achillea mille- folium). This plant community is found pri- marily on the west side of the refuge as well as the northern and southern ends. Haritat 6. — This plant community is char- acterized by low-growing shrubs with much bare, gravelly ground. It is confined to the eastern side of the refuge situated on the broad alluvial fans coming from the Maverick Springs hills. It is just above Habitat 1 in elevation on the eastern side and makes up 16% of the refuge area. A very dry area, this habitat is dominated by shadscale, spiny hop- sage (Grayia spinescans), and dwarf sage- brush (Artemisia arbuscula). Grasses and forbs are scarce in this area, those seen most often including bottlebrush squirreltail and peppergrass (Lepidium lasiocarpum). This study was conducted during the periods of June-September 1986 and April- October 1987. Its major goal was a general inventory of all small mammals inhabiting the principal habitat types on the refuge. This information was used in the compilation of a wildlife checklist (USFWS 1987). Several data-collecting techniques were used and all possible habitat types were in- ventoried in compiling this general inventory. For the previously described six habitat types, a total of 100 Sherman live traps were used for a total of 2,050 trap nights. In each habitat the live traps were placed in two paral- lel transects of 50 stations each. Single traps were placed 15 paces apart and baited with rolled oats. Each habitat type was trapped at least three consecutive nights. Also used on selected habitats were 15 Tomahawk live traps for squirrels and rabbits and 40 pitfall traps for shrews. Mist nets were utilized on five separate occasions to capture, identify, and release species of bats. All observations of mammals seen or signs of mammals, tracks, and scats were recorded as to species and habitat. For each animal captured on the trap lines, the sex, age, and reproductive condition were recorded. The majority of the animals were released after identification. Those that died in the traps were preserved and are now housed in the museum of NNCC. Data col- lected here were analyzed by species, num- ber of captures for each species, and number of captures per 100 trap nights. The relative frequency was calculated for each species in each of the six habitats, and the total number of species captured or observed was tabulated according to habitat. Relative frequency is the number of individuals captured per species divided by the total individual captures of all species, multiplied by 100. Results A total of 11 species of small mammals were sampled from the six habitat types in which trap lines were used. The composition of small mammal communities and the relative fre- quency of each species differed among habitat types (Table 1). 126 Great Basin Naturalist Vol. 49, No. 1 Table 1. Small mammals trapped along transects from six habitat types on the Ruby Lake National Wildlife Refuge. C = total captures during period June-September 1986 and April-October 1987. C/T = captures per 100 trap nights. RF = relative frequencies of captures within each habitat type. Habitat Habitat Habitat Habitat Habitat Habitat 1 2 3 4 5 6 C C/T RF% C C/T RF% C C/T RF% C C/T RF% C C/T RF% C C/T RF% Peromyscus maniculatus 75 23 67 98 46 86 29 6 37 48 16 84 60 11 32 1 0.6 1 Eutamius minimus 7 2.2 6 6 2.8 5 14 2.5 7 1 0.6 1 Dipodomys ordii 19 6 17 5 1.0 2 1 0.6 1 Dipodomys microps 8 2.5 7 28 19 33 Microdipodops megacephalus 2 0.6 2 Perognathus parvus 1 0.3 1 8 3.8 6 1 0.2 1 3 1 5 110 20 59 49 33 57 Perognathus longimembris 5 3.3 6 Reithrodontomys megalotis 2 1.0 2 Microtus montanus 36 7.4 46 5 1.7 9 Sorex vagrans 1 0.5 1 12 2.5 16 1 0.3 2 Onychomys leucogaster 1 0.6 1 Totals 112 34.6 114 54.3 78 16.1 57 19 186 33.5 86 57.1 Number of species 6 5 4 4 4 7 Peromyscus maniculatus and Perognathus parvus made up 76% of the total captures within the six habitat types. These two species occurred in all six habitat types, with P. maniculatus being most common in Habitat 1 (greasewood and big sagebrush); the smallest number of captures was in Habitat 6 (shad- scale and spiny hopsage). Perognathus parvus occurred in the largest numbers and in ap- proximately equal frequencies in Habitat 5 (big sagebrush and antelope bitterbrush) and Habitat 6 (shadscale, spiny hopsage, and dwarf sagebrush), while in all other habitats this species was represented in very small numbers (Table 1). Other habitats that contained species in relatively large numbers were specialized habitats tending toward mesic conditions (Microtus montanus and Sorex vagrans in Habitat 3), sand dunes and sandy soils (Di- podomys ordii in Habitat 1), and dry habitats with low-statured shrubs (Dipodomys mi- crops in Habitat 6) (Table 1). Our data, based on species composition among the six habitat types, indicate that Habitats 3 (hay meadows) and 4 (springs) were identical in the species of small mammals present. In addition, both had the lowest number of captures of all the habitat types (Table 1). The greasewood-big sagebrush habitat (1) and the big sagebrush-antelope bitterbrush habitat (5) were also similar be- cause of the presence of species that find both greasewood and big sagebrush suitable habi- tats. Habitat 1 is enriched with the addition of two heteromyids, Dipodomys microps and Microdipodops megacephalus, not present in Habitat 5 (Table 1). The remaining habitats, 6 and 2, were dis- similar in species composition with those de- scribed above. Habitat 6 had the highest num- ber of species (seven) with two, P. parvus and D. microps, making up 90% of the total cap- tures. Habitat 2 (greasewood and grass), al- though similar in vegetational structure to Habitat 1, lacked three of the heteromyid spe- cies and contained two specialized species, Reithrodontomys megalotis and Sorex va- grans. Dipodomys ordii did occur on the pe- riphery of this habitat in a specialized habitat of sandy soils and rubber rabbitbrush, as sug- gested by kangaroo rat tracks found on the road each day. None were captured on the transect, however. Besides P. maniculatus and P. parvus, three other species of mammals proved to be almost as ubiquitous on the refuge. Eutamius minimus was most common in the big sage- brush-antelope bitterbrush habitat but was also present in Habitats 1, 2, and 6. Dipo- domys ordii preferred any habitat containing sandy soils, which included the sand dunes of Habitat 1 and the roadsides of Habitats 5 and 6. Finally, Sorex vagrans, an inhabitant of mesic hay meadows (Habitat 3) and springs (Habitat 4), also occurred in the much drier habitat of greasewood and grass (Habitat 2) (Table 1). January 1989 Ports, Ports: Small Mammals 127 Table 2. Small mammals trapped or observed on the Ruby Lake National Wildlife Refuge and associated plant communities, 1986-1987. Habitat Species 3 Marsh- lands Cliff sites Home- steads Sorex vagrans Myotis evotis Myotis leibii Mustela frenata Mustela vison Taxidea taxus Lepus californicus Sylvilagus nuttelli Sylvilagus idahoensis Eutamius minimus Spermophilus lateralis Spermophilus townsendii Spermophilus beldingi Thymomas talpoides Dipodomys ordii Dipodomys microps Microdipodops megacephalus Perognathus parvus Perognathus longimembris Peromijscus maniculatus Reithrodontomys megalotis Onychomys leucogaster Neotorna cinereus Microtus montanus Ondatra zibethicus Total species Id 10 Although information on sex, reproductive status, and age by pelage was recorded for most individual captures, only data for P. maniculatus and P. parvus proved significant enough for comparison within Habitats 1, 2, and 5. Table 2 shows the occurrence of all 25 mam- mal species trapped or observed during the study period. Included are the six habitat types used in Table 1 as well as three others where trap lines were not used. These three habitats include the large cattail (Typha sp.) and bulrush (Scirpus sp.) marsh, cliffsites and carbonaceous rock caves around the periph- ery of the refuge, and homestead that includes the refuge headquarters and Gallegher Fish Hatchery (USFWS 1987). Two common spe- cies of bats in the area include a bachelor roost of Myotis evotis and Myotis leibii from the Cave Creek grotto and a series of lactating females from the Maverick Springs area. Small carnivores captured or observed are listed according to habitat (Table 2). These include Mustela frenata, Mustela vison, Canis latrans, and Taxidea taxus. Ground squirrel species were limited to colonies and to a short season of activity in both years. These species include Spermophilus town- sendii, S. beldingii, and S. lateralis. Leporids included Lepus californicus, Sylvilagus nutel- lii, and S. idahoensis (Table 2). The pygmy rabbit was closely associated with stands of big sagebrush around the periphery of the sand dunes in Habitat 1, while Nuttell's cottontail occurred in dense cover along sagebrush roadsides, cliff sites, and homesteads. Large rodents that can be considered habitat spe- cialists include Ondatra zibethica in the marsh lands, Neotorna cinerea in cliff sites and homesteads, and Thomomys talpoides in Habitats 2, 3, and 4 (Table 2). Discussion Ruby Lake has been shown to have a large number of diverse plant communities and to be one of the most perennial and mesic of the pluvial lake basins still existing in Nevada to- day (USFWS 1987). Mifflin and Wheat (1979) suggest that the flora and hydrologic makeup of Ruby Lake as seen today may be reminis- cent of what the drier pluvial lakes of west 128 Great Basin Naturalist Vol. 49, No. 1 central Nevada looked like during the more mesic late Pleistocene era. This may account for the similarities in small mammal commu- nities seen at Ruby Lake, as compared to the Carson Sink basin in western Nevada (Hall 1946) and the Mono Lake basin of California (Harris 1982). Both of these pluvial lakes are much drier and contain more seasonal, ephemeral wet- lands. The only major difference between these pluvial lakes and the western Great Basin desert in general (Kenagy 1973, Jor- gensen and Hayward 1965) is that the small mammal communities are dominated by four to five species of heteromyid rodents, whereas the more eastern pluvial lake mam- mal communities are dominated by only two species, P. maniculatus and P. parvus, both adapted to a wide range of habitats but less frequent in hot desert environments (O'Farrell 1974). River valleys and Pleistocene lake con- nections allowed easy dispersal routes for wetland- and shrubland-adapted mammals (Hall 1946). This would account for the many similarities in species composition between pluvial lake basins in the Great Basin. While we recorded 24 species of small mammals at Ruby Lake, Young and Evans (1980) list 22 species for pluvial Lake Gilbert in Grass Val- ley of central Nevada, Hall (1946, 1981) lists 24 species for the Carson Sink desert of west- ern Nevada, and Harris (1982) lists 26 species for the lower elevation habitats of the Mono Lake basin. All of these basins contain a ma- jority of the same species, with only a few differences due primarily to the influx of Mo- jave Desert fauna into the western Great Basin (Jorgensen and Hayward 1965). Few studies have been done which permit a quantitative comparison of small mammal communities in pluvial lake basins. O'Farrell (1986) studied five habitat types in Whirlwind Valley, a nonpluvial basin, 120 km NW of Ruby Lake. Here O'Farrell trapped 11 spe- cies of small mammals, compared with 14 spe- cies in Grass Valley (Evans and Young 1986), and 11 species for Ruby Lake (Table 1). Spe- cies composition between these three locali- ties varied only slightly; P. maniculatus and P. parvus made up the majority of the individu- als for each locality in all habitats sampled. Kenagy (1973) and Harris (1982) found that heteromyid rodents were the dominant spe- cies in the western deserts and at Mono Lake, but P. maniculatus and P. parvus were present and widespread. The contribution and relative frequency of each species varied among habitat types at Ruby Lake. We found that the greasewood-big sagebrush habitat had the largest number of species present during the study period. This was also the case in Grass Vallev (Evans and Young et al. 1986) and in Whirlwind Valley (O'Farrell 1986). Ro- dent species diversity has been correlated with resource abundance (Whitford 1976) and with vegetation structure and diversity of a habitat (Rosenzweig and Winakur 1969). Our data agree with these studies in reference to the complexity and relative abundance of forbs and grasses in the greasewood-big sage- brush habitat. Data recorded for the reproductive activity of P. maniculatus in the greasewood-big sage- brush habitat show that the sex ratios did not differ significantly from a 1:1 ratio. Pregnant females and scrotal males, some in juvenile pelage, were common in both May and mid- August, suggesting at least two litters for this habitat during the two years. This would indi- rectly suggest a habitat rich in resources. The big sagebrush-antelope bitterbrush habitat has been attributed with a community complexity and richness capable of supporting a small mammal community of 7 (Lent and Eckert 1982) to 12 (O'Farrell 1974) species. At Ruby Lake we captured only 4 species of small mammals in a habitat that seemed to be struc- turally diverse and rich in forbs and grasses. One possible explanation for the low diversity in this habitat could be the comparatively low number of trap nights. Another possibility is that the high frequency of P. parvus (59%) may limit the numbers of other rodent spe- cies, such as Onychomys leucogaster and Lagurus curtatus, which have been captured from this locality and habitat by Borrell and Ellis (1934). Also P. parvus has been shown to have an intricate relationship with the an- telope bitterbrush in collecting and caching its seeds (Evans et al. 1982). From our data we found that sex ratios for P. parvus did not differ significantly from a 1:1 ratio. The great- est reproductive activity occurred in mid- June, and the largest number of juveniles ap- peared in August. This suggests only one litter for each of the two seasons we sampled in this habitat and agrees with the findings of O'Farrell et al. (1974) in a similar habitat. January 1989 Ports, Ports: Small Mammals 129 The shadscale-spiny hopsage habitat was unusual in that its low stature and low plant complexity nonetheless maintained a high species diversity of seven species. O'Farrell (1986) found that species composition of any one of the five habitats in Whirlwind Valley experienced seasonal changes in species com- position. We suspect that this is also the case with the shadscale habitat at Ruby Lake, which was trapped only twice during the late summer. Four of the seven species in this habitat were probably wandering from adja- cent greasewood habitats. Two dominant spe- cies within this habitat, P. parvus and D. mi- crops, were also found to be co-dominants at Grass Valley (Young et al. 1980) and Whirl- wind Valley (O'Farrell 1986) in a similar shad- scale community. Perognathus longimembris was found only in the shadscale habitat at Ruby Lake. It did not occur in Whirlwind Valley (O'Farrell 1986) but was found in greasewood habitat in Grass Valley (Young et al. 1980). We found four of the five captured individuals to have interesting patterns of pure white spots on the dorsum, the backs of the ears, and the anterior flanks. These spots are white down to the integument and vary from a coverage of 1/3 of the pelage to minute spots. This pelage has been recorded else- where in specimens of P. longimembris from the desert of Millard County, Utah (Durrant 1952). Dipodomys microps has been shown to have a dietary reliance on the leaves of Atriplex rather than on seeds. This accounts for the large numbers of this heteromyid in the shadscale habitat (Kenagy 1973). Other small mammal associations on Ruby Lake suggest habitat expansions and histo- rical changes in status of some species. At Ruby Lake we found Eutamius minimus in four of the six habitat types, being most com- mon in the big sagebrush-bitterbrush habitat. O'Farrell (1986) found the least chipmunk re- stricted to a greasewood habitat by the pres- ence of the antelope ground squirrel (Am- mospermophilus leucurus), which occurred in three habitat types. At Grass Valley the least chipmunk was found both in greasewood and in sagebrush but not in all sagebrush habi- tats or in shadscale because of the presence of the antelope ground squirrel. Robey et al. (1986) showed the least chipmunk to occupy a narrow niche breadth in a desert shrub community shared with the antelope ground squirrel, a species occupying a wide niche. These observations suggest that the least chipmunk at Ruby Lake has expanded into all available shrub habitats in the absence of competition with the antelope ground squir- rel, a species not known from Ruby Valley. Brown (1978) describes another instance of habitat expansion in three species of montane chipmunks whose irregular distributions on mountain ranges in eastern Nevada result in atypical habitat utilization by these species and expansion into habitats they normally do not occupy. Borrell and Ellis (1934) trapped and col- lected data on mammals from the western flank of the Ruby Mountains along the west edge of Ruby Lake. They found both the mink (Mustela vison) and the muskrat (Ondatra zibethicus) present in 1927, years before the lake came under government management. This suggests that both are native and possibly dispersed south from the Columbia basin into Mary's River and the wetlands of Pleistocene Clover Lake. This wetland possibly had a Pleistocene connection with pluvial Franklin Lake and Ruby Lake (Mifflin and Wheat 1979). The same route of dispersal could be ap- plied to the vagrant shrew and the montane vole, both of which occur in the Mary's River basin, the Humboldt River basin, the South Fork and North Fork of the Humboldt River, and the western flank of the Ruby Mountains, all below 1,590 m in elevation (Ports, unpub- lished data). The presence of Sorex preblei, a rare shrew of shrublands and wetlands, in the vicinity of the Great Salt Lake (Tomasi and Hoffmann 1984) and in the Mary's River basin (Ports, unpublished data) suggests the possi- ble occurrence of this species in Clover Valley as well as Ruby Valley. The pluvial lake basins of the Great Basin provide many mosaic and diverse plant com- munities for the development of small mam- mal communities. Similar in species composi- tion and plant communities, these basins have undergone similar climatic changes and have provided easy dispersal routes for populations of wetland- and desert-adapted mammals. Unlike their contemporaries on nearby iso- lated mountain ranges in subalpine and alpine habitats, the lower-elevation populations of small mammals can be expected to experience a wide range of population mixing and a wider range of dispersal than previously supposed, 130 Great Basin Naturalist Vol. 49, No. 1 as seen in the vagrant shrew (Ports, unpub- lished data). However, we must consider not only the effects of long-term climatic changes on these populations but also their effects on species composition. Studies of greater detail and of longer duration are necessary to under- stand the influence man has made on the plu- vial basin plant and animal communities of the Great Basin. Acknowledgments We wish to thank our sons, Boger and Mark, for their enthusiastic service in the field, as well as Sara Brown of the Buby Lake Wildlife Befuge for her generous cooperation and suggestions. Charles and Elaine Knight provided us with more than ample facilities while in the field, for which we are grateful. Literature Cited Borell, A. E . AND R Ellis. 1934. Mammals of the Ruby Mountains region of northeastern Nevada. J. Mammal. 15: 12-44. Brown, J H 1978. The theory of insular biogeography and the distribution of boreal birds and mammals. Great Basin Naturalist Mem. 2: 209-227. Cronquist, A , A Holmgren, N Holmgren, and J Re- veal. 1977. Intermountain flora: vascular plants of the Intermountain West, USA. New York Botani- cal Gardens, New York. Vols. 1-VI. Durrant, S 1952. Mammals of Utah. University of Utah Press, Salt Lake City. Evans, R A., J. A Young, G J. Cluff, and J K McAdoo. 1982. Pages 195-202 in Bitterbrush and cliffrose symposium, Salt Lake City, Utah. Hall, E R 1946. Mammals of Nevada. University of California Press, Berkeley. Harris, J H 1982. Mammals of the Mono Lake-Tioga Pass region. Kutsavi Books, Lee Vining. Johnson, D N , S H Bouffard, B M Paisecki, L Rau, C Yde, and K L. Lauritsen. 1981. Plant species of the Ruby Lake National Wildlife Refuge. U.S. Fish and Wildlife Service. Kenagy, G.J 1973. Daily and seasonal patterns of activity and energetics in a heteromyid rodent commu- nity. Ecology 54: 1201-1219. Lent, P C . and R. E Eckert, Jr. 1984. Progress report for the 1983 Saval Ranch Research and Evaluation Project. Bureau of Land Management and the University of Nevada, Reno. Mifflin, M D , and M M Wheat 1979. Pluvial lakes and estimated pluvial climates of Nevada. Nevada Bureau of Mines and Geology. Bull. 94. O'Farrell, M. J. 1974. Seasonal activity patterns of rodents in a sagebrush community. J. Mammal. 61:589-605. O Farrell, M. J., and W. A. Clark 1986. Small mammal community structure in northeastern Nevada. Southwest Nat. 31: 23-32. Robey, E H . Jr., H D Smith, and M C Belk 1986. Niche pattern in a Great Basin rodent fauna. Great Basin Nat. 47(3): 488-496. Rosenzweig, ML, and J Winakur 1969. Population ecology of desert rodent communities: habitats and environmental complexity. Ecology 50: 558-572. Tomasi, T E., and R. S Hoffman. 1984. Sorex preblei in Utah and Wyoming. J. Mammal. 64: 708. United States Department of the Interior. U.S. Fish and Wildlife Service. 1987. Wildlife checklist. Ruby Lake National Wildlife Refuge. Whitford, W. G. 1976. Temporal fluctuations in den- sity and diversity of desert rodent populations. J. Mammal. 57: 351-369. Young, J A , R A Evans, J A Brown. B A Roundy, J. K. McAdoo, R. Elston. and A. L. Lesperance. Physical, biological, and cultural resources of the Gund Research and Demonstration Ranch, Nevada. 1980. US DA Science and Education Admin, ARM-W-11. Young, J A , R A Evans, B. A. Roundy, and J. A. Brown 1986. Dynamic landforms and plant communities in a pluvial lake basin. Great Basin Nat. 46(1): 1-32. UTAH CHUB (GILA ATRARIA) FROM THE LATEST PLEISTOCENE GILBERT SHORELINE, WEST OF CORRINE, UTAH Stuart B. Murchison Abstract. — A bulk sampling of gastropods collected near Great Salt Lake, Utah, revealed several fish bones of Gila atraria (Girard). The early date of this find, coupled with the relationship of the Gilbert lake transgression and successive regression of approximately 11,000 years B.P., reveals a death assemblage induced by a series of saline inundations into the freshwater paludal environment. A large death assemblage of late Pleis- tocene gastropods, lying in an exposed road- cut along State Highway 83, 17.5 km north- west of Corrine, Utah (UTM coordinates 0390500E/4606400N, Public Shooting Grounds, 7.5 minute quadrangle), was radiocarbon dated by Miller (1980), twice by Currey (1988), and by the author. The respective dates of 10,920 ± 150 C-14 years B.P. (W-4395), 11,990 ± 100 C-14 years B.P. (Beta-16912, 11,570 ± 100 C-14 years B.P (Beta-16913), and 10,990 ± 110 C-14 years B.P. (Beta-22431) imply a period of low lake level before the Gilbert rise of 11,000 vears B.P. (Currey and Oviatt 1985). Currey (1980), Miller et al. (1980), Scott et al. (1983), and Currey and Oviatt (1985) use data from this site to assign the Pleistocene-Holocene boundary to Great Salt Lake and its prede- cessor. The discovery of Gila atraria at this site provides a unique opportunity for under- standing late Pleistocene and early Holocene environmental conditions and lake level fluc- tuations. Gastropods from the Public Shooting Grounds were collected in 1987 for further identification of representative genera, spe- cies, and habitat. The gastropod genera Amni- cola, Helisoma, Lymnaea, and Physella (Table 1) were excavated from a sand unit overlain by an organic marsh deposit. The fossil gastro- pods and Gila atraria were sampled at an altitude of 4,232 feet (1,290 m) a.s.l. Labora- tory cleaning of the shells consisted of ultra- sonic washing in deionized water and air dry- ing. Several unexpected rib, pharyangeal, vertebral, and maxillary bones oiGila atraria (Utah chub) were discovered within the ma- trix of this gastropod-rich sand unit. The bones were identified in 1987 by Mark Rosen- feld of the Department of Biology at the Uni- versity of Utah. Todays Utah chub are native to Utah, a small part of Nevada, and Idaho and are com- mon in several rivers draining into Great Salt Lake (Rawlev 1980) in the Bonneville basin. Stokes et al.' (1964) and Smith et al. (1968) report gastropod species and Gila atraria, with respective dates of 13,000 years B.P. (estimate) and 12,860 ± 100 years B.P. (W-2000), from two sites above 4,440 feet (1,350 in) on the margins of a regressive Lake Bonneville. This euryphagic species inhabits pelagic and littoral epilimnion areas (Varley and Livesay 1976). Stratigraphy The oldest exposed sediments from the Public Shooting Grounds site are the post- Provo to pre-Gilbert red beds (calcareous muds and minor sands). The red beds were reddened off-site and deposited basinward on mudflats and sandflats of the newly exposed regressive Bonneville basin about 13,000 years B.P. (Currey et al. 1988). The red bed deposition continued for about 1,000 years due to sediment washing into lower basin ar- eas. This regressive stage led to the precipita- tion of Glauber's salt (Na2S04 . 10 H20) at the deepest portion of the present Great Salt Lake (Currey and Oviatt 1985). A transgressive epi- sode of green, muddy sands, which have a minimum limiting date of 12,000 years B.P. 'Department of Geography, 270 Orson Spencer Hall, University of Utah, Salt Lake City. Utah 84112. 131 132 Great Basin Naturalist Vol. 49, No. 1 Table 1. Gastropod species and total percentage con- centration. Sources: Chamberlin and Jones (1929) and fieldwork (1987-88). Species Habitat Concentrations (%) Helisoma trivolvis Quiet to stagnant 5 fresh water Physella utahensis Ponds and streams 7 Lymnaea stagnalis Ponds, lakes, and 8 streams often attached to plants Amnicola limosa Streams, rivers, and 80 more quiet bottom waters (Beta-16912-3, Currey et al. 1988), uncon- formably overlies the red beds and grades upward into the Gilbert shoreline deltaic sed- iments. Currey (1980), Scott et al. (1983), and Currey and Oviatt (1985) suggest that the Gilbert shoreline represents a fluctuating stand followed by a regressive interval to lower lake stages. The green, muddy sand exposure is common to ancient lake basins and represents a reducing environment with a high ferric iron content (Reeves 1968). Suc- cessive layers of fine, clean, silty sands and marshy deposits of the Bear River and possi- bly the Malad River lie conformably over the green, muddy sands. The first layer consists of clean, fine sands and is interpreted as a 10- to 20-foot (3- to 6-m) minor transgression. As the lake regressed, a dark organic layer, indicat- ing high humic concentrations, appears pre- dating the death assemblage. This organic layer is overlain by lacustrine silts and repre- sents the second transgressive saline inunda- tion of the fluctuating Gilbert stand. Three additional episodes of saline, fossil-rich sedi- ments and subsequent clean, fine sand de- posits are recorded in this banded exposure. During the third minor regression, four gen- era of gastropods and Gila atraria migrated basinward previous to 10,990 ± 110 C-14 years B.P. (Beta- 22431), a date that is based on a species-specific sample of Lymnaea stag- nalis shells. Gastropods are typically found in freshwa- ter streams and ponds and usually feed on aquatic plants and organic detritus (Chamber- lin and Jones 1929). Gila atraria, at the Public Shooting Grounds site, would have taken ad- vantage of this newly formed marsh habitat that existed after previous saline water inun- dations. These faunas are suspended in a death assemblage matrix of trangressive lacus- trine sands and silts. The beginning of the next transgression probably killed the remain- der of freshwater organisms in this paludal margin, laying down the final, thick shell layer containing Gila atraria. This final, highest fossiliferous layer is 7.5 inches (19 cm) thick and is overlain by light brown, fine sands. Two organic marsh layers of 1.75 to 3 inches (4.4 to 7.6 cm) overlie the death assemblage. These undated layers are thought to be the last marsh deposits prior to the transgression that geomorphically marks the highest stage of the Gilbert shoreline. A fine, poorly sorted, near- shore sand layer deposited by this Gilbert high stage and a 4- to 5-inch (10- to 12.7-cm) modern soil comprise this exposure (Fig. 1). Discussion The paleoenvironment of Gila atraria was probably an ephemeral, shallow, freshwater marsh that is evident today at lower eleva- tions in the area. Aquatic plants, pioneer- ing organisms, insects, and detritus would have provided a minimum of sustenance for the euryphagic Gila atraria and gastropod species. As the lake transgressed, gastropods died and were deposited in 3- to 7.5-in (7.6- to 19-cm) layers within weak sand matrices. It is inferred from the thickness of these layers that the paludal margins were inundated by saline water depositing the faunal remains in small depressions. Due to the quality of the re- mains, it is hypothesized that the fossils were covered by nearshore sands rather quickly. Pleistocene-Holocene transitions in Great Salt Lake marginal deposits tend to have had faunal assemblages that are restricted to gas- tropods. The discovery of Gila ataria in these deposits, with radiocarbon date association, suggests greater faunal diversity than previ- ously thought. Acknowledgments Support facilities were provided by the Vasyl Gvosdetsky Laboratory for Environ- mental and Paleoenvironmental Studies and the Quaternary Research fund of the Depart- ment of Geography at the University of Utah. January 1989 Murchison: Pleistocene Utah Chub 133 -•■.. .•*.•>. . jo '. ; Gila atraria lodern soil Nearshore sand ^Marshy organic sand Gastropod-rich sand / 10,990+/- 110 yr BP. * Beta 22431 S f larshy organic sand Marshy organic sand Marshy organic sand Marshy organic sand Transgressive green muddy sand 12,000 yr BP. Pre-Gilbert red beds Fig. 1. Generalized stratigraphic column containing Gila atraria; radiocarbon date on Lymnaea s. I acknowledge Dr. Jeanne Kay and Dr. Don- ald Currey for their valuable comments and improvements on this manuscript. Literature Cited Chamberlin, R V , and D T. Jones. 1929. A descriptive catalog of the Mollusca of Utah. Bulletin of the University of Utah, Biological Series, Vol. 1, No. 1. Currey. D R 1980. Coastal geomorphology of Great Salt Lake and vicinity. Pages 69-82 in J. W, Gvvynne, ed., Great Salt Lake — a scientific, historical, and economic overview. Utah Geological and Mineral Survey Bulletin 116. 1988. Personal communication. Department of Geographv, University of Utah, Salt Lake Citv, Utah 84112. Currey, D R . M S Berry, S A Green, and S B Murchison 1988. Very late Pleistocene red beds in the Bonneville Basin, Utah and Nevada. Geo- logical Society of America Abstracts with Pro- grams 20: 55. Currey. D R , andC G Oviatt 1985. Duration, average rates and probable causes of Lake Bonneville ex- pansions, stillstands and contractions during the last Deep- Lake Cycle 32,000-10,000 years ago. Geographical Journal of Korea 10: 1085-1099. Miller, R D 1980. Surficial geological map along part of the Wasatch Front, Salt Lake Valley, Utah. USGS Map MF-1198, two 1:100,000 maps and pam- phlet, p. 13. Miller, R D . R Van Horn. W E Scott, and R M Forester 1980. Radiocarbon date support con- cept of continuous low levels of Lake Bonneville since 11.000 years B.P. Geological Society of America Abstracts with Programs 12: 297-298. Rawley. E V. 1980. Wildlife of the Great Salt Lake. Pages 287-303 in J. W. Gwynn, ed., Great Salt Lake— a scientific, historical, and economic overview. Utah Geological and Mineral Survey Bulletin 116. Reeves, C C , Jr 1968. Page 78 in Introduction to pale- olimnology. Elsevier Publishing Company, New York. Scott. W E . W D McCoy. R. R Schroba. and M Ru- bin. 1983. Reinterpretation of the exposed record of the last two cycles of Lake Bonneville, western United States. Quaternary Research 20: 261-286. Smith, G R., W L Stokes, and K F Horn 1968. Some late Pleistocene fishes of Lake Bonneville. Copeia 4: 807-816. Stokes, W L., G R Smith, and K. F. Horn 1964. Fossil fishes from the Stansbury level of Lake Bon- neville, Utah. Utah Academy of Science, Arts, and Letters Proceedings 41: 87-88. Varley, J. D., and J. C. Livesay. 1976. Pages 4-5 in Utah ecology and life history of the Utah chub, Gila atraria, in Flaming Gorge Reservoir, Utah- Wyoming. Utah Division of Wildlife Resources Publication 76-16. MEDIATION OF NUTRIENT CYCLING BY ARTHROPODS IN UNMANAGED AND INTENSIVELY MANAGED MOUNTAIN BRUSH HABITATS T. A. Christiansen1, J. A. Lockwood1, and J. Powell" Abstract. — The role of arthropods in mediating nutrient cycling on a community level was examined in a mountain shrub habitat that was managed by mowing brush to a 20-cm stubble, applying aerially 2,4-D butyl ester, or burning sixteen 4-ha study sites. Malathion and carbaryl were used to decrease arthropod populations. Higher nutrient concentrations occurred in the litter and foliage than in the soil of unmanaged habitats. Arthropods decreased nutrient concentrations in litter and foliage in unmanaged and herbicide-sprayed sites. Arthropod populations increased nutrient concentrations in mowed and burned sites. Nitrogen was consistently affected by both arthropods and brush management in all habitats. Regulation of nutrient cycling by arthro- pods appears to be a function of the frequency and severity of habitat disturbance (Schowal- ter 1986). Arthropod responses to pertur- bances appear to stabilize ecosystem produc- tivity through regulating plant, soil, nutrient, and light relations by changing plant structure and plant species biomass (Mattson and Addy 1975). In addition, the severity of a distur- bance may be reduced by increasing its reli- ability. Plant age structure as well as plant biomass is changed as old, nonproductive plants and plant parts are eaten or reduced to litter by arthropods. This process also has an effect on nutrient release and containment in a habitat (Schowalter 1986). Phytophagous insects require proteina- ceous nitrogen for their life cycles. When the nitrogen level in plants increases, insect as- similation and growth efficiencies increase (Mattson 1980). Thus, although high nitrogen levels can be detrimental to some insects (Stark 1965), evidence suggests that insects will generally seek and respond to high nitro- gen levels (Prestidge and McNeill 1965). Higher available nitrogen levels can be found in annual plants as compared to perennial plants. Being short-lived, annuals do not com- mit high levels of energy to defensive chem- istry but allocate most of their energy to re- production. Therefore, insect feeding can result in an increase of nutrient and energy flow in annual plants as they recover from nitrogen loss (Grimes 1979). Other inorganic elements, such as calcium, magnesium, phos- phorus, potassium, and sodium, are vital to diets of many insects (Dadd 1977, 1985), and these elements are likely regulated to some extent in the environment by insects. The role of arthropods in forest nutrient cycling has been studied by Cornaby (1977), Crossley (1977), and Webb (1977). However, few studies have been conducted on nutrient cycling in the sagebrush/bitterbrush system. Because of the lack of research in this area, we undertook this study with two objectives in mind. The first was to determine if arthropods were a mediating factor in nutrient cycling, on a community level, in a sagebrush habitat. The second objective was to determine if habitat disturbance would influence the role of arthropods in nutrient cycling. Materials and Methods This study was conducted on a sagebrush (Artemisia tridentata) and bitterbrush (Pur- shia tridentata ) habitat located at an elevation of 2,400 m, 12 km southeast of Saratoga, Car- bon County, Wyoming. Precipitation aver- aged 480 mm per year, mostly in the form of snow. Temperatures ranged from 21.0 to 27.0 C during the 100 days of the summer study period though the mean annual temperature is 10.2 C. Soils are the North Park Formation of brown sandy loams developed on loess, limestone, sandstone, and tuff. 'Department of Plant, Soil and Insect Sciences, University of Wyoming, Laramie, Wyoming 82071. 2Department of Range Management, University of Wyoming, Laramie, Wyoming 82071. 134 January 1989 Christiansen et al.: Arthropod Nutrient Cycling 135 The study site consisted of 16 blocks of at least 4 ha each. These blocks were randomly chosen from sites that had similar vegetation, soil chemistry, and soil texture characteris- tics. Habitat manipulation in May 1986 con- sisted of either mowing four 4-ha blocks to a 20-cm stubble height or applying 2,4-D butyl ester in water at an aerial rate of 0.91 kg per hectare to four 4-ha blocks. In the fall of 1986, four 4-ha blocks were burned. These treat- ments are used as sagebrush management practices in Wyoming. Control blocks con- sisted of four 4-ha, unmanaged, high-density shrub areas. Arthropod populations were reduced using two insecticides, carbaryl (1.68 kg/ha) and malathion (1.4 kg/ha). These compounds were alternately applied every two weeks from early May through August of 1987 to half of each block in the managed and unmanaged areas. The other half of each block was left untreated as a control. To determine effectiveness of insecticide treatments, we estimated arthropod densities using 100 sweeps of a 38-cm diameter sweep net to collect arthropods along three 100-m transects in both closed (covered by brush canopy) and open (not covered by brush canopy) microhabitats in each managed and unmanaged split-plot every 10 days. Foliage, litter, and soil samples were col- lected in open and closed microhabitats in early September 1987 from fifteen 0.25-m" quadrats located along three 100-m transects in each split-plot. Samples were placed in paper bags, returned to the laboratory, and dried for three days at 75 C. All foliage and litter samples were ground in a Wiley Mill with a 40-mesh screen. Total nitrogen for litter, plant, and soil ma- terial was determined by the use of a block digestion method and analysis (Jones 1971). Concentrations of magnesium, calcium, phos- phorus, potassium, and sodium were deter- mined by the Havlin and Soltanpour (1980) method of nitric digestion. All nutrient con- centrations except for phosphorus, which was determined by the Olsen and Dean (1960) colorimetric method, were determined by the use of plasma spectrometry. Soil pH, electri- cal conductivity, and soluble cation analysis were evaluated with methods of Richards (1980). Soil organic matter was determined by combustion (Grewling and Peech 1965). Tex- Table 1. Vegetation parameters as determined for the baseline vear (1985), treatment year (1986), and recovery year (1987). Treatment Year 1985 1986 1987 Herbaceous biomass (g/m2)a Unmanaged 7.0a 7.2a 5.7a Herbicide 7.3a 16.7b 29.1c Mowed 7.4a 24.0b 23.0b Burned 7.2a 7.4a 4.3b Litter biomass (g/m2)a Control 78.0a 79.6a 87.1a Herbicide 75.8a 77.3a 73.1a Mowed 79.0a 291.2b 287.6b Burned 76.4a 77.7a 2.3b Sagebrush density (%) Control 54.7a 54.8a 55.9a Herbicide 47.9a 21.8b 18.0b Mowed 63.7a 32.0b 38.0b Burned 60.2a 59.7a 5.7b Bitterbrush density (%)b Control 54.0a 55.0a 57.0a Herbicide 14.4a 45.7b 64.0c Mowed 41.4a 51.4b 52.0b Burned 47.4a 48.1a 6.3b Total shrub cover (%) Control 28.1a 28.1a 29.0a Herbicide 29.6a 17.5b 17.0b Mowed 31.6a 24.0b 23.0b Burned 30.3a 29.8a 6.2b "Means for a parameter within a treatment between years followed by the same letter are not significantly different (P .05; LSD Test [Fisher 1949]). N 60 samples Means for a parameter followed by the same letter are not significantly different (P .05; Chi-square). N 60 samples. ture analysis applied the theory of particle fractionation as used by Day (1965). The results were analyzed as a split-plot design, and Fisher's (1949) protected least sig- nificant difference test was employed to com- pare specific treatment effects on nutrient content of foliage and litter. Chi-square analy- sis was used to assess differences in the pro- portions of shrub cover between management areas. In all statistical tests, differences were considered significant at P < .05. Results and Discussion Disturbance of vegetation by brush man- agement practices resulted in significant de- creases in density of the major overstory plant species, sagebrush and bitterbrush (Table 1). Herbaceous biomass increased dramatically in both mowed and herbicide-sprayed blocks during and after the year of management, as 136 Great Basin Naturalist Vol. 49, No. 1 Table 2. Nutrient concentrations (ppm) of soil, litter, and foliage in sagebrush/bitterbrush habitats before man- agement procedures were applied. Nutrient" Habitat Soil Litter Foliage Nitrogen Closed 21.00a 96.00c 112.00e Open 23.50ab 104. OOd 149. OOf Potassium Closed .48a 2.54cd 10.75e Open .54ab 1.78bc 9.68de Magnesium Closed .70a 1.45d l.Olabc Open .84ab .84ab 1.90e Sodium Closed .02a lid .04b Open .09c .02a .02a Phosphorus Closed .69ab .54a 1.19c Open .74b .72b .92b Calcium Closed 4.98ab 8.49d 4.17a Open 5.23abc 5.65bc S.lOd "Means for a nutrient followed by the same letter are not significantly different (P .05; LSD Test [Fisher 1949)). N 60 samples. Closed foliage indicates shrubs and open foliage indicates grass. often occurs with the decline of the shrub layer (Barbour et al. 1980). Litter biomass significantly increased during and after the year of mowing. Both herbaceous and litter biomass decreased significantly after being burned. In undisturbed sites the vegetation compo- nent contained higher concentrations of nitro- gen, potassium, and magnesium than the lit- ter, which had higher concentrations than the soil (Table 2). Phosphorus levels were higher in shrub vegetation than in any other compo- nent. This was probably due to long-term ac- cumulation of this element in the woody por- tion of shrubs, as is known to occur in forest habitats (Horn 1974). Calcium was highest in the litter of closed microhabitats and in grasses. Sodium concentrations were highest in litter located under shrubs. Sodium is read- ily leached from open litter (Forth and Turk 1972). Thus, sodium concentration under shrubs may have been higher due to more protection from precipitation and leaching. Insecticide applications reduced arthropod densities by 82% in unmanaged areas, 76% in mowed areas, 74% in herbicide-applied areas, and 78% in burned areas (Table 3). All orders were clearly diminished by the insecticides, although Acari generally fared better than the insect orders. The nutrient levels in unmanaged habitats with intact arthropod communities compared well with previous measurements in sage- brush (Gough and Erdman 1980). Arthropods in open microhabitats of unmanaged blocks significantly increased nitrogen, magnesium, and calcium in both litter and foliage (Table 4). Phosphorus was significantly decreased in foliage of open microhabitats by arthropods. Foliage potassium significantly increased in insecticide-treated, open areas. Arthropods significantly increased nitrogen in both litter and foliage of closed microhabitats and signifi- cantly increased potassium in litter. Arthro- pods significantly decreased magnesium and sodium in both litter and foliage in closed microhabitats and significantly decreased phosphorus in foliage of closed microhabitats. Decreases in the level of a nutrient follow- ing insecticide applications indicate that arthropods have some direct or indirect effect on the release of these nutrients. In particu- lar, arthropods are clearly acting to increase plant nitrogen, since this element decreased significantly in foliage and litter of open and closed microhabitats after insecticide treat- ments. There is some evidence that insect feeding may stimulate plant growth (Walms- ley et al. 1987), and this effect may account for the impact of arthropods on plant nitrogen levels. Since arthropods function in unman- aged sagebrush habitats to increase the levels of nitrogen in plant material, the indirect impacts of widespread applications of broad- spectrum insecticides on western prairies for control of grasshoppers should be considered (Hewitt and Onsager 1983). Insects may also be functioning to directly mobilize nutrients; aphids are a good example of how insects may mediate nutrient cycling. As a group, aphids are generally inefficient at energy conversion of plant sap; only 5% of the potential dietary energy is utilized, and the remainder is excreted as honeydew (Hagen et al. 1951). This product is nutritionally impor- tant to a number of animals and fungi (Wilson 1971). As such, honeydew production func- tions as an important and rapid return of en- ergy and nutrients to the local habitat. Soluble nutrients in the phloem are absorbed by aphids and provide food to other organisms through honeydew or indirectly through pre- dation. Therefore, aphids and other arthro- pods may act as nutrient-storage organisms or sinks for plant nutrients in a small scale (Way and Cammell 1970). Joy (1967) reported that local amino acid synthesis can be induced from nutrients contained in the aphid sink. The reduction of arthropod population den- sity by insecticides in herbicide-treated sites January 1989 Christiansen etal.: Arthropod Nutrient Cycling 137 Table 3. Effects of insecticide treatments on arthropod population densities (no./m") in unmanaged, mowed, and herbicide-managed blocks averaged over the study. Habitat Unmanaged Mowed Herb icide Order Treated Control Treated Control Treated Control Acari foliage 0 0 2 0 2 0 Araneae foliage 6 1 2 0 3 1 Homoptera foliage 218 41 35 10 39 7 Coleoptera foliage 6 1 15 1 6 3 Diptera foliage 16 2 3 0 2 1 Hymenoptera foliage 62 10 72 29 85 24 Acari litter 495 95 640 170 677 175 Araneae litter 9 2 6 0 7 1 Homoptera litter 11 0 11 2 12 4 Coleoptera litter 13 0 9 0 10 2 Diptera litter 24 5 13 1 25 7 Hymenoptera litter 16 0 12 0 13 0 Collembola litter 14 3 93 18 63 16 Thysanura litter 6 1 8 0 3 0 % decrease 82% 76% 74% Table 4. Arthropod effects on nutrients (ppm) in unmanaged sagebrush/bitterbrush habitats treated with in- secticides. Habitat Fol iage Litter Nutrient3 Treated Untreated Treated Untreated Nitrogen Closed lOO.OOcde 112. OOf 69.00a 96.00cd Open 87.00cb 149. OOe 82.00b 104.00ed Magnesium Closed 1.29cd 1.01 abed 194g 1.45def Open 0.92cab 1.90g 0.75a 0.84ab Calcium Closed 4.17a 4.17a 8.87ed 8.49d Open 4.78ab 8.10d 4.78ab 5.65c Sodium Closed <.01a 0.04abc 0.14e O.lld Open 0.04abc 0.02ab <.01a 0.02ab Potassium Closed ll.51g 10.75g 1.58ab 2.45cd Open 8.51f 5.59e l.OOab 0.54a Phosphorus Closed 1.61g 1.19ef 0.56abc 0.54abc Open 1.15e 0.92d 0.57abc 0.72abc aMeans for a nutrient followed by the same letter are not significantly different (P .05; LSD Test [ Fisher 1949]) N 60 samples resulted in a significant decrease of all litter nutrients except sodium (Table 5). However, only phosphorus was significantly decreased in foliage. Again, the decrease of nutrients when arthropod populations were reduced in herbicide-treated areas could be the result of arthropods being storage components in the community as in undisturbed sites. The appli- cation of herbicide effectively eliminated the closed microhabitat. In herbicide-treated mountain brush habi- tats, arthropods functioned to generally in- crease litter nutrient levels. Whether insects directly (via death and decomposition) or indi- rectly (via secretions and excretions) elevate litter nutrient levels is yet to be determined. The general lack of change in foliage nutrient levels suggests that the loss of nutrients in litter following insecticide treatments did not result in the effective mobilization of the nu- trients for uptake by grasses. Many of the annual grasses may not have been in a pheno- logical stage suitable to exploit the available resources as they became available in mid- to late summer. Application of insecticides to mowed sites caused significant increases of litter nitro- gen, potassium, and phosphorus in open mi- crohabitats (Table 6). The increase in phos- phorus could be an artifact from application of the organophosphorous insecticide. Canopy- covered litter in insecticide-treated plots had significant increases in nitrogen, magnesium, and calcium. There were significant decreas- es of foliage magnesium and phosphorus in 138 Great Basin Naturalist Vol. 49, No. 1 Table 5. Arthropod effects on nutrients (ppm) in herbicide-treated sagebrush/bitterbrush habitats treated with insecticides. Habitat Foliage Litter Nutrient2 Treated Untreated Treated Untreated Nitrogen Open 142.00b 142.00b 55.00a 146.00b Magnesium Open 0.91ab 0.86ab 0.91ab 1.03c Calcium Open 6.62b 6.30b 4.05a 6.42b Sodium Open <01a 0.04a 0.32b <.01a Potassium Open 7.12c 7.12c 0.47a 1.12b Phosphorus Open 0.98b 1.24c 0.24a 0.94b aMeans for a nutrient followed by the same letter are not significantly different (P .05; LSD Test [Fisher 1949]) N 60 samples. Table 6. Arthropod effects on nutrients (ppm) in mowed sagebrush/bitterbrush habitats treated with insecticides. Habitat Foliage Litter Nutrienta Treated Untreated Treated Untreated Nitrogen Closed 122.00c 120.00c 117.00c 99.00b Open 114.00c 139. OOd 119.00c 85.00a Magnesium Closed 1.13cd 1.48e 1.48e 0.68a Open 1.17cd 0.96bcd 0.93bc 0.88ab Calcium Closed 4.71a 6.78abcd 9.05g 6.78ef Open 5.94abcde 5. 12abc 5.07ab 5.63abcd Sodium Closed <.01a 0.03a <.01a 0.30b Open <01a 0.03a <.01a 0.04a Potassium Closed 11.13g 10.79g 2.05abcd 0.85ab Open 0.74a 6.24ef 5.34e 0.93abc Phosphorus Closed 0.98f 1.40g 0.53abc 0.33ab Open 0.58cd 0.66de 0.70e 0.32a aMeans for a nutrient followed by the same letter are not significantly different (P = .05. LSD Test [Fisher 1949]) N 60 samples. Table 7. Arthropod effects on nutrients (ppm) in burned sagebrush/bitterbrush habitats treated with insecticides. Habitat Foliage Litter Nutrient2 Treated Untreated Treated Untreated Nitrogen Open NA NA 215.00a 204.00a Magnesium Open NA NA 3.60a 2.85b Calcium Open NA NA 8.65a 7.69a Sodium Open NA NA 0.04a <.01a Potassium Open NA NA 43.57a 29.52b Phosphorus Open NA NA 2.30a 1.83b aMeans for a nutrient followed by the same letter are not significantly different (P = .05, LSD Test [Fisher f949]). N 60 samples. NA = not applicable. canopy-covered microhabitats. Foliage in insecticide-treated, open microhabitats had significant decreases in nitrogen, potassium, and phosphorus (Table 6). In mowed mountain brush habitats, arthro- pods functioned to restrict the flow of nutri- ents from foliage to litter or, conversely, to accelerate the flow of nutrients from litter to foliage. Earlier work on sagebrush habitats showed that insects are critical components in the process of litter decomposition (Chris- tiansen and Lockwood, unpublished data). Thus, we suggest that the application of insec- ticides decreased the breakdown of litter by arthropods and thereby reduced the nutrients available to plants. The fact that arthropods apparently had the opposite effect on mobi- lization of nutrients in mowed and herbicide- treated habitats may have been a result of the functional differences in these treatments; the overstory plants were rapidly recovering fol- lowing mowing but were effectively elimi- nated by herbicide applications. Burning a habitat dramatically altered the plant architecture, leaving no closed micro- habitats. In burned sites treated with insec- ticides, magnesium, potassium, and phos- phorus were all significantly increased in litter (Table 7). As with mowing, significant January 1989 Christiansen etal.: Arthropod Nutrient Cycling 139 nutrient increases in litter indicate that in- sects are important in mobilizing nutrients. Burned habitats had properties of herbicide- treated habitats (elimination of the canopy) and mowed habitats (ongoing recovery of the overstory plants, although insufficient for sampling). Plants under various stresses respond differently (Hale and Orcutt 1987), and it is not unexpected that the interacting impacts of arthropods and management prac- tices resulted in different alterations in nutri- ent cycling. Acknowledgment We thank C. C. Burkhardt for assistance with pesticide applications and P. Lew for field and laboratory assistance. This project was partially funded bv USDA-CSBS grant 85-CSBS-2-2702. Literature Cited Barbour, M. G , J. H Burk, and W D. Pitts 1980. Ter- restrial plant ecology. The Benjamin/Cummings Publ. Co., Inc., Menlo Park, California. 604 pp. Cornaby. B W. 1977. Saprophagous organisms and prob- lems in applied resource partitioning. Pages 96-100 in W. J. Mattson, ed., The role of arthro- pods in forest ecosystems. Springer- Verlag, New York. 104 pp. Crossley, D A , Jr 1977. The role of terrestrial sapro- phagous arthropods in forest soils: current status of concepts. Pages 49-56 in W. J. Mattson, ed., The role of arthropods in forest ecosystems. Springer- Verlag, New York. 104 pp. Dadd, R H 1977. Qualitative requirements and utili- zation of nutrients: insects. Pages 305-346 in M. RechCigil, ed., Handbook series in nutrition and food. CRC Press, Cleveland, Ohio. 561 pp. 1985. Nutrition: organisms. Pages 313-390 in G. A. Kerkutand L. I. Gilbert, eds., Comprehen- sive insect physiology, biochemistry and pharma- cology. Vol. 4. Regulation: digestion, nutrition, excretion. Pergamon Press, New York. 639 pp. Day, P. R. 1965. Particle fractionation and particle size analysis. Pages 425-442 in C. A. Black, ed., Methods of soil analysis. Part I. Agronomy 9. Amer. Soc. of Agron., Madison, Wisconsin. 1,188 pp. Forth, H. D., and L M Turk 1972. Fundamentals of soil science. John Wiley and Sons, New York. 454 pp. FlSHER, R A 1949. The design of experiments. Oliver and Boyd, Edinburgh. 456 pp. GOUGH, L. P., AND J A. Erdman 1980. Seasonal differ- ences in the element content of Wyoming big sagebrush. J. Range Manage. 33: 374-378. Grewling, T., and M Peech 1960. Chemical soil tests. Cornell Univ. Agric. Expt. Sta. Bull. No. 960. 362 pp. Grimes, J P 1979. Plant strategies and vegetation pro- cesses. John Wiley and Sons, New York. 222 pp. Hagen, K S., R H Dadd, and J Reese 1984. The food of insects. Pages 79-112 in C. B. Huffaker and R. L. Rabb, eds.. Ecological entomology. Wiley and Sons, New York. 844 pp. Hale. G M , and D M Orcutt 1987. The physiology of plants under stress. John Wilev and Sons, New York. 206 pp. Havlin, J L , and P L Soltanpour 1980. A nitric acid plant tissue digest method for use with inductively coupled plasma spectrometry. Comm. Soil Sci- ence and Plant Analysis 11: 969-980. Hewitt, G B . and J A Onsager 1983. Control of grasshoppers on rangeland in the United States — a perspective. J. Range Manage. 36: 202-207. Horn. H S 1974. The ecology of secondary succession. Ann. Rev. Ecol. System. 5: 25-37. Jones, J B 1971. Laboratory procedures for the analysis of soils, feed, water, and plant tissues. Soil Testing and Plant Analysis Laboratory, Athens, Georgia. 534 pp. Joy, K W 1967. Carbon and nitrogen sources for protein svnthesis and growth of sugar-beet leaves. J. Exp. Bot. 18: 140-150. Mattson, W J , Jr 1980. Herbivory in relation to plant nitrogen content. Ann. Rev. Ecol. Syst. 11: 119-161. Mattson. W. J. Jr., and N D Addy 1975. Phytophagous insects as regulators of forest primary production. Science 190: 515-522. Olsen, S R . and L A Dean 1965. Phosphorus. Pages 403-430 in C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White, F. E. Clark, and R. C. Dinauer, Methods of soil analysis. Amer. Soc. Agr. Inc., Madison, Wisconsin. 1,159pp. Prestidge. R A , and S McNeill 1983. The importance of nitrogen in the ecology of grassland Auchenor- rhvncha (Homoptera). Symp. Brit. Ecol. Soc. 7: 117-124. Richards, L A 1980. Diagnosis and improvement of saline and alkali soils. Agric. Handbook No. 60. USDA, Washington, D.C. 158 pp. Schowalter. T. D 1986. Adaptations of insects to distur- bance. Pages 235-251 in S. T. A. Pickett and P. S. White, eds., The ecology of natural disturbance and patch dynamics. Academic Press, New York. 472 pp. Stark, R W 1965. Recent trends in forest entomology. Ann. Rev. Entomol. 10: 303-324. Walmsley. M. R , J L Capinera, J K. Detling, and M I Dyer 1987. Growth of blue grama and west- ern wheatgrass following grasshopper defoliation and mechanical clipping. J. Kansas Entomol. Soc. 60: 51-57. Way, M J , and M. Cammell. 1970. Aggregation be- haviour in relation to food utilization by aphids. Pages 229-247 in A. Watson, ed., Animal popula- tions in relation to their food resources. Blackwell, Oxford. 477 pp. Webb, D. P 1977. Regulation of deciduous forest litter decomposition by soil arthropod feces. Pages 57-69 in W. J. Mattson, ed., The role of arthropods in forest ecosystems. Springer- Verlag, New York. 104 pp. Wilson, E O. 1971. The insect societies. The Belknap Press, Cambridge, Massachusetts. 548 pp. LOCALITY, HABITAT, AND ELEVATION RECORDS FOR THE DESERT SHREW, NOTIOSOREX CRAWFORDI Russell Davis1 and Ronnie Sidner1 Abstract. — Two specimens of Notiosorex crawfordi (Coues) were obtained from the Rincon Mountains in south- eastern Arizona. Elevations were 2,438 and 2,618 m. At the lower-elevation site the habitat was a meadow in a ponderosa pine forest. The desert shrew, Notiosorex crawfordi (Coues), is known to occur in a wide variety of habitats ranging from desert gravel to pon- derosa-pinyon pine woodlands, and at eleva- tions from sea level to 2, 100 m (Armstrong and Jones 1972). On 14 August 1985, while trapping rodents on top of Spud Rock (elevation 2,618 m by bench mark) in the Rincon Mountains of Saguaro National Monument near Tucson, Arizona, we found a dead specimen of this shrew. While there is not much vegetation on the top of Spud Rock itself, it is surrounded below by a fir forest to the north and east and a ponderosa pine (Pinus ponderosa) forest to the south and west (Marshall 1956). The dead shrew was lying exposed on a large rock, and the skin was dry with some minor damage (the skull was partially ex- posed). This suggested the possibility that this individual might have been caught by an owl or other predator in a lower-elevation habitat and carried to this site. A 32-km radius should encompass virtually the entire area from which the prey of raptors would be ob- tained, and most prey are taken within 5 km of a site (Harris 1977). But in this case, be- cause of local topography in this small moun- tain range, these distances (whether consid- ered horizontally or vertically) could extend through habitats ranging from desert shrub to montane conifer forest. During early summer of 1985 we placed 25 plastic can traps in a grass-fern meadow that was surrounded by ponderosa pines near Manning Camp (elevation 2,438 m by Park Service sign), about 2 km south of the Spud Rock site. These can traps, with holes in the bottoms for drainage, were checked in late summer 1985 and early summer 1986 without success. They then remained unattended un- til removed on 2 June 1988. At that time one of the traps, located adjacent to a rotting log in a dry, grassy portion of this meadow, contained a decomposed specimen of Notiosorex craw- fordi, the only mammal caught in these traps throughout the three-year period. This specimen establishes the fact that a small population of the desert shrew does occur at least in one meadow within a pon- derosa pine forest in these mountains. If an owl (or other predator) were responsible for the other specimen found dead on Spud Rock, it may not have been carried for any apprecia- ble distance nor was it necessarily obtained from some habitat different from that occur- ring near its discovery site. Findley (1969) has shown that Tamias dor- salis occurs from the bottom to the top of certain southwestern mountains in the ab- sence of other species of chipmunks (as is the case in the Rincons), but that it is restricted to lower portions when these species are present. Perhaps the elevational and habitat distribution of Notiosorex crawfordi expands comparably in the absence of other species of shrews. Rodents collected at both study sites in- clude Sigmodon ochrognathus (see Davis and Ward 1988), Peromyscus boylii, and Neotonia mexicana. Whitetail deer (Odocoileus virgini- anus) and 14 species of bats occasionally occur in the meadow, pocket gopher mounds are abundant, and Sciurus aberti, Sperm- ophilus variegatus, and Tamias dorsalis have been seen in the vicinity. Except for the two 'Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721. 140 January 1989 Davis, Sider: Desert Shrew 141 specimens reported here, no other shrews are known from the Rincon Mountains. Acknowledgments S. P. Cross suggested the can traps, and O. G. Ward helped in putting them out. E. L. Cockrum verified the identification of the two specimens. This study was possible only with logistic help from Park Service per- sonnel at Saguaro National Monument. Literature Cited Armstrong. D. M., and J. K. Jones, Jr. 1972. Notiosorex crawfordi. Mamm. Species 17: 1-5. Davis. R , and O. G. Ward. 1988. A vacant "Microtus niche" now occupied by the yellow-nosed cotton rat (Sigmodon ochrognathus) on an isolated moun- tain in southeastern Arizona. J. Mamm. 69: 362-365. FlNDLEY. J. S 1969. Biogeography of southwestern bo- real and desert mammals. Univ. Kansas Mus. Nat. Hist. Misc. Publ. 51: 113-128. Harris, A H 1977. Wisconsin age environments in the northern Chihuahuan Desert: evidence from the higher vertebrates. Pages 23-52 in R. H. Wauer and D. H. Riskind, eds., Transactions of the symposium on the biological resources of the Chihuahuan Desert region. United States and Mexico. Natl. Park Serv. 3: 1-658. Marshall, J T 1956. Summer birds of the Rincon Moun- tains, Saguaro National Monument, Arizona. Con- dor 58: 81-97. NOTICE TO CONTRIBUTORS Manuscripts intended for publication in the Great Basin Naturalist or Great Basin Natural- ist Memoirs must meet the criteria outlined in paragraph one on the inside front cover. The manuscripts should be sent to Stephen L. Wood, Editor, Great Basin Naturalist, 290 Life Science Museum, Brigham Young University, Provo, Utah 84602. Three copies of the manuscript are required. They should be typewritten, double-spaced throughout on one side of the paper, with margins of at least one inch on all sides. Use a recent issue of either journal as a format, and the Council of Biology Editors Style Manual, Fifth Edition (AIBS 1983) in preparing the manuscript. An abstract, about 3 percent as long as the text, but not exceeding 200 words, written in accordance with Biological Abstracts guidelines, should precede the introductory paragraph of each article. Authors may reduce their typesetting costs substantially by sending a copy of their article on a 5.25-inch floppy disk prepared with WordPerfect software in addition to three copies of the manuscript. Illustrations and Tables. All illustrations and tables should be made with a view toward having them appear within the limits of the printed page. Glossy photoprints of illustrations should accompany the manuscript. Illustrations should be prepared for reduction by the printer to either single-column (2« inches) or double-column (5| inches) width, with the length not exceeding l\ inches. Costs Borne by Contributor. Contributors to the Great Basin Naturalist should be prepared to donate $40 per printed page toward publication of their article. No reprints are furnished free of charge. Excessive or complex tables requiring typesetting may be charged to the author at cost. Great Basin Naturalist Memoirs costs must be borne in full by the author. Reprint Schedule for the Great Basin Naturalist 100 copies, minimum cost for 2 pages, $26. Each additional 2 pages, $6. Each additional 100 copies, $4 for each 2 pages. Examples: 300 copies of 10 pages = $82; 200 copies of 13 pages = $86. Great Basin Naturalist Memoirs No. 1 The birds of Utah. $10. No. 2 Intermountain biogeography: a symposium. $15. No. 3 The endangered species: a symposium. $6. No. 4 Soil-plant-animal relationships bearing on revegetation and land reclamation in Ne- vada deserts. $6. No. 5 Utah Lake monograph. $8. No. 6 The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. $60. No. 7 Biology of desert rodents. $8. No. 8 The black-footed ferret. $10. No. 9 A Utah flora. $40. No. 10 A reclassification of the genera of Scolytidae (Coleoptera). $10. No. 11 A catalog of Scolytidae and Platypodidae (Coleoptera), part 1: bibliography. $30. No. 12 Research in the Auchenorrhyncha, Homoptera: a tribute to Paul W. Oman. $30. TABLE OF CONTENTS On the distribution of Utah's hanging gardens. Stanley L. Welsh 1 Changes in mule deer size in Utah. Dennis D. Austin, Robert A. Riggs, Philip J. Urness, David L. Turner, and John F. Kimball 31 White-tailed prairie dog (Cynomys leucurus Merriam) diggings in western harvester ant, Pogonomyrmex occidentals (Cresson), mounds. William H. Clark and Cynthia J. Clark 36 Amphibians of western Chihuahua. Wilmer W. Tanner 38 Observations on recruitment and ecology of razorback sucker: lower Colorado River, Arizona-California-Nevada. Paul C. Marsh and W. L. Minckley 71 Competition between adult and seedling shrubs of Ambrosia dumosa in the Mojave Desert, Nevada. Richard Hunter 79 Aquatic insects in Montezuma Well, Arizona, USA: a travertine spring mound with high alkalinity and dissolved carbon dioxide. Dean W. Blinn and Milton W. Sanderson 85 A new Haliplus from Warm Springs, Nevada (Coleoptera: Haliplidae). Samuel A. Wells 89 On the genus Paracarinolidia (Cicadellidae: Coelidiinae: Teruliini). M. W. Nielson .... 92 Two genera and two new species of Teruliine leafhoppers (Homoptera: Cicadellidae: Coelidiinae). M. W. Nielson 96 A new species oiAsclepias (Asclepiadaceae) from northwestern New Mexico. Kenneth D. Heil, J. Mark Porter, and Stanley L. Welsh 100 Effect of timing of grazing on soil-surface cryptogamic communities in a Great Basin low-shrub desert: a preliminary report. James R. Marble and Kimball T. Harper 104 Size and overlap of Townsend ground squirrel home ranges. Nicholas C. Nydegger and Donald R. Johnson 108 Pisolithus tinctorius, a Gasteromycete, associated with Jeffrey and Sierra lodgepole pines on acid mine spoils in the Sierra Nevada. R. F. Walker Ill Spatial and temporal variability in perennial and annual vegetation at Chaco Canyon, New Mexico. Anne C. Cully and Jack F. Cully, Jr 113 Associations of small mammals occurring in a pluvial lake basin, Ruby Lake, Nevada. Mark A. Ports and Lois K. Ports 123 Utah chub (Gila atraria) from the latest Pleistocene Gilbert shoreline, west of Cor- rine, Utah. Stuart Murchison 131 Mediation of nutrient cycling by arthropods in unmanaged and intensively managed brush habitats. T. A. Christiansen, J. A. Lockwood, and J. Powell 134 Locality, habitat, and elevation records for the desert shrew, Notiosorex crawfordi. Russell Davis and Ronnie Sidner 140 »o; HE GREAT BASIN NATURALIST )lume49 No. 2 30 April 1989 Brigham Young University v T SEP 18 1989 HA ■TY GREAT BASIN NATURALIST Editor. Stephen L. Wood, Department of Zoology, 290 Life Science Museum, Brigham Young University, Provo, Utah 84602. Editorial Board. Kimball T. Harper, Chairman, Botany and Range Science; Ferron L. An- dersen, Zoology; James R. Barnes, Zoology; Hal L. Black, Zoology; Jerran T. Flinders, Botany and Range Science; Stanley L. Welsh, Botany and Range Science. All are at Brigham Young University. Ex Officio Editorial Board Members include Bruce N. Smith, Dean, College of Biological and Agricultural Sciences; Norman A. Darais, University Editor, University Publications; Stephen L. Wood, Editor, Great Basin Naturalist. The Great Basin Naturalist was founded in 1939. The journal is a publication of Brigham Young University. Previously unpublished manuscripts in English pertaining to the biological natural history of western North America are accepted. The Great Basin Naturalist Memoirs series was established in 1976 for scholarly works in biological natural history longer than can be accommodated in the parent publication. The Memoirs appears irregularly and bears no geographical restriction in subject matter. Manuscripts for both the Great Basin Naturalist and the Memoirs series will be accepted for publication only after exposure to peer review and approval of the editor. Subscriptions. Annual subscriptions to the Great Basin Naturalist are $25 for private individuals and $40 for institutions (outside the United States, $30 and $45, respectively), and $15 for student subscriptions. The price of single issues is $12. All back issues are in print and are available for sale. All matters pertaining to subscriptions, back issues, or other business should be directed to the Editor, Great Basin Naturalist, Brigham Young University, 290 Life Science Museum, Provo, Utah 84602. The Great Basin Naturalist Memoirs may be purchased from the same office at the rate indicated on the inside of the back cover of either journal. Scholarly Exchanges. Libraries or other organizations interested in obtaining either journal through a continuing exchange of scholarly publications should contact the Brigham Young University Exchange Librarian, Harold B. Lee Library, Provo, Utah 84602. Manuscripts. See Notice to Contributors on the inside back cover. 7-89 650 40815 ISSN 017-3614 The Great Basin Naturalist Published at Provo, Utah, by Bricham Young University ISSN 0017-3614 Volume 49 30 April 1989 No. 2 SNAKE CREEK BURIAL CAVE AND A REVIEW OF THE QUATERNARY MUSTELIDS OF THE GREAT BASIN EmileeM. Mead13 and Jim I. Mead'25 Abstract — Snake Creek Burial Cave (SCBC), east central Nevada, is a unique paleontological deposit. The cave is the first natural trap excavated in the Great Basin and one of the few localities describing a valley-bottom community. The recovery of extinct Camelops sp. (camel) and Equus spp. (horse), in addition to radiometric dates, indicates at least some of the deposits to be of late Pleistocene age. Eight mustelid species have been identified from SCBC, including three species not previously reported from the late Rancholabrean of the Great Basin: Mustela nigripes (black-footed ferret), M. nivalis (least weasel), and Gulo gulo (wolverine). A review of late Pleistocene deposits indicates that there are more species of mustelids recovered from Snake Creek Burial Cave than from any other locality in the Great Basin. The Great Basin of western North America encompasses some 390,000 km2 and, although centered in Nevada, also extends into several adjoining states (Fig. 1). This arid region is generally characterized by linear, north-south trending mountain ranges, separated by closed-drainage valleys (Hunt 1967). A num- ber of archaeological and paleontological sites have been excavated in the Great Basin. Few, however, have contained lengthy, well- dated, stratified sequences extending from the Pleistocene, across the critical Pleis- tocene-Holocene boundary, and into the Holocene. Futhermore, those few sites meet- ing these criteria have generally been re- stricted to the mountainous peripheries (Grayson 1987), leaving the valley-bottom communities largely undescribed. Since small mammals are generally sensitive to en- vironmental change, their recovery in archae- ological and paleontological sites can provide important paleoenvironmental data. Snake Creek Burial Cave (SCBC), White Pine County, east central Nevada, has recently been excavated; an estimated 30,000 fish, am- phibian, reptilian, avian, and mammalian bones have been recovered. Of particular in- terest are the many carnivore remains, in- cluding numerous mustelids (Order Car- nivora: Family Mustelidae; weasel family). This paper provides a brief description of SCBC, including recovered mustelids, as well as a timely review of other localities in the Great Basin reporting fossil or subfossil mus- telids. A detailed osteological report of all car- nivores recovered from SCBC is in prog- ress. Additional reports describing the avian and remaining mammalian remains are forth- coming. Snake Creek Burial Cave Setting Snake Creek Burial Cave is a unique pale- ontological deposit. The natural trap cave, with a sinkhole depression and a 17-m, verti- cal, free-fall entrance (Fig. 2), is located at an elevation of 1,731 m on a small limestone 'Ralph M Bilby Research Center, Box 6013, Northern Arizona University, Flagstaff, Arizona 86011-6013. department of Geology, Box 6030. Northern Arizona University, Flagstaff, Arizona 8601 1-6030. Quaternary Studies Program, Northern Arizona University, Flagstaff, Arizona 8601 1-6030. 143 144 Great Basin Naturalist Vol. 49, No. 2 Oregon Idaho ^ , 12 1 7 \ 8 f~ ■*. s ~* 5 y ^\ \ l3 "i 9 C\r-| 10 \ \ 4 >Ay\ 3 / - " Y 6 17 15 16 / X ^\ 18 ^ / V Utah t ' — V— / \ Nevoda , Colifornio Ar izona Fig. 1 . Location map indicating the extent of the Great Basin (heavy line; after Kiiehler 1966); fossil and subfossil locality numbers correspond to those in Table 1. ridge (Devonian-age Guilmette Formation) in the midst of a bajada developed from the drainage of Snake Creek, southern Snake Range (Figs. 1, 3). When pluvial Lake Bon- neville was at its highest stand during the Wisconsinan glacial (16,000 to 13,000 yr B.P. ; Thompson et al. 1986), the beach was within 5 km to the northeast of the cave. SCBC lies directly in a Pleistocene north-south trending habitat corridor that extends from Idaho to northern Arizona and is bordered by numer- ous mountain ranges, including the Snake Range, and by the Bonneville lake system. SCBC allows a description to be made of the late glacial vertebrate community that oc- curred in a valley bottom along the lake mar- gin, an as yet inadequately described biome. The remains are from a natural trap cave; no other caves of this type have been excavated in the central Great Basin. A high proportion of carnivores was recovered from SCBC, in addition to an excellent record of other small and large animals. Natural trap cave localities have traditionally recorded a different faunal sequence from those in walk-in caves, open- air sites, raptor roosts, or packrat middens (White et al. 1984). Methods Just as SCBC is currently accumulating fau- nal material, so too in the past did animals fall into the natural trap and die. It is certain that accumulation has been occurring since at least late Pleistocene (Rancholabrean) time. In- cluded in the 30,000 bones are remains of the extinct Camelops sp. (camel) and Equus spp. (horse). The deposits are of an undeter- mined depth; at 120 cm below the floor sur- face, faunal remains were still recovered in good quantity. Five 1-m units were excavated, all to a depth of approximately 1 m below present floor surface (Fig. 4). An initial 1 x 1-m test pit was excavated in 1984, and four additional units were excavated in 1987. All units were excavated in 10-cm arbitrary levels since the natural stratigraphic levels were all deeper than 10 cm. All material was screened through both 5-mm and 1-mm screens. Stratigraphy. — SCBC is commonly vis- ited by spelunkers. Consequently, some dis- turbance of the sediments has occurred as the cavers have dug in search of additional cave passage. Unit I of the excavated sediment is not of primary deposition. Rather, it retains the mottled appearance of mixed sediment and represents the backdirt of the cavers' dig- gings from nearby areas in the cave. Speci- mens recovered from this unit (I) are used only in a presence-absence status, not in any chronosequence of events. A bat guano layer was encountered below the disturbed unit and is referred to as Unit II. This dung layer is 100% bat guano with no mixing; therefore, we infer that primary deposition begins with Unit II, that is, the bat guano and all sediment below it (Unit III) are in natural stratigraphic sequence. Three radiometric dates have been secured on materials in situ from the excavation. A conventional radiocarbon date of 7,860 ± 130 yr B.P. (Beta-22169) was obtained directly on the bat guano (Unit II). The Accelerator Mass Spectrometry (AMS) radiocarbon technique was used on a small piece of wood recovered from the very top of Unit III; the date is reported as 9,460 ± 160 yr B.P. (Beta-24643, ETH-3688). A U-Th (uranium-thorium; per- formed by Dr. Richard Ku, University of Southern California) series date on an Equus second phalanx recovered from near the bot- tom of our excavations in Unit III is 15,100 ± 700 yr B.P. Figure 5 is a schematic strati- graphic section illustrating the proposed stratigraphic relationships including the ra- diometric dates. April 1989 Mead, Mead: Snake Creek Burial Cave 145 Fig. 2. Entrance to Snake Creek Burial Cave Recovered remains. — Eight species of mustelids have been recovered from SCBC, including three species not previously re- ported from late Rancholabrean localities in the Great Basin: Mustela nigripes (black- footed ferret), M. nivalis (least weasel), and Gulu gu/o (wolverine). The implications of these mustelid reports are discussed. The di- versity of identified mustelid species from SCBC is higher than that from any other local- ity in the Great Basin. Table 1 provides a list of those localities in the Great Basin reporting fossil or subfossil mustelid remains. Reports from a majority of the sites include only one or two mustelid species (Hidden Cave and Smith Creek Cave both report six, the most after SCBC). However, some 18 localities are now reporting fossil or subfossil mustelids. Figure 1 illustrates the geographic locations of these 18 sites; the arbitrary site numbers corre- spond to the numbers assigned in Table 1. Also included is a review of the fossil and subfossil mustelid localities in the Great Basin. Descriptive Accounts Maries americana American marten Map localities. — 2, 3, 5, 18. Discussion. — Maries americana generally prefers mature conifer or mixed-forest stands with greater than 30% canopy cover, although meadows are often used in the summer if more food is available there. The American marten is an opportunistic feeder that takes advantage of local and seasonal abundances. While Clethrionomys spp. (red-backed vole), Microtus spp. (meadow vole), Lepus ameri- canus (snowshoe hare), and Tamiasciurus spp. (tree squirrels) are favored foods, fruits and insects can play a significant role in the diet (Ewer 1973, Strickland et al. 1982, Clark etal. 1987). Martes americana is reported from four fos- sil localities in the central Great Basin. How- ever, one of the two specimens from Bronco Charlie Cave (Spiess 1974) has been reas- signed to M. nobilis (Grayson 1987) and is 146 Great Basin Naturalist Vol. 49, No. 2 Fig. 3. SCBC is located by the arrow, background. The valley bottom (pluvial Lake Bonneville) is clearly visible in the discussed in that section. The other specimen remains undetermined. Crystal Ball Cave, Deer Creek Cave, and Snake Creek Burial Cave are all outside the modern range of M. americana. Figure 6 identifies the fossil sites of M. americana, M. nobilis, and Maries sp. in the arid West and illustrates the modern geographic range of M. americana. The American marten currently ranges in the Sierra Nevada and the Bocky Mountains, but it does not occur on mountain ranges within the Great Basin. However, un- til more precise environmental and chrono- logical guidelines are found that distinguish M. americana from M. nobilis localities, most paleoecological interpretations should be con- sidered tentative. Maries nobilis Extinct noble marten Map localities. — 2, 8, 17, 18. Discussion. — Maries nobilis has been re- ported from three fossil localities in addition to the reassignment of the Bronco Charlie specimen (Fig. 6). Anderson (1970) provides measurements and morphological characters of the cranium and dentition used to distin- guish the species. Anderson (1970: 85) states, "I do not believe that M. nobilis was related to M. americana, and that competition with the American marten, a warming climate, and perhaps the activities of man caused the ex- tinction of the noble marten. Until recently, M. nobilis was thought to have become extinct at the end of the Pleistocene and perhaps to have adapted to cooler conditions. Grayson (1987), however, reports three localities in the arid West, two from the Great Basin, with noble marten remains dating into the late Holocene: (1) Dry Creek Bockshelter, south- western Utah, 3,270 ± 110 yr B.P.; (2) Hid- den Cave, Nevada, 3,600-3,700 yr B.P.; and (3) Bronco Charlie Cave, Nevada, 3,500 yr B.P. If we assume these dates are accurate in their associations, we see no clear reason why M. nobilis survived terminal Pleistocene en- vironmental changes and then became extinct in the late Holocene. Clearly, additional re- search is required. Only at SCBC is there an indication that both species may have oc- curred together in the Great Basin. April 1989 Mead, Mead: Snake Creek Burial Cave 147 /<2> ' to lower / cave orea Chimney Fig. 4. Plan view of SCBC with the excavation units enlarged. Martes sp. is reported from two localities, Hidden and Smith Creek caves (8, 17) (Fig. 6). The Hidden Cave specimens are all post- cranial elements, and Grayson (1985) does not definitely distinguish them as either M. amer- icana or M. nobilis. The Smith Creek Cave specimen was first identified by Goodrich (1965) as Martes sp. and further reported by Miller (1979), who made no further identi- fication. Mustela erminea Ermine Map localities. — 15, 17. Discussion. — Ermines generally inhabit a variety of boreal habitats, although they tend to avoid dense coniferous forests and deserts. The diet of M. erminea consists pri- marily of small mammals, especially Microtus spp., Blarina spp. (short-tailed shrew), and Peromyscus spp. (deer mouse) (Ewer 1973, 148 Great Basin Naturalist Vol. 49, No. 2 Table 1. Fossil and subfossil mustelid localities reported from the Great Basin. Site numbers are arbitrary designations and correspond to the numbers on Figure 1. X = fossil/subfossil, in vicinity today; * = extirpated; ! = extinct. « -p "H = Locality Locality t £ number name § § § --. < 1 Amy's Rockshelter 2 Bronco Charlie * I1 X 3 Crystal Ball Cave * X 4 Danger Cave X 5 Deer Creek Cave * 6 Gatecliff Shelter 7 Hanging Rock Shelter X 8 Hidden Cave 1 X X 9 Hogup Cave X 10 Juke Box Cave 11 Kaehina Cave 12 Last Supper Cave X 13 Mineral Hill Cave 14 O'Mallev Shelter 15 Owl Cave #1 X 16 Owl Cave #2 17 Smith Creek Cave ! * X X 18 SNAKE CREEK BURIAL CAVE * ! X U E-. ■g w ~_ Z- r» SI ;i K5 --. y. j-. Primary relerence(s) Miller 1979. Meadetal. 1982 ! X Spiess 1974 Heaton 1985 Jennings 1957, Grayson 1988 Ziegler 1963 X X Grayson 1983 Grayson and Parmalee 1988 X X X X Grayson 1985 Aikens 1970 Jennings 1957 Miller 1979, Meadetal. 1982 X X X Grayson 1988 McGuire 1980 Fowler et al. 1973 Turnmire 1987 Turnmire 1982 X X Miller 1979, Meadetal. 1982 X X X X X X X* X X X X X X X X This report 'See Grayson 1987. Possibly M. erminea; see discussion in te SCHEMATIC STRATIGRAPHIC SECTION Covers' backdirt I ("disturbed") 7860 ±130 rguano; TT Bat guano 9460 ±160 (wood) White to tan-pink, very soft, fine, highly fossiliferous 15 100 ±700 (Equus sp ) :':':':\ (NN^ T!x*S>i-ii. S:;:;*::::;:v:i. o V:|: / 1 7° I1 !yJFi\— — ___ U o ;:f:::::::::xl ° Fig. 6. Modern distribution of Maries americana Fig. 5. Schematic stratigraphic section from SCBC, (shaded area) with fossil and subfossil localities of M. with associated radiometric dates. americana (O) and M. nobilis (•); A = SCBC. April 1989 Mead, Mead: Snake Creek Burial Cave 149 HV / • / • 7l ~--~J\ •J S:-:;:':'::'-'-*-y^:'*^v -4™ 5. V¥ ;...., "^